IPTW 2001

Proceedings of the International Phosphorus Transfer Workshop 2001








Connecting Phosphorus Transfer from Agriculture to Impacts in Surface Waters








28 August to 1 September 2001,

Robbins Centre,

Plymouth University

Plymouth, Devon, England
















Proceedings edited by:

P. M. Haygarth, L. M. Condron, P. J. Butler and J. S. Chisholm

Background and Introduction

This is the 3rd International Phosphorus Transfer Workshop, following the precedents established at Johnstown Castle, Wexford in 1995 and Greenmount College, Antrim in 1998. This workshop seeks a scientific rationale for connecting the environmental effects of phosphorus (P) pollution from agriculture to eutrophication and impacts on aquatic communities. Point sources of P entering surface freshwaters are now relatively easily isolated and abated, but agriculture persists as the major diffuse source. Farms often import excess P via animal feeds and fertilizers and discharges in waters from plot studies have shown that total P loads often exceed 3 kg ha-1 per year, with concentrations of >100 μg L-1. In many cases, the link between land and water has been made anecdotally, with passions running high. In the USA, Congress has been called recently to strengthen the Clean Water Act, and polluted runoff has been described as the "most pervasive problem in the coastal marine environment". Examples of problems downstream of livestock production are internationally distributed and include Chesapeake Bay (North Eastern USA), Gippsland Lakes (Victoria, Australia), and Slapton Ley (Devon, UK).

However, despite such a high international profile, the scientific basis for connecting on-farm practices to impacts of P loss on surface water quality presents many challenges. These arise from:

We have designed an exciting workshop format that we hope will provide for an interactive and innovative think-tank to make progress on this challenging issue.

Phil Haygarth,

Institute of Grassland and Environmental Research,

North Wyke Research Station,



United Kingdom.






Workshop Organising Team

Workshop Organiser:

Philip M. Haygarth, IGER North Wyke, UK

Workshop Secretary:

Ms. Jo Chisholm, IGER North Wyke, Okehampton, Devon, EX20 2SB, UK

Tel + 44 1837 883503 Fax +44 1837 82139 email: IPTW.2001@bbsrc.ac.uk

Web page: http://www.iger.bbsrc.ac.uk/igerweb/NWNew/IPTW/iptw2001.htm

Scientific Committee:

B. L. Turner (United States Department of Agriculture, Idaho, USA)

A. L. Heathwaite (University of Sheffield, Sheffield, UK)

L. M. Condron (Lincoln University, Canterbury, New Zealand)

M. McHugh (Soil Survey and Land Research Centre, North Wyke, UK)

Local Organising Committee:

P. Butler, A. Joynes, S. C. Jarvis, P. Whitehead, F. Wood (Institute of Grassland and Environmental Research, North Wyke, UK)

A. I. Fraser and T. Harrod (Soil Survey and Land Research Centre, North Wyke, UK)

N. Griffen (University of Plymouth, Plymouth, UK)

K . Chell and R. Barrett (Slapton NNR and Field Centre, UK).


Workshop Function and Operation


The workshop is addressing specific questions (see Workshop Questions) which are arranged in four ‘Themes’, increasing in scale from source to impact. These are to be discussed both indoor and outdoors, throughout the formal and social duration of IPTW. Each Theme involves invited lecture(s), followed by offered oral and poster presentations. Delegates are arranged into three discussion groups, around which they discuss the Workshop Questions, with the aim of reporting back in Workshop Consensus Sessions.

The Role of ‘Theme Teams’

For each Theme, an Invited Speaker will give a critical overview of the topic aimed at provoking discussion. A Chair and a Reporter have been allocated to each Theme. The Reporter is required to work closely with the Chair to help reach a consensus and present the conclusions to each Theme on the final day of the Workshop.

Invited Speaker: The Invited Speaker will give an incisive and critical overview of the Theme, with the aim of generating discussion and debate. Some Themes have more than one Invited Speaker.

Chair: The Chair is the team leader and will co-ordinate discussion towards the workshop questions during the indoor presentation and presentation session and during the ‘Consensus Sessions’. The Chair leads and co-ordinates the team.

Reporter: The Reporter will summarise the conclusions and points raised during all the Theme discussion sessions, and present these during the summary session on the final Workshop day, aimed at answering the Workshop Questions.


Groups aim to report their view back to the workshop. Each group has an allocated leader and also a local guide to assist. It is the role of the leader to inspire and co-ordinate the group. It is the responsibility of the group leader to make time on the field workshop (and at other times, evenings, lunches, etc.) to co-ordinate and rationalise your group views.

Theme Teams






Tom Sims

University of Delaware, USA

Ben Turner, USDA USA

John Quinton, Cranfield University, UK

David Nash, NRE, Australia

Ben Turner, USDA, USA




Keith Beven*

Lancaster University,




Christian Stamm* ETH, Switzerland

Ian McKelvie,

Monash University, Australia

Darren Baldwin, CSIRO, Australia






Graham Harris,

CSIRO, Australia



Bob Foy, DANI, Northern Ireland




Rachael Dils, Environment Agency, UK






Peter Costigan,




Tim Burt,

University of Durham, UK



Wim Chardon,

Alterra, The Netherlands

Group Leaders

Owen Carton, TEAGASC, Ireland

Leo Condron, Lincoln University, New Zealand

Emmanuel Frossard, ETH, Switzerland


*In the Concensus Session Dr Stamm will chair in place of Professor Beven



Workshop Questions

Connecting Phosphorus Transfer from Agriculture to Impacts in Surface Waters


What farm management practices and agricultural policy measures do we recommend to reduce

of P?



Is it necessary to distinguish between surface and subsurface pathways when studying P mobilisation?

What is the usefulness of studying P forms during movement from source to impact?


What scientific evidence is there to link the observed aquatic impacts of P to the farm?

Can we adequately quantify agricultural (versus point source) contributions to the total P load in a catchment?


Integrate the findings of themes 1, 2 & 3 and propose achievable and economical remediation options for reducing the aquatic impacts of P from agriculture?





Day 1 – Wednesday 29th August - Robbins Centre & Lecture Theatre



Introduction and Welcome to IPTW, Professor Steve Jarvis, IGER.

Connecting P transfer from Agriculture to Impacts in Surface Waters - an introduction to the workshop concept

P. Haygarth, M. McHugh, B. Turner, L. Condron & L. Heathwaite



What farm management practices and agricultural policy measures do we recommend to reduce

  • solubilisation transfers
  • physical transfers
  • incidental transfers

of P?

Chair: Professor Tom Sims, University of Delaware, USA


Chair’s introduction


Benjamin Turner, USDA, Idaho USA and Richard McDowell, AgResearch, New Zealand - Alchemy within agriculture: Current perspectives on phosphorus solubilisation


David Nash, Agriculture Victoria Ellinbank, Austrailia - The structure of phosphorus exports (invited paper)


John Quinton - Cranfield University, UK- Detachment of phosphorus: processes and interactions (invited paper)


Coffee and Theme 1 Posters (session 1 of 2)


Marianne Bechmann, Jordforsk, Norway - Effects of extreme weather on losses of phosphorus and suspended solids from agricultural catchments


Robin Hodgkinson, - ADAS Gleadthorpe, UK - The significance of tile drain flow for sediment associated P transfer at the catchment scale


Rory Maguire University of Delaware, USA - Observations on leaching and subsurface transport of P on the Delmarva Peninsula


Gurpal Toor, Lincoln University, New Zealand - Incidental phosphorus loss from a grassland soil following application of dairy shed effluent


Regis Simard SCRC, Canada - Phosphorus transfer as affected by timing of manure application


Lunch in the Robbins Centre






Is it necessary to distinguish between surface and subsurface pathways when studying P mobilisation?

What is the usefulness of studying P forms during movement from source to impact?

Chair: Professor Keith Beven, University of Lancaster, UK


Chair’s Introduction


Christian Stamm - Institute of Terrestrial Ecology, Switzerland - Hot spots and short cuts causing P losses into open waters (invited paper)


William Gburek - USDA-ARS, University Park, PA, USA. - Phosphorus transport in upland catchments - Connectivity of field and stream


Jim W. Cox - CSIRO Land and Water, Australia - Pathways for phosphorus loss off pastures in South Australia


Ian Foster - Centre for Environmental Research and Consultancy, Coventry University, UK. Modelling sediment delivery from agricultural land to the fluvial system via sub-surface drainage


Tea and Theme 1 Posters


Ian McKelvie, Monash University, Australia - Phosphorus in fluvial systems: speciation, transformations and dynamics (invited paper)


Alice Melland - Institute of Land and Food Resources, University of Melbourne, Australia - Runoff from high and low fertility sheep pastures: P concentrations and forms


Darren Baldwin - Murray-Darling Freshwater Research Centre, Australia. - Sediment- phosphorus interactions: a synthesis


Declan Ryan, - TEAGASC, Oak Park Research Centre, Ireland - Using water table and weather data to identify fields prone to overland flow























Day 2 – Thursday 30th August - Field Workshop Day

Organized and coordinated by Marianne McHugh, Soil Survey and Land Research Centre, North Wyke, Devon

Field Workshops for Themes 1, 2 and 3 at Slapton Ley.

Group Leaders and Guides for the Field Workshops will be:

Blue Group: Owen Carton, TEAGASC, Ireland (with Trish Butler as local Guide)

Red Group: Leo Condron, Lincoln University, New Zealand (with Adrian Joynes as local Guide)

Green Group: Emmanuel Frossard, ETH, Switzerland (with Marianne McHugh as local Guide)


Introduction and Background

On August 30th, Workshop Day 2, delegates will leave Plymouth to spend the day at Slapton Ley, a mixed agricultural catchment in Southeast Devon.

Slapton Ley is a National Nature Reserve situated in the South Devon Area of Outstanding Natural Beauty. Its main feature, the freshwater Ley, is separated from the sea by a narrow shingle bar and is the largest natural freshwater lake in Southwest England. The reserve is renowned for the richness of its plant, fungi and bird life and is managed to enhance and conserve this wildlife while promoting education and research through the Field Centre, which was established in 1959.

Slapton, which means Slippery Place, has been populated since the Bronze Age and had a population of 200 at the time of the mid-12th century Domesday Survey. During the second World War, some 750 residents of Slapton and the surrounding parishes were evacuated to make way for US troops preparing for the D-Day landings on the Normandy Beaches. Many troops lost their lives over the next year and a half, either from ‘friendly’ fire or from a surprise attack by German e-boats during Operation Tiger. Today, a memorial on Slapton Beach and a Sherman tank, raised from the seabed and on display at Torcross, remind visitors of Slapton’s difficult past.

Further information on Slapton Ley is available at www.slapton.org and from www.english-nature.org.uk.

The Field Day

The following table outlines the structure of the IPTW Field Day. On each coach, theme leaders and the local guides will discuss the Field Day and offer insights into the South Devon countryside.

At no point will delegates enter farmland or will they come into contact with livestock.

Throughout the Field Day, delegates will continue to discuss the three IPTW 2001 themes of sources, movements and impacts of P, with particular focus on the Workshop Questions. It is the role of Group Leaders to co-ordinate and target discussion and present (or nominate someone to present) the main conclusions arising from these discussions, under topic headings, to the workshop on Friday (August 31st).



Activity and Location



Depart Plymouth for Slapton Ley.



Delegates will be introduced to the field day format (M. McHugh) and to Slapton Ley Field Centre & National Nature Reserve (Keith Chell, Slapton Ley Field Centre)


Field Component 1

While the RED group discusses the catchment’s Sources of Phosphorus with T. Harrod and P. Whitehead, the GREENs deal with Hydro-connectivity (A. Fraser, F. Wood and R. Smith). The BLUE group, meanwhile, will discuss Impacts of P loss within Slapton Ley (L. Heathwaite + T. Burt, P. Johnes, I. Foster).




Field Component 2

Delegate groups rotate to deal with the second of the three field day issues

REDs go to Hydro-connectivity

GREENs go to Impacts

BLUEs go to Sources



All delegates reconvene in Field HQ


Field Component 3

Delegates complete their third and final field session

REDs deal with Impacts of P loss

GREENs consider Sources of P

BLUEs discuss Hydro-connectivity



Day 3. Friday 31st August - Robbins Centre & Lecture Theatre













Consensus Session, Theme 1:

What farm management practices and agricultural policy measures do we recommend to reduce

  • solubilisation transfers
  • physical transfers
  • incidental transfers

of P?

Chair: Professor Tom Sims, University of Delaware, USA

Chair’s Introduction

Blue Group Reports

Red Group Reports

Green Group Reports

Open Discussion











Workshop Consensus Session, Theme 2:

Is it necessary to distinguish between surface and subsurface pathways when studying P mobilisation?

What is the usefulness of studying P forms during movement from source to impact?

Chair: Dr Christian Stamm, ETH, Switzerland

Chair’s Introduction

Blue Group Reports

Red Group Reports

Green Group Reports

Open Discussion



What scientific evidence is there to link the observed aquatic impacts of P to the farm?

Can we adequately quantify agricultural (versus point source) contributions to the total P load in a catchment?

Chair: Professor Graham Harris, CSIRO Land and Water, Australia.


Chair’s introduction


Bob Foy, Department of Agriculture, Northern Ireland - Defining the impact of agricultural P on aquatic systems (invited paper)


Berit Arheimer -SHMI Research and Development, Sweden - A model for biogeochemical processes in lakes to be used in eutrophication management


Coffee and Theme 2 Posters


Rachael Dils - Environment Agency, National Centre for Ecotoxicology and Hazardous Substances, UK. - Freshwater Eutrophication - Linking nutrient levels and biological impact


Andrew Wade -Aquatic Environments Research Centre, University of Reading, UK. -Towards modelling the transport and fate of phosphorus in river systems


Patricia Chambers - Environment Canada -Phosphorus losses from agriculture: Effects on Canadian aquatic ecosystems


Lunch in the Robbins Centre












Consensus Session, Theme 3:

What scientific evidence is there to link the observed aquatic impacts of P to the farm?

Can we adequately quantify agricultural (versus point source) contributions to the total P load in a catchment?

Chair: Professor Graham Harris, CSIRO Land and Water, Australia.

Chair’s Introduction

Blue Group Reports

Red Group Reports

Green Group Reports

Open Discussion


Integrate the findings of themes 1, 2 & 3 and propose achievable and economical remediation options for reducing the aquatic impacts of P from agriculture?


  • the key omission(s) in our technical knowledge of P transfer and future research area(s)
  • a theme for the next P workshop in three years time.

Chair: Dr Peter Costigan, Department of Environment, Food and Rural Affairs, UK


Chair’s introduction.


Tim Burt - Dept Geography, University of Durham, UK. - Connecting fields to the river: the need for a spatially distributed approach to modelling phosphorus transport in agricultural catchments. (invited paper)


Malcolm McGechan - Scottish Agricultural College - Modelling through soil losses of phosphorus to surface waters


Murray Hart - Summit- Quinphos (NZ) Ltd, New Zealand - Assessing the risk of phosphorus loss from land to surface water - a farm management tool


Roger Smith - AESD, Dept Agricultural and Rural Development, Belfast - Reversing the upward trend in soluble phosphorus losses in drainflow from a grassland catchment


Tea and Theme 3 and 4 posters


Reporter Theme 1 – Ben Turner, USDA, Idaho, USA (5 minute summary presentation + open discussion)


Reporter Theme 2 – Darren Baldwin, Murray Darling Freshwater Research Centre, Australia (5 minute summary presentation + open discussion)


Reporter Theme 3 – Rachael Dils, Environment Agency, UK (5 minute summary presentation + open discussion)


Reporter Theme 4 – Wim Chardon, Alterra, The Netherlands (5 minute summary presentation + open discussion)


Editorial Note of Extended Abstracts

An assembly of extended abstracts offered in advance of the Workshop follows. Contributors were asked to report completed research or substantial interim results or conceptual thinking. The editors do not assume responsibility for content of the abstracts.

Authors were responsible for having the abstracts internally refereed prior to submission. The editors did not review the abstracts. However, it was necessary to undertake some minor changes, in respect to clarity and consistency. Thus, presentation of the work does not reflect scientific merit per se, and therefore cannot be considered a refereed publication. Rather, the purpose is to communicate to peers in the workshop, for the purpose of facilitating discussion.

P. M. Haygarth, L. M. Condron, P.J. Butler and J. Chisholm



Theme 1 - Sources of phosphorus (P) on the farm and the initiation of P transfer

- Oral papers


Alchemy within agriculture: current perspectives on phosphorus solubilisation

Benjamin L. Turner1 and Richard McDowell2

1USDA–ARS, Northwest Irrigation and Soils Research Laboratory, Kimberly, Idaho, USA

2AgResearch, Invermay Agricultural Centre, Mosgiel, New Zealand

Phosphorus (P) solubilisation describes a complex series of interactions between chemical P forms, release mechanisms and the soil micro-environment, which provide the first key step in the transfer of soluble P from soils to watercourses. Solubilisation is strictly defined as the release of particles <1 nm in size, but is defined operationally as the release of particles <0.2 or 0.45 µm. This means that particle-associated P released through detachment mechanisms can be included, which is further confounded by chemical speciation based on molybdate reaction. Both biological and chemical release mechanisms are involved, each contributing to the solubilisation of organic and inorganic phosphorus. Chemical mechanisms releasing inorganic orthophosphate are relatively well understood, but detailed knowledge of organic P solubilisation and biological release mechanisms remains elusive, despite their potential importance. Phosphorus solubilisation can be investigated at the molecular level using a variety of analytical techniques, which provide important information on the processes involved. The challenge is to link molecular level solubilisation to observed patterns of P movement at larger scales.

The structure of phosphorus exports and the role of critical incidents

D. Nash, D. Halliwell and M. Hannah

Agriculture Victoria Ellinbank, RMB 2460 Hazeldean Road Ellinbank, Vic. 3821, Australia


Phosphorus (P) exports from agricultural land are a worldwide problem. Exports are generally initiated when P is entrained in moving water as a result of dissolution or physical processes. The amount of P exported depends on the source quantity, mobilisation rate and transport properties. Critical incidents, such as spreading fertiliser and animal wastes or grazing, increase P availability, P concentrations and P exports. Partitioning P concentrations in exported water into base (systematic) and incremental (incidental) components provides a useful research and management framework for addressing nutrient management at a farm scale.

A framework for fnalysing phosphorus export data

When water is exported from land P is exported with it. Given the episodic nature of weather events and nutrient loads being the product of export volume and concentration, it is not surprising that export volumes rather than export concentrations primarily determine P loads per event. Unfortunately, at a farm scale the land manager often has little if any control over export volumes, to reduce P exports they must rely on reducing P concentrations when runoff occurs.

For a given hydrology the concentrations of P exported can be partitioned into a base or systematic component that varies with land-use, and an incidental component resulting from decisions, premeditated or not, made by the land manager.

Using this framework and non-linear modelling techniques, P concentrations can be attributed to incidental factors and to a lesser extent systematic factors (Equation 1). While these techniques do not demonstrate causality and noise in field data may reduce levels of statistical significance, the physical interpretation of the equations provides useful management information.

Equation 1 refines an earlier model applied to an Australian grazing system (Nash et al., 2000) in which P export concentrations were related to the time (days) between fertiliser application and runoff (DF), grazing and runoff (DG) and total runoff volume (TF). The ‘base plus increments’ model given by Equation 1 (adj. R2 = 0.75) characterises the effects of DF and DG as additive exponential decay terms. Hydrological (dilution) differences are allowed for by the multiplicative flow term (TF).


The half-life of fertilisers in this example was in the order of two days (95% confidence interval 1 and 6 days). By incorporating a year factor into the analyses (k i) the longer-term systematic effects of changing background soil fertility can be compared with the short-term, incidental impacts of fertiliser and grazing. The dilution effect of TF on TP was weak (a = 0.10 ± 0.067) but when removed the model yielded some non-sensical results presumably due to the effects of a few large storms.

At a farm level these data have been used as justification for rescheduling fertiliser application to times when runoff is unlikely and the development of a withholding period for irrigation water following fertiliser application. At a research level these data are justification for developing fertiliser compounds that reduce P exports by optimising P release for the hydrology of the irrigated or rain-fed system in which they are to be used.

An additional benefit of partitioning nutrient exports in this way is that it provides the benchmark against which we can compare land-uses within particular landscapes. Numerous authors have demonstrated the considerable between- and within- year variation in P export concentrations. The constant (k i) is an estimate of the minimum P concentration in runoff that might reasonably be expected with our current technology.

Concluding comments

Small decisions often have significant impacts. Using a simple ‘base plus’ framework and appropriate data and statistical techniques helps us identify how to most effectively and efficiently reduce P exports. We can manage incidental P exports. Future gains will be made where strategic research brings more of the systematic component under our management control (ie. converting it to an incidental component).


Nash, D., Hannah, M., Halliwell, D and Murdoch, C. 2000. Factors affecting phosphorus exports from a pasture based-grazing system. Journal of Environmental Quality 29: 1160-1166.

Detachment of phosphorus: processes and interactions

J. N. Quinton

Institute of Water and Environment, Cranfield University at Silsoe, Bedfordshire, MK45 4DT

This paper will review the processes controlling the detachment of particulate phosphorus (P) from soils into overland flow and its subsequent transport and deposition, and will highlight some of the interactions between these processes.

Overland flow is an important source of the P reaching surface waters and contributes both dissolved and particulate P. Solubilisation processes are the main controls on solution losses, however, the short contact time between the water, soil surface and sediment particles entrained in the flow suggests that losses of dissolved P will be low. This is normally the case with most of the P moving in this pathway doing so while attached to sediment or organic matter (Figure 1).

Before particulate P can be transported it must first be detached. We can describe at least three mechanisms: physical detachment caused by the impact of raindrops and the flow of water over a soil surface, and the physio-chemical process of dispersion. The physical processes of detachment are well described by erosion scientists and mathematical descriptions are included in a number of process based soil erosion models. However, physio-chemical detachment has largely been ignored by erosion scientists, and the interactions between physical and physio-chemical detachment and its implications for P loss have been ignored my almost everyone. If we want to understand, and predict, P transfer then it becomes important to know not only how much soil is detached by these processes, but what its characteristics are: is it still an aggregate or is it a primary particle? Is the P from the outer edge of an aggregate, where P concentrations may be higher, or from an aggregate centre, where P concentrations may be lower? What is the particle’s size and density? All these factors will influence how much and how far the detached P will be transported, and not only require us to have a better understanding of the detachment process, but also to know what the initial distribution of P within the surface of the soil is. Without this information we cannot predict how much P will be detached since we do not know how much P is associated with particular aggregates or soil particles.

Once detached, sediment may be transported by overland flow. The generation and description of overland flow mechanisms has also been explored in depth by hydrologists and erosion scientists. Overland flow is variable, in space and time, and is often driven by the variability in rainfall patterns. Overland flow may be generated on a hillslope only to infiltrate further down the slope when the rainfall stops. Flow conditions can also change within a storm: at the storm’s peak the flow may be able to transport particles of all sizes, but during periods of lower discharge only the smallest particles may be transported. This variability in flow conditions leads to selective deposition of P, since lower flow velocities and depths will encourage the deposition of coarser particles, with some colloidal material only being deposited when all the flow has infiltrated. Flow conditions will also influence the connection of P bearing overland flow to surface waters, since travel distances from some contributing areas may be high requiring long flow durations.

The interaction between the overland flow and soil detachment processes can produces unexpected results. For example Quinton et al (2001) have shown that smaller events transfer proportionally larger amounts of phosphorus (Figure 2) and that frequent small erosion events accounted for more phosphorus transfer than infrequent large events. This suggests that small events should not be ignored when targeting control measures and that the phosphorus transport associated with sediments may be more widespread than previously thought.


Quinton, J.N., Catt, J.A. and Hess, T.M 2001 The selective removal of phosphorus from soil: is event size important? Journal of Environmental Quality 30

Effects of extreme weather on losses of phosphorus and suspended solids from agricultural catchments

M. Bechmann and N. Vagstad

Jordforsk - Centre for Soil and Environmental Research, Frederik A. Dahls vei 20, 1432 Aas, Norway

Many countries has experienced periodical extreme weather conditions during the last years. In Northern Europe, the autumn 2000 had extremely high precipitation and high temparatures. In southern Norway precipitation up to five times the 30-year-average was recorded from October to December. Since precipitation has a major influence on erosion and phosphorus (P) losses – the question arises of how these extreme weather conditions influence the losses of suspended solids (SS) and P from agricultural land. This abstract presents preliminary results from measurements in southern Norway in the autumn 2000.

Results were based on The Agricultural Environmental Monitoring Program in Norway (Bechmann and Våje, 2001). Four catchments (65-680 ha) and two field scale studies (4-6 ha) in the south eastern Norway were included. Water discharge and losses of P and SS were measured in the stream outlet of the catchments and for the fields both surface and subsurface runoff were measured.

Losses of phosphorus and suspended solids

The results presented here are preliminary and cover only the first part of this extreme event. Results for the whole autumn period will be available for presentation at the workshop.

The runoff in autumn 2000 was from 2 to 14 times higher than the earlier measured runoff in October-November (Table 1). The effect of high precipitation on loss of SS and P in the catchment studies were higher, 7 to 27 times for SS and 4 to 20 times normal losses of P.

The two field scale studies (Vandsemb and Bye) showed insignificant surface runoff during the whole runoff event. The transport of SS and P resulted mainly from subsurface water for both fields. The subsurface runoff increased 5 and 6 times, for Vandsemb and Bye, respectively, while subsurface losses of SS and P only increased to three times compared to the mean losses in October-November the previous 7 years.

Erosion in the streams may explain a part of the increased transport of SS and P measured in the catchment studies.

During this continuous runoff period of one and a half month, the concentration of total P in subsurface water declined significantly at Vandsemb field and stabilized at around 0.13 mg L-1. The concentration of total P in subsurface water at Bye field decreased to 0.007 mg L-1. The soil P content measured as P-AL was 17 and 7 for Vandsemb and Bye, respectively. Soils may have reached their minimum P concentration in leaching water.

Table 1. Runoff (mm), loss of SS (kg ha-1) and P (g ha-1) from October 6th to November 12th in mean for the 90’s and for the year 2000.



Loss of SS

Loss of P



kg ha-1

g ha-1



















































Vandsemb, subsurface










Bye, subsurface










Results show that extreme precipitation in autumn leads to increased losses of SS and P in subsurface runoff, while no increase were observed for losses in surface runoff in the field scale studies. Losses of SS and P increased less at field scale than at catchment scale, accordingly the contribution from other sources, i.e. stream erosion, needs further evaluation. Stream erosion may explain some of the differences in 2000/mean ratio between catchments. The concentration of P in subsurface runoff from the fields declined during this event and stabilized at a higher level for the field with highest soil P content.


Bechmann, M., and Våje, P. 2001. Monitoring erosion and nutrient runoff losses from small basins representative for Norwegian agriculture. IAHS Red Book (in press).


The significance of tile drain flow for sediment associated phosphorus transfer at the catchment scale

R.A. Hodgkinson1, D.E. Walling2, I.D.L. Foster3, M. Russell2 and A.S. Chapman 3

1ADAS Gleadthorpe, Meden Vale, Mansfield, Nottinghamshire, NG20 9PF, UK

2 School of Geography and Archaeology, University of Exeter, Exeter, EX4 4RJ, UK

3 Centre for Environmental Research and Consultancy, Coventry University, Priory St., Coventry CV1 5FB


In the UK research has shown that transfer of particulate P in tile drain flow occurs in agricultural catchments, (Hodgkinson et al., 1998). The importance of tile drains as a conduit for sediment, and hence particulate P, loss has also been recognised in many other European countries (Laubel et al., 1999). This paper reports on a project, which specifically addressed two issues: firstly, the importance of sediment and associated P loss in tile drain discharge on catchment sediment budgets and secondly source identification. In this paper the main emphasis will be on sediment budgets and sediment sources at the catchment scale. The paper also deals with possible mitigation options for areas where the risk of sediment-associated P loss through tile drains is high.


The study utilised two existing instrumented catchments ADAS Rosemaund (a 150 hectare, mainly arable, on silty clay loam soils) and two subcatchments of the Gilwiskaw Beck near Ashby de La Zouch (mixed arable/grassland on mainly clay soils). The two subcatchments studied were Smisby (260 ha rural) and Cliftonthorpe, (90 ha agricultural).

Statistically verified composite fingerprints of potential sources (Walling et al., 1999) and mixing/unmixing models were used to assess the relative importance of potential sediment sources/pathways and to identify the source of particulates transferred in tile drain flow. Sediment budgets were constructed from traditional measurements and 137Cs measurements. The components quantified were sediment mobilisation; sediment transfer through drain systems; sediment storage and sediment yield at the catchment outlets.

A national risk assessment for losses of sediment, and associated P, in tile drainflow was undertaken by overlaying vulnerable HOST classes with cropping, drainage and climate data to produce a total risk index.

Results and discussion

The results from Rosemaund indicated that the most important sediment source in this catchment was discharge from the tile drains which accounted for ca. 55 % of total load compared to ca. 33 % from surface sources, and ca. 12 % channel banks. At Trent the tile drain contribution was only ca. 30% and bank erosion was only ca. 6% with surface erosion being the dominant pathway. In both catchments the tile drain sediment was predominantly surface derived. Despite the high sediment outputs (ca. 0.7-1.3 t ha-1 yr-1) in these headwater catchments the sediment delivery ratios ranged from 14-28%, which demonstrated that there is considerable retention of sediment. At Rosemaund arable cultivation was found to be the most important sediment source whereas at Cliftonthorpe and Smisby pasture fields had increased importance.

The interaction of land use, soils and landscape with sediment loss will be discussed based on the data from these two catchments with reference to the variable source area concept. National scale implications for sediment-associated P loss in tile flow will be discussed. Potential mitigation options to control sediment loss through tile drains will be discussed.



On soils with a good macropore structure sediment associated transfer in tile drain flow can form a major component of catchment output. Both arable and pasture systems can act as sediment sources.


Hodgkinson, R. A., Matthews, A. and Withers, P. J. A. 1998. Phosphorus losses from small rural catchments: factors influencing pattern and forms, field studies and an outline modelling approach In: Diffuse Pollution and Agriculture II Ed Petchey, T., D’Arcy, B. and Frost, A. SAC, 1998, 84-100

Laubel, A., Jacobsen, O. H., Kronvang, B., Grant, R. and Anderson, H. E. 1999. Subsurface drainage loss of particles and phosphorus from field plot experiments and a tile drained catchment. Journal of Environmental Quality 28: 576-584.

Walling, D.E., Owens, P.N. and Leeks, G.J.L., 1999. Fingerprinting suspended sediment sources in the catchment of the River Ouse, Yorkshire, UK. Hydrological Processes 13: 955-975





Observations on leaching and subsurface transport of phosphorus on the Delmarva Peninsula, USA

R. O. Maguire and J. T. Sims

1Department of Plant and Soil Sciences, University of Delaware, Newark, DE 19717-1303, USA


Concerns about elevated levels of phosphorus (P) in surface waters are longstanding and have recently focused attention on P loss from agricultural land. Much work has gone into developing tools such as the P Site Index, which rank agricultural fields according to their risk for P loss. Most of the research that has gone into the development of the P index concerns overland flow, as this is considered to be the major pathway for P loss from agricultural land (Gburek et al., 2000). However, generally flat topography and sandy soils predominate on the Delmarva Peninsula, and a large proportion of rainfall infiltrates rather than flowing over the soil surface, providing another possible pathway for P loss by subsurface transport.

The research presented here addresses the links between soil properties and P losses from agricultural topsoils by subsurface pathways.

Materials and methods

Five representative soil series were selected on the Delmarva Peninsula and for each series, fields were identified ranging from 'low' to 'excessive' in soil test (Mehlich-1) P, according to the University of Delaware Soil Testing Laboratory (Sims and Gartley, 1996). Twenty one intact soil columns (15 cm diameter by 20 cm deep) were collected from each soil series. These columns were taken to a greenhouse and leachate water was collected after adding the equivalent of 5 mm rainfall to the tops of the columns for four successive days. The leachate was passed through 0.45 µm filters and analyzed for molybdate reactive P. Soil samples collected before and after leaching were analyzed for pH, organic matter, water soluble P, iron strip P, Mehlich-1 P, and Mehlich-3 P, iron (Fe) and aluminum (Al).


For each soil series, the soil columns collected ranged from low to excessive in agronomic soil test P. Leachate P concentrations were frequently greater than those reported to cause eutrophication. For water soluble P, iron strip P, Mehlich-1 P, or Mehlich-3 P, a threshold was seen above which leachate P concentrations increased rapidly with increasing extractable P. This threshold occurred at different values of extractable P for each of the soil series studied. Unlike each of the soil extractable P measurements, soil P saturation, defined as the ratio of Mehlich-3 P:[Al+Fe], showed the same threshold for all of the soil series studied.


High concentrations of P were detected in leachate waters from the topsoils in many cases, indicating that the potential exists for significant subsurface loss of P from agricultural soils on the Delmarva Peninsula. However, other factors specific to each field, such as retention of P by subsoils and hydrologic connectivity to surface waters, must be considered for P loss via subsurface pathways from agricultural soils. All of the soil tests used showed a threshold above which P concentrations in leachate waters increased sharply. However the agronomic 'optimum' soil test P value was always below this threshold, indicating that it is possible to maintain sufficient soil P for agricultural purposes without greatly increasing the risk of P loss via subsurface pathways. The extractable soil P at this threshold varied between soil series, but the ratio of Mehlich-3 P:[Al+Fe] produced the same threshold for all soils. This indicates that Mehlich-3 P:[Al+Fe] may be an appropriate environmental soil P test and a useful tool to aid in the estimation of P loss by subsurface pathways. If we can keep the soil P or P:[Al+Fe] ratio below this threshold, we may be able to reduce P losses from agricultural soils to surface waters via subsurface flow.


Gburek, W.J., Sharpley, A.N., Heathwaite, L., and Folmar, G.J. 2000 Phosphorus management at the watershed scale: A modification of the phosphorus index. Journal of Environmental Quality 29: 130-144.

Sims, J.T., and Gartley, K.L. 1996. Nutrient management handbook for Delaware. Cooperative Bulletin No.59. University of Delaware.

Incidental phosphorus loss from a grassland soil following application of dairy shed effluent

G. S. Toor, L. M. Condron,, H. J. Di, K. C. Cameron and T. Hendry

Centre for Soil & Environmental Quality, PO Box 84, Lincoln University, Canterbury, New Zealand


High concentrations of phosphorus (P) in soil resulting from long-term application of inorganic P fertilizers can lead to P loss which in turn may cause accelerated eutrophication of natural waters. Similarly, continuous applications of animal excreta in intensive dairy farming systems can result in enhanced P leaching, especially in free draining soils. This study reports incidental P leaching losses from a stony grassland soil immediately following application of dairy shed effluent (DSE).


Lysimeters (50-cm diameter, 70-cm depth) were collected from a free draining stony Lismore silt loam soil (Udic Haplustept) (Cameron et al., 1992). The experiment included 4 replicates of the following annual treatments: P45, P45+DSE200, P45+DSE400, P45+DSE400+U (P45 = 45kg P ha-1 as single superphosphate; DSE = dairy shed effluent at 200 or 400 kg N ha-1, U= urine at 1000 kg N ha-1). Flood irrigation (100 mm per application) was applied every three weeks between November and April. Leachate was collected following DSE application in 1999 (Aug, Nov) and 2000 (Feb, May, Aug, Nov). Dissolved reactive P (DRP) and total dissolved P (TDP) were determined in filtered (<0.45 m m) samples, and total P (TP) was determined in unfiltered samples.

Results and discussion

The amount of TP present in DSE varied from 28-62 mg L-1 (Table 1). Of the TP in DSE, dissolved (DRP and DUP) and particulate (TPP) fractions accounted for 49 and 51 %, respectively. For the dissolved fractions in DSE, 44 % was DRP while DUP constituted only 5 %. The average TP loss in leachate ranged from 144 (P45) to 950 - 1385 m g L-1 across DSE treatments; PUP was the main constituent in the leachate (70-79 %) followed by DUP (11-16 %), PRP (7-11 %) and DRP (2-5 %). This suggests that reactive forms of P (DRP) are adsorbed in the soil profile while P in unreactive fractions (DUP and PUP) is more mobile. The higher adsorption of P in reactive forms is particularly related to the higher amounts of Fe and Al oxides and hydroxides present at various depths in the Lismore soil (Sinaj et al., 2001).

Table 1. Concentrations of P in DSE and leachate following DSE application (mean of 6 drainage events).





DSE (mg P L-1)

19 (13-23)*

2 (0.3-6)

22 (8-38)

43 (28-62)

Leachate (m g P L-1)




7 (2-15)

23 (8-67)

12 (4-19)

102 (11-410)

144 (32-500)


21 (6-78)

114 (73-160)

63 (16-96)

752 (252-1699)

950 (351-1930)


49 (7-251)

140 (116-184)

133 (26-186)

958 (258-1765)

1280 (671-2129)


49 (3-266)

208 (36-640)

154 (52-293)

974 (389-1811)

1385 (490-2651)

lsd (5%)






*data in parenthesis are concentration ranges;

DUP = dissolved unreactive P (TDP – DRP); TPP = total particulate P (TP-TDP);

PRP = particulate reactive P (TRP – DRP); PUP = particulate unreactive P (TPP – PRP)


Dairy shed effluent contained equal proportions of dissolved and particulate P, although 70-79 % of the P loss from soil following DSE application occurred as particulate P (PUP).


Cameron, K.C., Smith, N.P., McLay, C.D.A., Fraser, P.M., McPherson, P.J., Harrison, D.F. and Harbottle, P. 1992. Lysimeters without edge-flow: an improved design and sampling procedure. Soil Science Society of America Journal 56: 1625-1628.

Sinaj, S., Stamm, C., Toor, G.S., Condron, L.M., Hendry, T., Di, H.J., Cameron, K.C., and Frossard, E. 2001. Phosphorus availability and loss from irrigated grassland soils. Journal of Environmental Quality (in press)

Phosphorus transfer as affected by timing of manure application

R. R. Simard1, I. Royer1, and G. M. Barnett2

1Agriculture and Agri-Food Canada (AAFC), Soils and Crops Research Centre, Sainte-Foy, Qc, Canada, G1V 2J3

2AAF, Dairy and Swine Research and Development Centre, Lennoxville, Qc, Canada, J1M 1Z3.


The transfer of phosphorus (P) from soil to surface waters is very significant in areas of concentrated livestock operations. Soils from those areas receive amounts of manure P in excess of crops exports resulting in a build up of soil P and high degree of soil P saturation (DSPS) near the surface. The extend of this transfer may be influenced by cropping practices such as timing of manure application and crop species. The objective of this work is to report the impact of the timing of manure application on P transfer in surface runoff and tile-drainage under corn (Zea mays L.).

This experiment was conducted on a site initiated in 1989 near Lennoxville, Quebec, Canada at the AAFC Research Centre on a Coaticook silt loam (loamy, mixed, mesic Typic Humaquept). The soil (0-20 cm) had a pH of 5.8, 53 g kg-1 organic matter and a Mehlich-3 extractable P content of 81 mg kg-1. The site has a 6 % slope and is equipped to collect surface runoff and tile-drainage waters. The experiment includes 5 treatments: (C) control; (M) inorganic fertilizers according to soil test and local recommendations (180 kg N, 7 kg P, 12 kg K ha-1 to corn); liquid hog (Sus scrofa) manure (HLM) at 360 kg total N ha-1 + inorganic fertilizers applied either all in spring at pre-seeding (S), 50 % in the spring and 50 % in the fall after harvest (SF) and all in the fall after harvest (F). The corn residues are chisel ploughed in the the fall. Soils were sampled in the fall of 1997 with a Giddings probe from the 0-5 and 5-20 cm layers. The amount of labile P was determined by extraction in water (Pw) and in the Mehlich-3 solution (M3P) and the soil P saturation degree was evaluated by the ratio of the extractable P to the extractable Al + Fe contents in ammonium oxalate. The surface runoff and sediments were collected in 1996 and 1997 in all significant rainfall events. Filtered (0.45 µm) and unfiltered water samples were analysed for molybdate-reactive P (MRP) and total P forms.

Soil M3P and DSPS are much larger in the manure treatment than in the mineral fertilizers or control plots in the 0-20 cm layer only. The DSPS were near recognised thresholds in the 0-5 cm layer. The water soluble P contents of the 0-5 cm layer from manure treated soils were 3 to 5 times larger than from mineral fertilizers. Most (90 %) of the total P (TP) in surface runoff was in particulate form (PP). Dissolved reactive P was 30 % of total dissolved P in drainage in 1996 and 44 % in 1997. The average P transfer was 10 kg ha-1 in 1996 and 2.6 kg in 1997. The P transfer in surface runoff was more important for fall manure application that for the spring, the split application or mineral fertilizers in 1996. No difference between treatments was noted in 1997. The average load of dissolved P was more important in tile-drainage from mineral fertilizers plots than from manure applied in the spring or in the fall in 1996 and 1997. Although the amount of P applied is much larger in the spring manure application than in the mineral fertilizer one, comparable amounts of P were transferred in these two treatments. The results from this study suggest that pre-seeding application of liquid manure or split applications are efficient strategies to reduce the P transfer from such soils under corn.

Theme 1 - Sources of phosphorus (P) on the farm and the initiation of P transfer

Phosphorus change point in heavily manured clay soils

Roy Anderson1

1Department of Agricultural & Environmental Science, The Queen’s University of Belfast and Department of Agriculture and Rural Development for Northern Ireland, Newforge Lane, Belfast BT9 5PX, UK

Materials and methods

In previous work it has been demonstrated that phosphorus (P) sorption-desorption behaviour in the surface layers (0-10cm) of animal manure-amended soils may be poorly correlated with standard agronomic measures of P, or with soil total P (TP), at higher P loading rates (Anderson and Wu, 2000). In subsequent work, P loading down the soil profile was compared with lysimeter-collectable molybdate reactive P (MRP), on plots treated, long-term, with animal slurries (Anderson and Xia, 2000). The present report takes this a stage further by considering P balance over the life of the experiment and the relationship between lysimeter-collectable P and agronomic measures of P. The work was conducted in the Long Term Slurry Experiment at the Agricultural Research Institute, Hillsborough, Co. Down, Northern Ireland (UK).

Plot treatments (three replicates) in the experiment were as follows: mineral fertiliser only, with 32 kg P applied ha-1 yr-1 (FERT); pig or cow slurry applied at low rates of 50 m3 ha-1 yr-1 (LPIG, LCOW); medium rates of 100 m3 ha-1 yr-1 (MPIG, MCOW); or high rates of 200 m3 ha-1 yr-1 (HPIG, HCOW). Olsen P (bicarbonate-extractable), total P and Q/I parameters derived from fitted Langmuir isotherms were determined down the soil profile, at 0-10 cm, 10-30 cm, 30-60 cm amd 60-90 cm. Molybdate RP was determined in soil micropore water collected at two depths, 35 and 90cm, using suction cup lysimeters. These data were used to determine the response of soils to applied P over the 30 years of the experiment (set up in 1970), and the following results and conclusions derived.

Results and conclusions

Olsen P had previously been regressed upon total P across plots and the two were found to be well correlated, with r2 = 0.918 at 0-10 cm and r2 = 0.814 over all depths to 90 cm. Olsen P was found to decline with depth to very small values below 40 cm in the profile regardless of treatment, even though the high slurry treatments had been receiving 130-180 kg P ha-1yr-1 for 30 years. Despite this lack of P in the soil profile below 30-40 cm, environmentally significant amounts of P had been collected from soil water by lysimeter at 35cm and to a lesser extent at 90 cm in the high slurry treatments but not the rest (Anderson and Xia 2000).

There was, furthermore, a significant difference in behaviour between the two types of manure used, cow slurry having a relatively high organic and dry matter content, whereas pig slurry had a much lower organic and dry matter content. Equivalent treatments had been roughly matched for P input from the inception of the experiment and P balance was found to be close over the thirty years since elapsed. Nevertheless the Olsen P in soils of the pig slurry treatment plots was significantly higher than in soils from the equivalent cow slurry-treated plots at equivalent levels of input. The MRP content of lysimeter-collected water from the HPIG plots was actually slightly lower than that from the HCOW plots. Plotting Olsen P against lysimeter P therefore indicated P change points for pig and cow treatments which differed substantially. For pig treatments the change point was above a P index of 4 (derived from Olsen P), whereas in cow treatments it was above a P index of only 2.

It had previously been demonstrated that agronomic measures of soil P, while good estimates of total P in the soil, could be misleading in assessing the risk of P leaching through the soil profile. This is further demonstrated in the present report where P release to micropore soil solution, while following expected trends from Q/I measurements, is not well correlated with Olsen P or total P levels present in the soil profile. It is concluded that the difference in behaviour between soils in the two main experimental blocks, where either cow or pig slurry was applied, must be due to the large difference in organic matter applied in the two slurry types. Northern Ireland is a foremost dairying region within the British Isles producing substantial volumes of cattle manure annually from winter housing of stock. Use of Olsen P and other chemical extraction methods may be delivering a misleading picture of the risk of P loss from soil where cattle manure is being applied in quantity to silage swards.


Anderson, R. and Wu, Y. 2000 Phosphorus quantity-intensity relationships and agronomic measures of P in surface layers of soil from a long-term slurry experiment. Chemosphere 42: 161-170.

Anderson, R. and Xia, L. 2000 Agronomic measures of P, Q/I parameters and lysimeter-collectable Pin subsurface soil horizons of a long-term slurry experiment. Chemosphere 42: 171-178.



Phosphorus losses in surface runoff response to successive rainfalls from agricultural land

E. Azazoglu1, P. Strauss2, I. Sisák3 and W. E. H. Blum1

1Institute of Soil Science, University of Agriculture, Vienna, Austria

2Institute for Soil and Water Managment Research, Federal Agency for Water Managment, Petzenkirchen, Austria

3 Department of Soil Science and Water Management, University of Veszprem, Keszthely, Hungary


Phosphorus (P) and Nitrogen (N) are both nutrients often associated with accelerated eutrophication of lakes and streams. However, P is most often the element limiting accelerated eutrophication. Therefore, minimizing lake eutrophication from agricultural nonpoint source pollution (NPS) often requires controlling P inputs to surface water. Sharpley et al. (1996), also identified the importance of developing P management strategies to limit surface water eutrophication from agricultural NPS. For intensively used agricultural areas erosion is a major pathway of P losses from agriculture to surface waters. In general this pathway leads to an enrichment of P because of selective transport of soil particles during the soil erosion process (Sharpley, 1985). Within any given year, soil erosion hazard exhibits a huge temporal variation due to dynamic changes of soil, plant and climatic characteristics. This may affect the interaction between soil loss and phosphorus transport. However, experimental studies about the effects of repeated rainfall on soil- and associated phosphorus losses are limited. We therefore started a rainfall simulation experiment to assess the effect of repeated rainfall on runoff, soil loss and associated P losses for different soils of Europe. At this occasion we report the experimental results for a Luvisol (sandy loam), located in Keszthely, Western Hungary.

Materials and methods

We etablished four 2 m by 5 m plots on a 7,5% slope which had been prepared to seedbed conditions five days before the experiment started. We used a rainfall simulator described by Strauss et al. (2000) to apply rainfall at a rate of 60 mm/h until steady runoff conditions were reached (typically after about 70 min). We repeated the same treatment at days 4, 11, and 18 after the first rainfall simulation. As a pretreatment at each experiment we covered the plots with a permeable mesh and applied 30 mm of rainfall to obtain equal water contents for the start of the main experiment. We measured runoff, sediment, total phosphorus and dissolved phosphorus (molybdate reactive P in filtrate after filtration of the soil water suspension at 0.45 μm). Results presented are mean values for each consecutive rainfall simulation.


Runoff and soil loss rates increased drastically with repeated rainfall application. Total runoff ranged from 4 l m-2 during first rainfall to 39 l m-2 for the fourth successive repetition of applied rainfall. Total amounts of sediment were 34 times greater for the fourth repetition compared to the first application of simulated rainfall. Final infiltration rates decreased with each consecutive rainfall and ranged from 49 mm h-1 (first rainfall) to 23 mm h-1 (fourth rainfall). The positive relationship between these driving forces and the amounts of total P losses were clearly apparent. Losses of dissolved P amounted between 0.7 and 5.3 mg P m-2 for the repeated rainfall applications. In contrast the dissolved P concentration was relatively constant, varing between 0.12 and 0.16 mg l-1 during each rainfall application. No significant effect of successive rainfall on dissolved P concentrations could be detected. Similar observations on the behaviour of dissolved P concentrations were reported by Torbert et al. (1995) and Baker and Laflen (1982).


This work was supported by the European Community under contract EVK1-CT-1999-00007.


Baker, J.L. and J.M. Laflen. 1982. Effects of corn residue and fertilizer managment on soluble nutrient runoff

losses. Transactions of the ASAE, 344-348.

Torbert, H.A., K.N. Potter and J.E. Morrison jr. 1996. Management effects on nitrogen and phosphorus losses in

runoff on expansive clay soils. Transactions of the ASAE. 39(1): 161-166.

Sharpley, A.N. 1985. The selective erosion of plant nutrients in runoff. Soil Sci. Soc. Am. J. 49: 1527-1534.

Sharpley, A.N., T.C. Daniel, J.T. Sims, and D.H. Pote. 1996. Determining environmentally sound soil

phosphorus levels. J. Soil and Water Cons. 51(2): 160-166.

Strauss P., J. Pitty, M. Pfeffer, and A. Mentler. 2000. Rainfall simulation for outdoor experiments. In P. Jamet

and J. Cornejo (eds.). Current research methods to assess the environmental fate of pesticides. Pp. 329-333, INRA Editions.

Phosphorus leaching from five heavily fertilized soils

E. Barberis1, F. Ajmone-Marsan1, M. Presta1, D. Sacco2, L. Zavattaro2, and L. Celi1

1University of Torino, DIVAPRA, Chimica Agraria, v. L. da Vinci 44 , 10095 Grugliasco (TO), Italy

2University of Torino, AgriSelviTer, v. L. da Vinci 44, 10095 Grugliasco (TO), Italy

The loss of phosphorus (P) from agricultural soils to watercourses has increased over the past few decades as a result of the excessive accumulation of P in soil. The transfer of P from agricultural land can occur through a variety of pathways, including overland flow, lateral subsurface movement and vertical seepage through the soil profile. Owing to the relatively high capacity of most soils to fix inorganic P, P leaching has often been perceived as negligible in relation to transfer by erosion and overland flow. However, recent studies have shown that P may be dissolved from the solid phase and transported vertically through the soil. Heckrath et al. (1995) demonstrated that the concentration of dissolved P in drainage water increased considerably when the P-status of the topsoil exceeded a certain level (a change point).

Until recently, most studies gave little attention to the movement of organic P fractions in solution and leachates even if organic phosphorus can represent a major part of the total P present in the soil solution. The role of organic forms in leaching of P can be particularly important in soils supplied with manure or slurry that contain significant amount of organic P (Gerritse and Vriesema, 1984).

The purposes of this work were (1) to verify if vertical deep leaching can represent a mechanism for P transfer from soil to waters in an area of high livestock density, and (2) to evaluate the relative importance of dissolved organic P.

This research was conducted on an area of intensive cattle and pig breeding in the western Po Plain (Italy). Previous studies had shown that cattle farming produces high P surplus, the input being higher than the uptake of as much as 94 kg P ha-1 (Grignani, 1996). Five soil profiles were studied: a Mollisol and four Inceptisols, representative of the soil types of the area. Their texture was coarse loamy or fine loamy. The organic matter content ranged from 1.6 to 4.7 % in the topsoil and sharply decreased below the ploughed layer. Four profiles were located in farms that used liquid and solid cattle manure, while one profile received only inorganic P fertilizers. Two soils were cropped with winter cereals, one with It. ryegrass-maize rotation, one with It. millet and one was a grass ley. The water table was very shallow in the Mollisol (<0.7 m), ranged between 1 and 3 m in three Inceptisols and was always deeper than 3 m only in one case. Because of the shallow water table the soil water content was often high, and this facilitated water and solutes movement through the soil profiles.

Soils were sampled at 2 months interval from June 1999 to June 2000, at depths corresponding to the diagnostic horizons. Soil CaCl2 extracts were obtained and reactive P (RP) and total soluble P (TP) were determined on filtered (<0.45 m m) samples. The difference between TP and RP was called unreactive soluble P (UP). RP represents inorganic orthophosphate and P weakly bound to soluble organic matter that may also be released into solution; UP is generally considered to represent organically bound P, such as phosphate esters, although condensed inorganic P forms, such as pyrophosphates, are also included.

Mean concentrations of TP from the five soils ranged from 0.02 to 1.38 mg L-1 being always more than 0.38 mg L-1 in Ap horizons. In those soils where the Ap horizons was divided in two sub layers, in the deeper one the TP content was approximately halved, indicating that most of P supplied through fertilization is retained in the upper part of the soil. In the deepest horizons TP resulted relatively high (0.045-0.081 mg L-1). RP was high in the top horizons, then decreased and was almost absent below 60 cm. In all profiles UP was the most abundant form even in the profile that did not receive manure or slurry. Unreactive P represented 61% of TP in the ploughed horizons and 92% of TP in the deeper horizons, indicating that this form can significantly contribute to P load to groundwater. Unreactive P was significantly correlated (r=0.878, n=108) to the soil organic matter content.

Both RP and UP showed maximum levels in June and minimum levels in January. This could be related to the fertilizer supply and to biological processes, which result in the mineralization or immobilization of P, combined with seasonal changes in the solubility of organic matter.

The five studied profiles showed a potential for P leaching especially in the organic form: the composition of soil solution is controlled by several processes: amorphous iron oxides and organic matter can retain a large proportion of RP in the Ap horizons, biological processes and changes in the solubility of organic matter can regulate the composition of soil solution.


Gerritse R.G. and Vriesema R. 1984. Phosphate distribution in animal waste slurries. J. Agric. Sci. 102: 159-161.

Heckrath G.W., Brookes P.C., Poulton P.R. and Gouldin K.W.T. 1995. Phosphorus leaching from soils containing different concentrations in the Broadbalk experiment. J. Environ. Qual. 24 : 904-910.

Grignani C. 1996. Influenza della tipologia di allevamento e dell'ordinamento colturale sul bilancio di elementi nutritivi di aziende padane. Rivista di Agronomia 30 (3 suppl.): 414-422.

Phosphorus in wastewater from pig farms

Andelka Belic and Sima Belic

University of Novi Sad, Faculty of Agriculture, Institute for Water Management, 21000 Novi Sad, Trg Dositeja Obradovica 8, Yugoslavia

Agriculture is, and always has been, an activity involving a close interaction with the environment. Recent decades have seen great changes agricultural production in many parts of Yugoslavia. Thus there are many pressures to modernize and intensify agricultural systems, and to minimize the effect of such systems on the environment.

Intensive development of livestock breeding in the northern part of the Yugoslavia (Vojvodina Province) has been characterized by construction of large-capacity pig breeding farms involving wet discharge technologies, and most often, by inappropriate solutions to the treatment and disposal of the generated wastewater. In the majority of cases such wastewaters are temporarily stored in lagoons. Sometimes, wastewater from lagoons is disposed of onto surrounding plots.

The experimental field was established to obtain a complete insight into the changes of groundwater and surface water in the vicinity of a lagoon. Wastewater from the lagoon, surface and groundwater samples were taken four times a year during a three-year period and analysed for phosphorus (P). Wells were drilled on the experimental area in order to monitor the changes in composition of the groundwater. The analyses of wastewater composition in the lagoons also included P, to elucidate the source of water pollution.

Results are that changes of P content in surface and groundwater are influenced by the wastewater from the lagoon. Movement of P through soil and groundwater were analysed in a theoretical way, and models are being developed for simulation of spatial and time changes in the concentration of P in groundwater.

Do different sowing and ploughing dates of cover crops have an effect on the phosphorus loss into groundwater?

K.E. Böhm, H. Spiegel and J. Hösch,

Federal Office and Research Centre for Agriculture, Spargelfeldstraße 191, 1226 Vienna


Fallow is one of the greatest contributory factors to nutrient loss. One possibility to minimise the risk of nutrient loss is intercropping. The choice of sowing and ploughing dates is of particular importance. Early sowing with sufficient water supply means that the amount of biomass which develops is high and as a result larger quantities of water will be withdrawn. This effect might not be a problem in areas with high precipitation rates but in semi-arid areas the water will be lacking and influence the growth of the next main crop. Choosing an early sowing date means the amount of percolation will be low, although the concentration of substances in the leakage might be higher.

With an early ploughing date (at the beginning of the winter) the formation of a stable crumb structure due to frost effects is possible. On the other hand the risk of increased nutrient losses might rise too. A ploughing date in spring can cause an inferior emergence of the following crop which will take up less nutrients and finally result in yield depression. Further effects may be higher nutrient losses in later spring.

Catch-crop growing is also one important measure within the framework of the OEPUL (Austrian environment programme). Therefore an assessment of the effects of suitable sowing and ploughing dates and repercussions on nutrient losses is necessary. The results are obtained by lysimeter studies.


Lysimeters have the advantage that they are easier to monitor than field experiments. Results obtained integrate the processes that are normally measured separately in the laboratory. The lysimeter station of the Federal Office and Research Centre of Agriculture in Vienna was designed, in autumn 1995, to study long-term effects of agricultural practice on soil and water budget. Water movement mechanisms and the dynamics and pathways of nutrients can be observed in relation to groundwater quality.

The lysimeter station is located in a zone with a temperate continental climate (cold winters; hot and dry summers). The annual amount of precipitation is about 550-600 mm and the mean annual temperature is 9,5°C. The lysimeter station is stationed 160 m above sea level. The lysimeter station consists of 18 lysimeters, containing three different soil types, a Calcic Chernozem, a Calcaric Phaeozem and a Gleyic Phaeozem with six replicates each.

The containers of the lysimeters are made of stainless steel and have a cylindrical shape. The depth is 2,5 m and the surface area measures 3 m2. Lysimeters drain by gravity into a water reservoir which is emptied at least once a week. Each lysimeter is equipped with a tipping bucket to register the amount of percolation water, suction cups in different depth steps to collect soil solution, tensiometers to measure the water tension and with a TDR system to detect the volumetric water content of the soil. Each lysimeter is equipped with a data logger, in which information on water tension, water content and soil temperatures is registered.

The concept of the recent lysimeter experiment, which started in 1998, is to observe the influence of different ploughing dates of cover crops on the nutrient loss into groundwater. Results should provide practical advice for agriculture and new ways to prevent groundwater contamination.


The effect of cover crops on the phosphorus (P) preservation in top soils and the resulting lower P loss into the ground water will be shown by the examples of three different soil types found in a continental climate. One key question will be the scale of different sowing and ploughing dates as an efficient control mechanism to avoid phosphorus loss into deeper soil layers and consequently into ground water. Furthermore the described effects will be shown in comparison to an experimental year with sufficient precipitation and also with an experimental year without sufficient precipitation. The research results will show the operational area in a region with continental climatic conditions for agriculture and the sustainable protection of ground water quality.

Potentially leachable phosphorus as affected by fertilization in some cultivated Swedish soils

K.Börling1 and E. Otabbong1

1Department of Soil Sciences, Swedish University of Agricultural Sciences, Uppsala, Sweden


The objective was to study potentially leachable phosphorus (P) in different soil types as affected by fertilization, and to establish whether or not leachable P is related with sorption capacity of the soils. Soils were sampled in 6 long-term fertility field experiments in southern and central Sweden (Fjärdingslöv, S. Ugglarp, Klostergården, Bjertorp, Högåsa and Ekebo). Phosphorus had been applied at four different levels including one depletion level without P addition. The soils had pH of 5.8-7.5 and contained 6.3-36.5% clay, 1.1-2.3% organic carbon (C) and 64.9-120.1 mmol kg-1 ammonium oxalate-extractable iron (Fe) plus aluminium (Al).

Sorption isotherms were constructed for soils from the unfertilized plots and P sorption maxima were calculated according to the Langmuir equation (Table 1). Phosphorus sorption maxima were well correlated with ammonium oxalate-extractable Fe plus Al (r2=0.82, p<0,05) and weakly correlated with organic C (r2=0.45, p>0.05). By contrast, P sorption maxima were not correlated with clay contents (r2=0.06, p>0.05).

Since desorption through extraction with CaCl2 has been shown to have good relationship with P in drainage water (Hesketh and Brooks, 2000), this extractant was used for estimating potentially leachable P. Phosphorus was extracted with 0.01 M CaCl2 at a soil:solution ratio of 1:3. Total P concentration (CaCl2-Pt) (Martin et al., 1999) and molybdate reactive P (CaCl2-Pr) were determined in the extracts. Molybdate unreactive P (CaCl2-Pur) was calculated as the difference between CaCl2-Pt and CaCl2-Pr. The ranges of CaCl2-extractable P for the four P levels are presented in μg L-1 extraction solution, in Table 1. CaCl2-Pt and CaCl2-Pr markedly increased in response to increasing fertilization. The increases in CaCl2-Pr were accentuated in soils possessing low P sorption maxima compared with soils possessing high P sorption maxima. The increase of CaCl2-Pur in response to fertilization was small and not consistent. For Fjärdingslöv and S. Ugglarp, CaCl2-Pur even decreased in the highest P level. The percentage CaCl2-Pur to CaCl2-Pt decreased with increasing P levels.

Table 1. P sorption maxima and ranges of CaCl2-extractable P


P sorption maxima




Mmol kg-1 soil

———————— μg L-1 ——————————

































Hesketh, N. and Brookes, P. C. 2000 Development of an indicator for risk of phosphorus leaching. J. Environ. Qual. 29:105-110.

Martin, M., Celi, L. and Barberis, E. 1999. Determination of low concentrations of organic phosphorus in soil solution. Commun. Soil Sci. Plant Anal. 30: 1909-1917.

Indicators of phosphorus leaching loss from soil to water

P.C. Brookes, S. Fortune and N. Hesketh

Rothamsted Experimental Station, Harpenden, Herts., AL5 2JQ, UK

The movement of phosphorus (P) from terrestrial to aquatic environments poses a threat to the quality of surface waters. Since eutrophication may start at P concentrations as small as 20 m g P l-1 water, P losses from soil to water, while insignificant in direct economic terms, may have very serious implications for the environment.

It is widely considered that P leaching from the surface soil in drainage water is usually insignificant as the P is so firmly fixed on the soil surface. Recently published results from the Broadbalk Continuous Wheat Experiment at Rothamsted have challenged this assumption (Heckrath et al., 1995). This experiment, now in its 155th year, has soil P concentrations which range from about 10 to more than 100 mg 0.5 M NaHCO3-extractable (Olsen) P kg-1 soil. At Broadbalk’s inception, field drains were set at 65 cm below the soil surface of each plot so that drainage waters could be collected. These drains were only replaced a few years ago.

Drainage water samples were collected at the drain outlets of plots selected to cover a wide range of P concentrations on five occasions between October 1992 to February 1994 and analysed for P. Full analytical details and results are given by Heckrath et al. (1995). Briefly, below about 60 mg Olsen P kg-1 soil total P concentrations in the drainage waters were small (about 0.15 mg P l-1). However, above about 60 mg Olsen P kg-1 soil there was a rapid linear increase in both Total P (TP) and Molybdate Reactive P (MRP) concentrations in the drainage waters up to the maximum Olsen P concentration of about 100 mg P kg-1 soil. Generally, MRP ranged from about 70 to 90% of TP, and TP could be more than 3 mg P l-1.

The Broadbalk results, we believe, uniquely imply that there is a certain soil P concentration below which there is little P leached in drainage water but that above this soil P concentration (which we term the ‘Change Point’) P leaches to drainage water in proportion to the soil P concentration. We need to know if the ‘Change Point’ is similar in different soils or does it differ depending upon soil type of management? Because of the almost complete lack of other suitable experiments we had to find an indirect method to predict if ‘Change Points’ differed between soils. Other work has indicated that the ratio between Laboratory 0.01 M CaCl2-extractable P and Olsen P might be a useful indicator of ‘Change Points’ under field conditions. We present data to support this hypothesis showing that ‘Change Points’ measured by CaCl2 were readily detectable in other soils and differed widely, ranging from 21 to 104 mg Olsen P kg-1 soil.

There was also, unexpectedly, a very close (r2 = 0.87) negative linear relationship between the ‘Change-Point’ measured in CaCl2 and total soil iron (Fe) concentration. This is exactly the opposite of that which we expected, implying that as soil Fe concentration increases, P is held less strongly, so the ‘Change Point’ decreases. The use of this empirical relationship as an indicator of P leaching losses from soil to water is illustrated and discussed.


Heckrath, G., Brookes, P.C., Poulton, P.R. and Goulding, K.W.T. 1995. Phosphorus leaching from soils containing different phosphorus concentrations in the Broadbalk experiment. J. Environ. Qual. 24: 904-910.

Temporal changes in Olsen phosphorus (P) concentrations when P fertiliser rates were applied to nine different soil types of varying P buffering capacity

L. L. Burkitt1, 2, C. J. P. Gourley1 and P. W. G. Sale2

1 Agriculture Victoria Ellinbank, Ellinbank, Victoria, Australia

2 Department of Agricultural Sciences, LaTrobe University, Bundoora, Victoria, Australia

Australian soils are generally deficient in phosphorus (P) in their natural state and they require the regular application of P fertiliser to satisfy the P requirements of productive plant species. Moreover, these soils have a wide range of P buffering capacities (PBC) and this influences the soil's capacity to buffer changes in plant available P concentration when P is added to or removed from soil (Ozanne, 1980). The availability of P fertiliser can be measured as the slope (D EP) of the relationship between the extractable P concentration (mg

kg-1) and the rate of P fertiliser applied (kg P ha-1). Knowledge of a soil’s D EP value is critical in order to improve the efficiency of P fertiliser use and is a key component of P fertiliser budgeting. Accurate P fertiliser budgets are essential to reduce the risk of P losses from land to water. This study examined the changes in D EP values of 9 different soil types over a 24 month period, when P fertiliser was withheld.

Field studies were established on nine different soil types used for pasture production in south-eastern Australia. Sites were selected to represent a range of PBCO&S (Ozanne and Shaw, 1967) values. Plots received a single application of triple superphosphate or single superphosphate at rates of 0 to 280 kg P ha-1. Soils were analysed for extractable P (Olsen) concentration 6 and 24 months after fertiliser was applied, to monitor the decline in P concentration.

A strong negative relationship was observed between PBCO&S and D EPOlsen (r2 = -0.73) after 6 months, with this relationship weakening 24 months (r2 = -0.46) after P fertiliser was applied. It was not surprising that highly buffered soils had lower D EPOlsen values, or lower increments of plant available (extractable) P per kg of P fertiliser applied after 6 months, and that the decline in D EPOlsen values between 6 and 24 months (Table 1) was also related to PBCO&S values (r2 = -0.59). By definition, PBC is a measure of a soil's ability to maintain a constant soil solution P concentration when P fertiliser is added. However, the negative relationship between D EPOlsen and PBCO&S, despite being somewhat weaker, remained when P fertiliser was withheld over a 24 month period. Weakly buffered soils (soils with PBCO&S values 10 mg kg-1) maintained D EPOlsen values 2 to 3 times higher than soils with PBCO&S values 45 mg kg-1, 24 months after receiving the same quantity of P fertiliser (Table 1).

These results suggest that the increase in extractable P concentration per kg of P fertiliser applied is greater both in the short and medium term, on low P buffered soils. This raises the issue of how long a single P fertiliser application will maintain elevated extractable P concentrations in low P buffered soils, which is important information for developing accurate P fertiliser budgets in these soils.

Table 1. Changes in Olsen phosphorus(P) concentrations (D EPOlsen), six and 24 months after P fertiliser rates were applied to nine different soil types, varying in PBCO&S value.

Soil type


D EPOlsen (x10-3)

at 6 months

D EPOlsen (x10-3)

at 24 months

Decline in D EPOlsen between 6 and 24 months (x10-3)


mg kg-1


Outtrim-loamy sand





Curdie Vale-loamy sand





Athlone-silty loam





Yarragon-silty loam





Strzelecki-silty clay loam










Glenormiston-red clay loam





Ellinbank-red clay loam











Ozanne, P.G. 1980. Phosphate nutrition of plants-a general treatise. p. 559-589. In F.E. Khasawneh et al. (ed.) The role of phosphorus in agriculture. ASA, CSSA, and SSSA, Madison, WI.

Ozanne, P.G., and T.C. Shaw. 1967. Phosphate sorption by soils as a measure of the phosphate requirement for pasture growth. Aust. J. Agric. Res. 18:601-612.

Aluminium chloride as an amendment to reduce soluble phosphorus concentration of land applied swine manure

R. T. Burns, F. R. Walker, D. R. Raman and R. E. Yoder

The University of Tennessee, P.O. Box 1071, Knoxville. Tennessee. USA 37901-1071

The environmental impact of phosphorus (P) in non-point source runoff is of increasing concern to the swine industry. Non-point source P runoff often occurs after swine waste has been land applied as fertilizer. Research in the southeastern United States, has demonstrated that the addition of aluminium sulphate (alum; Al2(SO4)3) reduces the amount of phosphorus in runoff from swine manure by 87%, as well as significantly reducing ammonia emissions. However, there are concerns that the hydrogen sulphide (H2S) emissions from swine facilities would increase as a result of increased sulphur being liberated from the alum used to treat the swine waste. Therefore, aluminium chloride (AlCl3) was tested as an alternative to alum. The objectives of this study were to 1) determine the stoichiometric dosing rate and dose timing of aluminium chloride to bind soluble P in swine manure in laboratory testing and 2) assess the practicality and effectiveness of aluminium chloride as a swine manure amendment under full-scale field conditions. In laboratory studies, soluble P reductions of greater than 97% were observed after treating a mixture of swine manure and supernatant from a swine waste holding pond with aluminium chloride. This soluble phosphorus reduction was obtained using 1.5 times more aluminium than measured soluble P in the swine slurry on a mole basis. Experiments comparing weekly versus daily dosing found no statistical difference between the two treatments. In a field scale experiment, over 565,000 litres of swine manure were treated with approximately 4,150 litres of aluminium chloride (64 % solution) immediately prior to land application. The soluble P concentrations in the land applied manure was observed to be reduced by an average of 65% compared to the untreated manure.

The effects of tillage and reseeding on phosphorus transfers from drained grassland

P. J. Butler and P. M. Haygarth

Institute of Grassland and Environmental Research, North Wyke, Okehampton, EX20 2SB, Devon, UK


There have been many studies of phosphorus (P) transfers from grasslands under steady state conditions (e.g. plot studies by Haygarth et al., 1998; lysimeter studies by Turner et al., 2000) but little is know about the effects of perturbing the land during pasture reseeding. This project aimed to study both the short term and long term effects of tillage and reseeding on both drained and un-drained grassland.


One hectare lysimeter plots, hydrologically isolated from neighbouring plots by gravel filled ditches, with a 30 cm impermeable clay horizon were used in this study at Rowden, North Wyke, Devon. There were two hydrological treatments, undrained plots with composite (surface plus inter-) flow and drained plots, which also had mole drains at 60 cm depth. The soil in the plots was clayey non-calcareous pelostagnogley (USDA typic haplaquepts) described previously in Haygarth et al. (1998). Two drained and two un-drained plots were studied following reseeding, and one drained and one un-drained plot were also studied under permanent (i.e. non–reseeded) grassland. Reseeding occurred in June 1998. Short term soil removal and subsequent P losses were estimated by survey. Longer-term sampling was also undertaken, with the collection of drainage waters from the plots. Samples were analysed for total P (TP) by acid persulfate digestion (Eisenreich et al, 1975).


Short term detachment

Sixteen days after reseeding, 94 mm rain fell, with 57 mm over 22 hours. This resulted in 7.5 tonnes of soil lost from one 1 ha plot, with an estimated load of 3.75 kg P ha-1 in 16 days on a reseeded treatment, compared to normal annual losses of around 0.8 kg ha-1 on non-reseeded treatments.

Long term solubilisation

Table 1. Ranges of total phosphorus (TP) concentration in water discharges following pasture reseed

Composite = Surface and subsurface flow to 30 cm; Drained = Flow from Mole drain to 60 cm


Permanent pasture

Reseed plot

Reseed plot

Undrained plots




Maximum TP m g L-1

Minimum TP m g L-1

Mean TP m g L-1










Drained plots







Maximum TP m g L-1

Minimum TP m g L-1

Mean TP m g L-1



















Table 1 shows that on undrained plots mean and maximum TP concentrations for the reseeded plots are higher than for permanent pasture. The drained plots show a reversal of this trend.

Discussion and conclusions

Short term soil detachment

Rapid P transfer can occur when reseeding is followed by intensive rainfall, particularly on undrained grassland. This may be especially common on grassland soils which often occur on sloping land.

Long term solubilisation

Following sward establishment on undrained soils, concentrations of P transfer are higher on reseeded grassland. Drained soils appear to respond differently to undrained soils; tilling and reseeding soil allows percolation of P into the soil, which has the effect of reducing P transfer dramatically.


Eisenreich, S. J., Bannerman, R.T., and Armstrong, D.E. 1975. A simplified phosphorus analysis technique. Environmental Letters. 9:43-53.

Haygarth, P. M. and Jarvis, S. C., 1999. Transfer of phosphorus from agricultural soils, Advances in Agronomy, 66, 195-249.

Turner, B. L. and Haygarth, P. M. 2000. Phosphorus forms and concentrations in leachate under four grassland soil types, Soil Science Society of America Journal, 64, 1090-1097.


Some molecular aspects of organic phosphorus dynamics in soil

L. Celi, F. Ajmone-Marsan and E. Barberis

University of Turin, DIVAPRA, Chimica Agraria, v. L. da Vinci 44 , 10095 Grugliasco (TO), Italy

Phosphorus losses from agricultural soil to water bodies are mainly related to build-up of soil available P, as a result of continued inputs of fertiliser P and manure in excess to crop requirements. In the evaluation of P loss, besides dissolved P, particulate forms must be considered. These are represented by the amount of P retained by erodible soil particles and are regulated by the dispersion/flocculation behaviour of soil colloids. This, in turn, depends on size and surface properties of particles and can be P-dependent (Barberis and Whithers, 1998).

Applications of manure and slurries can increase considerably soil organic P levels (Sims, 1992). Fresh animal manure contains from 3 to 15 g Kg-1 of organic P and 50-70% of it is represented by inositol phosphates (Sun and Zhang, 1992). In soils organic P can constitute from 20 to 80%, depending on several parameters, such as the area, the type of soil and the profile depth. Unlike inorganic P, which can be readily fixed in the soil, some organic P forms are more mobile and can move along the profile and relatively easily reach waters. In an area of intensive cattle breeding in the Po Valley (Italy) we observed that dissolved organic P accumulated in the subsurface horizons up to 92% of total P while the amount in the particulate form was less than 1%, probably due to the flocculation of colloids by CaCl2, used as extractant. Thus, dissolved organic P can constitute the largest P fraction in subsurface soil solution as observed also by Ron Vaz et al. (1993) and Chardon et al. (1997). However, the build-up of organic P in soils has been reported to be due to sorption of P mono esters, such as inositol phosphates, by soil minerals that hamper their biodegradation. Moreover, P has been shown to be also transported with dispersed soil particles in the percolation water (Hartikainen, 1978).

This work is aimed to evaluate the molecular aspects of organic P interaction with minerals while assessing the sorption extent and its effects on the particle charge and size. Inositol phosphate was used in this study, as it represents the main organic P form in soils. These results could help to better interpret the trend and distribution of P in the different forms in agricultural soils subjected to intensive fertiliser and manure application.

The adsorption extent depended on the nature and properties of the mineral surfaces. Iron oxides retained more inositol phosphates than phyllosilicates, such as kaolinite and illite. All these minerals showed a higher affinity for inositol phosphate than for inorganic phosphate as deduced by the Langmuir K values. In terms of P, a higher amount was retained in the organic form with respect to the inorganic one. The adsorption occurred through the phosphate groups of inositol hexaphosphate similarly to the free orthophosphate ion, forming a binuclear complex while the organic moiety affected the mechanism only in terms of conformational hindrance (Ognalaga et al., 1994). In calcareous soils inositol phosphate retention was governed by CaCO3 with complexation and precipitation of Ca salts at any inositol phosphate concentration. The sorption reaction with ferrihydrite and goethite was found to be nearly irreversible. The adsorption of inositol phosphate caused a net increase of the negative charge and particle dispersion of all minerals under investigation. This was noticeable also at the lowest inositol phosphate addition.

These results lead to two conclusions: a great inositol phosphate-fixing capacity of minerals that may affect its accumulation in soils and P bioavailability, and a considerable change of mineral electrochemical properties and particle size distribution that modify soil particle charge and colloidal stability. This therefore suggests that organic P can greatly react with soil minerals, such as iron oxides, clays, calcite. This results in a rapid removal from solution, but organic P sorption increases the mobility of erodible particles and hence P transfer as particulate through the profile and to water bodies.


Barberis E. and P. Withers. 1999. Influence of soil processes on detachment of P forms: A review of experimental data. COST action on Phosphorus meeting, Cordoba (Spain) May, 13-15 1999.

Chardon W. J., Oenema O., Del Castilho P., Vriesema R., Japenga J. and Blaauw D., 1997. Organic phosphorus in solutions and leachates from soils treated with animals slurries. J. Environ. Qual. 26: 372-378.

Hartikainen, H. 1978 Leaching of plant nutrients from cultivated soils. II Leaching of anions. J. Scient. Agric. Sco. Finl. 50: 263-269.

Ognalaga M, Frossard E and Thomas F (1994) Glucose-1-phosphate and myo-inositol hexaphosphate adsorption mechanisms on goethite. Soil Sci Soc Am J 58: 332-337.

Ron Vaz M. D., A.C. Edwards, C.A. Shand, and M.S. Cresser. 1993. Phosphorus fractions in soil solution: Influence of soil acidity and fertilizer additions. Plant Soil. 148: 179-183.

Sims, J.T. 1992. Environmental management of phosphorus in agriculture and municipal wastes. In: Sikora F.J. (ed). Future directions for agricultural phosphorus research. Natl. Environ. Res. Center, Muscle Shoals.

Sun, X. and Y.S. Zhang. 1992. Study on inositol phosphate in organic manure and paddy soils and its effects on rice growth. Acta Pedol. Sinica 29: 365-369.

Impact of phosphorus-based nutrient management regulations on farm operations: the Maryland (USA) situation

F. J. Coale1 and J. T. Sims2

1University of Maryland, College Park, Maryland, USA.

2University of Delaware, Newark, Delaware, USA


In the spring of 1998, the State of Maryland (USA) enacted new regulations that mandated phosphorus (P)-based nutrient management planning for nearly all of Maryland’s commercial agricultural operations. Conventional soil-test P levels and phosphorus site index (PSI) evaluations were specified as the primary tools to be used for determining the potential for P losses from agricultural land. The objective of the PSI was to identify soils, farm management practices, and specific locations within a farm where P losses in field drainage water may pose the potential for negative environmental impacts on nearby surface waters. At the time the regulations were enacted, a PSI tool had not been evaluated under Maryland conditions.

Development of the Maryland PSI

Assessment of the appropriateness of PSI site characteristics, scale factors, and final P loss rating categorization was made by reference to the published scientific literature and consensus of scientists experienced in agricultural P management and watershed hydrology of the region. In the end, a two-part decision making process was incorporated into the PSI to evaluate the independent effects of the quantity of P present (P source) and the potential for P transport from the field. The final outcome was the categorization of the assessed field into one of four final P loss rating groups: 'low', 'medium', 'high', and 'very high'.

During the final phase of testing, between 25 and 30 farm fields were identified in each of Maryland’s 23 counties that, when taken in aggregate for a given county, reflected the diversity and proportion of soil types, topographies, farming systems, and nutrient management practices prevalent in that county. Data were collected from 646 field sites beginning in the spring of 1999 and continuing through the spring of 2000.

Application of the Maryland PSI

The State of Maryland has defined an 'environmental threshold' for conventional soil-test P equal to 150 FIV (fertility index value), or three-times the agronomic critical level of 50 FIV. According to new Maryland nutrient management regulations, agricultural fields with conventional soil-test P greater than or equal to 150 FIV must be evaluated by the PSI to determine the nutrient management planning requirements for that site. The frequency distribution of PSI outcomes or 'performance' was evaluated for several sub-categories of the statewide data set. The various categories of PSI performance distributions were divided into groups: all data, fields with soil-test P (STP) less than 150 FIV, and fields with STP greater than or equal to 150 FIV (Table 1).

Table 1. Number and percent of evaluated fields in each P loss rating category for the Maryland P Site Index (PSI) evaluation. Soil-test P sub-categories include fields with soil-test P (STP) less than 150 fertility index value (FIV) or greater than or equal to 150 FIV.


P loss rating


Soil-test P


Number of fields




Very High






All fields






STP < 150






STP > 150






Currently available data can not determine if the Maryland PSI is 'right' or 'wrong'. The frequency distribution data presented represent how the Maryland PSI has performed on the sites evaluated. The proportion of evaluated sites that fell within each of the four final P loss rating for the statewide data set or any subdivision of the statewide data set was strictly determined by the numerical definitions of the P loss rating categories. Although the P loss rating categories were numerically defined based on best available data and professional scientific judgements, it may be necessary in the future to redefine the P loss categories as warranted by new research findings and more experience in using the PSI.

Variability in the loss of phosphorus and other elements from dairy pasture in South Australia

N. K. Fleming1 and J. W. Cox2

1South Australian Research and Development Institute, Urrbrae, South Australia

2CSIRO Land and Water, Urrbrae, South Australia


Dairying is an intensive grazing industry and in southern Australia is often located on texture-contrast soils. There was limited information available on the magnitude of nutrient loads from Australian pastures (Young et al. 1996). Nutrient movement from pasture is undesirable as eutrophication of most fresh waters is limited by phosphorus (P) inputs. This report presents the magnitude, forms and seasonal and annual trends in movement of P and other elements from a pastured dairy subcatchment in the Adelaide Hills, South Australia. The climate is Mediterranean, with hot dry summers and cool wet winters. Annual rainfall is 750 mm.

Material and methods

Dissolved (<0.45μm) and particulate (>0.45μm) P, carbon, aluminium, calcium, iron, magnesium, potassium, sodium, sulphur and zinc losses were measured in runoff (overland flow and A/B horizon interflow) from dryland dairy pasture (2.2 and 2.6 ha subcatchments) at Flaxley Agricultural Centre, Adelaide Hills, South Australia from 1996 to 1999. Infrastructure, soils and procedures are described in Fleming and Cox (1998).

Runoff and nutrient loads

Runoff ranged from 0.4 – 10% of annual rainfall. Over 90% of this was overland flow. Losses of P and other elements each year were highly variable and strongly related to rainfall patterns. Over the study period an average of 98% of P was lost in overland flow. The proportion of P present in the particulate form was highest in the wettest year.

Table 1. Annual rainfall, amount and form of P loads from Flaxley East subcatchment


Relative rainfall as a proportion of 30-year average (750 mm)

Total P load

Proportion as dissolved P


g ha-1




2 289














Pattern analysis

From each runoff event there was an associated load of each elemental fraction. Pattern analysis within each year was carried out by making a correlation matrix of all element fractions, using the element load data of all events within that year as described by Fleming and Cox (2001). In the wettest year (1996) almost all fractions separated clearly into one of two groups - Particulate (all particulate fractions) or Dissolved (all dissolved fractions plus chloride and nitrate). As annual rainfall decreased, there was a trend of more fractions being incorporated into the Dissolved group. In the driest year (1997) almost all fractions were associated with the Dissolved group.


The results of pattern analysis appear to reflect the interplay of two major transport processes - Dissolved and Particulate. In wetter years, the particulate process dominates the magnitude of element losses, but becomes an increasingly minor component of element loss as annual rainfall decreases. This appears to be a relatively simple mechanism for losses of a wide range of elements from dairy pasture.


Fleming, N.K., and J.W. Cox. 1998. Chemical losses off dairy catchments located on a texture-contrast soil: Carbon, phosphorus and sulphur and other chemicals. Aust. J. Soil Res. 36:979-95.

Fleming, N.K., J.W. Cox, D.J. Chittleborough, and C.B Dyson. 2001. Prediction of chemical loads and forms in overland flow from dairy pastures on duplex soils in South Australia. Hydrological Processes (in press).

Young, W.J., F.M. Marston, and J.R. Davis. 1996. Nutrient Exports and Land Use in Australian Catchments. J. Env. Manage. 47:165-183

Phosphorus status of soil and leaching losses: results of long-term lysimeter studies

F. Godlinski1, P. Leinweber1 and R. Meissner2

1Institute of Soil Science and Plant Nutrition, University of Rostock, Justus-von-Liebig-Weg 6, 18059 Rostock, Germany

2UFZ Centre for Environmental Research Leipzig-Halle GmbH, Institute of Soil Science, Lysimeter Station Falkenberg, 39615 Falkenberg, Germany


We present results of comprehensive soil and leachate investigations of more than 100 differently managed and long-term operated lysimeters in Germany. The objectives were (1) to compare the most important soil phosphorus (P)-tests used in the European Union (EU) for their extraction efficiency and applicability, (2) to derive relationships between P contents in soils and P-losses, and (3) to investigate how depth profiles of P contents develop in lysimeters due to long-term soil management.

(1) Comparison of soil phosphorus-tests and deduction of conversion factors

The analysis of nearly 300 data sets with three P-extraction methods widely used in the EU revealed significant correlation's between double lactate (DL)-P and Olsen-P, ammonium lactate-acetate acid (AL)-P and Olsen-P, and AL-P and DL-P. Factors for the conversion of P contents were calculated and compared to similar factors published by Sibbesen & Sharpley (1997) and Vanderdeelen (1999). We obtained the following conversion factors: Olsen-P ´ 1.38 = DL-P, Olsen-P ´ 2.75 = AL-P, and DL-P ´ 1.94 = AL-P, which were generally lower as published previously. This is explained by the different soil textures used in the studies. Our above factors are valid for loamy sandy soils. An improved comparison of results can be obtained by the application of regression equations, which will be shown in the poster.

(2) Soil phosphorus contents and phosphorus-leaching losses

Soil P-tests were assessed for their ability to estimate P-leaching losses. The P-concentrations of leachates varied from 0 – 1.2 mg/l with a mean of 0.065 mg/l. The highest values were found in lysimeters with sandy soils. We derived significant correlations between P in leachates and P-contents in topsoils, as determined by the various extraction methods. The AL-P (r = 0.45***, n = 174), oxalate-extractable P (r = 0.24***, n = 202) and degree of phosphorus saturation (DPS) were the best predictors for P concentrations in leachates. Olsen-P and DL-P gave insignificant relationships if the same data set was considered. An extended data set for the years 1991 to 1998 showed a significant correlation between P-concentration and DL-P (r = 0.19***, n = 974). Subdivision of the data sets according to different texture groups improved the correlations in some cases. The P-concentrations of the leachate and the P-losses (0 – 3291 g/ha, mean 78 g/ha) were closely correlated as well (r = 0.86***, n = 1049). In parts these data confirmed results by Leinweber et al. (1999), however, on the basis of a largely expanded data set. Residual variation and insecurity of estimates originate from factors other than P in topsoils, which may influence the P leaching. Probably, one of these factors is the vertical sequence of soil properties in the profiles.

(3) Depth distribution of phosphorus-fractions in selected lysimeters

Highly resolved (10 cm) depth distributions were determined for DL-, AL-, oxalate-extractable P and Olsen-P in 8 lysimeters. Generally, the long term management led to characteristic depth distributions of all these P-test values. The contents were largest at 5-35 cm depth, and decreased towards the subsoils. The decrease in P contents was steeper under arable than under grassland and afforestation. This and the extensive grassland lysimeters had the lowest P-contents in the top 30 cm and at the lysimeter basis (afforestation: 38 and 17 mg kg-1 DL-P, extensive grassland: 88 and 33 mg kg-1 DL-P). Surprisingly, these two land uses showed slight P accumulations directly at the lysimeter bottom. This is probably due to the construction of the lysimeter, where plant roots, and thus plant available P, may accumulate at the bottom. Larger variation in the whole subsoils were observed in the lysimeters with arable use. This was determined with all four P-extraction methods in good agreement. New evidence on P distribution in lysimeters subsoil will be used to an improved explanation of P leaching losses.


Leinweber, P., Meissner, R., Eckhardt, K.-U. and Seeger, J. 1999. Management effects on forms of phosphorus in soil and leaching losses. Eur. J. Soil Sci., 50, 413-424.

Sibbesen, E., and Sharpley, A.N. 1997. Setting and Justifying Upper Critical Limits for Phosphorus in Soils. p. 151-176. Tab. 7.2 In H. Tunney, O.T. Carton, P.C. Brookes and A.E. Johnston (ed.) Phosphorus Loss from Soil to Water. CAB international. Wallingford. UK

Vanderdeelen, J. 1999. Meeting of COST 832. WG 2. Cordoba. Spain. 13-15 May

Cattle grazing density impacts on soil phosphorus concentrations and runoff water quality

D. A. Graetz1, J .C. Capece2, K. L. Campbell1, K. M. Portier1, and P. J. Bohlen3

1University of Florida, Gainesville, Florida, USA

2Southern DataStream, Inc., LaBelle, Florida, USA

3Archbold Biological Station, Lake Placid, Florida, USA

Beef cattle ranches are the most extensive land use in the Lake Okeechobee watershed, the main water supply for south Florida. Due to a combination of wetland loss, improved drainage, and increased area for livestock production, phosphorus (P) loadings into the lake have increased substantially, leading to increased frequency of algal blooms and decreasing water quality. The University of Florida Institute of Food and Agricultural Sciences, Archbold Biological Station, South Florida Water Management District and the Florida Cattlemen's Association have initiated a project to develop sustainable environmentally sensitive cattle ranch management practices for beef cattle in central and south Florida.

The study site is located on the MacArthur Agro-Ecology Research Center, a 4,170 ha working cattle ranch within the Lake Okeechobee watershed in southern Florida. Sixteen hydrologically separated pastures, eight each of improved summer pasture and winter native wet prairie, have been constructed. The pastures are instrumented to continuously monitor volume and chemistry of surface water drainage. In the summer (May-Oct.), cattle are stocked on improved pastures that are planted with bahiagrass (Paspalum notatum). In the winter (Nov.-April) cattle are moved to wet prairies, which are dominated by native grasses (Andropogon virginicus, andropogon glomeratus, Paspalum laeve, Axonopus affinis) and are generally not fertilized. Summer and winter pastures are 20 and 32 ha, respectively and are stocked at four animal densities. Stocking densities differ between summer and winter pastures due to the carrying capacity of the two pasture types. Summer pastures were stocked at rates of 0, 0.74, 0.99 and 1.73 cow/calf pairs per ha. Stocking rates for the winter pastures were 0, 0.46, 0.62, and 1.08 cow/calf pairs per ha.

Soils were sampled to a depth of 30 cm in increments of 0-5, 5-10, 10-20, and 20-30 cm on a yearly basis. Each summer and winter pasture treatment was divided into five equal components. Two sampling locations were randomly selected within each of the components and marked (GPS) for future sampling. Soils were analyzed for water-soluble P (WSP). Pretreatment WSP concentrations for the summer pastures averaged over all stocking rates were 41, 5.6, 1.4 and 0.7 mg kg-1 for the 0-5, 5-10, 10-20, and 20-30 cm depth increments, respectively. Corresponding values for the winter pastures were 19, 4.5, 1.0 and 0.3 mg kg-1. The higher WSP concentrations, especially at the soil surface (0-5 cm), reflect the greater intensity of management, i.e., P fertilization, in the summer pastures relative to the winter pastures. WSP concentrations during the first treatment year did not change significantly from the pretreatment values.

Incremental runoff samples were collected by automated samples during each runoff event. Total P showed the most dramatic differences in water quality between the summer and winter pastures, with summer pastures exceeding winter pasture concentrations by a factor of five or more. Winter pasture TP concentrations were in the range of 0.10 mg L-1 while the summer pastures TP concentrations were in the range of 0.50 mg L-1. Nutrient loads were calculated by multiplying incremental runoff volumes by the nutrient concentration corresponding to the end of the runoff increment. Total P loads from the winter pastures were approximately 0.15 kg ha-1 compared to 0.75 kg ha-1 for the summer pastures. The higher TP concentration in the runoff was matched by the correspondly higher WSP concentrations in the surface soil of the summer pastures. Initial statistical analysis of both concentration and load results show only a difference between summer and winter pastures but not between cattle stocking rate treatments. These data represent only the first year of treatment implementation and we anticipate seeing differences between treatments as the study progresses.

Gradual desorption by water and calcium chloride solution in assessing potential release of soil phosphorus

Helinä Hartikainen

Department of Applied Chemistry and Microbiology, P.O. Box 27, 00014 University of Helsinki, Finland

Restriction of phosphorus (P) losses in agriculture requires identification of the fields of the highest risk. However, despite a lot of effort, no consensus on proper methodology has been obtained. There are indications that a risk assessment based on conventional chemical tests may lead to ambiguous conclusions (Sharpley, 1995; Hooda et al., 2000). One disadvantage of the chemical tests is that they give an estimate only for the P intensity or quantity.

Water extraction, that predominantly measures P intensity has been proposed also as an indicator for P loading risk (Yli-Halla et al., 1995; Gächter et al., 1998). Dilute CaCl2 solution has been extensively used for determination of P intensity as well. It is also a common background solution in sorption tests used to describe the P buffering properties of soils. Even if both sorption and desorption control the P buffering, they do not follow the same pathway but more or less hysteris can be found. Therefore, in risk assessment a desorption curve would be a more relevant tool to integrate information about factors contributing to the potential P release from soil than a sorption curve. Desorption obtained at progressively increasing solution to soil ratio (SSR) would simulate the P buffering reactions during surface runoff. A narrow ratio approaches the field conditions and wider ratios correspond to conditions prevailing during erosion process. Because the P desorption is dependent on the bathing solution, this study was undertaken to compare the sensitivity of water and dilute CaCl2 solution in reflecting fertilisation-induced changes in P intensity and quantity at various extraction ratios.

Materials and methods

Soil samples from four P fertilisation field trials were taken from blocks cultivated for seven years without P addition or amended annually with 30 or 60 kg of P per ha. Inorganic P reserves were analysed according to a modified Chang and Jackson's procedure. Degree of P saturation (DPS) was calculated as a molar ratio of Al and Fe bound P to oxalate soluble Al and Fe. For desorption curves, soil samples were extracted with water (Pw) or 0.01 M CaCl2 at solution to soil ratios of 2:1, 5:1, 10:1, 50:1, 150:1 and 400:1. For comparison, P status of the soils was investigated using Bray P1 test and an acid NH4-acetate method (AAAc) used in routine soil testing in Finland.

Results and discussion

Phosphorus fertilisation increased Al and Fe bound P and DPS, excluding one soil of very high DPS where added P had obviously leached to lower horizons. The P concentrations in CaCl2 extracts were much lower and less affected by SSR than in water extracts. Consistently with P fractions, Pw in a given soil rose with increasing accumulation of P, whereas CaCl2-P showed differences between the P treatments only in soils of high DPS. As expected, increasing SSR decreased the P concentration in solution. However, this hold truer for Pw than for CaCl2–P which was quite weakly dependent on the extraction ratio. CaCl2 extraction was also rather insensitive in reflecting fertilisation-induced changes in P reserves, whereas water extraction responded even to the presence of competing ligands (organic matter) contributing to P desorption tendency. The salt effect and the divalent cation reduced the sensitivity of CaCl2 solution in the determination of potential P release, especially at narrow SSRs. Also a single water extraction at conventional solution to soil ratio of about 50:1 may give an inaccurate estimate for P release potential. However, at the ratio of 400:1 water extracted P more than AAAc, and in some soils even to the same extent as Bray P1. The results indicate that a desorption curve with water at SSRs varying from very narrow to large might be a promising method to describe the P supplying power of soils and to simulate the P release from soil particles during surface runoff.


Gächter, R., J.M. Ngatiah, and C. Stamm. 1998. Transport of phosphate from soil to surface waters by preferential flow. Environ. Sci. Technol. 32:1865-1869.

Hooda, P.S., A.R. Rendell, A.C. Edwards, P.J.A. Withers, M.N. Aitken, and V.W. Truesdale. 2000. Relating soil phosphorus indices to potential phosphorus release to water. J. Environm. Qual. 29:1166-1171.

Sharpley, A.N. 1995. Dependence of runoff phosphorus on extractable soil phosphorus. J. Environm. Qual. 24:920-926.

Yli-Halla, M., H. Hartikainen, P. Ekholm, E. Turtola, M. Puustinen, and K. Kallio. 1995. Assessment of soluble phosphorus load in surface runoff by soil analyses. Agric. Ecosystems & Environm. 56:53-62.

Impact of farm management on phosphorus loss after pig slurry application to a drained clay soil

R. A. Hodgkinson1, B. J. Chambers1, P. J. A. Withers2, J. R. Williams3and R. Cross3

1ADAS Gleadthorpe, Meden Vale, Mansfield, Nottinghamshire, NG20 9PF, UK

2ADAS Bridgets, Martyr Worthy, Winchester, Hampshire, SO21 1AP, UK

3ADAS Boxworth, Battlegate Rd. Boxworth, Cambridgeshire, CB3 8NN


The loss of soluble and particulate phosphorus (P) from agricultural land receiving organic manures is a potential contributor to the eutrophication of inland waters. Losses arise as a result of both the accumulation of phosphorus in the soil on livestock holdings and following the application of manures to the soil surface. Research has shown that application of liquid manures such as pig slurry represents the main threat to water quality on drained clay soils (Hodgkinson et al., 1998). This paper reports the preliminary (first year) results of a project to study the interaction between P losses following the application of pig slurry and farm management practices.


Each of nine treatments was replicated 3 times using 27 hydrologically isolated plots at ADAS Boxworth. Pig slurry was applied in autumn (August-September), winter (November-January) and spring (February-March) to give a target rate of 60 kg ha-1 of P. The treatments tested are shown in table 1.

Table 1. Treatments

Autumn cultivations

Spring cultivations

Control (Farm N Nil P)

Control (Farm N Nil P)

Autumn slurry ploughed down

Autumn slurry top-dressed to stubble

Autumn slurry disc cultivated

Winter slurry top-dressed to stubble

Winter slurry top-dressed to growing crop

Spring slurry top-dressed to stubble

Spring slurry top-dressed to growing crop


Tile drain discharges were continuously monitored and water samples collected during runoff events using automated water samplers. Samples were analysed for total P, total dissolved P and molybdate reactive P.

Results and discussion

Data will be presented on the impact of the time of application, and method of incorporation, on P concentrations and loads observed in tile drain discharge. Cognisance will be given to the importance of soil moisture status at the time of application. The results will be discussed in light of any implications for current best management practices (MAFF, 1991) and provisional recommendations will be made as to how to minimise P loss following slurry applications to drained clay soils.


Hodgkinson R. A., Williams J.R. and Chambers B.J. 1998. Phosphorus loss from manure application to a drained clay soil In: Foy R and Dils R [eds] Practical and innovative measures for the control of agricultural phosphorus losses to water Proceedings of an OECD Sponsored Workshop.

MAFF. 1998. Code of Good Agricultural Practice for the Protection of Water. MAFF Publications, London (PB0585).

Getting to grips with phosphorus

M. Hoffmann1, K.A. Ivarsson2 and I. Rydberg3

1Federation of Swedish Farmers, S-105 33 Stockholm, Sweden

2Swedish Farmers Supply and Crop Marketing Association, Box 30192, S-104 25 Stockholm, Sweden

3Swedish Board of Agriculture, S-751 86 Uppsala, Sweden

The Swedish parliament has decided upon 15 new environmental goals. They cover all parts of society and are to be reached within one generation. One of these goals focuses on eutrophication and nutrient losses from agriculture. For this purpose a new campaign named 'Nutrients in focus' has been initiated. The aim of the project is to reduce nitrogen (N) leaching, ammonia emissions and phosphorus (P) losses by surface run-off and preferential flow.

The unique approach consists of a close collaboration between plant and animal advisors from many different agricultural organisations. The measures focus on less tillage, cultivation of catch crops, no autumn application of slurry, creating wetlands, more accurate protein level in cow feed and a higher P efficiency in pig feed. These measures are also co-ordinated with other initiatives among the farmers´ organizations.

Crop production according to Swedish Seal of Quality

Swedish Seal of Quality is a concept for environmentally sound and high quality assured crop production. In 2000 Swedish Seal consisted of around 35 000 ha of a total arable area of over 85 000 ha on 620 farms. These farms produced 240 000 tonnes for Swedish Seal. Crop production according to this concept consists of the following plant nutrition rules:

Nitrogen: A yearly field balance for input/output of N and goals depending on production of the individual farms.

Phosphorus and potassium: A farm gate balance every third year according to STANK (the Swedish Board of Agriculture’s input/output model) and goals related to soil map values. The arable soils of southern Sweden have received a P surplus of up to 1100 kg per ha since the Second World War.

Nitrogen efficiency: The Swedish Seal goal for the farm gate balance is a N efficiency of 80 % of input on cereal farms and 60 % on the most efficient livestock farms but farms with milk production usually have a balance of around 30 %. In 1999 the N efficiencies on the field level for different crops were: winter wheat 69 %, spring wheat 66 %, spring malting barley 74 %, oats 72 % and rye 72 % of input.

Phosphorus and potassium efficiency: The Swedish Seal goal for P is an optimum use of this finite resource and on-farm balance considering both soil map values and environmental risks. The use of potassium is also a resource problem, especially on sandy soils with high precipitation.

The recent farm gate values for the efficiency in % of input are:

Phosphorus: 103 % of input, standard deviation 31, min 21 % and max 278 %

Potassium: 101 % of input, standard deviation 41, min 30 % and max 377 %

Feed phosphorus efficiency and balanced animal density

One part of the ongoing programme (from 1988) to reduce plant nutrient losses, is legislation on livestock density (allowed number of animals per hectare). The limiting factor by Swedish legislation is the amount of P in manure, with a maximum amount of approximately 22 kg P ha-1 (50 kg P2O5).

A balance between the amount of animals on the farm (and thus the amount of manure) and the amount of land available for spreading manure is important. The manure produced should balance requirements for removal of plant nutrients and normal crop production. One advantage with using phosphorus instead of nitrogen as limiting factor, is that estimates of P content in manure are more reliable, since losses are low. P is also easily monitored through nutrient balances and there will be surplus of P long before there is a surplus of N. Additional nitrogen should then be supplied through mineral fertiliser.

Due to intensified production, pig-units usually exceed this limit of 22 kg P ha-1. There are two different solutions to achieve the intended limit. Either the amount of animals permitted is lowered which would mean reducing the approved herd size for pigs by 33%, or the P-content of the manure is reduced by better P efficiency. To decrease both nitrogen and P contents in manure by better feeding strategies is an interesting challenge and this will hopefully reduce the influence on the environment from manure.

Managed riparian buffers to control phosphorus runoff losses from maize fields

W. E. Jokela1, J. W. Hughes1, D. Tobi1, and D. W. Meals2

1University of Vermont, Burlington, VT, USA

2New England Interstate Water Pollution Control Commission, Wilmington, MA USA


The use of vegetative riparian buffers has been encouraged, and even mandated in some cases, as a means of controlling nutrient runoff from cropped fields. Sediment and total P losses have been reduced by buffer strips of grass or woody vegetation of varying widths in the USA and Europe (Dillaha et al., 1989; Clausen et al., 2000; Uusi-Kämppä, 2000). We established a paired watershed field study to compare the effectiveness of field buffer strips of grass-legume hay (7.5 and 15 m widths) against a control treatment of maize planted to the stream edge.


We employed a paired watershed approach using experimental units ('watersheds') much larger than those used in most plot studies. The basic approach requires a minimum of two similar watersheds - control and treatment - and two periods of study - calibration and treatment (Clausen and Spooner, 1993). The control watershed serves as a check for weather or other year-to-year variations. During calibration, the two watersheds are treated identically and data are collected to develop a regression equation between the two watersheds to account for inherent differences between the two. During the treatment period only the treatment watershed receives the treatment while the control watershed remains as before. A separate regression equation is developed during this post-treatment period and the calibration and treatment regression equations are tested for differences. This technique has historically been used to compare runoff from forested watersheds, but it has seen only limited use to evaluate agronomic practices (Clausen et al., 1996).

The field site was a 6-ha field on a dairy farm in NW Vermont, USA, bordered on two sides by a small stream. The soil was a Vergennes clay (Glossaquic Hapludalf) with slopes of 2 to 10% and moderate to high erosion and runoff potential. In our study, the 3 'watersheds' consisted of portions of a field, ranging in size from 1 to 2 ha, each of which has separate surface drainage and an outlet instrumented with an H-flume with a stilling well and Coshocton wheel proportional sampling device to monitor surface runoff. During the 20-month calibration period maize was planted up to the flumes at the field edge in all three sub-watersheds. Data were collected on an event basis to develop regression equations for each variable (runoff, sediment, and P and N concentrations) between the control watershed and each of the watersheds to be treated. At the start of the treatment period (late May, 1998) 7.5-m and 15-m buffers of mixed Fescue-ryegrass-red clover were established on two of the sub-watersheds, while the control watershed remained in maize to the field edge. Monitoring continued for the 2-year treatment period.


Total P concentrations ranged from <1 to 20 mg L-1 during the study and showed no consistent pattern during the calibration period. However, beginning in late June, 1998, when buffer vegetation had become established, concentration values were clustered by treatment – the 15-m buffer the lowest and the control (no vegetative buffer) the highest (a statistically significant difference). The total mass of P per runoff event showed a similar pattern but three events in summer of 1998 were much greater than all others. Cumulative TP mass over the monitoring period reflected two changes as the treatment period began: a) loss of P in runoff increased dramatically in all watersheds, a result of a series of unusually large precipitation and runoff events starting in late June and b) an effect of buffer treatment with greater reduction in P loss from the 15-m than the 7.5-m buffer strip. Results for total suspended solids show very similar trends, reflecting the association of phosphorus with the sediment fraction of the runoff.


Clausen, J.C., K. Guillard, C.M. Sigmund, and K. Martin Dors. 2000. Water quality changes from riparian buffer restoration in Connecticut. J. Environ. Qual. 29:1751-1761.

Clausen, J.C., W.E. Jokela, F.I. Potter III, and J.W. Williams. 1996. Paired watershed comparison of tillage effects on runoff, sediment, and pesticide. J. Environ. Qual. 25:1000-1007.

Clausen J. C. and J. Spooner. 1993. Paired watershed study design. US Environmental Protection Agency. EPA 841-F-93-009. Office of Water. Washington D. C.

Dillaha, T.A, R.B. Reneau, S. Mostaghimi, and D. Lee. 1989. Vegetative filter strips for agricultural nonpoint source pollution control. Trans.ASAE 32:513-519.


Reducing phosphorus leaching from agricultural soils by phytoremediation

G. F. Koopmans, R. A. A. Suurs, P. A. I. Ehlert, W. J. Chardon and O. Oenema

Alterra, Wageningen University and Research Center, Wageningen, the Netherlands


In the Netherlands, soils have been excessively enriched with phosphorus (P) in areas with intensive livestock farming. These soils have become a threat for surface waters, as (subsurface) leaching of P contributes to eutrophication. In some watersheds in the Netherlands, agricultural contribution to the total P load can be as high as 90%. Evidently, to reduce P losses, measures are needed regionally. Remediation of P leaky soils may be achieved by phytoremediation (Chardon et al., 1996). Due to plant uptake of P and harvesting of the crop, P is withdrawn from soil. If P removed from soil is not replenished, the (readily) available P content decreases rapidly, although the stable P content decreases more slowly. We characterized P extractability in soils in increasing stages of P depletion by P plant uptake using various methods, and evaluated whether plant uptake reduced the soil potential to release dissolved P.

Materials and methods

A loamy sand soil, enriched with P in the past, was depleted in a pot experiment by cropping grass during 32 months in a greenhouse on either a 5 or 10 cm soil layer. At 8 times, 2 pots of both treatments (5 and 10 cm) were sacrificed to sample the soil. Grass was harvested every 3 to 6 weeks, and total N and P contents were determined. After each harvest, soil was fertilized (N and K). The following soil extractions were used: water (P-1:2, 1:2 w:v; Pw, 1:60 v:v) and 0.01 M CaCl2 (P-CaCl2, 1:10 w:v), FeO-impregnated filter paper (Pi) and ammonium oxalate (P-ox). The phosphorus saturation degree (PSD) was calculated by: P-ox/a *[Fe+Al]-ox with P-ox and [Fe+Al]-ox in mmol kg-1 of soil and a =0.5.


Uptake of P by the above ground plant parts of grass removed a fair amount of P from soil; 329 and 229 mg P kg-1 of soil (37 and 26% of the initial P-ox content) was withdrawn from the 5 and 10 cm treatments, respectively. However, P uptake showed a large decrease in time, although dry matter production remained high. Total P contents decreased from 6.8 and 7.6 g P kg-1 of dry matter to 0.8 and 1.1 g P kg-1 in the 5 and 10 cm treatments, respectively. Removal of P by plant uptake caused a large decrease of water, CaCl2 and Pi extractable P in 32 months (Table 1). Decrease was largest in the 5 cm treatment, and ranged from 83 to 93%. In the 10 cm treatment, decrease ranged from 73 to 91%. On the other hand, the P-ox content decreased only by 48 and 32% in the 5 and 10 cm treatments, respectively. Uptake of P by the above ground plant parts amounted to 77 and 80% of the decrease in P-ox. After 32 months, P in the zod and root system represented 12 and 13% of the P-ox decrease.

Table 1. Results of soil analyses (mg P kg-1, average values of duplicate pots and remaining percentage).








initial soil

0-5 cm

0-10 cm

4.84 (100)

0.34 (7)

0.77 (16)

8.36 (100)

0.69 (8)

0.73 (9)

25.5 (100)

3.8 (15)

6.9 (27)

49.2 (100)

8.4 (17)

13.3 (27)

896 (100)

468 (52)

611 (68)

0.84 (100)

0.46 (55)

0.57 (67)

Discussion and conclusions

Our results suggest that the more labile P forms are readily removed from soil. Water, CaCl2 and Pi extractable P represent these labile P forms, and are related to the potential of a soil to release dissolved P to the environment. Risk of P leaching from P rich soils may, therefore, decrease rapidly when P is removed from soil by plant uptake. The more stable P, which is included in P-ox but not in water, CaCl2 and Pi extractable P, and which is the largest part of P in soil, is removed to a lesser extent. Therefore, risk of P loss via runoff of particulate P may still be high, because P-ox decreases slowly. The potential of P rich soils to release dissolved P may increase again after phytoremediation has stopped, due to buffering of the more labile P compounds by the relatively large amount of P-ox. Mineralization of P in the zod and root system may further increase risk of P leaching. In that case, continuous cropping and zero P application may be required to maintain a low environmental risk. Field testing of phytoremediation is necessary.


Chardon, W.J., Oenema, O., Schoumans, O.F., Boers, P.C.M., Fraters, B., and Geelen, Y.C.W.M. 1996 Exploration of options for management and restoration of phosphorus leaking agricultural soils (in Dutch). The Netherlands Integrated Soil Research Programme, Wageningen, vol. 8.

Diffuse phosphorus pollution from re-wetted peat soils

P. Leinweber1, A. Schlichting1,R. Meissner2, H. Rupp2, S. Robinson3, E. Otabbong4, A. Sapek5, I. Litaor6, and M. Shenker7

1University of Rostock, Faculty of Agricultural and Environmental Sciences, Institute of Soil Science and Plant Nutrition, Rostock, Germany.

2UFZ Centre of Environmental Research, Leipzig-Halle GmbH, Department of Soil Science, Falkenberg, Germany.

3The University of Reading, Department of Soil Science, Reading, UK.

4Swedish University of Agricultural Sciences, Division of Plant Nutrition and Soil Fertility, Department of Soil Sciences, Uppsala, Sweden

5Institute of Land Reclamation and Grassland Farming, Falenty, Poland

6Tel-Hai Academic College, Department of Biotechnology and Environmental Sciences, Upper Galilee, Israel.

7The Hebrew University of Jerusalem, Faculty of Agricultural Food & Environmental Quality Sciences, Department of Soil & Water Sciences, Jerusalem, Israel


It is hypothesised that re-wetting and restoration of peatlands (synonym: Histosols) may be accompanied by a mobilisation of phosphorus (P), its transport to adjoining aquatic ecosystems leading to an accelerated eutrophication and deterioration of water quality. This presentation gives an overview on the PROWATER-project (Program for the prevention of diffuse water pollution with P from degraded and re-wetted peat soils). This project is integrated and financed within the key action 'Sustainable management and quality of water' of the program 'Energy, Environment and Sustainable Development'. The consortium of PROWATER consists of 10 partners from 5 countries who operate 13 experimental sites.

Description of work and objectives

PROWATER is organised in four connected workpackages (WPs). In WP 1 the study areas in the Hula Valley (Israel), the Droemling Experimental Catchment and the Trebel Valley (Germany), the Biebrza Valley (Poland), Kristianstad (Sweden) and in the Somerset Levels and Moors (U.K.) were characterised. These study areas represent fenlands of different genesis and use, soil, hydrological and climatic conditions and socio-economic problems, which will be briefly described and illustrated in the presentation. Starting in spring 1999, experimental sites were equipped with automated field stations, and a co-ordinated field research (ground water table, soil moisture, redox, temperature) and sampling (soil solution, ground and surface water) programme was initiated. We will report major characteristics of the sites and the resulting differences in the data obtained from continuous field measurements, highlighting the redox potential as a major factor of P-mobilisation. One-year data sets on P concentration in soil solution indicate the different degree of P mobilisation and P losses in Histosols of various degradation, land use and re-wetting status.

In WP 2, detailed investigations of the P status at the experimental sites are carried out by state-of-the-art analytical methods (ad-/desorption isotherms, P forms in sequential fractionation, 31P nuclear magnetic resonance and mass spectrometry). Furthermore, relationships between external factors (e.g. redox, temperature) and P mobilisation will be quantified in microcosm experiments under controlled conditions. Selected examples will be given, pointing to the detailed contributions of project partners. In WP3, the quantified relationships between P status, external factors and P mobilisation shall improve conceptual models of P transformations in Histosols. On this basis, existing numerical models of P mobilisation and transport will be tested, validated and adapted. The expertise with different model approaches will be combined to obtain a synergistic effect by the application of advanced modelling to all of the study areas. This could allow to simulate reliable scenarios of re-wetting strategies and their effects on diffuse P transfer from Histosols to freshwater. In WP 4 a Decision Support System for Histosol restoration and sustaining of good water quality will be developed. This requires the integration of scenarios from WP 3 with the ecological, political, and socio-economic targets in study areas, participating countries and the EU. PROWATER will finish with the application and evaluation of the Decision support system in case studies of the consequences of re-wetting measures for end-users and stakeholders, who are represented by three Assistant Contractors.

The presentation gives a general introduction to the project and presents recent results on soil P chemistry, dynamic changes of soil conditions and P leaching in Histosols. An integration and better understanding of selected individual contributions of project partners shall be achieved.

The geochemistry of phosphorus in the histosols of the Hula valley, Israel

Litaor, M.I.1, Reichmann, O.2, Nishri, A.3, Auerswald, K.4, Shenker, M.2

1Department of Biotechnology and Environmental Sciences, Tel-Hai College, Upper Galilee, 12210

2Department of Soil Science, Hebrew University of Jerusalem, Rehovot

3Kinneret Limnological Laboratory, POB 345 Tiberias

4Department of Grassland Science, Techn. Univ. Muenchen, Am Hochanger 1, D-85350 Freising-Weihenstephan, Germany

The drainage of Lake Hula and surrounding swamps in the late 1950s resulted in numerous environmental problems, including significant land subsidence and re-flooding of the subsided areas. To mitigate some of these problems, the regional groundwater level was raised, and a shallow lake, named Lake Agmon, was created. The elevated groundwater resulted in highly anoxic conditions (-200 mV) in the formally aerobic pedosphere. The anaerobic conditions yielded elevated concentrations of dissolved ferrous (x = 44, SD = 27 mg L-1, n = 31) and variable amounts of dissolved phosphorus (P) (x = 410 SD = 320 m g L-1, n = 31) across the study site. This could lead to an increase in P flux from the pedosphere to the drainage system, depending on the geochemical attributes of the peat horizons and calcareous layers that typified the area around Lake Agmon. Sequential extraction experiments have shown that the highest amount of P was associated with NaOH and 1M HCl extracts, which implied that P is mainly associated with Fe-oxides and Ca-P solid phases. The sorption experiments indicated that the oxidized peat horizons have high sorption capacity, probably due to their iron oxide mineral assemblages, whereas the calcareous layers are characterized by low to moderate sorption capacity. These results suggest that even under reducing conditions, peat horizons may still serve as a sink for P, thus limiting the leaching of P. On the other hand, calcareous layers that exhibit relative high equilibrium P concentration (EPC0) may allow P leaching, especially if the flowing interstitial water contains P concentrations lower than the EPC0. Although the anaerobic conditions yielded a considerable amount of ferrous and dissolved P, there was no spatial correlation between ferrous and P across the geochemical transect. We attribute the lack of spatial correlation to a myriad of processes that affect P solubility, such as adsorption-desorption, co-precipitation, and chemical equilibria of several chemical systems (Ca-P, Fe-P, and Fe-S-P).


Investigation of the mechanisms of phosphorus movement using a triaxial soil slope

C. A. Llewellyn 1, J. N. Quinton 1, A. J. Reynolds 2 and T. G. Myers 3

1 Institute of Water and Environment, Cranfield University at Silsoe, Bedfordshire, MK45 4DT

2 Institute of Agri-Technology, Cranfield University at Silsoe, Bedfordshire, MK45 4DT

3 Applied Mathematics and Computing Group, Cranfield University, Bedfordshire, MK43 0AL

Research has taught us that phosphorus (P), is the limiting element in the eutrophication of fresh waters and as discharge regulations have changed in recent years, the contribution of P from diffuse sources has become increasingly apparent. Either naturally occurring or applied in artificial forms, P is largely immobile in the soil and is readily adsorbed to clay. Identification of environmental conditions under which clay particles move to surface waters is key to the improvement of land management practices and surface water quality, however, the selective nature of the erosion processes controlling particle detachment, transport and deposition are only poorly described. In particular, we know little about the effects of changes in slope on the transport and deposition of sediment. This has partly been due to the lack of equipment available for conducting experiments under laboratory conditions.

To address this situation, a unique soil erosion simulation facility has been constructed within the soil erosion laboratory at Cranfield University. Designed and constructed in-house, as part of a Biotechnology and Biological Sciences Research Council funded project (63/MAF12260), the facility is hydraulically operated and offers a working soil surface 6 m long, 3 m wide and 0.3 m deep. The working surface is comprised of three linked sections, each measuring 2 m long and 3 m wide, which enable the formation of complex slopes, and the hinged sub-frame facilitates the simulation of converging flow. Reverse osmosis grade water may be applied to the facility both as run-on and as rainfall. Water, transported solutes and particulate matter can be collected as runoff, sub-surface flow and drainage. The facility also boasts water table control.

Initial experiments focuses on the effect of slopes, simple and complex on the selective detachment, transport and deposition of soil particles and aggregates, on a Loamy Sand soil and will provide information to test a two dimensional mechanistic model being developed in the project.

The suitability of soil test phosphorus for prioritizing fields receiving manure

J. A. Lory1, P. C. Scharf1, B C. Joern2 and D. H. Pote3

1Department of Agronomy, 210 Waters Hall, University of Missouri, Columbia, 65211

2Department of Agronomy, Purdue University, West Lafayette, IN

3 Dale Bumpers Small Farms Res. Center, 6883 South State Hwy. 23, Boonville AR 72927-9214

Soil testing has traditionally focused on identifying fields where phosphorus (P) is a limiting nutrient for crop production. Recent concerns about the effects of runoff P on water quality have led to the successful use of soil test P (STP) to identify fields with high potential for P runoff. This has spurred interest in using STP thresholds as the sole or partial criteria for limiting manure applications. We reviewed relationships reported in the literature between P applications (manure or fertilizer), STP, and runoff P to evaluate the hypothesis that moving manure applications from fields with high STP to soils with lower STP will improve water quality (reduce P in runoff) within a watershed. Our review suggests that such strategies may not be effective for soils containing up to 400 mg kg-1 Bray-I or Mehlich III P, except on some very low testing soils. This result emphasizes the importance of runoff potential as the core component for making manure application decisions on these soils. Additional research is needed at STP levels above 400 mg kg-1 Bray-I or Mehlich-III P and to evaluate the effects of STP on the equilibration rate of applied P with the soil.

The influence of soil physical and chemical factors on phosphorus release in overland flow

N. J. Miller and J. N. Quinton

Institute of Water and Environment, Cranfield University at Silsoe, Bedfordshire, MK45 4DT

Overland flow is the movement of water over the soil surface and down slope. It can lead to problems with soil erosion and sediment transport. The loss of phosphorus (P) associated with overland flow is often considered to be greater than that likely to occur via leaching because of the ability of soil to sorb P from infiltrating water. Overland flow transports both dissolved and particulate P (PP) but it is the PP associated with eroded soil particles which accounts for the majority of the P transported.

The efficiency of this pathway for P transport is dependent upon many soil and rainfall characteristics. Numerous studies have focused on the role individual soil physical and chemical properties play in influencing P release to overland flow. However, few studies have looked at how these soil physical and chemical factors interact together and the role they play in influencing P release. This study will investigate these interactions and the importance of both soil physical and chemical characteristics with the aim of developing a new procedure for P risk assessment. With this is mind the 9 metre indoor gravity fed rainfall simulator at Cranfield University, Silsoe was used to compare the potential for P transfer in overland flow from 24 soils under different land management practices and agricultural P inputs. The simulated rainfall rate was 60 mm hr-1 and water and sediment samples were collected at intervals for 30 minutes. Samples were analysed for total P (unfiltered), total P (<0.45 m m) and reactive P (<0.45 m m). The amount of P in samples was then related to soil physical and chemical characteristics.


Development and validation of a phosphorus index for pastures

P. A. Moore, Jr.1, P. B. DeLaune2, D. C. Carman3, T. C. Daniel2 and A. N. Sharpley4

1USDA/ARS, Poultry Production and Product Safety Research Unit, Fayetteville, AR, USA

2Dept. of Crop, Soil and Environmental Sciences, University of Arkansas, Fayetteville, AR, USA

3USDA/NRCS, Little Rock, AR, USA

4USDA/ARS, University Park, PA, USA


Phosphorus (P) runoff from pastures fertilized with animal manure can lead to accelerated eutrophication of adjacent waterbodies. In the U.S., many agencies are assuming that high levels of P runoff from pasture and croplands is the result primarily of high soil test P levels. Consequently, these agencies are establishing upper cutoff limits or thresholds for soil test P, above which manures cannot be applied. However, recent research suggests that other factors, such as P application rates, amounts of soluble P in manure, and hydrology may play a more important role in P runoff than soil test P. Another approach to managing P on farms that includes these factors is the Phosphorus Index (P.I.). Although the original P.I. was superior to approaches using soil test P alone, the weighting factors were based primarily on professional opinion. The objective of this research was to develop a P.I. for pastures based on data collected from rainfall simulation studies.

Methods and materials

Rainfall simulation studies were conducted on 72 runoff plots cropped to tall fescue to determine the relative contribution of different variables to P loading in runoff. Simulations were conducted to determine the effects of the following variables: (1) soil test P, (2) soluble P in manure, (3) P content of poultry diets, (4) fertilizer type (commercial vs. manure), (5) manure application rate, and (6) application timing with regards to weather. Rainfall simulators were used to produce a 5 cm hr-1 storm of sufficient duration to cause 30 minutes of runoff. Runoff samples were analyzed for soluble reactive P (SRP) and total P. Under these conditions, most (90-95%) of the P is in the soluble fraction. Multiple regression techniques were used to analyze the data in order to determine the relative importance of each variable and develop the P.I.

Once the index was developed, validation studies were conducted at two scales; (1) on small plots using rainfall simulators, and (2) on field-scale plots with natural rainfall. The small plot validation studies were conducted on six poultry/beef farms located in Arkansas and Oklahoma on soils considered by USDA/NRCS to be 'benchmark' soils. Twelve runoff plots were constructed at locations within each field at locations needed to insure a wide range in soil test P. Six rainfall simulations were then conducted on each of the 72 plots; three of the events were prior to manure applications, the other three were after application. Validation was also carried out using data from paired watersheds (0.5 ha) for which runoff data (P load) had been collected for the past six years.


The results from the initial research indicated that the P runoff from pastures was more affected by the soluble P content of the fertilizer source, whether commercial or manure, than any other variable. Phosphorus concentrations in runoff were highly dependent on soil test P levels when manure had not been applied, however, after applications the soluble P content of manure determined P loads in runoff. Decreasing P contents in diets with high available P (HAP) corn or phytase additions lowered the amount of total P in manure. However, these treatments resulted in higher soluble P in manure which resulted in higher P runoff. Phosphorus runoff was much higher from commercial fertilizer, such as triple superphosphate fertilizer (0-46-0) than swine or poultry manure, even though it was applied at the same rate of P. This was due to higher soluble P in triple superphosphate than in the manures. Phosphorus runoff was also highly correlated to manure application rates and the time after manure application until the first runoff event.

The P.I. that was developed had four components; (1) P sources, (2) P transport, (3) BMPs, and (4) rainfall. The P.I. value is a predicted P load in runoff in units of lbs acre-1 year-1 (Imperial units were used because this is what USDA uses for farms in the U.S.). Validation studies conducted on small plots showed that P runoff was poorly correlated to soil test P, but highly correlated to the P.I. value. Validation data on the small watersheds showed that P loads in runoff were highly correlated (R = 0.91) to the P.I. for those fields, with a slope of near one and an intercept of near zero (y = 1.16x – 0.21). These data indicate that this index predicts the risk of P runoff at the field-scale under natural rainfall conditions. Phosphorus indices that predict P loading would be superior to indices which predict a relative risk factor for watersheds where TMDLs (total maximum daily loads) have been established for P.

Development of a phosphorus assessment tool for Virginia: a mechanistic approach

G. L. Mullins1, W. L. Daniels1, L. W. Zelazny1, M. R. Brosius1, M. A. Beck1, W. L. Wolfe2, A. Vincent2 and J. W. Pease3

1Virginia Tech, Department of Crop and Soil Environmental Sciences, Blacksburg, Virginia

2Virginia Tech, Department of Biological Systems Engineering, Blacksburg, Virginia

3Virginia Tech, Department of Agricultural and Applied Economics, Blacksburg, Virginia

Public concern over the effects of animal manure and other organic nutrient sources on water quality in Virginia and other Mid-Atlantic States in the eastern U.S. has increased markedly in recent years. Protecting water quality from nutrient pollution will require the development and implementation of farm-level management tools that can be used by Virginia producers to minimize environmental impacts resulting from the land application of nitrogen (N) and phosphorus (P) nutrients in organic and inorganic fertilizers. One promising P management tool is the P Index. The U.S. Natural Resources and Conservation Service has developed a P indexing tool that can be used to assess the potential for P losses from agricultural fields to the environment. This tool was developed to include the solubility of soil P, the type, rate, and method of P source management, and transport factors which can affect P losses from agricultural fields to surface and ground water supplies. Adaptation of the P Index as a nutrient management tool for Virginia will require tailoring of the index to soil, management and transport conditions specific to varying Virginia soil landscapes.

Development of a working P Index that can be used for nutrient management planning in Virginia is being approached through an interdisciplinary effort at Virginia Tech. The first phase of this project included a statewide sampling of Virginia soils, focusing on three counties characterized by intensive animal and row crop agriculture. Samples were collected from Accomack (Eastern Shore), Amelia (South Central) and Rockingham (Shenandoah Valley) counties. Sites represented a wide range in soil test P levels and varied management of organic amendments. Samples were subjected to soil testing procedures to characterize soil P forms including the relative degree of soil P saturation. Samples were collected in the fall of 1998 from 19 fields in Accomack County, 15 fields in Amelia County and 28 fields in Rockingham County.

Total soil P and the distribution of soil P among different chemical forms differed considerably for the three counties and among fields within a given county. In general, soils impacted by long term applications of animal manure (Rockingham County) or high applications of commercial fertilizer for vegetable production (Accomack county) contained the highest levels of water-soluble and soil test P. Mehlich 1 extractable P ranged from 42 to 779 mg kg-1 in Rockingham, 31 to 1083 mg kg-1 in Amelia, and 2 to 330 mg kg-1 in Accomack counties. Several fields (74% of the fields sampled) in all three counties (highest proportion for Rockingham County) had soil test P levels that were high enough to maintain crop production for several years with little or no additional P fertilizer application.

The soil P data suggest several practical implications for Virginia producers. Fields with a history of receiving long-term annual applications of P from organic and/or commercial P fertilizers had elevated levels of total soil P and soil test P. Therefore, applying P in either fertilizer or organic forms at levels above crop P needs will result in further buildup of soil P levels. For these Virginia soils, the relative degree of soil P saturation increased as the level of soil test P increased (r2 of 0.97, 0.68 and 0.92 for Amelia, Accomack and Rockingham counties, respectively). In addition, water-soluble soil P increased with increasing soil test P. Since higher levels of water-soluble soil P have been reported to be associated with greater potential for P losses in surface runoff, this data set suggests that long-term over-application of P in organic-P sources at rates in excess of crop needs increases the potential for P losses in surface runoff from these soils.

Investigating the relationships between soil P levels and potential P losses in surface runoff using portable rainfall simulators has expanded the laboratory phase of the project. Soil data collected as part of the statewide survey and data from the rainfall simulators are being used to calibrate a model for predicting P losses from agricultural fields on a watershed basis. Work is on going to improve a process-based model of P fate and transport. The ANSWERS-200 model is a continuous, distributed parameter, watershed-scale model that simulates rainfall runoff, erosion, sediment delivery, and N and P transformation processes and transport. Updated algorithms are being incorporated into the model and the detailed data from the statewide survey and rainfall simulations are being used in testing the model. Further work will estimate the economic consequences of practices altered by manager use of the P-index for nutrient management planning.

Potential phosphorus losses in runoff from New Zealand pastoral soils receiving superphosphate and reactive phosphate rock applications

Long Nguyen1 and Bert Quin2

1National Institute of Water & Atmospheric Research Ltd. (NIWA), P. O. Box 11-115, Hamilton, New Zealand.

2 Summit Quinphos (NZ) Ltd., P. O. Box 24-020, Auckland, New Zealand

Potential phosphorus (P) losses in runoff from four fertilized New Zealand sheep-grazed pastoral topsoils with long-term (> 20 years) annual applications of superphosphate (SP) or reactive phosphate rock (RPR) were investigated using microplots (1x 0.5 x 0.1 m) and rainfall simulation. Collected topsoils (0-5 cm) were sieved (<5 mm), packed at a field bulk density (1g/cm3) into microplots, treated with SP, RPR and nil fertiliser (non-fertilized) treatments and subjected to simulated rainfall events on day 3, 10 and 32 (D3, D10 and D32 respectively) after the application of these treatments. Runoff particulates (>0.45 m m) and runoff solution were determined for dissolved reactive P (DRP), particulate P (PP), and total dissolved P (TDP). Particulate obtained from each runoff event and pastoral soils collected from each microplot on D3 prior to rainfall simulation were sequentially fractionated into water-soluble P, readily-exchangeable (labile) P, P associated with aluminum (Al) and iron (Fe) oxides and calcium P (Ca-P) using sodium chloride (NaCl), sodium bicarbonate (NaHCO3), sodium hydroxide (NaOH) and sulphuric acid (H2SO4), respectively. Regardless of previous topdressing histories, approximately 41-61 % of P in runoff from SSP-treated microplots was inorganic (DRP), while this fraction accounted for <5 % of P in runoff from RPR and non-fertilized plots. DRP concentration in the first simulated runoff event 3 days after fertiliser applications was 70 times higher in SSP-treated microplots, compared to RPR-treated microplots. By D 32 after fertilizer applications, DRP concentration in the third simulated runoff event was still 14 times higher with SSP than with RPR-treated microplots. Particulate P concentration measured in the first simulated runoff was 1.8-3.9 times higher from SSP, compared to RPR-treated microplots. By D32 after fertiliser applications, PP concentration in runoff was still 2.2 times higher with SSP than with RPR. Soils with long-term annual applications of RPR had similar P levels in NaCl, NaHCO3 and NaOH-extractable fractions to soils that had annually received long-term SSP applications. However, they had 37-42% higher H2SO4-extractable P than soils with previously-applied SSP. Applications of SSP and RPR significantly increased resin-extractable soil P and readily-desorbed P and mineral P fractions associated with soil Al, Fe and Ca (NaCl, NaHCO3, NaOH and H2SO4-extractable fractions) in runoff. Approximately 1.5-4.9 times more P was extracted by resin from SSP-treated microplots compared to the non-fertilized microplots, while the corresponding P values in RPR-treated plots were only 30-90% higher than those in the non-fertilized microplots. Applications of SSP significantly increased NaCl and NaOH-extractable soil P by 21-32 times and 21-49%, respectively over the non-fertilized microplots. In contrast, RPR applications mainly increased the H2SO4-extractable P fraction (apatite). Regardless of previous topdressing histories, runoff particulate P in NaCl, NaHCO3 and NaOH-extractable fractions measured in a runoff event 3 days after fertilizer application was higher (60-150, 14-27, and 10-25 times, respectively) in SSP than RPR-treated microplots, indicating that the sum of these fractions potentially represent mobile P in runoff. Residual effects of RPR and particularly SSP on runoff P losses probably last more than 32 days after fertilizer applications, since the increase in NaCl, NaHCO3 and NaOH-extractable P fractions in runoff particulate obtained from SSP, compared to RPR-treated microplots were still prevalent on D32 after 3 simulated runoff events.


Organic phosphorus mineralisation in wetland peat soils under alternating waterlogged and drying conditions

A. Niedermeier and J. S. Robinson

Department of Soil Science, The University of Reading, Whiteknights, PO Box 233, Reading, RG6 6DW, UK


The restoration of wetland habitats by the flooding of agricultural fields is becoming increasingly common. Through prescribed water level management, the land encourages a variety of wetland wildlife during the winter and spring, and allows low density grazing during the summer and early autumn. From a water quality point of view, one of the main problems associated with flooding former agricultural land is the release of phosphorus (P) to the overlying water column and to adjacent water bodies.

In many wetland soils, organic P (Po) makes up a significant proportion of the total P. Because of the existence of Po in different forms and with varying degrees of availability, it is hypothesised that the mineralisation of Po can make an important contribution to the release of inorganic P to the overlying water column. In wetlands that are subjected to alternating waterlogged and drying conditions (seasonal wetlands) the soil switches back and forth between aerobic and anaerobic conditions. In such an environment, the potential for P release from soil Po to water may be modified through seasonal changes in i) the forms and stability of the soil Po and ii) the turnover of the microbial biomass. The general aim of the current research is to evaluate the short-term dynamics of soil Po and microbial biomass P in peat soils of seasonal wetland.


Peat samples were taken during the summer (August 2000) and winter (January 2001) from the surface (0 – 20 cm layer) of a recently created wetland in the Somerset Levels and Moors. Formerly, the site comprised agricultural fields that had received high inputs of P fertiliser for the production of vegetable crops. Owing to its current water management regime, the land is drained in the summer, but the soil surface is still moist, depending on rainfall and temperature conditions. In the winter, the whole site is flooded. A preliminary evaluation of the amounts of readily mineralisable soil Po for the two sampling periods was conducted by performing a short, sequential fractionation scheme, involving 0.5 M NaHCO3 followed by 0.1 M NaOH. In upland soils, Po turnover studies have shown that NaHCO3-extractable Po is rapidly mineralised (Bowman and Cole, 1978) whilst NaOH extracts more biologically resistant forms of Po that are involved in the long-term transformations of P (Batsula and Krivonosova, 1973). Solution-state 31P-NMR spectroscopy is currently being used to identify and quantify the distribution of the major forms of P (inorganic ortho P, ortho-P monoesters and ortho-P diesters) and relate these to the chemical fractions.

Preliminary results and discussion

At both sample times, NaHCO3-Po constituted a very small proportion (< 5 %) of soil total P, indicating the highly humified nature of the peat. However, NaHCO3-Po in the soil sampled in August (40 mg kg-1) was 25% higher than that in the following January (32 mg kg-1). The amounts of Po subsequently extracted in NaOH were not significantly different between the two sample dates. It is likely that the warm, moist soil conditions during the summer stimulated the microbial decomposition of soil organic matter, thus increasing the release of P into the labile pool. Also, the increased rates of microbial activity in the summer may have accounted directly for the higher NaHCO3 – P through an increase in biomass decomposition.

Future work will evaluate the contribution of microbial biomass to the labile Po pool as a function of seasonal wetting and drying cycles. More detailed and frequent soil sampling, combined with laboratory microcosm studies, will provide a fuller evaluation of the short-term changes in the fractions and molecular forms of Po in the soils of seasonal wetlands. It is also expected that the relationship between Po forms and phosphatase enzyme activities will be elucidated, for the purpose of improving our understanding of the role of enzyme activity in the cycling of Po in the peat soils of seasonal wetlands.


Batsula, A.A., and G.M. Krisonova. 1973. Phosphorus in the humic and fulvic acids of some Ukrainian soils. Soviet Soil Science 5:347-350

Bowman, R.A., and C.V. Cole. 1978. Transformations of organic phosphorus substrates in soils as evaluated by NaHCO3-extraction. Soil Science 5:49-54.

Phosphorus content and degree of phosphorus saturation in Danish soils

Gitte H. Rubæk1, Goswin Heckrath1, Jørgen Djurhuus1, Svend E. Olesen1 and Hans S. Østergaard2

1 Danish Institute of Agricultural Sciences, Research Centre Foulum, P.O. Box 50, DK-8830 Tjele, Denmark

2 Danish Agricultural Advisory Centre, Udkærsvej 15, Skejby, DK-8200 Århus N, Denmark


Phosphorus (P) has accumulated in agricultural soils in Denmark over the past decades. However, detailed knowledge of the P accumulation in various soil textural classes and farming systems is lacking. Neither is it known to which extent the phosphorus content has changed in subsoils. We therefore examined the soil phosphorus content and the degree of phosphorus saturation (van der Zee et al., 1990) as well as the changes of these parameters from 1986 to 1997/98 in 337 agricultural soils collected by a national sampling scheme for monitoring nitrogen leaching. Thirty-two soils from deciduous forests were also examined.

Total soil phosphorus content increased in agricultural top- and subsoil from 1986 to 1997/98

Between 1986 and 1997/98, in agricultural soils the total P concentrations increased on average by 34 mg P kg-1 soil at the 0-25 cm depth and 39 mg P kg-1 soil at the 25-50 cm depth. These increases occurred mainly on the sandy soil types, coinciding with comparatively large livestock densities. Correspondingly, soils with livestock densities of zero or less than one livestock units only had increases in total P of 30 mg P kg-1 and 34 mg P kg-1, while soils with animal densities larger than one livestock unit had increases of 63 mg P kg-1. The average increase in total P concentrations corresponded to an annual accumulation of c. 25 kg P ha-1. This value is somewhat larger than, but still in good agreement with, the surplus of P additions to agricultural soils obtained by balance sheet calculations. Our results show that substantial amounts of the accumulating P were transferred from the plough layer to the subsoil of the agricultural soils.

Degree of phosphorus saturation in agricultural soils

In 1997/98 the average degree of P saturation (molar ratio of P and half the sum of iron and aluminium extractable by ammonium oxalate) in agricultural soils varied between 24 and 39% at the 0-25 cm depths and between 16 and 26% at the 25-50 cm depths across the different soil textural classes. The highest degrees of P saturation were found in the sandy soils coinciding with comparatively large livestock densities. Many top- and subsoils had degrees of P saturation above a critical level of 25% that previously had been associated with increased risk of P loss.

Agricultural soils contained much more phosphorus than deciduous forest soils to 75 cm depth

The P contents were substantially larger in agricultural soils than in soils from deciduous forests, and correspondingly the degrees of P saturation were approximately twice as high in the agricultural soils compared with the forest soils to a depth of 75 cm (Table 1). These results indicate that substantial amounts of the phosphorus added to agricultural soils were transported to the subsoil through time.

Table 1. Total phosphorus and degree of phosphorus saturation in soils from deciduous forest and agricultural land.

Soil depth

Total phosphorus

Degree of phosphorus saturation

Deciduous forest


Deciduous forest



—————— mg P kg-1 ——————

—————— % ——————






















Rubæk, G.H., Djurhuus, J., Heckrath, G., Olesen, S.E. and Østergaard, H.S. 2000. Are Danish soils saturated with phosphorus. In: Jakobsen, O.H. and Kronvang, B. (eds.) Phosphorus loss from agricultural areas to the aquatic environment. DIAS report nr 34, Plant production. p. 17-30. (In Danish with English abstract).

van der Zee, S.E.A.T.M., van Riemsdijk, W.H. and de Haan, F.A.M. 1990. Het protocol fosfaatverzadigde gronden, Wageningen, Landbouwuniversiteit, Vakgroep Bodemkunde en Plantevoeding.

Differential pollution risks from soils receiving organic and inorganic sources of phosphorus

M. T. Siddique and J. S. Robinson

Department of Soil Science, The University of Reading, Whiteknights, PO Box 233, Reading, RG6 6DW, UK


In spite of the growing evidence for the accumulation of phosphorus (P) in agricultural soils that receive large and/or continuous applications of organic wastes, there is still relatively little information in the UK on the P reactions in waste-treated soils that are relevant to subsurface P losses. In view of the wide range of organic wastes that have been, and continue to be, responsible for the environmentally unacceptable levels of P in many soils, it is important to study how the measurements of P adsorption-desorption and plant availability are influenced by the nature and properties of the organic waste materials that are applied to the soil. Future waste and P fertiliser management decisions for the protection of surface waters against P inputs requires this information. A laboratory incubation study was conducted to evaluate the P sorption and availability characteristics of incubated soils that were mixed with different organic wastes on a P-equivalent basis.


Five topsoils (0 – 20 cm) were incubated for 20 d with sewage sludge, poultry litter, cattle slurry or KH2PO4 at rates equivalent to 100 mg P kg-1 soil. The soils were selected to represent typical, continuously cultivated loam soils that display a range of Olsen-extractable P values. After the incubation period, P sorption and availability parameters were determined, including values for EPCo obtained from P sorption isotherms, P sorption index (PSI), and amounts of P extracted in 0.01 M CaCl2 and anion exchange resin.


Application of all the P sources decreased the PSI and increased the EPCo, CaCl2-P and resin P in all five soils. In all soils, cattle slurry and KH2PO4 consistently had a much larger influence on PSI, EPCo, CaCl2-P and resin P than did sewage sludge and poultry litter. For example, following slurry and KH2PO4 additions, the decreases in PSI ranged from 37 to 111 L kg-1, and from 29 to 85 L kg-1, respectively. Following sludge and litter additions, the decreases in PSI ranged from 16 to 36 L kg-1, and from 12 to 45 L kg-1, respectively. With regard to EPCo, following slurry and KH2PO4 additions, increases ranged from 0.07 to 2.8 mg L-1, and from 0.05 to 1.6 mg L-1, respectively. Following sludge and litter additions, the increases in EPCo ranged from 0.03 to 0.8 mg L-1, and from 0.01 to 0.5 mg L-1, respectively.


In the sewage sludge and poultry litter treatments, correlation analyses implied that soil solution and available P concentrations were regulated by large inputs of Ca, possibly through the formation of Ca-P precipitates. Also, the concomitant input of soluble low molecular weight (< 50,000) organic compounds may form complexes with the Ca and other metals; it has been shown that these complexes increase the sorptive area of soil material (Gerritse, 1981). The Ca contents of the sludge and litter were 42 and 14 g kg-1, respectively. Therefore, high concentrations of both Ca and P were likely to exist in the zone of application of these two high Ca-status materials. With regard to the higher solubility of P in the cattle slurry- than KH2PO4-treated soils, it was speculated that the organic acids released in the slurry treatments effectively blocked P sorption sites, thereby reducing the effective P sorption capacity (Reddy et al., 1980).

Clearly, there is a need for more detailed inorganic and organic analysis of animal manures and sewage sludges in order to identify and explain fully those properties of organic wastes that influence the availability and potential mobility of both native and applied P in the soil matrix. For a range of animal manures and biological sewage sludges, current research is attempting to evaluate the following:

  1. Simple relationships between the contents of polysaccharides, proteinaceous compounds and humic acids, and metal solubility.
  2. The fate of major classes of organic P (monoesters and diesters), that are found in animal manures and biological sludges, after they are released in to the soil matrix.


Gerritse, R. G. 1981. Mobility of phosphorus from pig slurry in soils. p. 347-366. In T.W.G. Hucker and C. Catroux (ed.) Phosphorus in sewage sludge and animal wastes. D. Reidel Publ. Co., Dordrecht, the Netherlands.

Reddy, K.R., M.R. Overcash, R. Khaleel, and P.W. Westerman. 1980. Phosphorus adsorption-desorption characteristics of two soils utilized for disposal of animal wastes. J. Environ. Qual. 9:86-92.

Phosphate losses from grassland fields used for dairy farming

Caroline van der Salm and Oscar Schoumans

Alterra Green World Research, PO BOX 47, 6700 AA Wageningen, The Netherlands


The present high levels of agricultural production require relatively high soil phosphorus (P) levels. Large uncertainty exists on the magnitude of the P surplus necessary to maintain these relatively high soil P levels. To formulate manure reduction plans that are acceptable from both an agronomic and environmental point of view more knowledge on the faith of P on fields used for dairy farming is essential. To obtain this information a monitoring program has been started in 1996 on four grassland soils in the Netherlands. Two of these sites are situated on relatively well drained sandy soils, one on a well drained clay soil and one site on a poorly drained peat soils. At each sites three plots were established that received different combination of N and P surpluses: a surplus of 180 kg ha-1 N and 0 kg ha-1 P, 180 kg ha-1 N and 10 kg ha-1 P and 300 kg ha-1 N and 20 kg ha-1 P

To obtain P budgets for the considered plots both the input of P by manure and fertilisers and the net output of P by grass, milk and meat production and losses by leaching have been quantified. To obtain leaching fluxes both meteorological, hydrological and soil solution data have been collected. To calculate hydrological fluxes rainfall has been measured daily at the farms, other meteorological data are obtained from weather stations close to the sites. Groundwater levels have been measured two-weekly at the different plots to validate the hydrological calculations with the model SWAP. Soil solution concentrations have been measured three times a year: in early spring before and after the first manure application and in autumn. Soil samples have been collected yearly and were analysed for various soil P pools to detect any changes in available P during the experiment.

During the first three years of the measurements no significant changes in the phosphate pools were found. This is mainly caused by the large size of the pools compared to the applied phosphate surplus. On basis of the size of the pools, significant changes over the period 1997-2000 may only be expected for the amount of easily available ('labile') P or in Pw values (water soluble P). However, these changes were obscured by a strong year to year variation in the measured values. Average P losses, due to leaching and runoff, ranged from 3.7 kg ha-1 P at one of the sandy sites to 0.4 kg ha-1 P at the clay site. Average leaching fluxes at the two other sites (peat and sand) were respectively 2.1 and 2.6 kg ha-1 P. Differences in P surplus were hardly reflected in the P leaching losses from the different plots due to the high amounts of adsorbed P. However, measured P losses were strongly related (R2adj. = 0.80) to the present P content (Pw level) at the bottom of the rootzone (20-30 cm) of the plots.

The obtained data set was used to validate the nutrient model ANIMO. The model simulated the order of magnitude of the measured soil solution concentrations quite well. However, the model did not always simulate the seasonal variation in measured concentrations accurately. Yearly simulated leaching fluxes of total P were quite well simulated by the model, except for the peat soil were simulated fluxes of organic P were higher than measured values. The simulated Pw values were close to measured data, in particular at the sandy soils. However, the model tended to overestimate the decline in Pw values at the plots with a P surplus of 0 kg ha-1 a-1 and slightly overestimated the rise in Pw values at plots with a P surplus of 20 kg ha-1 a-1.

The validated model was used to simulate the long-term effect of P surpluses of 0, 10 and 20 kg ha-1 jr-1 on both the P leaching and the changes in P levels. These calculations were carried out for both the monitoring sites and a number of characteristic Dutch sites to get an impression on the national variation in losses as a response to different manure scenarios.

Phosphorus in the soil and groundwater in a farmstead and its vicinity

B. Sapek

Institute for Land Reclamation and Grassland Farming at Falenty, 05-090 Raszyn, Poland

The agricultural activity could be a cause of the phosphorus (P) dispersion in to the environment (Haygarth and Sharpley, 1997). The P surplus, not used in agricultural production, is the main source of this dispersion. A farmstead and its vicinity is an integral part of the farm, where a constant flux of matter, including P, between the farmstead and rest of farm takes place. The goal of these investigations was to recognise the 'hot spots', which are mainly responsible for the P input from farm into the soil and groundwater.

The investigations were made on the basis of soil samples taken from 20 cm soil layers down to the 200 cm, as well as from 0-10 cm and 10-20 cm soil layer taken in the same places. The P content in the fresh soil samples was measured volumetrically and determined in the extract of 1% K2SO4 (more mobile form) and in 0.5 M HCl (potential pool of P) colorimetrically with Mo-blue reaction by the means of flow autoanalyser. The 1% K2SO4 soil extract is used to determine N-NO3 and N-NH4 content in the same soil samples (Sapek & Sapek, 1997). The P content in the soil was calculated in mg dm-3 of fresh soil or recalculated to kg ha-1. The groundwater samples were taken from the control wells installed at the similar monitoring points as the soil samples. The water samples from unused farm wells were also collected. The P concentration in water samples (RPunf, according to Haygarth and Sharpley, 2000) was determined without filtration, in the same manner as the soil extract. The monitoring points were installed at 12 demonstration farms in different part of Poland within the project 'Baltic Agriculture Runoff Program' (BAAP) and at one farm of experimental station at Falenty.

The highest P contents determined in the soil extract of 1% K2SO4 (about 300 kg ha-1 in 0-20 cm soil layer and about 700 kg ha-1 in 0-200 cm soil layer) were confirmed in the vicinity of manure storage, by the barn or slurry tank. The P content determined in 0.5 M HCl extract of soil from these places corresponded to about 1800 kg ha-1 in 0-20 cm soil layer and to 9300 kg ha-1 in 0-200 cm soil layer. The mean P concentration in groundwater in such places was about 2.6 mg l-1 and maximum concentration was 27.5 mg l-1. The increased P concentration could happen also in the water samples from unused farm wells (0.33-0.74 mg l-1).

The soil in poultry paddock in demonstration farms was normally enriched in P. The mean content of more mobile P form in 0-20 cm soil layer was about 40 kg ha-1 and maximum 200 kg ha-1. In comparison to that, P content in the soil at farm¢ s garden was about 6-17 kg ha-1.

The risk of farmstead effect on soil and water pollution with P depends closely on the manner of animal wastes management. The soil enrichment with P in farmstead and its vicinity in experimental station was also high but lower in comparison to demonstration farms (Table 1).

Table 1. The ranges of phosphorus (P in 1% K2SO4) content in the soil at different animal waste storage places in experimental station

P, kg ha-1

Soil layer, cm

Manure heap

Slurry tank

Calf paddock










Summarising, animal waste storage places, animal and poultry paddocks and housing as well as silage storage places and sewage tanks are the 'hot spots'. These could be the main sources of P dispersion and transfer of this element from the farm in to the environment.


Haygarth A.N., Sharpley A.N.2000. Terminology for phosphorus transfer. J. Environ. Qual. 1:10-14.

Sapek A. Sapek B. 1997. Methods of organic soils chemical analyse. Materialy Instruktazowe 115:80p. Falenty IMUZ Publication (in Polish).

Sharpley A.N., Rekolainen S., 1997. Phosphorus in agriculture and its environmental implications. p.1-53. In H. Tunney at al. (ed.) Phosphorus loss from soil to water. CAB International.

Effect of phosphate sorbing amendments on phosphate availability in two soils

M. Schärer1, S. Sinaj1, P. Moosbauer1, G. Favre1, T. Vollmer2, C. Stamm2 and E. Frossard1

1Institute of Plant Science, Plant Nutrition, Swiss Federal Institute of Technology, Switzerland

2Institute of Terrestrial Ecology, Soil Physics, Swiss Federal Institute of Technology, Switzerland


Diffuse phosphorus (P)-losses from agricultural land can cause eutrophication in surface waters. Particularly permanent grassland soils in areas of intensive livestock rearing are susceptible to such losses since large P-surpluses have accumulated in the topsoil. In this study, we tested different types of soil amendments as models for drastic remediation measures to reduce P-losses from P-enriched grassland soils to acceptable levels.

Study sites and experimental set-up

Soil samples (0-20 cm) were collected at two different field sites. The soil from site I was a neutral Cambisol (calcaric, mollic) (0.01 M CaCl2: pH 6.7). The soil from site II is an acid Luvisol (0.01 M CaCl2: pH 4.8). The cation exchange capacity (CEC) were of 58 and 28.8 cmolc kg-1 for soil I and II respectively, while its saturation with basic cations (Ca2+, Mg+, Na+ and K+) was 99.8 % and 45.3 % respectively. The P status of the two soils is listed in Table 1.

Table 1. Phosphorus-status of the topsoil (0-2cm) from sites I and II.

Total P

Organic P

Inorganic P


Olsen P

¾ ¾ ¾ ¾ ¾ ¾ ¾ ¾ ¾ ¾ ¾ ¾ ¾ ¾ ¾ mg P kg soil-1¾ ¾ ¾ ¾ ¾ ¾ ¾ ¾ ¾ ¾ ¾ ¾ ¾ ¾ ¾

Site I






Site II






¶: isotopically exchangeable P within 1 minute (Frossard and Sinaj, 1997)

Four different amendments were used in this study: hematite Fe2O3 (98.5%), aluminium-oxide Compalox AN/V-801 (Alusuisse Martinswerk, D) and lime CaCO3 (76%). Subsamples of 800 g field moisture soil (60% of waterholding capacity) were mixed with different amounts of amendments (control, 1:0.1; 1:0.01 and 1:0.002 soil/amendment weight/weight ratios). Samples taken immediately after mixing, 1 and 4 weeks of incubation were analysed for desorbable P according to Freese et al. (1999) and exchangeable P according to Frossard and Sinaj (1997). For every treatment we conducted a pot experiment with 33P labelled soil to estimate the effect of the amendments on P availability for Lolium multiflorum.

Preliminary results

For some amendments a strong effect on isotopic exchange parameters (n, Cp) was observed immediately after mixing soil and amendment. Aluminium-oxide application to soil at 1:0.01 ratio increased the P sorption capacity (n) from 0.29 to 0.46, at site I and from 0.36 to 0.41 at site II. With the 1:0.1 soil/amendment ratio n values increased to 0.57 and 0.58 respectively. For the 1:0.01 soil/amendment ratio the P concentration in the soil solution (Cp) decreased from 0.42 to 0.27 mg P L-1 and from 0.26 to 0.15 mg P L-1 for the site I and II respectively, and to 0.05 and 0.1 for the ration soil/amendment 1:0.1. The lime amendment had a strong influence on Cp of the soil from site II: 1:0.01 soil/ CaCO3 ratio lead to a 70 % decrease of P in the soil solution. With hematite the effect was smaller and only large addition (1:0.1 ratio) increased n from 0.36 to 0.49 and decrease the Cp from 0.26 to 0.13 mg P L-1. For the soil from site I no strong effect was observed for the hematite and the lime amendment. The application of aluminium and iron oxides lead to a higher P-fixing capacity by increasing the number of sorption sites. For the lime amendment the decrease in CP could be explained by the precipitation of P as calciumphosphate-species and the increase in P-sorption capacity (n) due to a stabilisation of metal-P-bonds by Ca2+ ions.

For P-rich grassland areas with a high risk for P-losses to surface waters, application of lime and aluminium-oxide may be considered as remediation measures but further investigations should be conducted researching the environmental effects of amendments in field conditions prior to application.


Freese, D., Weidler, P.G., Grolimund, D. and Sticher, H. 1999. A Flow-Trough Reactor with an Infinite Sink for Monitoring Desorption Processes. Journal of Environmental Quality 28: 537-543.

Frossard, E., Sinaj, S. 1997. The Isotope Exchange Kinetic Technique: A Method to describe the Availability of Inorganic Nutrients. Applications to K, P, S, and Zn, Isotopes Environ. Health Stud. 33: 61-77.

Sequentially extracted phosphorus (P)-fractions as indicator for P-losses from re-wetted peat soils

A. Schlichting, and P. Leinweber

1University of Rostock, Faculty of Agricultural and Environmental Sciences, Institute of Soil Science and Plant Nutrition, Rostock, Germany


European-wide efforts to re-wet and restore fenlands require a critical evaluation of eutrophication risks arising from the mobilisation and release of phosphorus (P). The chemistry of P compounds and P turnover in degraded and re-wetted peat soils is not fully understood. The sequential P fractionation (Hedley et al., 1982) is a suitable method for assignment of P compounds to pools of different turnover and availability; however, it was seldom applied to peat soils. We studied 12 soil profiles, representing typical European fenlands in the U.K., Sweden, Poland and Germany. The objectives were (1) to disclose land-use effects on the P-status, (2) to improve the methodology of pretreatment and sample processing, and (3) investigate the concentrations of P fractions in four Histosol profiles of widely different geographic origin by using the improved method.

Description of work

(1) Land use effects were studied in two areas (Droemling, Bode-Selke Valley) in Saxony-Anhalt/Germany. Being thoroughly Histosols in the past, only five of the eight profiles were classified as Histosols (most of them are Ombri-Sapric Histosols; FAO, 1997), and three were degraded to Gleysols due the long-term drainage and intensive cultivation. Total P (Pt) was always enriched in topsoils and ranged from 320 to 2488 mg kg-1. The P concentrations decreased with soil depth (87 to 515 mg kg-1) except for relic peat horizons. Sequential P fractionation showed great variation in the proportions of labile P (resin P + NaHCO3-Pi,o) and Ca-bound P (1 % to 63 % of Pt). The Histosols generally differed from the stronger degraded Gleysols in larger amounts and proportions of organically bound P and residual P. There were also effects of land use within groups of Histosols and Gleysols. For instance a degraded Histosol in alder forest had the largest proportions of labile NaHCO3-Po. However, we realised that slight differences in sample pretreatment and processing perhaps influenced the proportions of sequentially extracted P fractions.

(2) Consequently, we subjected an originally field-moist low-moor peat soil to two drying (air-, freeze-) and three storing techniques (fresh, frozen, cool) prior to the Hedley P fractionation. The P concentrations in nearly all fractions differed among treatments. The proportions of resin-P and NaHCO3-P decreased in the order 'fresh' > 'frozen' > 'cool stored'. Drying led to great increases in the proportion of residual P, most of which was assigned to organic P forms. Organic P extracted by NaHCO3 and NaOH disappeared or was lowered in proportion due to drying. It was concluded that the sequential P fractionation should be carried out with fresh peat samples to avoid undesired changes in the fractions. As a consequence we recommended a protocol for processing peat samples before and during fractionation (Schlichting and Leinweber, 2001). Hence the samples should not be dried before fractionation and the resin extraction must be carried out in a modified way.

(3) Very recently we finished a first set of fractionation for fresh peat samples from four sites of different geographic locations in Europe. The results showed great differences in labile and potentially mobilisable P fractions according to the differently degraded peat soil under various use and management, and in re-wetting stages. The proportions of resin P ranging from 2 % (Sweden) to 13 % (Germany) of Pt and for NaHCO3 Pi,o from 5 % (U.K.) to 16 % (Germany). The Pt concentrations in the top soils are on average high for all sites (1257 to 1789 mg kg-1). Currently, a continuously sample collection from ground, surface, and seepage water and soil takes place. Phosphorus forms in soil solution (Pt, Pi) have been detecting since 2000, and standard soil P tests have been proving too. At least correlation to the P fractions and their changes will be investigated. We point to the large gaps in our knowledge about the P compounds of the residual, and give an insight in the latest approaches and results of our work.


Food and Agricultural Organisation S. 1997. Soil map of the world, revised legend, Rome.

Hedley, M. J., Stewart, J. W. B., and Chauhan, B. S. 1982. Changes in inorganic and organic soil phosphorus fractions by cultivation practices and by laboratory incubations. Soil Sci. Soc. Am. J. 46 : 970-976.

Schlichting, A. S., and Leinweber, P. L. 2001. Pretreatment effects on sequentially extracted phosphorus fractions from peat soils. Comm. Soil Sci. Plant Anal. submitted.


Assessing and managing the risk of subsurface phosphorus transport to surface waters

J. T. Sims1, R. O. Maguire1, Frank J. Coale2, Regis R. Simard3

1University of Delaware, Newark, Delaware, USA.

2University of Maryland, College Park, Maryland, USA

3Agriculture and Agri-Food Canada, Quebec, Canada


Environmental concerns about the impact of nonpoint source pollution of surface waters by phosphorus (P) from agricultural soils have resulted in policies and legislation that will restrict land application of manures and fertilizers in some U.S. states and Canadian provinces. In some cases these restrictions will be based solely on a soil P measurement (e.g., soil test P), while in others a more holistic approach, commonly referred to as the Phosphorus Site Index, will be used. The Phosphorus Site Index approach integrates P site and transport factors with P management practices to assess the risk of P loss to water. With respect to P transport, the major concern for nonpoint P pollution has always been, and continues to be, the losses of soluble and particulate P by soil erosion. In some settings, however, the leaching of P to shallow ground waters or artificial drainage systems (e.g. tiles, drainage ditches) and subsurface flow of P to surface waters is also an important P transport mechanism. In response to these water quality concerns some state and national field-scale risk assessment indices have been developed, or proposed, to characterize the relative risk of P loss by subsurface flow. This paper: (i) provides an overview of the major approaches used in the U.S. and Canada today to integrate subsurface P transport into the Phosphorus Site Index; and (ii) describes best management practices used, or under consideration, to minimize subsurface P losses from soil to water.

Current risk assessment efforts for subsurface phosphorus transport

A review of all state, provincial and national Phosphorus Site Indices used, or proposed, in the U.S. and Canada is now being conducted to determine the methods used to integrate subsurface P transport into comprehensive risk assessments for P loss. Results to date indicate that most risk assessment protocols for subsurface P transport include some combination of soil drainage class, depth to mean high water table, indicators of preferential flow, the presence of artificial drainage systems, the extent and depth of soil P saturation, and nutrient management practices related to the method and timing of P application. A comprehensive summary of the various protocols, analyzing the similarities and differences will be presented along with recommendations for standardization of this component of the Phosphorus Site Index.

Best management practices for controlling subsurface phosphorus transport

All U.S. states and Canadian provinces that directly incorporate subsurface P transport into a Phosphorus Site Index are being surveyed to determine the best management practices (BMPs) specifically recommended to reduce P loss by subsurface pathways. Results of this survey will be summarized and presented to: (i) provide information on the rationale for the recommended BMPs, and (ii) summarize key research findings that justify the BMPs that are recommended, or required, to minimize the potential for subsurface P loss.

Chemical nature of phosphorus in leachate from a grassland soil

G. S. Toor, L. M. Condron, H. J. Di, K. C. Cameron and T. Hendry

Centre for Soil & Environmental Quality, PO Box 84, Lincoln University, Canterbury, New Zealand


Intensive grazing and associated application of fertilizers and manures increase phosphorus (P) accumulation in soils and losses to water bodies which can result in eutrophication. Studies have indicated that some soils high in P are vulnerable to leaching losses and P concentrations up to 10 mg P L-1 have been reported (e.g. Sims et al., 1998). This paper provides information on amounts and forms of P leaching losses from an irrigated grassland soil.


Lysimeters (50-cm diameter, 70-cm depth) were collected from a free draining stony Lismore silt loam soil (Udic Haplustept) (Cameron et al., 1992). The experiment included 4 replicates of the following annual treatments: P45, P45+DSE200, P45+DSE400, P45+DSE400+U (P45 = 45kg P ha-1 as single superphosphate; DSE = dairy shed effluent at 200 or 400 kg N ha-1, U= urine at 1000 kg N ha-1). Flood irrigation (100 mm per application) was applied every three weeks between November and April. Leachate was collected after irrigation or a significant rainfall event. Dissolved reactive P (DRP) and total dissolved P (TDP) were determined in a filtered (<0.45 m m) sample, and total reactive P (TRP) and total P (TP) were determined in an unfiltered sample.


Phosphorus loss in reactive fractions (DRP and PRP) constituted only 10 % of total P in leachate while 90 % of P loss was accounted for in unreactive fractions (DUP and PUP). It was observed that PRP loss was higher than DRP in all treatments, except P45 (Table 1). For unreactive fractions, 58-61 % of total P loss occurred as PUP compared with 28-33 % for DUP (except P45). The P loss was significantly higher for DSE treatments. The annual TP loss for the 1999-2000 year were 930, 1303, 1728 and 2139 g ha-1 for the P45, P45+DSE200, P45+DSE400, P45+DSE400+U treatments, respectively.

Table 1. Concentrations of P (m g P L-1) in reactive and unreactive fractions determined in leachate collected during the 1999-2000 season (mean of 25 drainage events).







7 (1-12)*

4 (1-15)

61 (1-239)

43 (2-410)

115 (9-500)


9 (1-24)

10 (1-80)

70 (2-237)

120 (1-1699)

209 (14-1930)


10 (1-23)

19 (1-176)

79 (1-231)

167 (1-1765)

275 (18-2129)


4 (1-10)

28 (2-293)

86 (7-285)

189 (9-1811)

307 (33-2396)

lsd (5%)






*data in parenthesis are concentration ranges;

PRP = particulate reactive P (TRP – DRP); DUP = dissolved unreactive P (TDP – DRP);

PUP = particulate unreactive P (TPP – PRP); TPP = total particulate P (TP-TDP)


Results of this study showed that 90 % of the P in leachate was present in unreactive particulate and dissolved forms which are believed to be organic P.


Cameron, K.C., Smith, N.P., McLay, C.D.A., Fraser, P.M., McPherson, P.J., Harrison, D.F. and Harbottle, P. 1992. Lysimeters without edge-flow: an improved design and sampling procedure. Soil Science Society of America Journal 56: 1625-1628.

Sims, J.T., Simard, R.R., and Joern, B.C. 1998. Phosphorus loss in agricultural drainage: Historical perspective and current research. Journal of Environmental Quality 27:277-293.

Soil test phosphorus and measured concentrations of phosphorus in water from grassland

H. Tunney1, R. H. Foy2 and P. M. Haygarth3

1 Teagasc, Johnstown Castle Research Centre, Wexford, Ireland

2 Agricultural and Environmental Science Division, Department of Agriculture and Rural Development, Newforge Lane, Belfast, BT9 5PX, Northern Ireland

3Institute of Grassland and Environmental Research, North Wyke Research Station, Okehampton, Devon, EX20 2SB, UK

Eutrophication, effectively phosphorus (P) enrichment, is the most serious environmental problem impacting on lakes and rivers in Ireland and the Environmental Protection Agency of Ireland estimates that about 50% phosphorus (P) loss to water originates from agricultural sources. In 1998 legislation was passed requiring local authorities to devise and implement regulations which, when adopted, will ensure that current trends of diminishing water quality will be reversed and lakes and rivers restored a good ecological status. A target water P concentrations is set at less than 20 mg P L-1 for salmonid waters and 35 mg P L-1 for other waters. To meet this target, measures are being adopted to reduce point source discharges of P from sewage treatment works and to reduce the P content of household detergents but deciding on the most appropriate measures for reducing the P contribution from agriculture remains contentious. The water quality target of 20 mg P L-1 is a an onerous one as, under Irish conditions, it implies that catchment transfers of P from land to water should be less than 0.25 kg P ha-1. This paper presents results showing the relationship between soil test P (STP) and concentrations of P in water obtained from plot and farm drain scale under Irish conditions. This data is compared with data from studies in other countries. The implications of the implied relationship are considered for water quality and for the sustainability of productive agriculture in Ireland. There have been a number of recent publications that have attempted to address these issues (OECD, 2000; Tunney et al., 1997).

In Ireland the Morgan test is used to estimate STP. Recent field experiments in at Johnstown Castle show a good relation between STP and dissolved reactive P (DRP) measured in all the surface runoff, during one year, from four grazed grassland fields. Results from four farm drains elsewhere in Ireland were also consistent with this relationship. Published data from Denmark, UK and USA on the STP versus water P relationship has employed the Olsen and Mehlich 3 tests to determine STP. These test have been converted to equivalent values of Morgan STP, using the equations: Olsen = 5.96 Morgan0.773 and Mehlich 3 P = 8.52 x Morgan P0.85 (Tunney, 2001). When converted to Morgan STP values, data from the other studies are consistent with the Irish relation of STP to water P. The overall relation is DRP = 0.002(Morgan STP)2, R2 = 0.92, where DRP is the annual mean concentration (mg P L-1) and Morgan STP is expressed as mg P L-1.

The relation is non-linear, but DRP consistent with for good water should be feasible at the lower end of the range of STP of 3 to 6 mg P L-1 for optimum grassland production. Indeed it would appear that good water quality might only be possible if soil P were reduced to these levels. However, STP is only one factor in determining P losses to water. Data from a continuously monitored drain site in Northern Ireland, with a Morgan STP equivalent of 8 mg P L-1, shows a wide variation in mean annual DRP concentrations and is consistent with many other studies (Lennox et al., 1997). These high values reflect the impact of manure P additions, followed by drain-flow and winter grazing of animals. Other studies in England have also argued the need to account for these ‘incidental’ losses of manure P as well as detached soil particles and colloids as part of a holistic approach to the problem (Haygarth and Jarvis, 1999).

The National Program for Sustainable Development in Ireland includes a target to reduce fertiliser P use by 10% per year, over five years. This will help reduce the current P surplus in Irish agriculture, which imports large amounts of P in animal foodstuffs, and perhaps reverse the national increase in soil P. In recognition of the importance of incidental losses of P, attention is increasingly being focused on measures that will reduce their impact by avoiding applications when the hydrological transport potential is highest in winter.


Haygarth, P. M. and Jarvis, S. C. (1999) Transfer of phosphorus from agricultural soils, Advances in Agronomy, 66, 195-249.

Lennox, S.D., Foy R.H., Smith R.V. and Jordan C. (1997) Estimating the contribution from agriculture to the phosphorus load in surface water. In: Tunney et al., (1997), Pages 55-75.

OECD (2000) Papers presented at OECD Workshop Antrim,Northern Ireland, June 1998. Journal of Environmental Quality, 29:1-176.

Tunney, H. (2001) Phosphorus needs of grassland soils and loss to water. Publ: International Association of Hydrological Sciences Journal, Wallingford, UK. (In press and due to be published in early 2001).

Tunney, H., Carton, O.T., Brookes, P.C. & Johnston, A.E. (1997) Phosphorus Loss from Soil to Water. Publ: CAB International, Wallingford, Oxon. UK. 467.


Variation in leaching of phosphorus in relation to soil phosphorus status

S.M. Vandsemb1, Marianne Bechmann1, Tore Krogstad2

1Center for Soil and Environmental Research (Jordforsk), Ås, Norway

2Department of Soil and Water Science, Agricultural University of Norway (NLH), Ås, Norway

The aim was to study the effects of the soil phosphorus (P) status on leaching of P in tile drains in one Norwegian agricultural catchment in the county of Akershus (Vandsemb, 2000). Phosphorus in the water samples was compared to soil P content in samples collected from the corresponding catchment area.

The study is based on 20 soil samples and 52 water samples. The soil samples were taken from silt and clay soils with and without use of manure. Water samples were collected from drainage pipes throughout seven drainage events.

The water samples were analysed for water soluble P (PWater). Soil P were measured as PAL (extracted by ammonium acetate lactate), POlsen (extracted by 0,5 M NaHCO3), PBray (extracted by 0,03 M NH4F and 0,025 M HCl) and PCaCl2 (extracted by 0,01 M CaCl2, soil:solution ratio of 1:20).

A slightly better correlation was obtained when PAL was correlated to PBray (r2 = 0,52) as when correlated to PCaCl2 and POlsen (r2 = 0,47). Comparing P concentrations in tile drain with the P status in soils gave some indications of what can be expected of P loss from different soils and use of farmyard manure. There were tendencies that drainage water from soils with high PAL and POlsen values and soils with manure gave high concentrations of PWater. The highest concentrations of PWater were obtained from silt soils.

Through a 14 day experiment P was extracted with 0,01 M CaCl2, using a soil:solution ratio of 1:10. The 0,01 M CaCl2 were daily separated from the soils and fresh liquids were refilled. The availability for P leaching over time was estimated. Soils with high PAL values and soils with farmyard manure generally gave more PCaCl2 during this leaching experiment. Subsoils gave only small amounts of PCaCl2. The results suggest that high soil P content can cause high risk of P-leaching.


Vandsemb, S.M. 2000. Variation in leaching of phosphorus compared to soil phosphorus status in soils in the Mørdre catchment. Master thesis at Department of Soil and Water Science, Agricultural University of Norway (NLH).

Remediation measures to reduce phosphorus losses from overfertilised grassland soils – phosphorus losses via surface runoff

T. Vollmer1, M. Schaerer2, C. Stamm1, S. Sinaj2, E. Frossard2 and H. Fluehler1

1Institute of Terrestrial Ecology, Soil Physics, ETH Zurich, Switzerland

2Institute of Plant Sciences, Plant Nutrition, ETH Zurich, Switzerland


As in many developed countries, programs are planned in Switzerland to reduce diffuse phosphorus (P)-losses from agricultural land to surface waters. Permanent grassland soils in livestock-breeding areas are especially critical since large P-surpluses have accumulated in the topsoil due to high (liquid) manure application rates.

However, the efficiency and practicability of the proposed measures are mainly hypothetical. Data, demonstrating the consequences of the measures as well as the time needed to observe the desired effects, are lacking. Therefore, we have established two field trials, each of them consisting of 15 study plots of 65 m2 with the goal to investigate the processes influencing P-losses at the plot scale.

Study sites and experimental setup

The study sites are located on two slopes in the watershed of Lake Greifensee in the Swiss Plateau, 25 km north-east of Zurich. The grassland soils of the study sites are rich in P with 1.9 g total P kg-1 soil in the top two cm at site I and 1.4 g kg-1 total P at site II, respectively. The background concentrations in the lower parts of the profile are 0.3 g total P kg-1 soil at site I, and 0.4 g total P kg-1 soil at site II, respectively. The availability of P in the topsoil is high with 58 and 27 mg P kg-1 soil that is isotopically exchangeable within 1 min (E1min) and Olsen P values of 126 and 69 mg P kg soil-1. These values decrease rapidly with depth. At site I, the soil cover consists of neutral to slightly acidic soils with soil texture sandy-clayey loam (calcaric Regosols, calcaric Cambisols, Phaeozems), whereas at site II acid soils (pH 4.2 – 5.3 in the topsoil) with soil texture sandy clay (dystric Cambisols, Luvisols) predominate.

The experimental measures under study will include (i) a control with normal P-input, (ii) zero P-input, (iii) a single tillage operation with zero P-input and (iv) zero P-input + a single tillage operation + amendments (to increase soil P sorption capacity). The monitoring of P-losses includes sprinkling experiments inducing surface runoff and interflow in the topsoil on 15 subplots of 2.25 m˛ as well as runoff measurements under natural rainfall conditions (5 subplots of 20 m˛).

The remediation measures will be applied to the plots of 65 m2 with three or four replicates per measure at both sites, starting in spring 2001. In the year 2000, the 'status quo' of P-losses from the studied plots was evaluated.


Results from the first monitoring year show that the chosen study sites are potential source areas for the eutrophication of surface waters with respect to P-mobilisation by runoff water: Sprinkling with deionized water on all 15 plots at each study site produced runoff (combined surface runoff and shallow interflow) containing up to 1.0 mg total P L-1 at site I (mean ± standard deviation: 0.8 ± 0.2 mg P L-1) and up to 1.5 mg total P L-1 (mean ± standard deviation: 0.9 ± 0.3 mg P L-1) at site II. These concentrations agree fairly well with those in runoff induced by natural rainfall, where we measured total P concentrations of up to 0.8 mg P L-1 (mean ± standard deviation 0.6 ± 0.5 mg P L-1, means of 5 events) at site I and up to 1.6 mg total P L-1 (mean ± standard deviation 1.0 ± 0.6 mg P L-1, means of 4 events) at site II. Dissolved reactive P was the main P fraction in all analysed samples of runoff water (sprinkling experiments and natural runoff).

At site I, concentrations of dissolved reactive P (DRP) in runoff induced by artificial and natural rainfall were positively correlated to soil P availability (Olsen P, E1min), whereas at site II this was not the case. Furthermore, the DRP concentrations in runoff were slightly higher at site II than at site I, despite the lower contents of available P (Olsen, E1min) at site II. At both sites, DRP concentration in runoff induced by artificial rainfall was negatively correlated to the time that passed since the grass on the experimental plots was cut.

These observations show that not only P-concentration and –availability in the soil, but also the (small scale) processes of water flow in the topsoil, the interaction of water and soil P at the pore-scale and management measures like grass cutting strongly influence P-losses from soils to water. This will be examined in more detail in the future.


Phosphorus release kinetics as affected by aqueous solutions

1D. T. Westermann, D. L. Bjorneberg, B. L. Turner and J. K. Aase

1NWIRSRL, USDA-ARS, Kimberly, Idaho, USA

Irrigation runoff phosphorus (P) concentrations are affected by spatial, temporal and hydrological factors, as well as physical and chemical interactions between soil and water. This study evaluated the P concentration in aqueous solutions of different ionic strengths and composition across time. Western U.S., irrigated soils were selected to span a wide range of soil P availabilities and other physical and chemical properties, while the aqueous solution varied from distilled water to relatively salty irrigation water. In addition, the solution:soil ratio was varied to bracket that expected in irrigation runoff. Filtered (0.2 μm) solution samples were taken from rotating plastic drums held at constant temperature (20oC) and analyzed for P and other ionic constituents by ICP-OES. In general, 75% of the final phosphorus concentration occurred after the first 15 minutes of soil-solution contact. Near equilibrium was achieved after two hours contact. Increasing salt concentration decreased phosphorus concentration compared with distilled water, with divalent cations having more effect than monovalent. Phosphorus released per unit soil mass decreased as solution:soil ratios decreased. Selected release kinetic parameters will be related to soil and water properties, and discussed in relationship to P transfer processes for different soil/water conditions.

Dairy slurry effects on phosphate retention and release in four UK grassland soils

P. Whitehead1, D. R. Chadwick1, P. M. Haygarth1 and J. S. Robinson2

1Institute of Grassland and Environmental Research, North Wyke, Okehampton, Devon, EX20 2SB, UK.

2Department of Soil Science, The University of Reading, Whiteknights, PO Box 233, Reading, RG6 6DW, UK


With a surplus of phosphorus (P) in most UK grassland systems , there is potential for P to accumulate in, or on the surface of soils, thus increasing the potential for transfers to watercourses. On UK dairy farms, 65% of manure produced is stored and spread as dairy slurry most of which is applied to grassland soils . This is the main pathway for recycling P back to soil. For such a vast resource, the knowledge of interactions between soil and dairy slurry are limited, particularly the retention and release of P.


This paper reports the findings of a laboratory investigation into the role of dairy slurry on the retention and release of phosphate in four UK grassland soils. Phosphate, as determined by the molybdenum blue method , was measured because it constitutes the greatest source of P in dairy slurry. Diluted dairy slurry (DSM – 10x dilution and sieved < 1 mm) and 0.01M CaCl2 (CCM) solution matrices were used to measure differences in phosphate retention. The solution matrices were added to the soils, at a 10:1 solution:soil ratio (v/w), to supply a range of phosphate concentrations (40 – 500 mg L-1 by the addition of KH2PO4). Equilibrium between soil and matrix solution was accomplished by gently shaking 72 hours in the presence of a microbial inhibitor at 20şC (± 2şC). The soils include: a silty clay soil (USDA - Typic Haplaquepts), a clay loam soil (USDA – Dystochrepts), a sandy loam over chalk (USDA – Hapludalfs), and a sandy soil (USDA – Udipsamments). Differences in the patterns of phosphate retention were investigated by constructing quantity / intensity (Q/I) curves which were modelled using the Freundlich equation. A measure of the availability of retained phosphate was made subsequently by performing three 0.01M CaCl2 sequential extractions.

Results and Discussion

In all soils, phosphate retention was significantly greater (P<0.01 measured by parallel curve analysis) from the DSM compared to the CCM. The derived values, after applying the Freundlich equation to Q/I data suggest that the buffering capacity for each soil is greater for phosphate retained from the DSM. This was shown in the values for a constant, which can be related to the buffering capacity of a soil which showed a 34, 125, 66 and 121% increase for the silty clay, clay loam, sandy loam over chalk and sandy soil respectively. At lower phosphate additions (<200 mg P l-1) all four soils retained a greater proportion of the added phosphate from the DSM than CCM. At the lowest phosphate concentration (40 mg l-1) the silty clay, clay loam, sandy loam over chalk and sandy soils showed a 10, 19, 20 and 7% increase, respectively, in the amount of phosphate retained from the DSM compared to the CCM.

During the subsequent, sequential 0.01M CaCl2 extractions, less phosphate was extracted in the soils treated with the DSM compared to the CCM, in spite of there being more phosphate retained in the former treatment. The enhanced phosphate retention in the four grassland soils from the dilute dairy slurry may be a result of the much higher ionic strength of the solution from which phosphate has been retained . The presence of some of the constituents of dairy slurry (e.g. Fe, Al, Ca), which have been shown to be responsible for some phosphate retention in soils, may also play a role.


These data suggest that in all the four soils, the phosphate retention capacity, and buffer strength for phosphate applied as dairy slurry are greater than when added as an inorganic source. Investigations are continuing to ascertain the mechanisms that give rise to enhanced phosphate retention in these four soils.


Theme 2 - Hydrochemical connectivity

- Oral papers


Hot spots and short cuts causing phosphorus losses into open waters?

C. Stamm.1, A. Fraser 2, P. Lazzarotto 1, 3, V. Prasuhn 3

1 Institute of Terrestrial Ecology, Swiss Federal Institute of Technology Zurich, Grabenstr. 3, CH-8952 Schlieren

2 SSLRC North Wyke, (Cranfield University), Okehampton, Devon EX20 2SB, UK

3 Swiss Federal Research Station for Agroecology and Agriculture Reckenholzstr. 191, CH-8046 Zurich

Phosphorus (P) losses from agricultural soils to surface waters are highly variable in space and time. This paper focuses on relevant factors and mechanisms controlling these heterogeneities. It has been recently demonstrated that natural and artificial preferred flow paths, acting as hydrological short cuts, may be important in determining P transfer. Results supporting this view are presented in this paper in addition to evidence indicating that this flow mechanism is of limited relevance under certain conditions. For example, the connectivity of preferred flow paths through the entire flow domain. In order to improve our understanding of the importance of different hydrological pathways we address the interplay between soil structure, climatic conditions and P status of selected soils. For this purpose the dynamics of P losses from two agricultural study sites (CH, UK) under different conditions are compared. At the Swiss study site (Lake Sempach region), a few large flow events dominated annual P losses. The results from the Rowden Experiment in Devon revealed the opposite in that the many smaller events were of major importance.

This paper also focuses on an evaluation of the spatial distribution of P losses, to identify whether only a few hot spots, or critical source areas, within a catchment determine P transfers or if many fields contribute to the overall P export. Results from an ongoing monitoring study in Switzerland sheds some light on this aspect. In this project we collect data for assessing the risk for P losses due to natural conditions, agricultural practices individual events (manure application, precipitation) for each of the 270 fields in the catchment.

Phosphorus transport in upland catchments - connectivity of field and stream

W. J. Gburek, B. A. Needelman and M. S. Srinivasan

Pasture Systems and Watershed Management Unit, USDA-ARS, University Park, PA, USA

Development of strategies for assessment and management of nutrient loss from agricultural land should consider and incorporate the critical source-area (CSA) concept. In the case of phosphorus (P) loss specifically, CSAs result from the co-location of areas of high available P source (i.e., high soil test P and/or high rates of fertilizer and animal manure application) and high potential for P transport to a water body (catchment areas generating runoff and erosion). Much research has been conducted at the laboratory, plot, and field scales to characterize the source aspects of P availability to runoff as a function of soil test P levels and amounts and methods of fertilizer and manure application. However, only limited research results are available to quantify the transport aspects of P loss incorporating the spatial and temporal variability of surface runoff generation and its relationship to streamflow.

Here, we address the transport aspect of the formation of phosphorus CSAs – water movement at the landscape/catchment scale – with the primary hydrologic control on P transport considered to be surface runoff and its associated water-borne sediment. Although subsurface P transport can exist under limited conditions, we will not address it here. We view the P transport potential as being composed of two factors: 1) initiation of runoff at specific and identifiable locations over the landscape, and 2) the 'connectivity' of these runoff-producing areas with a downgradient water body where the impact of P loss from the landscape may be manifested.

The catchment-oriented research community has come to generally accept the concept of variable-source-area (VSA) hydrology. Its basic premise is that there is a dynamic runoff-contributing subcatchment within the topographically defined catchment. The contributing subcatchment expands and contracts seasonally, as well as during a storm, as a function of precipitation, topography, soil, geology, ground water levels, and catchment moisture status. Much of the research related to VSA hydrology has focused on saturation-excess runoff as the dominant flow component. When concerned with P transport across the landscape however, we cannot neglect the possible impact of catchment areas that generate runoff by the infiltration-excess mechanism. While perhaps not as widespread areally and temporally as is saturation-excess runoff under the humid-climate conditions addressed here, these areas still have the potential to develop and contribute to streamflow under the proper combination of soil, precipitation characteristics, and connectivity to the stream.

We are attempting to link the concept of VSA hydrology with expressions of causative precipitation, landscape topography, soil properties, and ground water status, to express both the generation of surface runoff over the landscape, and the 'connectivity' between these areas of surface runoff generation (e.g., fields or actual points) and the downgradient water body potentially affected by P loss. Here, we report our efforts to characterize VSA hydrology by field-, landscape-, and catchment-scale studies of surface runoff generation processes and the connectivity of runoff source-areas with the stream. These studies encompass both mechanistic hydrology suitable for incorporation in process-based catchment models (e.g., AnnAGNPS, SWAT), and design hydrology more applicable to user-oriented tools for P management (e.g. P Index). To the extent possible, the latter research thrust considers extrapolation of landscape-specific results to the catchment scale based on design storms, interpretive soil survey maps, available DEMs, and catchment-scale runoff and geomorphic characteristics.

Locations, along with time and space scales, of runoff source-areas are illustrated with specific research results from the Mahantango Creek Watershed in east-central Pennsylvania, USA. Results of studies examining runoff mechanisms as controlled by precipitation, initial moisture status, topography, detailed soil properties, and catchment-scale ground water status are presented. The results of these studies are also extended to interpretation of spatial and temporal patterns of P concentration in the stream during storm hydrographs to illustrate the connectivity aspect of P transport within the catchment. Additionally, using design storm characteristics and catchment-scale runoff and geomorphic characteristics, we illustrate how these catchment-specific results can be extended to the ungaged catchment setting in such a manner that they might be applicable to a user-oriented tool such as the P Index. Finally, we consider the implications of these results, both mechanistic and design oriented, for development of land-management strategies to mitigate P loss.

Pathways for phosphorus loss off pastures in South Australia

J. W. Cox1, N. K. Fleming2, D. J. Chittleborough3, J. C. R. Varcoe3 and D. P. Stevens1

1CSIRO Land and Water, Urrbrae, South Australia

2South Australian Research and Development Institute, Urrbrae, South Australia

3Department of Soil and Water, Adelaide University, Urrbrae, South Australia


Phosphorus (P) is thought mobile only in sandy or P-saturated soils or those where the active sites for P adsorption are coated with organic matter. The findings presented are from pasture catchments with texture-contrast soils (loams or sands over clays) and a Mediterranean climate (winter dominant rainfall). Variability in P mobility in these catchments was due to specific chemical and physical properties of the texture-contrast soils, which determined the dominant pathway (lateral or vertical) for water.

Importance of lateral mobility of phosphorus

Perched aquifers did not always develop within the A horizon (on the subsoil clay) but occurred within the B or C horizon. Preferential horizontal pathways (with high hydraulic conductivity) often occurred within the B or C horizon clays, resulting in substantial throughflow and P mobility at depth. Also, in some years, deep, perched watertables rose into more permeable horizons, contributing to lateral flow and P mobility.

In about half the catchments monitored, overland flow was the dominant pathway for water and P with approximately equal amounts of throughflow in the A and B horizons (Stevens et al., 1999). These catchments generally had shallow loamy topsoils with high Ksat and B horizons with high macroporosity and preferential lateral flow. About 20% of the catchments had little overland flow and most lateral flow was subsurface flow. These catchments usually had deep, sandy topsoils with very high Ksat and B horizons with high macroporosity and preferential vertical and lateral flow. About 30% of the catchments had predominantly overland flow. The topsoils in these catchments were usually compacted and B horizons heavy textured. However, the soil profile had to be saturated before flow commenced. High production, intensive dairy catchments often fell into this category (Fleming et al., 2001).

Major findings with regards to lateral P mobility included: (i) the importance of overland flow and P mobility significantly increased when annual rainfall increased above 500 mm; (ii) there was a strong curvilinear relationship between annual P mobility and annual rainfall for a given soil P concentration; (iii) at the annual scale, particulate P dominated the total loss in runoff in wet years (though dissolved P, < 0.45 m m, was a high proportion of the total) and dissolved P dominated in dry years; (iv) at the seasonal scale, dissolved P remained fairly constant throughout the season whereas particulate P decreased over the year. Particulate P was only dependent on runoff volume early in the season; and (v) at the storm scale dissolved P was dependant on land slope and runoff volume whereas particulate P was dependant on peak flow and storm intensity.

Importance of vertical mobility of phosphorus

Phosphorus isotherms, which are commonly used to determine those soils that impede P mobility, were found to be a poor indicator of P mobility in most texture-contrast soils (Cox et al., 2000). Soil chemical properties (pH, electrical conductivity, exchangeable cations, adsorption isotherms, cation exchange capacity, and exchangeable sodium percentage) could at best only explain 24% of the variance in P mobility. Soil physical properties (bulk density, particle size, and saturated hydraulic conductivity) explained only 21% of the variance in P mobility Residence time (which is a function of hydraulic conductivity of the pathway) was an important factor in determining P mobility. The residence time of P was reduced by macropore flow because some macropores were sufficiently wide that P did not contact soil adsorption sites. Thus macropore flow can effectively facilitate the movement of P through a soil even when the soil contains a significant amount of clay and has a high P adsorption capacity. About 90% of the vertical mobility of P could be explained by combining the chemical and physical parameters of sodicity, saturated hydraulic conductivity, total exchangeable cations, and macroporosity.


Cox, J.W., C.A. Kirkby, D.J. Chittleborough, L.J. Smythe, and N.K. Fleming 2000. Mobility of phosphorus through intact soil cores collected from the Adelaide Hills, South Australia. Australian Journal of Soil Research 38:973-990.

Fleming, N.K., J.W. Cox, D.J. Chittleborough, and P. Dyson 2001. Prediction of chemical loads and forms in overland flow from dairy pastures on duplex soils in South Australia. Hydrological Processes (in press).

Stevens, D.P., J.W. Cox, and D.J. Chittleborough 1999. Pathways of phosphorus, nitrogen and carbon movement over and through texturally differentiated soils, South Australia. Australian Journal of Soil Research 37:679-693.

Modelling sediment delivery from agricultural land to the fluvial system via sub-surface drainage

I. D. L. Foster1, A. S. Chapman1, J. A. Lees1, R. J. Hodgkinson2 and R. H. Jackson1

1Centre for Environmental Research and Consultancy, School of NES, Coventry University, Coventry, UK, CV1 5FB

2ADAS Gleadthorpe, Meden Vale, Mansfield, Notts, UK, NG20 9PF

The delivery of sediment and particulate phosphorus (P) to surface waters from agricultural land is an increasing cause for concern. Sub-surface drains (referred to as field drains or tile drains) can deliver large quantities of sediment to the fluvial system under appropriate conditions. The presence of soil cracks, wormholes and gaps left by plant roots, collectively known as macropores, can provide a network of channels within the soil profile which enables rapid delivery of sediment and water to the drainage system while undergoing little physiochemical change. This study, based on two field sites in the English Midlands, uses a range of environmental tracers and a Simplex linear algorithm to identify the provenance within the soil profile of drain sediment collected from four field drains.

At the most intensively monitored land drain (Foxbridge, at ADAS Rosemaund, Herefordshire) sediment yields approached 100 t km2 yr-1 in 1998 and 1999. At all four sites topsoil was the dominant source of sediment, forming between 75% and 85% of the total sediment load. In one site contributions from an individual soil series varied from 0% to 45% of sediment transported. Variation in topsoil sources over the monitoring period was attributed to greater soil cracking caused by excessive soil moisture deficits during the summer of 1998. Given the propensity of PP to be concentrated in the upper ca. 30 cm of the soil profile, the high proportion of topsoil sediment present in the drain samples suggests that sub-surface drainage can be a significant pathway linking agricultural land to surface waters.



Phosphorus in fluvial systems: speciation, transformations and dynamics

I. D. McKelvie.1, N. Amini1, P. Gardolinski3, M. R. Grace1, M. E. Hindle1,2, B. Lovell1, and P. J. Worsfold3

1Water Studies Centre, School of Chemistry, Monash University, Clayton, Victoria 3800, Australia

2Present address: Environment Protection Authority, Victoria, Southbank, Victoria 3006, Australia

3Department of Environmental Sciences, Plymouth Environmental Research Centre, University of Plymouth, Plymouth PL4 8AA, UK

Environmental interest in phosphorus (P) derives mainly from its role as a limiting nutrient in the process of eutrophication. Orthophosphate is known to be the most biologically available phosphorus (BAP) species, and is readily utilisable by phytoplankton and bacterioplankton . Measurement of filterable reactive phosphorus (FRP) is therefore often used as a de facto estimator of the concentration of bioavailable phosphorus present.

This paper considers the nature and behaviour of P in fluvial systems, with an emphasis on the ultimate bioavailability of the various forms of P. Phosphorus species can be classified according to their chemical characteristics (orthophosphates, condensed and organic phosphates), their biological activity (bioassays, or enzyme reactivity), or according to their size or molecular mass. Once in the water column, P in its various forms may undergo abiotic adsorption or biotic uptake, processes that have typically been investigated by adsorption studies on both unaltered and sterilized sediments. Organic and condensed phosphates may also undergo conversion to orthophosphate either through abiotic or microbial hydrolysis or arguably even through photochemical oxidation reactions.

The dynamic behaviour of P in fluvial systems may involve a combination of uptake, storage, release and transport. In small streams this may be characterized by the spiral length obtained from so-called 'nutrient spiralling' experiments . In larger rivers and estuaries there is also the potential for P release from the sediments under anaerobic conditions . Investigations of this behaviour may typically involve the use of in-situ benthic chambers, the measurement of phosphate fluxes from manipulated sediments in sediment core reactor studies, or sorption-desorption experiments in the laboratory. The use of peepers and diffusive gradient techniques can also provide valuable information on the potential for phosphate release from sediment pore waters through diffusion, bioturbation or other disturbance.

The emphasis of most studies of P transport and dynamics to date has been on the behaviour of the readily bioavailable orthophosphate species. However, there is evidence that dissolved organic P compounds can be utilised directly by phytoplankton , but because measurement of this fraction or its components is often complicated, it has largely been ignored. Recent work from our own laboratory suggests that substantial amounts of various organic phosphates may be released under estuarine mixing conditions or conditions of low redox potential. The importance of organic phosphorus as a source of BAP and as a major component of the aquatic phosphorus cycle will be discussed.


Runoff from sheep pastures in south-west Victoria: phosphorus concentrations and forms

A. R. Melland1, R. E. White1, M. R. McCaskill2 and D. F. Chapman1

1Institute Of Land and Food Resources, The University Of Melbourne, Parkville 3010, Australia

2Pastoral and Veterinary Institute, Department of Natural Resources and Environment, Hamilton, Victoria 3300, Australia


Meat and wool producers in south-west Victoria, Australia, are increasing phosphorus (P) fertiliser application rates to boost production. There is concern, however, that this may lead to higher nutrient levels in surface runoff and catchment waterways. An excess of nutrients, especially P, in streams and reservoirs increases the likelihood of problems such as blue-green algal blooms. Dryland pasture is the dominant landuse in many catchments and its management can have a large influence on surface water quality. This prompted an investigation into the impact of local fertiliser and grazing management practices on the quality of surface runoff water.

Materials and methods

Surface runoff was isolated from four 0.5 ha phalaris/subterranean clover paddocks located 55 km NNW of Hamilton,Victoria, Australia. Paddocks were grazed by Merino ewes using either set-stocked (SS) or rotational grazing (RG) management. Fertiliser application rates, and the Olsen extractable P values measured in spring, are shown in Table 1. Runoff was measured from August 1998 until November 2000 using large tipping buckets connected to a datalogger. Composite flow-weighted subsamples were collected and filtered (<0.45 mm) within 24 h of a runoff event. These were analysed (immediately in 2000 or after storage at -15oC in 1998) for molybdate-reactive P, representing the dissolved inorganic P. An unfiltered subsample represented the total inorganic P (TRP) fraction. Unfiltered and filtered samples were stored at –15oC prior to analysis of total P (TP) and total dissolved P (TDP), respectively. These were measured using a persulfate autoclave digestion and flow injection colorimetric analysis.

Results and discussion

Table 1. Soil fertility details, and quantity and quality data for runoff in 1998 and 2000


Fertiliser applied kg P ha-1

Olsen P

mg P kg-1 soil



TP conc‡

mg L-1


















Low P, SS













High P, RG













High P, SS













High P, RG













† Conservative estimates were made when events exceeded the monitoring capacity of the tipping bucket

‡ Flow-weighted mean TP concentration for total annual runoff

Total rainfall in mm was 589, 550 and 575 in 1998, 1999 and 2000, respectively. No runoff occured in 1999.

Total P concentrations in runoff ranged from 0.19 to 1.28 mg P L-1 across all treatments (Table 1) and were 1.5 to 5 times greater in runoff from paddocks receiving the higher P rate compared with the low. All concentrations were at least four times above the standard for healthy streams in Victorian inland waters (0.05 mg PL-1). Runoff from the three pastures with highest Olsen P values tended to have a greater proportion of dissolved and inorganic P than the low fertility pasture. This trend was consistent with results for runoff from other fertile pastures (Nash and Murdoch, 1997). Up to 80% of the total flow in each year was generated from waterlogged land which comprised approximately 2% of the total hillslope area. The variability in flow volumes across the hillslope suggested reductions in P loss may be achieved through targeting runoff source areas for nutrient management.

In was concluded that surface runoff was a significant pathway for P movement from grazed pastures in south-west Victoria, with TP concentrations high enough to be of concern if transferred to receiving waters without dilution. The dominance of dissolved P in runoff from the more fertile pastures would reduce the potential for vegetative buffer strips to entrain particulate P in runoff. Attention should be focussed on reducing P availability in areas yielding high volumes of runoff.


Funding for this project was provided by the Wool and Science Quality programs of the Victorian Department of Natural Resources and Environment.


Nash, D., and C. Murdoch. 1997. Phosphorus in runoff from a fertile dairy pasture. Aust. J. Soil Res. 35:419-429

Sediment – phosphorus interactions: a synthesis

D.S. Baldwin1, A. M. Mitchell1, J. Olley2 and G. Rees1

1Murray-Darling Freshwater Research Centre and, The CRC for Freshwater Ecology, PO Box 921 Albury Australia 2640

2CSIRO Land and Water, PO Box 1666, Canberra ACT 2601, Australia

Sediments are complex mixtures of minerals, organic material and biota found suspended in the water column and at the bottom of water bodies. They can vary in size from coarse sands and gravels through to very fine clay-rich muds. Although ostensibly derived from soils and rocks, sediments often behave very differently from their terrestrial counterparts. These differences can be attributed to extensive microbial activity reworking organic and inorganic constituents of the sediments, and oxidation and reduction reactions resulting in changes to the surface coatings of the sediment grains.

Sediments play a number of key roles in the movement and transformation of phosphorus (P) derived from agricultural practice. Phosphorus can form strong associations with sediment particles so that the fate of the P is often intimately linked with the fate of the sediment. Sediment deposition may be an important pathway for the removal of the P from the water column. Conversely, sediment re-suspension and transport can be an important mechanism for the movement of P through an aquatic ecosystem. However, sediment-P interactions often go beyond simple sorption and transport processes. Biotic and abiotic processes associated with sediments may transform phosphorus compounds. Clearly, the interplay between the sediment and P can be quite complex. In this presentation we would like to present a synthesis of how the dynamics of P derived from agricultural practises is moderated by both suspended and bottom sediments.

In this presentation we argue that P sorption onto sediments followed by sediment deposition and then slow re-release of the pollutant back into the water column can be seen as buffering the potential impact of episodic pollution events. However, this buffering also serves to prolong the impact of the P in the aquatic environment. For example, removal of P inputs into lakes often does not result in a noticeable decrease in P concentrations in lake water, because of re-release of P from the sediments. Therefore, sediments can be seen as temporally shifting pollutant impacts. Interaction with sediments can also shift the impact of a P spatially. A pollutant can enter a waterway, become attached to sediment particles and be transferred a long distance downstream before the sediments are deposited and re-release of the P occurs.



Using water table and weather data to identify fields prone to overland flow

D. Ryan1 and J. Bennet2

1Teagasc, Oak Park Research Centre, Carlow

2Teagasc, Johnstown Castle Research Centre, Wexford


Data gathered by the EPA strongly suggests that Agriculture is the biggest source of phosphorus (P) in Irish rivers and lakes (McGarrigle 1999). Phosphorus is carried from fields into watercourses by overland flow (Sherwood and Fanning, 1981). The risk of pollution is particularly high if slurry or fertiliser lie on the surface. There is a need for an economic method to identify land prone to overland flow so that the land can be managed to minimise risk. Previous research was based on large models requiring expensive data to calculate overland flow (Lewis & McGechan, 1999; Provolo, 1995). In this study easily obtainable weather and soil water measurements are used in a simple model (WT) to identify problem sites.


Field measurements were made during seven months on three sites in the South East of Ireland. All three fields had drainage problems, and two of them had seepage or spring water. A network of water- table tubes was installed and monitored every three to four days. Rainfall and evaporation data were gathered every day of the period. A V-notch flow meter and collecting berm provided continuous measurements of overland flow. The water table and weather data were entered into the WT model. This model is based on a water balance on the air-filled pore space in the soil. When the pore space is full, the water table is at the surface and overland flow takes place. In a related trial maximum water level indicators, consisting of a polystyrene float in a plastic tube, were assessed. The indicators were compared to tape measurements in the laboratory. In the field, neither tape measurements nor recording by logger were feasible, so two measurements from adjacent indicators were compared to one another.


The WT model calculates the volume of air-filled pore space in the soil. The other parameters are calculated from this. A value for water table level was calculated for each of the 212 days of the period. This data matches closely the field measurements giving a regression coefficient ‘r’ for all three sites of 0.92 (p< 0.001). Overland flow values for all the pipes in a site were combined to yield total flow from the site. Model output is compared to measurements in Table 1

Table 1: Regression of overland flow by the WT model on measured values at 3 sites (mm)





















This is good agreement in this kind of measurement. The WT model has dealt successfully with downward and upward flow regimes. It should be tested further on other sites and in other hydrological conditions.

In the preliminary assessment, the max level indicators in the laboratory differed from tape measurements by 9 +/- 1.7 mm (p< 0.05). In the field, the difference between two indicators was 2.3 +/- 13.5 mm (p<0.05). An error closer to 3.5 mm was expected. This did not seriously affect the operation of the model. The use of max level indicators may allow fieldwork to be reduced. Tests with other labour saving devices are ongoing.


McGarrigle, M. 1999. Mimeograph report. EPA, Johnstown Castle Estate, Co. Wexford, Ireland

Theme 2 - Hydrochemical connectivity

Phosphorus transformations down two farm drains

K. Barlow1,2, D. Nash1, R. Grayson2 and H. Turral2

1Agriculture Victoria Ellinbank, RMB 2460 Hazeldean Road Ellinbank, Vic. 3821, Australia

2Department of Civil and Environmental Engineering, University of Melbourne, Parkville, Vic. 3052, Australia


Phosphorus (P) exported in overland flow from irrigated agriculture contributes to the decline in water quality of inland water systems. Investigations of P transfer tend to have focussed on the paddock scale and in-stream processes with little work investigating the transformations that occur as P moves down surface drains from the paddock to reuse ponds, community drains and river systems. This paper investigates the potential of surface drains to affect P export from irrigated dairy farms in the Macalister Irrigation District (MID) of south-eastern Australia. In particular it focuses on whether (i) P concentration changes significantly along farm drains, and (ii) drain design affects the changes in P concentration.


Two surface drains in the MID (38° 00’ S, 146° 54’ E) were used to investigate P transfer processes in farm scale drains. The first drain was managed as part of a paddock containing a perennial ryegrass and white clover pasture, this drained into a second recently cleared bare earth drain. The grassed drain had a 1:370 slope and it was 1.5 m wide by 180 m long. The earthen drain had a 1:800 slope and it was 3.5 m wide by 180 m long. For experimental purposes the drains were isolated from other water sources.

Channel water (<0.1 mg P L-1) was piped into the top of the grassed drain at a rate of 21± 1 L sec-1. Flow was monitored at the top of the grassed drain and bottom of the earthen drain using 200 mm RBC flumes and Wesdata 392 Capacitance probes (Dataflow systems Pty. Ltd., Aust). The intersection of the two drains was monitored using a 300 mm RBC flume with an ISCO storm monitoring system comprising a model 3700 auto sampler and a 4230 bubbler flow meter (ISCO inc., USA).

Water samples were collected at the wetting front at predefined points down each drain. Samples were also collected behind the wetting front at specific time intervals. Water samples were filtered through 0.45 m m filters and analysed for Total P (TP), Total Filterable P (TFP) and Filterable Reactive P (FRP).

Results and discussion

Phosphorus concentrations changed (p<0.01) in the wetting front as it moved down the drains (Table 1), increasing down the grassed drain and decreasing down the earthen drain. The increase in P concentration along the grassed drain was consistent with trends on flood irrigation bays, where presumably the wetting front takes up readily soluble P from live and decaying pasture soil and faeces. Fertiliser was not expected to contribute as the paddock and drain had not been fertilised in over 10 months. A decrease in wetting front P concentrations at the intersection of the drains was probably a response to the storage and dilution of the wetting front as a result of the monitoring equipment. Changes in the proportion of TFP to TP at the intersection suggests that the settling of colloidal material may also be contributing to the decrease in TP observed. The decrease in P concentration down the earthen drain was most likely due to the uptake of P by drain sediments. Whether this uptake is a temporary or permanent process requires further investigation.

Phosphorus concentrations declined in an exponential fashion behind the wetting front. The asymptote of the decay at each sampling point suggests that similar trends (increasing P along the grassed drain and decreasing P along the earthen drain) were observed in and behind the wetting front.

Table 1. Total phosphorus concentration in the wetting front with distance down drains A and B


Distance (m)







Drain A


(mg P L-1)








(mg P L-1)






Drain B


(mg P L-1)








(mg P L-1)








31Phosphorus nuclear magnetic resonance to trace organic dung-phosphorus forms in a temperate grassland soil

R. Bol1, W. Amelung2, L. Haumaier2 and B. Glaser2

1 Institute of Grassland and Environmental Research (IGER), North Wyke, EX20 EB, United Kingdom

2 Institute of Soil Science and Soil Geography, University of Bayreuth, D-95440 Bayreuth, Germany


A large proportion of total soil phosphorus (P) found in agricultural soils is present in organic forms, but the chemical nature of this soil organic P remained largely undefined. However, with the use of 31P nuclear magnetic resonance (NMR) spectroscopy it has been shown that several different inorganic and organic P species were present in soils and soil extracts. In this study, we investigated if 31P NMR could be used to dynamically trace the incorporation of the organic dung-P forms in a temperate grassland soil. Though dung-P inputs are important, highly localised, hot-spot P inputs to grassland systems are expected, but not much is known about their fate in the agricultural soils.

Materials and methods

Soil samples (1-5 cm depth) were collected 7, 14, 28, 42, 70 and 150 days after dung application on the 28th October 1997. The original applied dung was also kept for analysis. The experimental plots site was a long-term pasture situated at the Institute of Grassland and Environmental Research (IGER) North Wyke, Devon, S. W. England (50o45' N and 4o53' W). This site has a mean annual temperature of 10.5oC and a mean annual precipitation of 1035 mm. The soil type is classified as a clayey non-calcareous pelosol, corresponding to a Gleyic Cambisol. Plots had been maintained under grass pasture for at least 10 years prior to the initiation of the experiment and had not been grazed or fertilised with N since 1995 (Bol et al., 2000). All samples were extracted and analysed by 31P NMR spectroscopy according to the method by Sumann et al. (1998). The dialysis step in this method greatly reduces the inorganic P in the final extracts, with limited losses of the organic P, thereby allowing better quantification of the organic P structures by 31P NMR spectroscopy.

Results and discussion

In the dung sample diester P was the major P structure (with a d ppm of 0.2), with two other important peaks present, one representing monoester P (4.6 ppm) and the second being an unidentified P constituent (1.5 ppm). In the soil samples the P composition was more varied. The composition was dominated by monoester P, but diesters, pyrophosphate (-4.6 ppm), phosphonates (19-20 ppm) and inorganic orthophosphate (6.3 ppm) were also present. The higher proportion of diester P than monoester P (1.49:1) found in the dung rapidly diminished in the soils after its application, it being 0.40:1 after 7 days, 0.38:1 after 14 days and 0.28:1 after 150 days. Most of these P forms, including domination of organic P by monoesters were also found in other soil studies of cultivated or uncultivated land (Hawkes et al., 1984, Condron et al., 1990, Sumann et al., 1998). In our study of tracing the dung-derived P we did not examine the first 1 cm of soil under the dung pat, therefore an even clearer pattern of the disappearance and transformation of diester P might be observed when using 31P NMR analysis in this uppermost soil layer. However, our study indicated that diester P structures are more labile in the soil environment than monoester P, in line with observations by Hawkes et al. (1984) in relation to various P management strategies (i.e. P fertiliser and ploughing).


Bol, R, Amelung W., Friedrich C. and Ostle, N. 2000. Tracing dung-derived carbon in temperate grassland using 13C natural abundance measurements. Soil Biology & Biochemistry 32: 1337-1343.

Condron, L.M., Frossard, E., Tiessen, H., Newman, R.H., Stewart, J.W.B. (1990). Chemical nature of organic phosphorus in cultivated and uncultivated soils under different environmental-conditions. Journal of Soil Science 41: 41-50.

Hawkes, G.E., Powlson, D.S., Randall, E.W. and Tate, K.R. 1984. A 31P nuclear magnetic resonance study of the phosphorus species in alkali extracts of soils from long-term field experiments. Journal of Soil Science 35: 35-45.

Sumann, M., Amelung, W., Haumaier, L. and Zech, W. 1998. Climatic effects on soil organic phosphorus in the North American Great Plains identified by phosphorus-31 nuclear magnetic resonance. Soil Science Society of America Journal 62: 1580-1586.




Decision support system for phosphorus management

F. Djodji1, H. Montas2, A. Shirmohammadi2, L. Bergström1 and B. Ulén1

1Swedish University of Agricultural Sciences, Division of Water Quality Management, Box 7072, S – 750 07 Uppsala, Sweden

2University of Maryland, Biological Resources Engineering, College Park, Maryland 20742, Maryland, USA

Phosphorus (P) is one of the main nutrients limiting algal production in aquatic systems. Large part of the total P loads comes from Non-Point Sources (NPS), and has its origin in agriculture. Proper management of P in agricultural production systems can greatly enhance our ability to combat the pollution problem in aquatic environments. However, different fields within a watershed do not contribute equally to P export from the watershed. Consequently, the positive effects of the Best Management Practices (BMPs) are highest if applied on these P-export sensitive fields. Therefore, it is very important to identify the most contaminant generating fields for resource allocation and pollution abatement. To address this issue, a Decision Support System (DSS) consisting of Maryland P Index (PI), diagnosis Expert System (ES), prescription ES and a non-point source pollution model, GLEAMS, was developed and applied for Vemmenhog watershed, an agricultural watershed located in southern Sweden. This system is capable of identifying 'critical source areas' regarding phosphorus losses within the watershed, make a diagnosis of probable causes, prescribe the most appropriate BMPs, and test the environmental effects of the applied BMPs. The development of PI and ESs within Geographic Information System (GIS) environment enhanced processing of spatially varying input data and facilitated the evaluation of received results. PI calculations identified small parts of watershed as Critical Source Areas (CSAs).

Maryland P Site Index combines P loss potential due to site and transport characteristics and P loss potential due to management practice and source characteristics. Soil erosion, runoff class, subsurface drainage, leaching potential, distance from the edge of field to surface water and priority of receiving water are the subcategories of site and transport characteristics used to describe the potential for P removal both with surface runoff and through subsurface drainage. Accordingly, the Soil Test P Fertility Index Value, P fertilizer/manure application rate and methods of application are combined to describe P source characteristics. Following the approach that high risk for pollution occurs when both P source and an effective transport mechanism are simultaneously present at the same site, only 10.4 % of the total watershed area in 1995 and 5.2 % of the total watershed area in 1996 were classed as the 'high potential P movement' class.

Diagnosis ES includes seven different probable causes: high P level in soil, excessive P fertilization, high soil erodibility, low soil cover, stream proximity, moderate erodibility and subsurface drainage. For Vemmenhog watershed, four probable causes were identified by a diagnosis ES: high P level in soil, excessive P fertilization, stream proximity and subsurface drainage. Prescription ES consists of 10 BMPs: riparian buffer strips, no tillage, crop rotation, grassed waterway, reduced fertilizer application, P fertilizer incorporation, contour strip cropping, constructed wetlands, terracing and conservation tillage. Three BMPs (riparian buffer strips, reduced P fertilizer application and P fertilizer incorporation) were recommended by a prescription ES. Both diagnosis and prescription ES were applied only on fields that were categorized as the 'high potential P movement' class. GLEAMS simulations conducted for one selected CSA field for a 24-yr period showed that the recommended BMP reduced runoff P losses by 55 % and sediment P losses by 71 %, if applied from the first year.

All these results show that using DSS may enable us to realize the beneficial impact of BMPs on a long-term basis and select a proper BMP implementation strategy. The very good performance of DSS regarding the understanding, classification and processing of spatially variable data, augmented with the ability to account for temporal variations, would be extremely useful in many aspects and it would facilitate not only our understanding of the problem but also bring us closer to the solution of the problem.


Incorporating uncertainty into a phosphorus transfer model

E. G. Hope1, M. J. Whelan1 and K. Fox 2

1Department of Environmental Science, University of Stirling, FK94LA, UK

2Unilever Research, Port Sunlight, Wirral, CH633JW, UK


A simple model of phosphorus (P) transfer from agricultural land to surface waters is described which incorporates the effects of spatial variability in catchment properties and uncertainty in model parameter values. The model is appropriately-scaled for simulating processes in small to medium sized catchments and is parameter-efficient. It represents an improvement on previous approaches to modelling catchment-scale P transfer in providing greater temporal and spatial resolution than annual export-coefficient models and lower parameter and input data requirements than more complex, event-based, models.

Model concepts

TOPMODEL concepts (Beven and Kirkby, 1979) are used to estimate water, solute and sediment fluxes to water bodies. A daily time step is used which is consistent with the resolution of most meteorological observations and ensures that the model can be widely applied. Root zone-atmosphere interactions are simulated using the Penman-Moneith equation although a simpler evapotranspiration scheme may be used if meteorological data are restricted to daily temperature and precipitation. The spatial distribution of water table depth and saturation-excess overland flow are predicted on the basis of catchment topography and two calibrated model parameters. Dissolved P (DP) transfer is assumed to occur vertically in the unsaturated zone and laterally in the saturated zone. Readily soluble P is assumed to decrease exponentially with soil depth (after Haygarth et al., 1998). Particulate P (PP) transfers are modelled by estimating overland flow discharge and associated sediment transport capacity using a simple power law (after Kirkby and Cox, 1995). There is currently no attempt to simulate dynamic changes in the size of the available soil P store. Variations in model output are thus largely the result of hydrological processes and this weakness may have to be redressed in future developments. Uncertainty in the distribution of soil surface P concentrations and those parameters controlling the mobility of soil P are incorporated using Monte Carlo simulation, in which values are drawn randomly from predefined probability density functions in a large number of iterations. Model results can, therefore, be displayed as distributions (reflecting uncertainty) rather than single values.

Preliminary results and conclusions

The model is still undergoing development and has not yet been calibrated or validated. However, it is useful at this stage to examine some preliminary output and evaluate the potential utility of incorporating state variable and parameter uncertainty using Monte Carlo procedures within a deterministic model framework. Predicted losses of DP tend to be well correlated with modelled sub-surface discharge, show a marked seasonality and tend to be highest in areas close to the stream where the water table is near to the surface. Predicted PP fluxes, on the other hand, tend to be highly episodic and highest on concave mid-slopes where saturation excess overland flow is intermittent but where slope angles are sufficiently high for significant sediment transport to occur. The incorporation of a realistic estimate of uncertainty in model parameters generates wide uncertainty intervals in model output. These data highlight the need for better identification of model parameter values and provide an explicit recognition of the variability and lack of knowledge in the processes contributing to P transfers from agricultural land to surface waters.

Although the model requires further development and has yet to be validated, the idea of combining a deterministic process-based model core with a stochastic generation of uncertain state variables and parameter values, along the lines described, is attractive since it embraces variability and uncertainty whilst maintaining a synthesis of our understanding of the system dynamics.


This project is being financed by Unilever and the University of Stirling.


Beven K. and Kirkby M.J. 1979. A physically-based, variable contributing area model of basin hydrology. Hydrological Sciences Bulletin 24: 43-69

Haygarth P.M., Hepworth L. and Jarvis S.C. 1998. Forms of phosphorus transfer in hydrological pathways from soil under grazed grassland. European Journal of Soil Science 49: 65-72.

Kirkby M.J. and Cox N.J. 1995. A climatic index for soil erosion potential (CSEP) including seasonal and vegetation factors. Catena 25: 333-352

Phosphorus fractions in Irish karst aquifers

G. Kilroy, C. Coxon and N. Allott

Centre for the Environment, Trinity College Dublin, Ireland


Identification of phosphorus (P) movement via all hydrological pathways has been identified as a key research need in addressing the eutrophication of surface waters (Sharpley et al., 2000). Groundwater P movement has received less attention than other transfer mechanisms; however, P can occur in groundwater at environmentally significant levels, and P leaching has been reported in sandy soils, high organic matter soils, and soils with high P concentrations (Sims et al., 1998). This paper examines groundwater P fractions in three catchment areas in the west of Ireland: the Fergus catchment in County Clare, a highly karstic Carboniferous limestone aquifer; the Robe catchment in County Mayo, a moderately karstic Carboniferous limestone aquifer; and County Limerick, containing catchments with non-karstic aquifers, predominantly fissured Carboniferous limestones and Devonian sandstones. This work forms part of a larger project to determine the circumstances in which groundwater discharges contribute to surface water eutrophication.


Over 500 samples were taken periodically from 56 springs and 67 wells in the three areas between July 1998 and March 2000. All samples were analysed for total P (TP) and 30% of the samples were also analysed for total dissolved P (TDP) and dissolved reactive P (DRP). Dissolved organic P (DOP) and particulate P (PP) were derived by calculation, i.e., DOP=TDP-DRP and PP=TP-TDP.


Mean TP levels for the Fergus, Robe and Limerick catchments were 65, 68 and 76 µg l-1 respectively. Median TP levels for the same catchments were 26, 29 and 20 µg l-1 respectively. The disparity between the means and medians highlights the skewed nature of groundwater P data. Extreme values of over 1000 µg l-1 were reported in all catchments at a small number of sites experiencing pollution events. Thus medians are a more appropriate summary statistic and indicative of background P levels. Table 1 provides a summary of median levels of each P fraction for springs and wells in each catchment.

Table 1. Median phosphorus fraction levels in different Irish groundwater site types.


Site Type (n)

Dissolved Reactive P

Dissolved Organic P

Particulate P


(µg l-1)

(µg l-1)

(µg l-1)

Fergus (highly karstic)

Spring (58)





Well (5)





Robe (moderately karstic)

Spring (46)





Limerick (non-karstic)

Spring (6)





Well (37)






Dissolved reactive P was the dominant fraction for springs and wells in all catchments. Thus, in situations where groundwater P is contributing to surface water loading, it is doing so in the most biologically available form. The karst aquifers had higher DRP, DOP and PP levels than the non-karstic Limerick aquifers. This could reflect the influence of point recharge from surface waters via swallow holes, combined with less attenuation of P in karst systems due to rapid conduit flow. Springs had more PP than wells probably due to more turbulent flow. Karst aquifers pose a particular threat by providing rapid transfer of both dissolved and particulate P from land to surface waters.


Sharpley, A., Foy, B., and Withers, P. 2000. Practical and innovative measures for the control of agricultural phosphorus losses to water: an overview. Journal of Environmental Quality 29: 1-9.

Sims, J.T., Simard, R.R., and Joern, B.C. 1998. Phosphorus loss in agricultural drainage: historical perspective and current research. Journal of Environmental Quality 27:277-293.

Development of a quantitative method to study hysteresis behaviour of dissolved phosphorus transport in surface runoff

J. Langlois and G. R. Mehuys

McGill University, Department of Natural Resource Sciences, 21,111 Lakeshore Road. Sainte-Anne-de-Bellevue, Canada

Relationships between surface runoff discharge and dissolved phosphorus (P) loads are complicated by a hysteresis effect. Until now, researchers have investigated hysteresis loops using qualitative tools describing their direction (e.g., clockwise or counterclockwise). The quantitative study of these loops would ease comparison of results of dissolved P losses in surface runoff between events and therefore help to evaluate the impacts of hydrologic characteristics, such as soil moisture conditions, on P transfer from fields to water bodies. The objectives of the present study were to 1) develop a technique to quantify the hysteresis of P transport in runoff water and 2) use this technique to compare hysteresis behaviour on an event basis.

Since fall 1998, measurements have been performed on two undrained raised beds (290 m x 33 m each) separated by a ditch along their length. Trenches were dug on each side of the ditch and waterproof polyethylene membranes were laid in them to intercept runoff from the adjacent beds. At the end of each trench, continuous monitoring of runoff volume was carried out using tipping buckets. Manual sampling of runoff water in both trenches was performed every 5 minutes for the first hour after initiation of a runoff event and at 15- minute intervals thereafter. Aliquots of runoff water were filtered (0.45 m m) and analyzed for dissolved P. Finally, rainfall volumes were measured using a recording raingage.

For each rain event, P concentrations were coupled with discharges during the rising limb and the falling limb of the runoff hydrograph. The resulting hysteresis curves were characterised by 1) computing a regression equation for each part of the hydrograph and 2) calculating an index (H) as the ratio of the intregral of the equation of the rising limb to that of the falling limb. Soil antecedent moisture conditions were estimated by using the ratio of runoff to rainfall volumes. The results show that soil antedecent moisture conditions affect hysteresis loops. Indeed, as the soil prior to a rainfall event is wetter, the P hysteresis behaviour of the loops shifts from a counterclockwise (H<1) to a clockwise direction (H>1). Quantitative assessment of hysterist behaviour contributes to a better understanding of the effects of hydrology, with its complexities of temporal variability, on P transfers from soils to water bodies.

Assessing potential phosphorus-loss in runoff from grassland – an example from Switzerland

P. Lazzarotto and V. Prasuhn

Swiss Federal Research Station for Agroecology and Agriculture (FAL), CH-8046 Zurich

The loss of agricultural phosphorus (P) into soil due to surface runoff, soil erosion and leaching leads to unintended accumulation of nutrients in waters. In grassland-dominated regions with a high density of cattle and agricultural use coupled with manuring, surface runoff of dissolved phosphorus (DP) is seen as the major source of diffuse water pollution with P in surface waters. In 1993, a package of ecological measures among other things to reduce water pollution by agriculture has been introduced in Switzerland. They call for a biological and sustainable agricultural management and also for introducing of ecological compensation areas. One aim is to reduce agricultural P-loads into surface waters by 50% until 2005 (reference period 1990-92) (Forni et al., 1999). The effectivity of these ecological measures is investigated in the Lippenrütibach catchment (central Switzerland) which discharges into Lake Sempach, the lake which is most suffering from eutrophication in Switzerland.


The catchment (334 ha) is situated between 500 and 800 m.a.s. and mean annual precipitation is about 1000 mm. The agricultural area of 255 ha consists of 80% grassland and 20% arable land. Hourly P- and runoff-data of the creek Lippenrütibach are available. The creek discharged between 1986 and 1998 in average 0.86 t P (0.32 t DP and 0.54 t sediment bound P) into Lake Sempach.

Phosphorus-loss into waters due to surface runoff are estimated by an empirical-statistical modelling approach similar to the P-index of Gburek et al. (1996). Factors for the assessment of the risk for P-runoff due to natural conditions of individual plots (soil type, topography, distance to gully drains), due to agricultural practices (land use, livestock density, agricultural use, measures against P-loss into fields) and due to indvidual events (manuring, precipitation) are collected for each of 270 fields in the catchment. Connections between these factors and the measured P- and runoff-data in the creek can thus be related to each other. The assessment of total risk for P-runoff for each plot results from the links between these described factors. The risk for P-runoff is classified into four classes (low, medium, high, very high). According to classified total risk for P-runoff, an area-specific P-loss-value is introduced for each plot.

Results and conclusions

Results for 1998 show on 21% of all plots no or low, on 26% medium, on 48% high and on 4% very high total risk for P-runoff into surface waters (Braun et al., 2001). Since in 1998 mean annual precipitation was comparatively low (950 mm), classes of total risk may partially move towards higher classes by modelling additional years with average or above-average mean annual precipitation.

Calculated potential of P-reduction for 1998 amounts to 94 kg P, respectively 19% of agricultural P-loads (Braun et al., 2001). This reduction can be attributed to the introduction of ecological measures. However, it is still far smaller than the planned P-reduction of 50%. A scientifically reliable evidence for the efficency of ecological measures will only be possible after analyzing additional years. These investigations will be continued until 2005. In addition to the statistical modelling approach presented here, more physically based modelling approaches will be developed.


Braun, M., Aschwanden, N. und Wüthrich-Steiner, C., 2001: Evaluation Oekomassnahmen: Abschwemmung von Phosphor. Agrarforschung 8 (1), 36 – 41

Forni, D., Gujer, H. U., Nyffenegger, L., Vogel, S. und Gantner, U., 1999: Evaluation der Oekomassnahmen und Tierhaltungsprogramme. Agrarforschung 6 (3), 107 – 110

Gburek, W. J., Sharpley, A. N. and Pionke, H. B. 1996: Identification of critical source areas for phosphorus export from agricultural catchments. – Advances in Hillslope Processes Vol. 1, 263 - 281

Leaching of phosphorus through structured clayey soil: from batch kinetics to landscape scale.

Magid J.1, Jensen M. B.2, Hansen S.1 and Hansen H.C.B.3

1Dept. Agricultural Sciences, Royal Veterinary and Agricultural University, Thorvaldsensvej 40, 1871 Frederiksberg C, Denmark

2 Research Institute for Forest and Landscapes, Hørsholm Kongevej 11, 2970 Hørsholm, Denmark

3Laboratory for Soil and Environmental Chemistry, Chemistry Dept., Royal Veterinary and Agricultural University, Thorvaldsensvej 40, 1871 Frederiksberg C, Denmark

Our experimental approach embodied three scales of study. Firstly, a microscale in which the kinetics of ortho-phosphate sorption were characterized for materials from topsoil, bulk subsoil, cracks and biopores, that showed widely differing sorption characteristics depending on the materials association with flowpaths. Secondly, an intermediate scale involving intact soil mesocosms (14 by 20 cm) from three intensively monitored sites of the catchment to evaluate the plowlayer as a source for phosphors (P)-leaching. Thirdly, large intact soil blocks (0.5 m diameter, 0.5 m or 1 m length) were excavated and studied intensively under controlled conditions. Among other things, this work clearly demonstrated that substantial amounts of P can be leached through saturated or near saturated soil, when cattle feces is located on the soil surface. This was confirmed by catchment studies of stream and drainwater flow. The soil materials used in the process studies mentioned above, originated from the same catchments.

This paper will present unpublished data on large intact soil blocks loaded with aqueous ortho-P solution of defined ionic strength in order to examine the P carrying capacity of the soils. Hereby the effects of dissolved organic substances are nullified, and while these conditions may less closely resemble real life occurrences the data should allow an attempt to integrate the large scale with the batch kinetic scale studies performed and reported previously. In the presentation we will examine the possibility for making a model interpretation of the intact block study, using the kinetic parameters derived from batch studies, as well as information on large block hydrological characteristics. The Daisy model has recently been developed to allow description of preferential flow. We will extend the discussion to the landscape scale with reference to the original catchment studies and the available literature.





Modelling of subsurface phosphorus-leaching losses of a re-wetted site in the Droemling fen area, Germany

R. Meissner1, S. Pudenz2, H. Rupp1, P. Leinweber3, A. Schlichting3

1 UFZ Centre of Environmental Research, Leipzig – Halle, Department of Soil Sciences, Falkenberg, Germany

2 Institute of Freshwater Ecology and Inland Fisheries, Berlin, Germany

3 University of Rostock, Faculty of Agricultural and Environmental Sciences, Rostock, Germany


Global experience has shown that intensive agricultural use of Histosols (synonyms: peat soils, fens) results in decreased soil fertility, increased oxidation of peat and corresponding CO2-emissions to the atmosphere, nutrient transfer to aquatic ecosystems and accelerated losses in total area of these native wetlands. To prevent these negative environmental effects and to restore some of the wetland ecosystems, set-aside programs and re-wetting measures have been promoted in several countries, e.g. in the fen area 'Droemling', which is with an area of 280 km2, one of the most important water protection areas in North-East Germany (Kalbitz et al., 1999). The objective of the study is to select a P-leaching model and to use it for the calculation of flow and transport processes during re-wetting.

Research attempt

To ensure an environmentally harmless land use and Histosol re-wetting the amounts, forms, and pathways of P losses will be determined in the frame of the PROWATER- project funded by the EC. The recently developed model MORPHO (Modelling of Regional PHOsphorus Transport) is tested in context of this project to predict P- loads in surface water which are affected by diffuse pollution (Pudenz, 1999). For the model evaluation long-term data sets obtained on a representative re-wetted site equipped with an automated field research station and devices for ground water observation and soil solution sampling are used. Data sampling for the identification of basic P sorption and desorption parameters of the peat started in the year 2000 for the investigated site.


The following compartments and processes are considered in the code of the model MORPHO:

land surface: rainfall, evapotranspiration

soil/unsaturated zone: water flux, advection, dispersion, sorption/desorption

groundwater: water flux, advection, dispersion, sorption/desorption

An essential requirement to describe the flow and transport processes in both, in the unsatured zone and in groundwater is an adequate description of the P desorption. Here, non-linear Langumir, two site Langumir and kinetic sorption models are implemented and tested on several soil types.


First modelling results about the variation of subsurface P-leaching losses for a selected site (former intensively used grassland in natural succession with a rising ground water table) will be presented and conclusions for the calculation of re-wetting scenarios, risk assessments and future work will be derived.


Kalbitz, K., Rupp, H. Meissner, R. and Braumann, F. 1999. Folgewirkungen der Renaturierung eines Niedermoores auf die Stickstoff-, phosphor- und Kohlenstoffgehalte im Boden und Grundwasser. Zeitschrift für Kulturtechnik und Landentwicklung 40: 22 – 28.

Pudenz, S. 1999. Modellierung der regionalen Phosphorverlagerung im Boden und Grundwasser. Wissenschaftliche Schriftenreihe Umwelttechnik. Band 8. ISBN 3 – 58574 – 332 - 1. Verlag Dr. Köster, Berlin, Germany.

Dynamics of dissolved and particulate phosphorus during storm events in two rural river basins and in two headwater streams

M. Salvia-Castellvi1,2, J. Mosig2, J.F. Iffly2, P. Vander Borght1 and L. Hoffmann2

1Fondation Universitaire Luxembourgeoise, 185 av. de Longwy, B-6700 Arlon, Belgium

2CRP-Gabriel Lippmann, 162A av. de la Faïencerie, L-1511 Luxembourg


Most of the total phosphorus (P) in streams appears to be transported during storm events and is associated with suspended solids transport (Verhoff et al., 1979). This particulate fraction is not only quantitatively important (Prairie and Kalff, 1988), but also ecologically, because of the availability of the P released from the solid phase to the soil solution and the further transport to the water streams (Frossard et al., 2000). Thus, the proportion of particulate P (PP) exported annually from watersheds is fundamental in estimating annual lake P loadings (Lathrop et al., 1998). In this study, several characteristics of the behaviour of suspended solids (SS), total phosphorus (TP) and soluble reactive phosphorus (SRP) in four rural watersheds during flow events are exposed.

Materials and methods

The studied basins are situated on Devonian schistous substrates in the Ardennes region (Belgium-Luxembourg). Two catchments, Sûre and Wiltz, have 306 respectively 424 km2, with significant point sources impact on the Wiltz basin. Two streams, Kuebefiels and Teischelt, have 0.8 respectively 1.2 km2, the first with agricultural and the second with coniferous forestry land use. There is no housing pressure for these small subcatchments. Intensive sampling during some storm periods was undertaken with ISCO autosamplers in 1991 (Sûre), 1993 (Kuebefiels, Teischelt) and 2000-2001 (Wiltz). Parameters measured were SS, SRP (GF/C filtered, molybdate method), TP (acid mineralised unfiltered sample at 380°C) and discharge (continuous flow monitoring stations).


The observed behaviour during high flow events of the different watersheds is quite similar despite the big differences of size and land use. However, greater absolute values of SRP, TP and SS are measured at the point polluted Wiltz basin. Surprising was the minor differences in SRP concentrations and TP content of the SS between the agricultural and forestry small streams (erosion was greater in the forestry basin due to their steeper slopes). The storm behaviour can be classified in two essential types: (1) small summer storms and (2) high flow events during the humid period.

The first type is characterized by P-rich particulate material and by a slight increasing of SRP concentrations (from 0.020 to 0.050 mg l-1 SRP) during high discharges. During autumn and winter runoff events, quite stable concentrations of SRP are observed. The second type is also characterized by a classical clockwise hysteresis behaviour (e.g. Whitfield & Schreier, 1981; Prairie & Kalff, 1988) for SS, TP and PP concentrations: they rise with discharge but the concentrations during the rising stages of the hydrograph are higher than those on the falling stage at equal discharge. Peaks in TP or suspended solids occur before the peak in discharge. Following a rapid but brief rise in concentration, P levels declined sharply while discharge was still increasing. In fact, a rapid exhaustion of P and particulate material is observed, high flows having a great erosive force and sediment carrying capacity very closely linked to the rain intensity. For each storm type, SS and TP are strongly correlated (power function with R2 around 0,80 in the two cases). This typical behaviour clearly demonstrates that even within a restricted geographical area, changes in TP with respect to water discharge cannot be described adequately by a simple concentration-discharge equation for the studied streams.


Verhoff, F.H., Melfi, D.A., and. Yaksich, S.M. 1979. Storm travel distance calculations for total phosphorus and suspended materials in rivers. Water Resour. Res. 15:1354-1360.

Lathrop, R.C., Carpenter,S.R., Stow, C.A., Soranno, P.A., and Panuska, J.C. 1998. Phosphorus loading reductions needed to control blue-green algal blooms in Lake Mendota. Can. J. Fish. Aquat. Sci. 55: 1169-1178.

Frossard, E., Condron, L.M., Oberson, A., Sinaj,S., and Fardeau, J.C. 2000. Processes governing phosphorus availability in temperate soils. J. Environ. Qual. 29: 15-23.

Prairie, Y. T., and Kalff., J. 1988. Particulate phosphorus dynamics in headwater streams. Can. J. Fish Aquat. Sci. 45:210-215.

Whitfield, P.H., and Schreier, H. 1981. Hysteresis in relationships between discharge and water chemistry in the Fraser River basin, British Columbia. Limnol. Oceanogr. 26: 1179-1182.

Phosphorus concentration in surface- and groundwater under managed and unmanaged peatland

Andrzej Sapek, Barbara Sapek, Piotr Nadany, Marek Urbaniak

Institute for Land Reclamation and Grassland Farming at Falenty, PL-05-090 Raszyn, Poland

It is commonly supposed, that phosphorus (P) compounds in peat soil are better soluble then in mineral soils. Therefore, a higher P leaching can be expected. There are some opinions, that the dewatering of peaty bog would reduce the solubility of P compounds and in opposite, the renaturisation of peatland through raising the groundwater level would result in increasing this solubility. The better solubility of P compounds could cause a risk of eutrofication of surface water below the peatlands. The aim of this work was to identify the changes in solubility of compounds in peats, which are exposed to some changes in hydrological regime.

The investigations were localised in great area of peatlands in Poland - the valley of Biebrza river. Particularly, the Biebrza National Park of Peatlands (about 170 000 ha), and Kuwasy bog used as a productive meadows (about 1000 ha). On all objects, the groundwater level was lowered due to dewatering the valley for the agricultural use of some parts of peatland. Now, there are making some attempts to heighten the groundwater level in National Park area as well as in agricultural used site. The peats occurring in Biebrza River Valley are rich in P due to phosphate flow from the mineral soils in surrounding watershed. The agricultural use peat soils in Kuwasy bog were heavy dressed with P during last 40 years.

In National Park, two trans-sections were established, where a set of plastic tube were installed in direction downward to the subsurface runoff. 9 tubes were installed on the Grobla Honczarowska trans-section and 8 on he Gugny trans-section. 10 tubes were installed on in Kuwasy bog. Water samples of groundwater and surface water from nearby streams were taken once a month, and P concentration were determined calorimetrically in unfiltered samples by use of molybdate blue method.

The P concentrations in stream water were much higher then usually observed in Polish river water. While, these concentrations were rather high in some samples of groundwater (Table 2). The highest concentrations were found in samples from Gugny trans-section, where the increasing groundwater level was greatest during last few years, in spite of that the P content in peat was the lowest (Table 1). The P fertilisation on Kuwasy object has no visible effect on its concentration in water. There are observed some regular changes in P concentration along the trans-section, but the number of measurements made at the preliminary stage of investigation does not allow to do any conclusion.

Table 1. Mean phosphorus content (mg P dm-3 of fresh sample) in 1% K2SO4 extract and 0.5 M HCl extract from peat samples taken from investigated objects

Sampling depth cm

1% K2SO4 extract

0.5 M HCl extract









n = 6

n = 6

n = 9

n = 6

n = 9

n = 6






















Table 2. Mean phosphorus (RPunfil.) concentration (mg P dm-3) in surface- and groundwater from peatland objects


Peatland object








































This work has been done in scope of Environment Project EVK1-1999-00121 'PROWATER'

Processes and pathways influencing phosphorus retention within a farm catchment

M. Silgram1, M. Hutchins1, R. Hodgkinson2, and E. I. Lord1

1ADAS Wolverhampton, 'Woodthorne', Wergs Road, Wolverhampton WV6 8TQ, UK

2ADAS Boxworth, Boxworth, Cambridge CB3 8NN, UK

Net retention of phosphorus (P) in the landscape and in small farm streams can represent a critical component in limiting losses from agricultural fields into large river systems. In recent decades, experimental research has focused on plot and field-scale studies increasing our understanding of the dynamic processes controlling the movement of P, but relatively less attention has been given to its environmental fate. Improved knowledge of the fate of this P has wider importance by helping focus mitigation strategies at field and farm scale, and has broader policy relevance as part of catchment management plans to be developed under the new EC Water Framework Directive.

Titley Court is a mixed 250 ha farm on silt loam and silty clay loam soils in Herefordshire (Grid reference SO 333 597) in the reaches of the River Arrow, a whole-river Site of Special Scientific Interest (SSSI) and an area of ecological significance under the EU Habitats Directive. The farm has undulating topography, with the most steeply sloping land (5-15 degrees) to the north and more gently rolling land (0-5 degrees) leading to the flood plain to the south. Under-drainage is present in most fields, although there are areas where this is no longer functioning efficiently. A drainage ditch from the northern fields merges with a spring-fed stream before bisecting the farm as it flows south via a small (0.3 ha) shallow lake to join the River Arrow. MAFF-funded research has covered a variety of scales, from plot studies using runoff traps to investigate effects due to slope length, tramlines and over-winter catch crops, to more spatially-integrated measurements in subsurface drains and the farm stream. Moving progressively downstream, sampling locations after the confluence with water from the spring include Priory Marsh (1), Culvert (2) sited shortly before the inlet to the shallow lake, Sluice (3) at the lake outlet, and Hunton (4) before the stream joins the River Arrow. An additional station was established to monitor losses from drains and runoff contributing to flow in the separate small stream running east along the edge of Railway South field (5). Instruments measure continuous flow and turbidity, with EPIC flow-proportional samplers used to monitor molybdate-reactive P (MRP), total dissolved P (TDP), total particulate P (TPP), and total P (TP) concentrations and loads in relation to discharge, location and season.

Overall budgets in the 1999/2000 hydrologic year at the most upstream gauging station indicate a sediment export of 770 kg ha-1 and a flux of 2.93 kg ha-1 TP, including 1.01 kg ha-1 as TDP with a corresponding MRP fraction of 0.86 kg ha-1. Results from runoff traps and subsurface drains suggest that very little surface runoff reaches the water course directly: most is retained and re-infiltrated (e.g. at breaks in slope or buffer strips/headlands) within fields, reaching the water course via the drainage system. The flow-weighted mean annual concentration was 0.38 mg L-1 TP, a level deemed to be of environmental concern. However, such concentrations systematically decrease from this level at site (1) to 0.19 mg L-1 at site (3), with the result that net export of P from the farm into the River Arrow is 1.41 kg ha-1 TP, a reduction of 52% over the c. 1 km course of the farm stream. Total flow increases little between site (1) and sites (2) and (3), indicating that this decrease in flow-weighted mean concentration cannot result from additional inputs of more dilute drain waters, but must instead derive from retention of P within the stream itself.

Evidence of a further retention mechanism has been revealed from geochemical analysis of seasonal grab samples of Turning Ways ditch, Church spring and confluence waters at site (1). The overall characteristics of the monitored waters indicate lower conductivity in winter months associated with seasonal dilution. This is reflected in the pronounced decrease in concentrations of individual cations and anions, with the exceptions of potassium and nitrate. Evidence from geochemical samplings indicate a marked decrease in the calcite saturation index from over 1 in summer to less than zero in winter. High dissolved iron and aluminium contents in sediment-laden winter waters, combined with neutral pH, suggest that these species were not in true solution but rather were present as colloids such as Fe(OH)3. When used as conservative tracers, both silicon and overall charge balance alkalinity provide comparable estimates of the contribution of spring and surface-derived drainage waters to the overall composition at Site (1). The spring contribution ranges from 35-42 % in summer to only 19-33 % in winter, with the remainder originating from surface sources. Inputs of MRP in spring water were characteristically low (c. 0.05 mg L-1) in both summer and winter months, whereas inputs at Turning Ways ditch were consistently higher at c. 0.3 mg L-1. However, samples from site (1) consisting of an integration of these two input waters demonstrate pronounced contrasts between summer and winter months, with summer levels of only 0.03 mg L-1 but around 0.25 mg L-1 during the winter. In the context of the evidence of seasonal calcite saturation dynamics, we hypothesise that co-precipitation of MRP with calcite (i.e. occlusion) may be a substantial retention mechanism during summer low flow conditions at this site, but may be of negligible importance during the winter season.


Calibration and sensitivity analysis of a catchment phosphorus model

Russell M. S. Smith1, Howard S. Wheater1 and Matthew J. Lees1

1Environmental Water Resource Engineering, Department of Civil and Environmental Engineering, Imperial College of Science Technology and Medicine, London, SW7 2BU

The quality of surface waters throughout Europe, in terms of risk of eutrophication, is a major current concern because of increasing concentrations of phosphorus (P). Current legislation has focused on point sources and in many cases this has not resulted in alleviating the eutrophic conditions (Foy et al., 1995). With the introduction of the Water Framework Directive a shift in emphasis towards diffuse sources is required and a more integrated approach in the management and control of eutrophication. This necessitates the use of catchment-scale models in quantify the contribution from point and non-point and therefore where expenditure is best targeted.

Many catchment-scale hydro-chemical models require a large number of spatially distributed parameters making application and calibration difficult and resulting in large uncertainties in the model output. However, simple export coefficients have been advocated as empirical tools to assess the annual export of total phosphorus (TP) in relation to land use (Johnes, 1996). Previous regression analysis has also indicated that dynamic phosphorus export can be approximated as a direct function of discharge (Lennox et al. 1997). Here a simple hypothesis to integrate these 2 approaches and produce a simple dynamic model (Daldorph et al. 2000) is tested against data flow and total phosphorus concentrations from the Rutland Water, UK. Annual loads derived using export coefficients are disaggregated using the relationship between the flow and TP load in headwater streams. The diffuse TP load can then be estimated as a direct function of the daily runoff volumes and the annual TP export.

Calibration and sensitivity analysis is conducted in a Monte Carlo framework with uncertain parameters sampled at random from a uniform distribution. The performance of each model (here ‘model’ refers to the parameter set model combination) is assessed against different aspects of the observed data. The results are analysed using the Monte Carlo Analysis Toolkit, a series graphical techniques developed at Imperial College, (MCAT, Wagener et al., 2001).

Results of the sensitivity analysis show there is a trade off between fitting the different objective functions with estimated parameter values derived from different areas of the parameter space. Points which lie along this ‘trade-off’ form the Pareto set, where: ‘a solution is said to be Pareto optimal (i.e. part of the Pareto set) if the value of any objective function cannot be improved without degrading at least one of the other objective functions’ (Chankong and Haimes, 1993). Model simulations using the Pareto set also failed to encompass important aspects of the observed data, suggesting error in the model framework. However the analysis highlights areas the model performance could be improved. The assumption that the TP available in the soil remains constant is clearly an oversimplification and the inclusion of an exhaustion effect would improve model performance.


Chankong, V. and Haimes, Y.Y. 1993. Multiple optimization: Pareto optimality. In Young, P.C. (ed.) Concise encyclopedia of environmental systems. Pergamon Press, Oxford, UK, 387-396.

Daldorph, P.W.G., Lees, M.J., Wheater, H.S. and Chapra, S.C. 2000. Integrated Lake and Catchment Phosphorus Model: A Eutrophication Management Tool. I Model Theory (In press).

Foy, R.H., Smith, R.V., Jordan, C. and Lennox, S.D., 1995. Upward trend in soluble phosphorus loadings to Lough Neagh despite phosphorus reduction at sewage treatment works. Water Research 29: 1051-1063.

Johnes, P.J. 1996. Evaluation and management of the impact of land use change on the nitrogen and phosphorus load delivered to surface waters: the export coefficient approach. Journal of Hydrology 183: 323-349.

Lennox, S.D., Foy, R.H., Smith, R.V.& Jordan, C. Estimating the contribution from agriculture to the phosphorus load in surface water. In P loss from Soil and Water (ed H.Tunney). Cabs International. 1997.

Wagener, T., Lees, M.J. and Wheater, H.S 2001. A framework for the development and application of parsimonious hydrological models. To appear in Singh, Frevert and Meyer (Eds.) Mathematical models of small watershed hydrology – Volume 2. Wat. Resour. Publ. LLC, USA, in press.


Prediction model and long-term trend of phosphorus transport from arable land in Sweden

Barbro Ulén, Göran Johansson, Katarina Kyllmar

Department of Soil Sciences, Division of Water Quality Management, Swedish University of Agricultural Sciences Box 7072, SE-750 07 Uppsala, Sweden


As a part of nation-wide monitoring of the impact of agriculture on the aquatic system, phosphorus (P) transport was measured from subsurface drained fields in Sweden. The measured transports was a mixture of P from surface, subsurface and groundwater contribution. Fourteen of these fields together with a similar experimental field have been studied for at least ten years and most of them for much longer. Transport of total phosphorus (TP) from four of the fields together accounted for 74% of the total transport. Five other fields together accounted for another 19%, while transport from the other five fields was more or less negligible. Based on factor combinations, a simple regression model for calculating TP concentration was used for prediction of the losses. The parameters included were livestock density (LD), HCl extractable P (P-HCl) in the topsoil, duration of high water flow (DHF) and soil specific areas (SSA). The model was found to satisfactorily predict the TP transport in the streams of 32 small agriculture-dominated catchments in different parts of the country, with a P loss from arable land that was calculated as varying from 0.03 to 0.50 kg ha-1 y-1. However, for a very few catchments, the model worked only for transport by drainage water and not for P transport measured in the stream water.

Eleven of the observation fields had been investigated since 1977 or earlier. During this period, a surplus of P added to the soil during the initial years was transformed to a slight negative balance of P. Transport from most of the fields was found to be relatively constant, but the two fields with the highest P transport showed an increased trend during the 1970s and 1980s.

Forms of phosphorus in runoff from a clayey soil in southern Finland

Risto Uusitalo and Eila Turtola

Agricultural Research Centre of Finland, FIN-31600 Jokioinen, Finland


In the clayey soils of south Finland, particulate phosphorus (PP) is typically the major phosphorus (P) fraction in runoff waters (Turtola, 1999). Most of the PP in these waters is considered non-desorbable, i.e., mostly non-bioavailable in aerobic environments (Ekholm, 1998; Uusitalo et al., 2000). However, there are no published data about the potential for P release from the suspended soil material in anoxic conditions. In this paper we report the transport of different P forms from a clayey field at Jokioinen, southwestern Finland, and assess the potential for release of PP both in aerobic and in strongly reduced environment.

Material and methods

The soil of the Jokioinen field was a very fine Typic Cryaquept, having 48 % clay, 2.5 % organic C, and 31-45 mg kg-1 Olsen-P (for a detailed description of the field, see Turtola, 1999). Flow-proportional sampling of surface and subsurface runoff was arranged by tipping buckets, allowing all runoff waters to be analysed for dissolved molybdate-reactive P (DRP; filtration through a 0.2 m m polycarbonate filter), and total P (TP; digestion with peroxodisulfate and sulfuric acid in an autoclave). Particulate P was taken as the difference between TP and DRP.

To find out how much of the PP is released when the environment is aerobic, we extracted 205 runoff samples by anion exchange resin (AER). The extraction followed the procedure of Sibbesen (1978), with minor modifications given in Uusitalo et al. (2000). Preliminary assessment of the potential for P release in a strongly reduced environment (less than –400 mV) was done for 37 samples by using a modification of bicarbonate-dithionite (BD) extraction step of the P fractionation of Psenner et al. (1984). Extraction of a 30-ml water sample was conducted on an orbital shaker (120 rpm, 15 min) at room temperature after additions of 1 ml 0.223 M NaHCO3 and 1 ml 0.431 M Na2S2O4 (freshly prepared solutions). After shaking, the sample was immediately filtered through a 0.2 m m filter, and the filtrate digested with peroxodisulfate and sulfuric acid in an autoclave in order to destroy excess dithionite which otherwise caused major interferences in P determination by a spectrophotometer.

Results and discussion

Annual P loss from the Jokioinen field during 1992-1998 averaged 920 g ha-1 (Turtola, 1999), and the proportion of PP was 87 % of TP, i.e. 800 g ha-1. According to the results of the AER and BD extractions, 6.6 % of PP was desorbable in aerobic environment, whereas 10-80 % (mean 31 %) of the PP in the samples studied was released when the redox potential was lowered by dithionite addition. When compared to the mean annual DRP loss of 120 g ha-1, we estimate that two third of the potentially bioavailable P load from the Jokioinen field was in dissolved form, as long as the eroded sediment remained in aerobic environment. The situation would, however, be very different if the eroded soil material would end up in a strongly reduced environment. Then, the suspended sediment would be a source of twice as much potentially bioavailable P as was transported as DRP, and little less than half of the TP would be potentially available for primary production in the receiving waters.


Ekholm, P. 1998. Algal-available phosphorus originating from agriculture and municipalities. Monographs of the Boreal Environ. Res. 11, 60 p.

Psenner, R. von, Pucsko, R., and Sager, M. 1984. Die fraktionierung organisher und anorganisher Phoshorverbindungen von Sedimenten. Arch. Hydrobiol./Suppl. 70:111-155.

Sibbesen, E. 1978. An investigation of the anion-exchange resin method for soil phosphate extraction. Plant Soil 50: 305-321.

Turtola, E. 1999. Phosphorus in surface runoff and drainage water affected by cultivation practices. Agricultural Research Centre of Finland and Univ. of Helsinki. 108 p. Diss. ISBN 951-729-555-3.

Uusitalo, R., Yli-Halla, M., and Turtola, E. 2000. Suspended soil as a source of potentially bioavailable phosphorus in runoff waters from clay soils. Wat. Res. 34: 2477-2482.

The effect of calcium soil amendments on phosphorus (and dissolved organic matter) mobility

J. C. R. Varcoe 1,3 , D. J. Chittleborough 1, J. W. Cox 2 and J. van Leeuwen 3

1 The University of Adelaide, Australia

2 CSIRO Land & Water, Australia

3 CRC for Water Quality & Treatment, Australia


Two significant contributors of water quality decline in southern Australian reservoirs are phosphorus (P) and dissolved organic matter (DOM). P is a key factor in periodic outbreaks of toxic algal blooms in domestic reservoirs of southern Australia. DOM may be an agent that enhances P mobility through soils. Following a preliminary field study, we hypothesised that calcium (Ca), in the form of gypsum may be effective in attenuating movement of these constituents. In this paper we present results from both field and laboratory investigations into the effects of Ca on P and DOM mobility. In addition to this we are examining the possible mechanisms by which Ca is able to reduce P and DOM mobility through soil. This research may lead to the development of agricultural amendments that reduce the movement of these constituents from soil to surface waters, not only improving the quality of drainage water but also improving the productivity of the soil.

Field Study

Batch study

The effect of Ca on P and DOM adsorption to a clay mineral (kaolinite) was investigated by a batch methodology. Solutions (150 ml) containing 1 mg L-1 P, at ionic strengths of 0.001, 0.01 and 0.05 M NaCl and with or with out DOM (at 18 mg L-1) and Ca (at 100 mg L-1) were adjusted to 3 pH values (5, 7 and 9) and were added to 500 mg of kaolinite. Three replicates of the 36 possible combinations of the variables were shaken for a period of 24 hours.

Results of the batch study indicated that, in the absence of DOM, Ca reduced the amount of P left in solution. This reaches a maximum at pH 9. The effect of ionic strength was dependent on the presence or absence of Ca. In the presence of Ca, P removal from solution decreased with increasing ionic strength. In the absence of Ca, the reverse trend was true. DOM appeared to have considerable effect on P removal. Ca has little effect on P removal in the presence of DOM.


If the P removed from solution in the batch trials was assumed to be by adsorption to kaolinite, then it appears that Ca acted as a cooperative cation in P adsorption. If this is the mechanism, then the different effect of variation in ionic strength in the presence or absence of Ca suggests different adsorption mechanisms (He et al., 1997). Increasing adsorption with increasing ionic strength in the absence of Ca suggests adsorption is dominated by inner sphere or specific adsorption mechanisms. The reverse pattern in the presence of Ca suggests outer sphere or non specific ion pair bonding mechanisms. Literature suggests however that a more likely removal mechanism in the presence of Ca maybe precipitation of Ca-phosphates (Hawke et al., 1989). While there was no evidence of Ca-phosphate precipitation in blank trials (no kaolinite) it is further suggested that this precipitation maybe enhanced by the presence of the mineral surface (Hawke et al., 1989). However with no indication by XRD of Ca-phosphates present in the kaolinite containing batches, and the significant competitive effect of DOM on phosphate removal in the presence of excess Ca, the mechanism of phosphate removal requires investigation.


Hawke D., Carpenter P.D., and Hunter K.A. 1989. Competitive adsorption of phosphate on goethite in marine electrolytes. Environmental Science and Technology 23:187-191.

He Z.L., Zelazny L.W., Baligar V.C., Ritchey K.D., and Martens D.C. 1997. Ionic strength effects on sulfate and phosphate adsorption on γ-Al and kaolinite: triple-layer model. Soil Science Society of America Journal 61:784-793.

Connectivity of phosphorus transport in a grassland catchment during storm flows

F. L. Wood1,2, P.J. Butler1, A.L. Heathwaite2 and P.M. Haygarth1

1Institute of Grassland and Environmental Research, North Wyke, Okehampton, Devon, EX20 2SB, UK

2Department of Geography, University of Sheffield, Winter Street, Sheffield, S10 2TN, UK


There has been a wealth of studies recording edge of field phosphorus (P) losses (e.g. Jordan and Smith, 1985), but uncertainties have remained about the degree of connectivity through whole catchment systems. The study presented here examined the concept of connectivity beyond these ‘edge of field’ losses.


An empirical approach was used to track the P transfers through a range of nested field sites in a predominantly grassland catchment in Devon, south west England. High temporal resolution storm samples were collected from subcatchments of the River Taw, ranging from 30 m2 to 862 km2 in scale. Over a 48 hour storm period, three-hourly samples were taken simultaneously using a mixture of manual and automated sampling. Spot discharge readings were made manually at the smallest sites, and obtained from calibrated weirs or Environment Agency data for the larger sites. Samples were analysed for total P (Rowland and Haygarth, 1997).

Discussion and conclusion

The study revealed a strong correlation between small and large-scale P transfers. There was also a clear progression in the timing of the peak P concentration as scale increased, consistent with the hypothesis that river fluxes are composed of local inputs plus those travelling in from upstream. We believe that P entrainment from channel banks is likely to be low, because soil P distribution and river erosion studies (Haygarth et al., 1998; e.g. Walling and Woodward, 1995) have shown that extractable P concentrations in the local soil decrease dramatically with depth, and that channel erosion in local rivers generally accounts for no more than 12% of the suspended sediment flux. This implies that edge of field transfers from agricultural land are likely to be a significant source of associated river P fluxes. These may not be the only source, as the total storm flux of P per unit area of catchment does vary considerably from site to site, and tends to be greater than the flux measured from our field scale sites. However, there is some evidence that higher transfers generated from other land types within the catchment might account for this variation between sites.

We therefore conclude that although the relationship between scales is not simple, there is a significant underlying connectivity of transport of P from diffuse agricultural sources to downstream water bodies.


Haygarth, P.M., Hepworth, L., and Jarvis, S.C. 1998. Forms of phosphorus transfer in hydrological pathways from soil under grazed grassland. European Journal of Soil Science 49(1): 65-72.

Jordan, C. and Smith, R.V. 1985. Factors affecting leaching of nutrients from an intensively managed grassland in County Antrim, Northern Ireland. Journal of Environmental Management, 20: 1-15.

Rowland,A.P. and Haygarth,P.M. 1997. Determination of total dissolved phosphorus in soil solutions. Journal of Environmental Quality 26(2): 410-415.

Walling, D.E. and Woodward, J.C. 1995. Tracing sources of suspended sediment in river basins: a case study of the River Culm, Devon, UK. Marine and Freshwater Research 46: 327-326.


We thank Marta Alfaro, Mark Butler, Jane Hawkins, Elaine Jewkes, Carly Kenny, Kevin McTiernan, and Neil Preedy for their help with sampling. This study was funded by the Ministry of Agriculture Fisheries and Food, UK, grant awards NT 1043 and AE 9199.

Theme 3 - Impacts of agriculture-derived phosphorus on water quality

- Oral papers

Defining the impact of agricultural phosphorus on aquatic systems

R.H. Foy1 and S. D. Lennox2

1 Agricultural and Environmental Science Division, Department of Agriculture and Rural Development, Newforge Lane, Belfast BT9 5PX, Northern Ireland

2 Biometrics Division. Department of Agriculture and Rural Development, Newforge Lane, Belfast BT9 5PX, Northern Ireland

For limnologists, the impact of phosphorus (P) on lakes is considered to be comparatively well understood. Firstly the relationship, derived by Vollenweider over 30 years ago, between P transfers from land to water and their impact on lake P concentrations remains an elegant and powerful tool for relating a catchment process, namely P loss, to an impact on a waterway. Secondly there is a broad consensus as to the concentrations of P in lakes that result in oligotrophic, mesotrophic or eutrophic conditions. These concentrations may also be applicable to a trophic classification for rivers.

Currently P may soon move up the environmental agenda as the adoption by the European Union of a Water Framework Directive requires that member countries adopt measures which will maintain and, if necessary, return lakes and rivers to a good ecological status. Although lakes that are naturally eutrophic can be found, they are not particularly common and for many lakes a restoration of water quality implies a reduction in P loadings. The extent to which the ecological status of a lake has been diminished through higher P inputs is examined for Lough Neagh, which is a large lake in north-east Ireland, currently hypertrophic with respect to P and chlorophyll concentrations, and draining a catchment devoted almost entirely to agriculture. A time-series of P concentrations in the lake has been reconstructed based on direct monitoring of the lake since 1970 and, for the century prior to that, changes in the accumulation rates of diatom frustules and chironomid head capsules preserved in the lake sediment. In the late 19th century the lake appears to have been mesotrophic. Enrichment of lake with P was greatest post 1950 and it appears that diffuse inputs of P began to increase only after that date. By comparison the degree of enrichment that occurred before 1950 was limited and can be accounted for by increased discharges of P from point sources. However the sensitivity of lakes to P is such that this limited enrichment resulted in the largest quantitative and qualitative changes in the diatom flora of the Lough. The extent to which the agricultural input of P has increased is compared with changes in the use of P by agriculture within the catchment.



A model of biogeochemical processes in lakes to be used in eutrophication management

B. C. Pers and B. Arheimer

Swedish Meteorological and Hydrological Institute (SMHI), Norrköping, Sweden


Eutrophication is a well-known problem in many Swedish lakes that causes poor drinking water quality, overgrowth, large algae blooms, etc. By stream inflow the lakes receive nutrients leached from soil and emitted from point sources. However, several measures can be taken to restore or preserve the water quality; measures can either be ecotechnological in the lake itself, or changed emissions and drainage-basin management. The choice between nutrient reducing strategies is a question of combining biological effects in the lake with other considerations. A decision support system simulating the effects and costs for water management is currently under development in Sweden within the VASTRA program (Wittgren, 1998). This paper presents how a lake model within this system can simulate the biogeochemical response to different management scenarios.

Description of the model

The newly developed biogeochemical model is linked to an existing physical lake model developed at SMHI (Svensson, 1998). The model is horizontally homogeneous, but vertically divided in a discrete number of layers (typical 20). The only exception is the variable macrophyte, which has a horizontal variation.

The most important variables for nutrient dynamics in lakes are the dissolved nutrient concentrations (nitrogen and phosphorus) and the phytoplankton. Ordinary, phytoplankton use dissolved nutrients to grow, but in phosphorus-rich lakes nitrogen fixation by algae may be important. Shallow lakes are dominated by macrophytes or phytoplankton and can switch between these states. Sediments can be a source or a sink for nutrients. Zooplankton feeds on phytoplankton, but has other food sources (e.g. detritus), which stabilises the population. Top-down control can also be processed from further up in the food web (i.e. from fish). All these variables are potentially important factors for the nutrient dynamics and are therefore included in the model. Totally 14 state variables are used.

Other important processes included in the model are inflow/outflow of water and nutrients through tributaries/discharge, denitrification, nitrification, degradation of detritus and organic matter in sediment, sedimentation of phytoplankton, detritus and macrophyte.

Model response to different input data

In an application where the model results are used for management of a specific lake, data from that lake should be included in the model before doing simulations. So far, the model is run with hypothetical data to investigate the model’s response, however, during the first half of 2001 the model will be evaluated against observed data in a highly eutrophied lake. The present base scenario was complemented with runs with increased nutrient input through inflow, inducing phosphorus (P) and nitrogen (N) limited primary production, and with increased fish. A long run has also been performed to look at long term levels and seasonal variations of different variables.

The base scenario was run for seven years. After the first year the seasonal pattern were similar from year to year. Nitrogen fixating algae dominated all year with a peak in August-September, other phytoplankton were much less abundant and peaked earlier in summer before being grazed down. Zooplankton was most abundant in autumn, while macrophyte peaked in spring.

A doubling of the N inflow increased the macrophyte, which started to take up N from the water when sediment N ran out. It also suppressed the development of N fixating algae compared to the base run, although they still were dominating during autumn due to their lower edibility for zooplankton. A doubling of the P inflow did not influence the macrophyte or phytoplankton except for N fixating algae, which became even more dominating using atmospheric N. An increase in planktivorous fish stock decreased the zooplankton, but increased the phytoplankton in a top-down control.


The model has shown to be able to simulate effects of changing nutrients and plankton dynamic, and even top-down control by fish. The macrophyte part of the model is still immature. It needs to be studied further and the model adjusted. Still, the model shows promising results to be used as an evaluation instrument for various measures to improve water quality in lakes.


Svensson, U. 1998. PROBE Program for Boundary Layers in the Environment System Description and Manual. Swedish Meteorological and Hydrological institute, Report RO No. 24, Norrköping.

Wittgren, H. B. 1998. Water management research towards catchment-based strategies for sustainable resource use. Vatten 54: 295-300.

Freshwater eutrophication - linking nutrient levels and biological impacts

R. M. Dils and S. S Leaf

Environment Agency, National Centre for Ecotoxicology and Hazardous Substances, Evenlode House, Howbery Park, Wallingford, Oxon, OX10 8BD, UK

Extent of the aquatic eutrophication problem in England and Wales

Aquatic eutrophication was historically regarded by some as an exacerbation of a natural phenomenon giving rise to local problems in a small number of areas in England and Wales. Recent evidence, particularly as regards the freshwater environment, indicates that it is more than a limited localised problem in England and Wales. From 1989 to 1997, for example, some 3,000 different freshwater bodies (mainly standing, but also running waters) have been affected by algal blooms. To date, 80 Sensitive Areas (Eutrophic) have been designated under the Urban Waste Water Treatment Directive, of which 62 are rivers or canals, 13 are lakes/reservoirs and 5 are estuaries. Under this initiative, waters receiving discharges from large sewage treatment works were assessed for the effects of eutrophication using chemical, biological and use-related criteria. Excessive weed growth and changes to macrophyte communities were the main ecological impacts detected, although benthic and floating mats of blue-green algae have caused problems in some slow flowing rivers. Eutrophication is also implicated in depreciating the conservation value of many stillwater Sites of Special Scientific Interest (SSSIs). In a survey commissioned by English Nature, 84% out of a sample of 102 SSSIs showed symptoms of eutrophication, and in 68% of cases this ‘had overtly affected the nature conservation interest’ (Carvalho and Moss, 1995).

Drivers for developing improved methods for assessing eutrophication

In response to the impacts and risks associated with aquatic eutrophication, and the range of statutory and international commitments requiring action to be taken to address such problems, the Environment Agency (of England and Wales) launched a eutrophication management strategy in August 2000 (Environment Agency, 2000). The strategy aims to strike a balance between the recognised need for further and improved management action and the uncertain benefits of control measures, stemming from a relatively poor, albeit improving, understanding of cause and effect. Supporting further research into improved methods for assessing and monitoring the impacts of eutrophication is a key component of the strategy. The need for increased use of biological indicators and ecological assessment is recognised, although the relationship between nutrient levels and the desired ecological objectives can be unpredictable and differs according to waterbody type. This shift in emphasis from predominantly physico-chemical parameters towards biological parameters is in line with the newly adopted EU Water Framework Directive, which specifies the need to assess ecological quality (including physico-chemical, biological and hydromorphological elements) and to achieve good ecological status.

Historically, there has been limited linkage between monitoring for nutrient levels and biological parameters. For standing waters, the OECD trophic classification is based on the observed relationship between total phosphorus and phytoplanktonic chlorophyll concentrations, but has limited application for some water bodies (e.g. shallow lakes). In recent years, biological indicator schemes for rivers, such as the trophic diatom index (TDI), and trophic ranking scheme for macrophytes (MTR) have been developed. These have been employed in assessing the extent of eutrophication for some locations (e.g. for the designation and review of Sensitive Areas (eutrophic) under the Urban Waste Water Treatment Directive). However, these methods may still need refinement, and biological classification systems for other water bodies are less well developed. Further research is required before nutrient conditions and biological quality indicators can be linked with an acceptable degree of reliability. The ability to link nutrient conditions with biological quality elements will significantly affect decisions taken regarding the Water Framework Directive, implementation of the Agency’s eutrophication strategy, UK Biodiversity Action Plans, Habitats Directive, Nitrates Directive and Urban Waste Water Treatment Directive.

In this paper, we will discuss current methods and best practice in England and Wales for assessing and monitoring eutrophication. We will also discuss the implications of the Water Framework Directive, in particular, the need to improve current understanding of the relationship between nutrient concentrations and biological quality for different water body types and different levels of ecological status.


Carvalho, L. and Moss, B. 1995. The current state of a sample of English Sites of Special Scientific Interest subject to eutrophication. Aquatic Conservation: Marine and Freshwater Ecosystems 5: 191-204.

Environment Agency. 2000. Aquatic eutrophication in England and Wales; a Management Strategy. Environment Agency, Bristol, 32pp.

Towards modelling the transport and fate of phosphorus in river systems

A. J. Wade1, P. G. Whitehead1 and G. M. Hornberger2

1Aquatic Environments Research Centre, Department of Geography, University of Reading, Reading RG6 6AB, UK

2Department of Environmental Sciences, University of Virginia, Charlottesville, VA 22903, USA


It is now widely accepted that phosphorus (P) is the major limiting nutrient in UK freshwater systems. As such there are concerns regarding the effects of increased P loads to lakes and river systems, given increased loads can enhance the nutrient status of a water body and lead to excessive phytoplankton, macroalgae and macrophyte growth. Mathematical models are needed to aid the understanding of P in aquatic systems because such models begin to link ideas of P transport and storage. Moreover models are required to quantify the potential impacts of changing P loads on the water quality and ecology of aquatic environments. Such quantification is especially important for assessing the consequences of legislation including the Urban Wastewater Treatment and Water Framework Directives. Thus, the aim of the work presented is to create a mathematical model of the principal mechanisms controlling the in-stream P and macrophyte dynamics in the river systems, and to add a land-phase component to create a distributed, catchment-scale model.

Modelling strategy

A new model of in-stream P and macrophyte dynamics has been developed (Wade et al., 2001a, b). Based on mass-balance equations, the model represents the interactions between phosphorus and the suspended and bed sediments, the uptake of P by epiphytes and macrophytes and the exchange of P between the water column and the pore water. The model simulates the total P (TP) and the soluble reactive P (SRP) concentrations observed in a reach of the River Kennet. Furthermore, the model simulates the generalised macrophyte growth patterns and total biomass observed in rivers throughout S. England. The reach, which is 1.5 km long, is immediately downstream of Marlborough Sewage Treatment Works, where P reduction by tertiary effluent treatment began in September 1997. The model is used to simulate the flow, water chemistry and macrophyte biomass within the reach, both before and after P removal from the effluent. A General Sensitivity Analysis, based on Monte Carlo simulations and parameter values derived from the literature, identifies the key parameters controlling the model behaviour when simulating macrophyte growth. Furthermore, the Kennet Model was applied to a reach of the River Kennet to investigate the impacts of changing flow conditions on macrophyte growth. The investigation was based on the assessment of two flow change scenarios, which both included the simulation of decreasing TP concentrations from a Sewage Treatment Works due to improved effluent treatment. In the first scenario, the precipitation and potential evaporation outputs from a climate change model (HadCM2 GGx) where input into the catchment model INCA to simulate the mean daily flow response within the reach. In the second scenario, the mean daily flows observed in a historically dry year were repeated as input to the in-stream model to simulate an extended low flow period over two years. Recent work has been done to model P dynamics in the plant/soil system based on mass-balance, and link plant/soil system and in-stream models.

Results and discussion

Model simulations of in-stream dynamics indicate that epiphyte smothering is an important limitation to macrophyte growth, and that higher stream and pore water SRP concentrations allow the earlier onset of growth for the epiphytes and macrophytes respectively. Higher flow conditions are shown to reduce the simulated peak epiphyte biomass, though at present, the effect of flow on the macrophyte biomass is unclear.

The simulation results suggest that changes in the seasonal distribution of flow were not detrimental to macrophyte growth. However, the simulation of extended periods of low flow indicates that a proliferation of epiphytic algae occurs, even when the in-stream P concentrations are reduced due to effluent treatment. This epiphytic growth was predicted to reduce the macrophyte peak biomass within the reach by around 80%. Thus, the model simulations suggest that flow was more important in controlling the macrophyte biomass in the River Kennet than the in-stream P concentrations, which are elevated due to agricultural diffuse sources. Preliminary results of the application of the integrated catchment-scale model to the River Bure in East Anglia will be reviewed.


Wade, A. J., Whitehead, P. G., Hornberger, G. M., Jarvie, H. P. and Flynn, N. 2001a. On modelling the impacts of phosphorus stripping at sewage works on in-stream phosphorus and macrophyte/epiphyte dynamics: a case study for the River Kennet. Sci. Tot. Env., In press.

Wade, A. J., Whitehead, P. G., Hornberger, G. M. and Snook, D. 2001b. On modelling the flow controls on macrophyte and epiphyte dynamics in a lowland UK catchment: River Kennet. Sci. Tot. Env., In press.

Phosphorus losses from agriculture: effects on Canadian aquatic ecosystems

P. A. Chambers1, M. Guy1, E. Roberts2 and G. Grove1

1Environment Canada, National Water Research Institute, 11 Innovation Blvd., Saskatoon, SK S7N 3H5, Canada

2Environment Canada, Environmental Quality Branch, 351 St. Joseph Blvd, Hull, Québec, K1A 0H3, Canada

Phosphorus application to Canada’s agricultural land

Approximately 68 million hectares, or 7% of Canada’s total land area, is in farmland with 35 million hectares in cropland. In 1996, 307x103 tonnes of phosphorus (P) as fertilizer were applied to agricultural lands in Canada at concentrations ranging from 10 kg P ha-1 in the Prairies to 32 kg P ha-1 in the Atlantic Provinces. In addition, an estimated 155 x103 tonnes P as manure were applied at concentrations ranging from 44 kg ha-1 P in Québec to 184 kg ha-1 P in British Columbia. Overall, Canada had a P surplus of 0.8 kg ha-1 for all agricultural land or 1.6 kg ha-1 for cropland in 1996 (Chambers et al., 2001). Although the average P surplus is comparatively low, regional P surpluses or deficits have occurred. For example, in Québec, soils under annual crops were over-fertilized in the late 1980s and early 1990s because of excessive manure application. Conversely, cropland in the Prairies and the Maritimes was, on average, under-fertilized in 1991 compared to recommended P values.

Ecological effects of phosphorus loading

Phosphorus losses from agricultural land in Canada have accelerated eutrophication of certain rivers, lakes and wetlands resulting in loss of habitat, changes in species biodiversity and, in some cases, loss of recreational potential. For example, P loading from agricultural activities to the Yamaska River in southwestern Québec has accelerated the growth of algae and rooted aquatic plants to such an extent that the aesthetic value of the river and its potential for recreational activities are impaired and fish kills have occurred due to oxygen depletion (Painchaud, 1997). A recent study of surface water quality in Alberta found that P concentrations often (> 85% of all samples) exceeded the interim provincial guideline of 0.05 mg L-1 total P for the protection of aquatic life, especially in areas of intensive farming but even in areas of low-intensity farming (CAESA, 1998).

Management of agricultural nutrient losses

To minimize the release of nutrients to the environment from agricultural activities in Canada, most provinces have recently instituted or are developing nutrient management strategies for managing the production, storage and utilization of agricultural nutrients. The aim of these strategies is to ensure that added nutrients (fertilizer, manure, biosolids, etc.) are applied as function of soil test results and crop nutrient requirements, that nutrient sources are properly contained during storage, and that environmental risk associated with agricultural nutrients is identified and mitigated.


CAESA [Canada-Alberta Environmentally Sustainable Agriculture Agreement]. 1998. Agricultural impacts on water quality in Alberta; an initial assessment. Lethbridge, Alberta, Canada.

Chambers, P.A., R. Kent, M.N. Charlton, M. Guy, C. Gagnon, E. Roberts, G. Grove, and N. Foster. 2001. Nutrients and their impact on the Canadian environment. Government of Canada report, Ottawa, Ontario, Canada.

Painchaud, J. 1997. La qualité de l’eau des rivières du Québec: etat et tendances. Ministère de l’Environnement et de la Faune, Direction des écosystèmes aquatiques, Québec, Québec, Canada.

Theme 3 - Impacts of agriculture-derived phosphorus on water quality

  • Poster papers


Assessment and prediction of nutrient loads from a river basin

M. Chandra Sekhar and S. Sathish Kumar

Department of Civil Engineering, Water & Environment Division, Regional Engineering College, Warangal - 506 004, INDIA

Monitoring and modelling are two complementary approaches necessary for the assessment of river water quality degradation, often generated by different sources. Monitoring and modelling of diffuse pollution is more complicated than modelling and monitoring point sources. One of the important tasks of water quality management is identification, determination and assessment of both and point and non-point sources. Non-point sources are difficult to identify and quantify making implementation of effluent limitations almost impossible. The non-point sources pollution reflects the amount of elements distributed in the basin and their leaching to watercourse. Water quality management in developing countries including India, is still confronted with identification, assessment and control of non-point source pollution. Since the majority of nutrient loads are of diffuse origin, an exact understanding of the loading pattern is a prerequisite to reduce their impacts on receiving waters. The present paper describes a methodology for assessment of nutrient loads from the river basin under study.

The Musi River is a tributary of the River Krishna. The river basin lies between North Latitudes 16o 45' and 17o 45' and East Longitudes 77o 45' and 79o 30'. The basin is approximately L shaped and the terrain is flat to gently undulating except for a few hillocks and valleys. The basin experiences rainfall in one of the three seasons (summer from March to May, the monsoon from June to November and the Winter from December to February) of the year. The rainfall in the non monsoon period is insignificant. The average annual rainfall of the catchment is 800 mm. The model formulated attempts to evaluate the different sources of pollution in the river basin including the point sources. The predicted loads are in agreement with observed loads, with a few deviations due to data constraints. At present, nearly all emphasis of pollution abatement in the basin has been placed on the construction of elaborate wastewater treatment plants to control point source pollution. The results emphasise the need for non-point source controls for effective water quality management.

Relationship between erosion and algal-available phosphorus in agricultural runoff

P. Ekholm1

1Finnish Environment Institute, P.O. Box 140, 00251 Helsinki, Finland

In Finland, crop production areas are located mainly in the southern parts of the country, where the most severe eutrophication problems of lakes, rivers and coastal waters are also encountered. In this region, soils tend to be fine, surface runoff the major transport pathway and particulate phosphorus (P) the dominant P fraction in agricultural nutrient loading. The current P control measures – applied largely by EU’s agri-environmental subsidies – have two basic aims. First, to reduce excessive soil P status by regulating P fertilization and manure application. And second, to reduce erosion e.g. by increasing winter soil cover, or to capture the soil particles which have already left the fields e.g. by wetlands and sedimentation ponds. In order to become available to algae – and to contribute to eutrophication – P in soil particles has to be released into a dissolved form. This release may occur immediately when surface soil is in contact with rain or snowmelt water, or later from eroded soil particles during their transport from field to receiving waters. It is widely accepted that the losses of dissolved reactive P (DRP) increase with labile soil P, which justifies the reduction of soil P status for eutrophication abatement. However, the effect of erosion control on the losses of algal-available P is a more controversial subject. This is partly due to the fact that the ultimate fate – and effect – of P in eroded soil particles remains largely unknown.

In laboratory water extractions, the P concentration in solution increases rapidly when the soil-to-water ratio is increased from low to moderate values, but upon further increase in the ratio, the P concentration tends to gradually level off (e.g. Yli-Halla et al., 1995). Such a pattern is attributed to sorption-desorption equilibrium between soil and water; with increasing concentrations of dissolved phosphate and other ions in water, the desorption of P per unit amount of soil decreases, leaving higher proportion of potentially desorbable P in the solid phase. In the surface runoff from a clay soil under wheat production, an analogous relationship between DRP and total suspended solids (TSS) was found (Ekholm et al., 1999). This relationship was of a form DRP = α TSS β (β < 1) and could be interpreted as follows: DRP in runoff originates largely from eroded soil particles and its concentration depends on the sorption-desorption equilibrium between eroded soil and surrounding solution. In addition, the exponential relationship between DRP and TSS might reflect the varying characteristics of eroded soil on increasing soil loss; at high TSS concentration range, the share of coarser particles increases. These particles may be lower in desorbable P than the finer ones, which dominate at lower TSS concentrations. However, a substantial proportion of DRP may originate from the bulk soil not removed by erosion. The concentration of DRP in surface runoff has been found to increase with the (effective) depth of the interactive soil layer (Sharpley et al., 1985). This depth, in turn, increases with soil loss (Sharpley, 1985). In this case, TSS in runoff would indicate the magnitude of the contact between water and soil in the field (and thus the release of DRP).

In this study, the relationship between P and TSS was examined in runoff from 7 agricultural catchments (0.12–1088 km2) in southern Finland. The data consisted of 2–11 years of frequent water quality and quantity observations. As a part of a joint project estimating the effects of EU’s agri-environmental subsidies on nutrient loads, the aim here was to estimate the potential effect of erosion control on algal-available P. In all the catchments, DRP tended to increase exponentially with TSS. However, occasional peak values of DRP were observed at low TSS concentrations (and usually at low flow). In these cases, the source of DRP cannot have been TSS but surface soil (incl. plant residues). The P content of TSS decreased with TSS, most strongly at low concentration range of TSS, confirming the hypothesis that with increasing erosion, the share of particles low in P increases. According to a stepwise regression analysis performed on the flow-weighted mean concentrations of each catchment, TSS concentration and the P content of TSS explained 71% of the variation in DRP. Assuming that erosion control cuts down TSS peaks, it appears not to reduce DRP very efficiently (due to the nonlinear relationship between TSS and DRP). However, with a decrease in TSS, particulate P will also decrease, although to slightly lower extent. The corresponding change in algal-available particulate P remains unknown, but here too the relative reduction is probably somewhat less than in the case of TSS. These hypotheses will be tested later, especially against data on algal-availability of particulate P as measured with an anion exchange resin.


Ekholm, P., K. Kallio, E. Turtola, S. Rekolainen, and M. Puustinen. 1999. Simulation of dissolved phosphorus from cropped and grassed clayey soils in southern Finland. Agric. Ecosyst. Environ. 72:271–283.

Sharpley, A.N. 1985. Depth of surface soil-runoff interaction as affected by rainfall, soil slope, and management. Soil Sci. Soc. Am. J. 49:1010–1015.

Sharpley, A.N., S.J. Smith, W.A. Berg, and J.R. Williams. 1985. Nutrient runoff losses as predicted by annual and monthly soil sampling. J. Environ. Qual. 14:354–360.

Yli-Halla, M., H. Hartikainen, P. Ekholm, E. Turtola, M. Puustinen, and K. Kallio. 1995. Assessment of soluble phosphorus load in surface runoff by soil analyses. Agric. Ecosyst. Environ. 56:53–62.

The role of sediment transport in the movement of phosphorus in lowland streams

D. J. Evans1,2, P. J. Johnes1 and D. S. Lawrence2

1Aquatic Environments Research Centre, Department of Geography, University of Reading, UK

2Postgraduate Research Institute for Sedimentology, University of Reading, UK

A monitoring program designed to identify the role of sediments in phosphorus (P) transport in river systems was conducted at two contrasting tributaries of the River Kennet, Berkshire, UK, between March 1999 and March 2000. Bedload material transported along the bed of the river was sampled using 'pit-type' sediment traps that were emptied weekly. Suspended sediment was extracted from the water column using a pump/centrifuge system.

Samples were wet sieved and individual size fractions were analysed for major ion composition. The finest fraction of bedload sediment (<0.038 mm) was generally found to contain the highest concentrations of P and accounted for an average of 20 and 27 percent, respectively, of the total sediment weight transported for the Rivers Enborne and Lambourn. The lowest concentrations of P occurred in the coarsest fraction (>0.250 mm). Suspended sediment had similar P content and particle size distribution to the fine bedload sediment. The samples were also analysed for grain size distribution, mineralogy, specific surface area and organic matter content to identify the contribution of specific sediment 'types' to instream P transport. Samples from both sites with a higher abundance of constituent minerals, a higher surface area and a lower particle size generally contained higher concentrations of P. However, coarse size fractions with a high percentage of organic matter also contained high concentrations of P.

Work using a novel digital overlay mapping method has identified the presence of P 'hotspots' within the mineral matrix. These regions of high P concentration are spatially correlated with elevated levels of calcium in the Lambourn sediments, and aluminium and iron in the Enborne sediments. This indicates the importance of co-precipitation and complexation reactions in removing soluble P from the water column to mineral forms in the bed sediments.

Temporal variations in P fractionation, pH and alkalinity, suspended sediment, cation, anion and oxygen concentrations were also determined under baseflow and stormflow conditions. These data have been evaluated to identify interactions between bed sediments and water column P. The environmental conditions under which the majority of instream P transport occur have been highlighted. In addition, the roles that sediment precipitation and re-suspension play in regulating soluble P concentrations in the water column have been clarified.

Theme 4 - Integrated catchment management and mitigation of phosphorus transfer

- Oral papers

Connecting fields to the river: the need for a spatially distributed approach to modelling phosphorus transport in agricultural catchments

Tim Burt

Department of Geography, University of Durham, Durham DH1 3LE, UK

All fields are not created equal when it comes to phosphorus (P) export: some are clearly more important sources of P than others. Although we still have much to learn about the factors that control P loss from a given field, we seem now to have reasonable understanding of the relevant chemical and erosional processes. What is lacking is an ability to upscale from the field to the catchment scale.

The 'critical source area' is an important new concept in relation to P transport in catchments. It acknowledges the importance of hillslope hydrology in driving the delivery of overland flow, eroded soil and, thus, a large fraction of the total P export to the river channel. Two models of overland flow production are commonly recognised in the hillslope hydrology literature. Betson’s partial contributing area model invokes Horton’s infiltration-excess overland flow mechanism; overland flow is only produced from certain fields, those where rainfall intensity exceeds the infiltration capacity of the soil. The partial area model seems particularly relevant to agricultural catchments, Horton’s studies having been originally undertaken in the US mid-west. Hewlett’s variable source area model was first defined in the forests of the southern Appalachians. Based on the saturation-excess overland flow mechanism, it shows that only those parts of a catchment where the soil is saturated produce overland flow. Such source areas may enlarge seasonally and during storms. Despite preoccupation with the VSA model in recent decades, the partial contributing area model appears rather more relevant as far as soil erosion and P delivery is concerned (although in some cases, for example erosion of thalweg rills, the VSA model may also be important). What remains lacking, however, is knowledge of how individual fields link to the stream channel. Studies that characterise the connectivity of the hillslope hydrological system will be crucial therefore. In terms of landscape sensitivity, it is not just the field itself that is important but also its location: site and situation. It must be added that research that indicates the role of other P delivery mechanisms, for example tile drains, will also be important for certain localities.

Alongside empirical research into P delivery mechanisms, it will be vital to develop spatially distributed models of P transport in catchments. Eventually, such models must be applicable to large river basins, but in the first instance it will be necessary to build models at a more modest scale, catchment areas of the order of several tens to a few hundred square kilometres. Such field-scale models should be capable of identifying critical source areas for P loss within the catchment system, not merely indicating the fields that are sources of P but, critically, showing which P sources will actually deliver significant amounts of mobilised P to the river. Such models are likely to be built within a GIS environment and will emphasise the importance of landscape structure on P pollution transport in agricultural catchments.

Not every field contributes equally to P export from a river basin; spatially explicit models can indicate where in the landscape the risk is highest. Best management practices can then be targeted at the most sensitive locations.





Modelling through soil losses of phosphorus to surface waters

M. B. McGechan

Scottish Agricultural College, West Mains Road, Edinburgh, EH9 3JG, UK

It is commonly assumed that phosphorus (P) losses to surface waters arise mainly due to transport in surface runoff (overland flow). Soil erosion arises due to rainfall impact with preferential sorption (enrichment) of P onto very fine soil derived sediments which move with overland water flows. However, in Northern Europe many agricultural fields (both grassland and arable cropped) have tile drains, and if these are functioning correctly surface runoff rarely happens. Nevertheless, substantial losses of P to surface waters do occur, which can only be explained in terms of sorbed P attached to sediments or colloidal material moving through the soil to tile drains. Phosphorus losses can be particularly high following spreading of livestock slurry (liquid manure) on wet ground. The objective of the work described here is to develop a model description of the process of through soil transport to surface waters of colloid associated P.

The dual-porosity soil water and contaminant transport model MACRO (Jarvis, 1994) has been tested for its suitability to represent water flows and leaching of P through field drains following spreading of slurry. These flows are characterised by very high loadings of P, mainly in particulate form, for about one week following winter spreading of slurry, with a rapid decline thereafter to the low background level. Use was made of a recently developed option in MACRO for representing colloid facilitated contaminant transport (Jarvis et al., 1999), previously applied to representation of water pollution by pesticides (Villholth et al, 2000). When applied to water pollution by P, both soluble and particulate (colloidally attached) P can be represented in simulated outputs. Calibration involved selecting soil hydraulic parameters and P sorption characteristics for two soils from measured and literature values. Both rapid, reversible sorption of P to surface sites, and the slow, time-dependent reactions which deposit P at depth below surfaces (as reported extensively in the literature, e.g. Barrow and Shaw, 1975, Barrow, 1983) have been represented. Sorption by such processes occurs both onto static soil components (which lock up P) and onto mobile colloidal particles from slurry or soil (which aggravate the pollution problem). The dual porosity feature of MACRO gives separate representation of processes in macropores and in micropores (soil matrix pores). There is representation of restrictions on colloid movement due to straining and filtration, which are severe in micropores and almost non-existent in macropores.

Reasonable agreement was found between simulated water and leached P flows compared to measurements reported by Hooda et al. (1999). Work with the model suggests that macropore flow through the soil to field drains of colloidally transported P is an important component of water pollution associated with slurry spreading, particularly that which occurs immediately after spreading. Results show large polluting P flows to drains when slurry is spread on wet soil (with water-filled macropores), but negligible flows when slurry is spread on fairly dry soil in which macropores contain no water.


Barrow, N. J. 1983. A mechanistic model for describing the sorption and desorption of phosphate by soil. J. Soil Sci. 34: 733-750.

Barrow, N. J. and T. C. Shaw. 1975. The slow reactions between soil and anions; 2. Effect of time and temperature on the decrease in phosphate concentration in the soil solution. J. Soil Sci.119: 167-177.

Hooda, P. S., M. Moynagh, I. F. Svoboda, A. C. Edwards, H. A. Anderson and G. Sym. 1999. Phosphorus loss in drainflow from intensively-managed grassland soils. J. Environ. Qual. 28: 1235-1242.

Jarvis, N. 1994. The MACRO model - Technical description and sample simulations. Reports and Dissertations 19, Swedish University of Agricultural Sciences, Department of Soil Sciences, Uppsala.

Jarvis, N., K. G. Villholth and B. Ulén. 1999. Modelling particle mobilization and leaching in macroporous soil. Eur. J. Soil Sci. 50: 621-632.

Villholth, K. G., Jarvis, N. J., Jacobsen, O. H. and de Jonge, H. 2000. Field investigations and modelling of particle-facilitated pesticide transport in macroporous soil. J. Environ. Qual. 29: 1298-1309

Assessing the risk of phosphorus loss from land to surface water – a farm management tool

M. R. Hart1, J. Petersen2, S. Elliott3, M. J. Stroud3, B. F. Quin1, A. B. Cooper3 and L. Nguyen3

1Summit-Quinphos (NZ) Ltd., P.O. Box 24-020, Auckland, New Zealand

2National Institute of Water and Atmospheric Research Ltd. (NIWA) P.O. Box 109695, Auckland, New Zealand

3National Institute of Water & Atmospheric Research Ltd. (NIWA) P.O. Box 11-115, Hamilton, New Zealand

Over the last decade or so there has been increasing worldwide interest in and concern over the transport of nutrient elements from agricultural land to surface water, which may lead to eutrophication and its concomitant problems. Phosphorus (P) in particular is seen as the main limiting nutrient in aquatic ecosystems, and there have been many papers published on measurements of various forms of mobile P, of different mechanisms of transport via various hydrological pathways, and under various farm management systems (e.g. reviews by Gillingham and Thorrold, 2000; Haygarth and Jarvis, 1999). However, while there is now a large body of useful data on the principles behind P losses in agricultural run-off, there have been relatively few attempts to come up with practical solutions to the problem, e.g. the P Index developed by the USDA (Lemunyon and Gilbert, 1993).

At the practical farm level there is a need to identify the location of 'high risk' P loss areas on a farm, reasons why they are high risk and therefore management strategies that could be used to reduce this risk. For such risk identification and management to be effective it must be simple, cost effective, widely applicable and use information that is readily available. Additionally the approach must have a sound scientific basis and be technically defensible. This paper describes an approach that is being developed by NIWA in association with a New Zealand fertiliser importer and distributor, Summit-Quinphos (NZ) Ltd., named Nutri-Save®. The output is a P loss risk index that is intended to represent, in a semi-quantitative way, the relative risk of P input to surface waters from management units of land, typically a farm paddock. It has been shown that the loss of P in surface run-off is greatly reduced when it is applied in the form of direct application phosphate rock compared to superphosphate forms. The information from the P risk model, in highlighting areas of a farm that pose a risk of losing environmentally significant quantities of P to surface waters, can therefore be used in recommending slowly-available or fully-soluble forms of P to farmers, as appropriate, to specifically identified areas.

The conceptual framework of the index is based on the constraints mentioned above, existing knowledge of the scientific principles of water movement and sediment transport and existing mathematical models of P loss transport processes and relationships. The development of the index is ongoing and iterative with the emphasis at this stage being on developing and trialling the approach.

Three driving factors of total P loss were identified: surface runoff; erosion and transport of particulate P; and extraction and transport of soluble P. Key sub-factors influencing each of these driving factors were also identified and put into the model. These sub-factors were combined to give a high, medium or low risk assessment to each driving factor and these were in turn were combined to give an overall risk assessment for total P loss.

The risk of surface runoff occurring was determined using a rainfall intensity index (based on the number of days in a year with over 50 mm of rain) and a soil infiltration index (based on the soil type). A soil erosion risk factor was calculated from the surface runoff factor, the assessed erodibility of the soil type, and slope. A particulate P loss risk factor was calculated from the soil erosion factor and soil Olsen P levels. A soluble P loss risk factor was calculated from the runoff factor, the soil P retention capacities, and the soil Olsen P levels.

The P risk assessment index is calculated in a Geographic Information System (GIS) environment by overlaying and combining grids (30 m x 30 m cell size) of each of the input variables. The output for viewing by fertiliser representatives and farmers is an ArcView GIS map of each farm identifying areas with a different relative risk of P loss. This Nutri-Save® map can be interrogated to find out the values and indices of each input variable and therefore the reasons why any area has been calculated to have a particular risk value. This information can then be used to help make fertiliser and farm management decisions that minimise the loss of P to New Zealand waterways.


Gillingham, A.G., and B.S. Thorrold. 2000. A review of New Zealand research measuring phosphorus in runoff from pasture. J. Environ. Qual. 29:88-96.

Haygarth, P.M., and S.C. Jarvis. 1999. Transfer of phosphorus from agricultural soils. Adv. Agron. 66:195-249.

Lemunyon, J.L., and R.G. Gilbert. 1993. The concept and need for a phosphorus assessment tool. J. Prod. Agric. 6:483-486.

Reversing the upward trend in soluble phosphorus losses in drainflow from a grassland catchment

R. V. Smith1, S. D. Lennox2 and J. S. Bailey1

1Agricultural and Environmental Science Division, Department of Agriculture and Rural Development, Newforge Lane, Belfast BT9 5PX

2Biometrics Division, Department of Agriculture and Rural Development.

Monitoring of rivers in the Lough Neagh catchment area in Northern Ireland has shown that over the period 1974-1992 soluble reactive P (SRP) loads have increased by rates close to 15 g ha-1 yr-1 (Foy et al., 1995). Although these losses to surface waters were only 0.15% of the phosphorus (P) accumulation rate in soils in the catchment they nevertheless represented an annual increase of 1.54± 0.54 mg P l-1 yr-1 in rivers which is significant in terms of lake eutrophication. Direct proof that increases in SRP transported in drainflow were a response to soil P accumulation came from temporal studies on a 6 ha grassland catchment located at Greenmount Agricultural College, County Antrim. Smith et al. (1995) showed that SRP had increased between 1981-1991 by 1.1 mg P l-1 yr-1 in drainflow from this catchment in response to a soil P accumulation rate of 24 kg ha-1 yr-1.

The aims of the present study were to identify whether the concentrations of SRP and other P fractions in drainflow had further increased in the period 1989-1997 and assess the response to a new management regime introduced in 1998 and 1999. From 1998, slurry was no longer applied to the site and only sufficient fertiliser P was applied to maintain each field at Index 2 (16-25 mg P kg-1) according to ADAS recommendations (MAFF, 1994). In effect, the changes imposed an average P deficit for the site during 1998 and 1999 of 15.0 kg P ha-1 yr-1 compared to a mean surplus of 23 kg P ha-1 yr-1 for the period 1989-1997.

Regressions of median concentrations of P fractions in land drainage waters against time showed significant increases for SRP and soluble unreactive P (SUP) but not for particulate P (PP) (Table 1). Results for SRP suggest that concentrations were increasing at a rate of 2.4± 0.65 mg P l-1 yr-1 for the period 1989-1997 compared to only 1.1 mg P l-1 yr-1 for the period 1981-1991 (Smith et al., 1995). This increase in SRP concentration is interpreted as reflecting an increase in equilibrium P concentration in soil solution resulting from the increase in soil P reserves over the study period.

By employing the SRP versus time regressions (Table 1) one can predict that the median concentrations of SRP in drainflow in 1998 and 1999 would be 50 and 52 mg P l-1, respectively. However, the observed concentrations in drainflow following the application in 1998 and 1999 of the new management regime to the catchment were 40 and 43 mg P l-1, respectively. This reduction in approximately 10 mg P l-1 is an encouraging result and suggests that responses to a lowering of P inputs to grassland soils is relatively rapid.

Table 1. Median concentrations of phosphorus fractions vs. time (Year 1989=1) regressions for land drainage waters from the Greenmount Agricultural and Horticultural College catchment for the period 1989-1997.

P Fraction


Standard error


Standard error




















Total P






** Significant at the 0.01 probability level.



Foy, R.H., Smith, R.V., Jordan, C. and Lennox, S.D. 1995. Upward trend in soluble phosphorus loadings to Lough Neagh despite phosphorus reduction at sewage treatment works. Water Res. 29:1051-1063.

Ministry of Agriculture, Fisheries and Food 1994. Fertilizer recommendations for agricultural and horticultural crops. RB 209. 6th ed. HMSO, London.

Smith, R.V., Lennox, S.D., Jordan, C., Foy, R.H. and McHale, E. 1995. Increase in soluble reactive phosphorus transported in drainflow from a grassland catchment in response to soil phosphorus accumulation. Soil Use Manage. 11:204-209.


Theme 4 - Integrated catchment management and mitigation of phosphorus transfer

Model resolution effects on the prediction of phosphorus transport from agricultural fields to freshwaters

E. O. Ampontuah, L. P. Simmonds and J. S. Robinson

Department of Soil Science, The University of Reading, Whiteknights, PO Box 233, Reading, RG6 6DW, UK


This study examines potential phosphorus (P) losses at selected fields in the United Kingdom using various P transport models, with emphasis on the impact of different scales of model input parameters for typical combinations of slopes, soils, vegetation and climate. Phosphorus transport from agricultural fields constitutes a grave concern worldwide due to its effect on eutrophication of surface waters. Prediction of movement of the nutrient within the soil environment is essential for its effective management. It is well established that the major pathway for P transport from agricultural soils is through soil particulate loss, particularly via runoff (Sharpley and Withers, 1994; Gillingham et al., 2000). The fractional groundcover and topography, through their effects on runoff and soil erosion, may therefore influence P export. Dense groundcover offers total protection of the soil surface against raindrop impact and intercepts overland flow thereby minimising the erosive force of the runoff. There is the tendency for more soil erosion to occur on steep slopes compared with gentle slopes.

Research priorities

Soil erosion models (e.g. CREAMS, WEPP) have been used in some studies to predict accurately P transport from field plots. Gillingham et al. (2000) observed that the CREAMS model accurately predicted P loss at the field scale but was less reliable at the catchment scale. Johnes and Hodgkinson (1998) used the export coefficient modelling approach for P export prediction, which is applicable to large catchments, but their model was capable of predicting only annual P exports rather than predictions for shorter time scales. There is the need for a means of extrapolating successfully applied models at the small scale to the larger scale.

The resolution of the input data set for model predictions is critical in obtaining accurate simulation of solute transport in the soil. While the controlling mechanisms may be well understood and input parameters for modelling P transport may be accurately measured at field plot scale, extrapolating this information so that it can be useful at the catchment scale may not be simple. Most research work on P transport has been conducted on uniform plots under site-specific conditions and understanding of how the amounts of P exports are modified as catchment size increases is very poor (Johnes and Hodgkinson, 1998). The collection of high-resolution data-sets for parameterising distributed process-based models may be costly and a cumbersome task, yet model predictions based on input parameters with low resolution may not accurately reflect the transport processes occurring in-situ (Inskeep et al., 1996). At fine resolutions the processes of flow are well estimated but as the resolution becomes coarser, macroscale characteristics tend to dominate solute transport. As a result, significant losses of P may be predicted using high spatial resolution whereas model predictions with input parameters measured at low spatial resolutions may indicate relatively small losses. The implications for utilisation of remote sensing techniques for estimation of groundcover and topography as input parameters for P transport models at various spatial scales are discussed.


Gillingham, A.G. and Thorrold, B.S. 2000. A review of New Zealand research measuring phosphorus in runoff from pasture. Journal of Environmental Quality 29: 88-96.

Inskeep, W.P., Wraith, J.M., Wilson, J.P., Snyder, R.D., Macur, R.E. and Gaber, H.M. 1996. Input parameter and model resolution effects on prediction of solute transport. Journal of Environmental Quality 25: 453-462.

Johnes, P.J and Hodgkinson, R.A. 1998. Phosphorus loss from agricultural catchments: pathways and implications for management. Soil Use and Management 14:175-185.

Sharpley, A.N. and Withers, P.A. 1994. The environmentally sound management of agricultural phosphorus. Fertilizer Research 39: 133-146.

Development of HBV-P – a modelling system for phosphorus transport in catchments

L. Andersson1, K. Persson2 and B. Arheimer1

1Swedish Meteorological and Hydrological Institute, SE-601 76 Norrköping, Sweden

2 Swedish University of Agricultural Sciences, Department of Soil Sciences, Division of Water Quality Management, SE-750 57 Uppsala, Sweden


A modelling system for phosphorus (P) transport from catchments is under development. The system will enable source apportionment and consequence analyses concerning transport from river basins to the sea. Attention is given to make the model appropriate to climatic and physiographic conditions in Nordic regions. To be applicable for large river basins, the model will only depend on regionally available databases of driving variables and geographical data.

Integrating field scale phosphorus-models with a hydrological river basin model

Dissolved and particulate P from surface runoff and dissolved P from the rootzone are estimated with ICECREAM (Rekolainen and Posch, 1993). PARTLE is used for simulation of transport of particulate P through macropores (Shirmohammadi et al., 1998). Hydrological and climatological components in these field-scale models have been substituted with those of the hydrological catchment-scale model HBV (Bergström, 1995).

The surface runoff routine is similar to the curve number based routine in ICECREAM. GIS-analyses of spatial distributions of soil- and land-use classes are used for upscaling. Routines for soil temperature frost are based on Lindström (2000), and considers accumulated air temperature, snow depth, and soil moisture. The surface runoff routine was tested with promising results for one experimental field.

Rootzone routines were calibrated and validated against data from two experimental fields. Water flow through the drainage pipes was in general similar or better estimated with the hydrological components of HBV, compared to when using the hydrological components of ICECREAM. Since soluble-P losses, modelled with ICECREAM, not are dynamic in the short time-perspective, a database of leaching estimates for different initial values of the N-pool, and different soil types will be included into the HBV-P model package. The model for particulate-P, PARTLE, is driven by measured mean concentration of suspended solids. For upscaling, it is necessary to find standard values for loss of suspended solids from drainage pipes depending on soil type.

Routines for soil loss with surface runoff have been simplified, in order to be applicable on a river-basin scale. Sediment loss is calculated as a function of the erosivity factor from the Universal Soil Loss Equation, estimated from Finnish data (Posch and Rekolainen, 1993), and of the daily surface runoff. The routine was calibrated and tested against surface runoff and sediment loss data from one experimental field. After calibrating the erosion model against a few more fields, a GIS will be used to model spatial variability of erosion, based on distributions of soils, soil cover, slopes, and distance to watercourses. For calculation of particulate P from surface runoff, a simplified routine for enrichment will be added. The development of the field components of the model will be finalised in the first half of 2001.

Phosphorus-transport from river basins

For assessments of P-transport from river basins, P losses from other land categories than agricultural land will bee included. In addition to land cover, soil type and topography will be considered. Information about livestock density, point sources, and emissions from rural households that will be used, exists in an available GIS-database. The sub-basin model will include functions that dynamically describe changes in transport due to exchange with the bottom sediments in streams. Simulated P-transformation in lakes will depend on exchange between water and sediments, and biological uptake. The lake model will be conceptual and include only a few variables, to be applicable for large river basins with limited input data. The HBV-P model will be tested for some larger (>1000 km2) river basins and its feasibility as a decision support tool will be further evaluated.


Bergström, S. 1995. The HBV model. In: Singh V.P. (ed). Computer Models of Watershed Hydrology. Water Resources Publications, Highlands Ranch, Colorado, pp. 443-476.

Lindström, G., Tjäle och avrinning i Svartberget – studier med HBV-modellen. SMHI Rep. Hydrologi: 84, Norrköping, Sweden.

Rekolainen, S., and Posch, M. 1993. Adapting the CREAMS model for Finnish conditions. Nordic Hydrol. 4:309-322.

Shirmohammadi, A., Ulén, B., Bergström, L. F., and Knisel, W. G. 1998. Simulation of nitrogen and phosphorus leaching in a structured soil using GLEAMS and a new submodel, 'partle'. Transaction of the ASAE 41:353-360.

Phosphorus transfer from field to river via land drains in England and Wales. A risk assessment using field and national data

A. S. Chapman1, I. D. L. Foster1, J. A. Lees1, R. J. Hodgkinson2 & R. H. Jackson1

1Centre for Environmental Research and Consultancy, School of NES, Coventry University, Coventry, UK, CV1 5FB

2ADAS Gleadthorpe, Meden Vale, Mansfield, Notts, UK, NG20 9PF

Pathways linking agricultural land to surface waters have been increasingly investigated during the last ten years, as the significance of diffuse phosphorus (P) loss has become apparent. Land drains, also referred to as field drains and tile drains, are installed to reduce the extent of waterlogging and are widespread across the UK, making it one of the most extensively drained countries in Europe. Land drains have the potential to be significant carriers of sediment and particulate P (PP) given appropriate soil and land use conditions.

The current study combines data from field scale studies, laboratory based rainfall simulator studies and national databases to describe the distribution in England and Wales of agricultural land most likely to be susceptible to sediment and PP loss through sub-surface drainage. A range of soil, climatic, drainage and land use variables have been identified as key factors in controlling sediment loss through drains. Their national distribution results in an uneven pattern of risk of sediment loss. The cereal growing regions of Eastern England appear to be at greatest risk of such losses, while central England is at moderate risk and Wales and the South West are at very low risk. At a national scale, the lack of homogeneity within such data is acknowledged and further investigation at a local scale is required in order to assess risk more accurately.



Compounded problems of scale and stationarity in integrated agronomic modelling

D. Collentine1 and M. Larsson2

1Department of Economics, Swedish University of Agricultural Sciences, Box 7013,

S-750 07 UPPSALA

2Department of Soil Sciences, Swedish University of Agricultural Sciences, Box 7072,


This paper investigates how models, which seek to integrate natural science and social science models to study agronomic activities, may compound problems associated with scale and stationarity. The integrated decision support model (LENNART) for nutrient management in agriculture under development in Sweden, is used to illustrate scale and stationarity problems in this type of model. This decision support model is an administrative interface with links to databases which combine three independent models (two natural science models and one social science model) into an integrated model. The model is designed to be used for modelling nutrient leaching from cultivated farmland including the economic costs to users for introducing specific best management practices to reduce leaching. The unit scale output of the integrated model is presented on a per hectare basis. The model has been developed to be used for modelling individual fields and small catchment areas. These small catchment areas are treated for modelling purposes as aggregations of independent individual fields.

Scale and stationarity problems of the independent modelling components of LENNART (a root-zone leaching model SOILN, an economic cost model BAK and a distributed hydrological model ECOMAG) are first described. The problems which may arise when using this integrated model for field based and small catchment based studies and how these problems may lead to compounded effects, is then taken up. The conclusion is that these compounded problems should be regarded as one of the key problems in integrated modelling.

Modelling phosphorus export from agricultural grasslands to watercourses

Jim Freer1, Keith Beven1, Phil Haygarth2, Andy Fraser3 and Adrian Joynes2

1 Department of Environmental Sciences, Lancaster University, Lancaster, UK

2 Institute of Grassland and Environmental Research, North Wyke Research Station, Okehampton, Devon, UK

3 The Soil Survey and Land Research Centre (Cranfield University), North Wyke, Okehampton, Devon, UK

Understanding of the processes controlling the export of phosphorus (P) from agricultural grasslands has been significantly improved in recent years. At least three modes of P transport are now recognised: dissolution and movement in soluble form, physical transport of P sorbed to colloids through the soil and material eroded and transported to stream channels, and direct incidental losses of fertiliser P during rainfall following application (Haygarth and Jarvis, 1999). Many problems remain, however, in making predictions of P transport for the different grassland management scenarios and the varied soil types that make up the catchment areas contributing to watercourses. There is a need for a robust modelling strategy, taking account of current process understanding that could be applied, using readily available GIS datasets.

This paper will report on a BBSRC funded project that will attempt to construct a model that does consider the various transport mechanisms in moving P from source sites to watercourses and will do so within an uncertainty estimation framework. The fuzzy disaggregation approach to modelling complex environmental processes developed at Lancaster University arises from a recognition that it is not possible to either define or calibrate a single unique model representation of such systems (Beven, 1993, 2000, 2001) Instead, we are increasingly forced into a recognition that there is an equifinality problem, that is that there may be many different model structures or parameter sets that can be considered to provide predictions compatible with the observational data available. Thus, any mapping of a unique landscape space, with all its complexities and heterogeneities into the model space of competing model representations will be a fuzzy (or uncertain) mapping. This approach is also a scale dependent in that any model representation considered should explicitly reflect the dominant differences in behaviour to be expected at any prediction scale.

The model will ultimately be implemented within a GIS framework that will allow applications at larger scales, taking account of the scale dependence of the processes and the availability of observational data in conditioning the predictive uncertainty. The implementation will allow an assessment of the risk of P concentrations at different scales and under different conditions, with the possibility of local conditioning with local observations. The model will initially be applied to the data collected by IGER (Joynes et al., this volume)


Beven, K.J., 1993, Prophecy, reality and uncertainty in distributed hydrological modelling, Adv. in Water Resource, 16, 41-51.

Beven, K J, 2000, Uniqueness of place and the representation of hydrological processes, Hydrology and Earth System Sciences, 4, 203-213.

Beven, K J, 2001, How far can we go in distributed hydrological modelling? Hydrology and Earth System Sciences, 5, 1-12.

Franks, S and Beven, K J, 1997, Estimation of evapotranspiration at the landscape scale: a fuzzy disaggregation approach, WaterResources Research, 33(12), 2929-2938.

Haygarth, P. M. and Jarvis, S. C. 1999 Transfer of phosphorus from agricultural soils. Advances in Agronomy 66, 195-249

Joynes, A, P.M. Haygarth, K. Beven , J. Freer, A.Fraser. Monitoring the transfer of phosphorus from agricultural grasslands to first order streams: parameters for a predictive model, this volume.


















Strategic management of non-point source pollution from sewage sludge

Louise Heathwaite1, Sean Burke1, Paul Quinn2, Steven Merrett2, Paul Whitehead3, Neil Preedy1, David Lerner4 and Adrian Saul4

1Department of Geography, University of Sheffield

2Department of Civil Engineering, University of Newcastle

3Department of Geography, University of Reading

4Department of Civil & Structural Engineering, University of Sheffield

Sludge to land - the current position

Sludge recycling to land is expected to double by 2006 but the security of this route is threatened by environmental concerns and health scares. Strategic investment is needed to ensure sustainable and secure sludge recycling outlets. The environmental risk linked to soil phosphorus (P) saturation recognised in the 1998 MAFF Water Code, currently limits nutrient, including sludge, inputs to all fields ³ P Index 3 and covers 56% arable and 30% grassland soils in the UK. Such restrictions are based on limited information and do not optimise the available capacity of the agricultural system when viewed in terms of an integrated catchment management approach.

The SEAL project

We suggest that not all land has an equal risk of contributing nutrients to receiving waters. The SEAL project: a 3-year EPSRC-funded research programme which started on 1st February 2001, is investigating whether it is possible to minimise nutrient loss by applying sludge to land outside critical source areas (CSAs) regardless of P Index status. Research is underway to develop a predictive and spatially-sensitive semi-distributed model of critical thresholds for sludge application that goes beyond traditional ‘end-of-pipe’ or ‘edge-of-field’ modelling, to include delivery to receiving waters land from non-point sources at the catchment scale. The initial research findings will be described in the presentation. It is intended that the research output will be synthesised in an advice matrix (the NERM) for end-users such as our Water Utility and Environment Agency collaborators, to determine the most appropriate form and frequency of spatially-sensitive sludge application to land to achieve sustainable sludge management without detriment to the environment and receiving water quality.


Our multidisciplinary approach is combining process-driven field experimentation with spatially-sensitive predictive modelling. Five integrated tasks are being undertaken, with current focus on the first 3 tasks, the initial output from which will be described in the presentation. 1. Sludge Characterisation using field and laboratory experimentation to establish nutrient saturation thresholds and potential nutrient release from sludge-amended land in two research catchments: the Eden in Cumbria and the Kennet in the Thames basin. 2. Contaminant Flowpath Tracing to assess nutrient transport risk using a nested field approach and combining soil nutrient indexing, sludge tracking and critical source area (CSA) hotspot modelling. 3. Flow Connectivity Modelling to upscale to the fieldscale and predict nutrient transport routing to receiving waters from sludge-amended land. 4. Scenario Modelling to enable further upscaling to the drainage basin scale using the semi-distributed INCA model and to allow incorporation of vulnerability mapping and scenario analysis of different sludge types, nutrient saturation thresholds and hydrological conditions. 5. Sustainable Sludge Loading Strategies the INCA interface will be used to devise a Nutrient Export Risk Matrix (the NERM) by combining sludge/soil and transport indices with a land use/topographic index and sensitivity index for receiving waters for sludge-derived non-point source losses.



Towards a national consensus on indicators of phosphorus loss to water from agriculture

A. L. Heathwaite1, A. I. Fraser2, E.I.Lord3, P. J. Withers4, M. Hutchins3, P.J. Johnes5 and D. Butterfield5

1Department of Geography, University of Sheffield, Winter Street, Sheffield, S10 2TN, UK

2National Soil Resources Institute, (Cranfield University), North Wyke, Okehampton Devon, EX20 2SB, UK

3Environmental Modelling and GIS Group, ADAS, Woodthorne, Wergs Road, Wolverhampton, WV6 8TQ, UK

4ADAS, Bridgets Research Centre, Martyr Worthy, Winchester, SO21 1AP, UK

5Aquatic Environment Research Centre, University of Reading, Whiteknights, Reading, RG6 6AB, UK


This research aims to develop a new decision-support tool to derive environmentally-sensitive indicators to predict phosphorus (P) loss to water from agriculture. The new tool should enable more accurate estimates of P loss from diffuse agricultural sources at river basin, geoclimatic region and national scales and more effective management of high-risk regions through targeted mitigation programmes. The approach uses a simple lumped/semi-distributed tool on an annual timestep with low data requirements applicable at national, geoclimatic regional and river basin scales. The basic structure of the new tool has three layers or modules:

land cover & management


climate, soil, topography


routing & connectivity to receiving waters

Proposed phosphorus loss-potential indicators at 1km2 scale

Key indicators are soil nutrient status, fertilizer inputs, and manure inputs. Key drivers are land cover (major arable crops, permanent/temporary grass, rough grazing, woodland, non-agricultural sources); management (livestock numbers/type, livestock waste timing/method of application/source), and amounts and form of applied bag fertiliser input.

Proposed phosphorus-transfer indicators at geoclimatic regional scale

Key indicators are hydrologically effective rainfall, erosion vulnerability, potential for lateral and vertical water discharge, and topsoil P concentration. Key drivers are climate (mean annual runoff, HER, rainfall duration and timing, intensity), soil (structure, texture, infiltration capacity, organic matter, calcium status, cohesion) and topography (slope angle, and proximity to waterbody).

Proposed phosphorus-delivery indicators at1km2 scale

The key indicator is delivery ratio (water volume, P/sediment delivery from edge-of-field, surface flow/drainflow and P loss from farm infrastructure). Key drivers are hydrology (drainage density, drainage status, stream order, return period and degree of connectivity) and ‘artificial’ flow routing (field drainage, routing along roads and tracks, direct runoff from farmyards, presence or absence of exit pathways)

Selecting an appropriate platform for developing the new tool

An adaptable, flexible and user-friendly platform is needed to allow future refinements and integration of revised datasets in the new tool. The new tool will use ArcView as it is user-friendly and well suited to the extrapolation and presentation of spatial data and is used in a number of governmental organisations. A small number of complex calculations may require processing external to the GIS but this can be achieved using customised compiled C++ programmes. Input data will be readily available for extrapolation at appropriate resolutions at both the geoclimatic regions and national scale.

Trial application of the initial 3-Layer new tool

A prototype of the new tool will be applied to a number of test catchments representing different geoclimatic regions. The test catchments have been evaluated in the initial sensitivity analysis phase of the research (Heathwaite et al., 2001). Application of the new tool in future phases of the work could incorporate validation/calibration procedures, national-scale application, and updating of the input parameters.


Heathwaite, A. L., Fraser, A. F., Johnes, P. J., Lord, E. I., Hutchins, M., Butterfield, D. and Withers, P. J. A. (June 2001) Interim report to the Project Advisory Group for DEFRA Project PE0105: Towards a National Consensus on Indicators of P Loss to Water from Agriculture, 24pp

Regional scale modelling of phosphorus losses

M. G. Hutchins1, P. J. A. Withers2, E. I. Lord1, S. G. Anthony1 and M. Silgram1

1ADAS Wolverhampton, Wergs Rd, Wolverhampton, WV6 8TQ

2ADAS Bridgets, Martyr Worthy, Winchester, Hampshire, SO21 1AP

The estimation of annual budgets for nutrient fluxes in major rivers is a requirement of Oslo-Paris Commission (OSPARCOM) objectives to which the UK is committed. In the EU, whilst a number of well-documented models have been applied for prediction of nitrate flux, progress towards reliable methods for simulation of phosphorus loss is less advanced. This discrepancy reflects the differences in processes influencing nitrogen and phosphorus dynamics. Phosphorus (P) fluxes are hard to predict, losses largely occurring during intense short-lived storm events, the magnitude of which depend on a combination of risk factors which can be difficult to quantify. Furthermore, unlike nitrate, P may be strongly adsorbed to soil particles, and processes of retention and conversion of mobilised P are likely to be significant both within the landscape and in the stream channel itself. Therefore, at the catchment-scale, researchers have adopted a pragmatic approach, concentrating on the development of input-driven export coefficient approaches, which focus on gross annual predictions. The seasonal variability of processes, mechanisms of P release and representation of position in the landscape (e.g. hillslope processes) are typically not represented explicitly, the effects of these factors being represented by a set of coefficients whose values vary between catchments.

The paper presents results from a regionalised application (to 1 km2 resolution) of a simple conceptual model developed to estimate total P loss at national scale. In addition to land use and P inputs, the model gives explicit representation to spatial variability in climate, soil and topography, and is designed to be applicable to any catchment without prior calibration. The method categorises the landscape in terms of woodland, upland rough grazing, permanent grassland (drained and undrained), and 3 classes of arable land (including ley grassland). Categorisation makes use of the ADAS Land Use database, as used in the MAGPIE policy Decision Support System (Lord and Anthony, 2000), which combines the ITE Land Cover Map of Great Britain with MAFF Agricultural Census data. Results presented here used the 1995 census. Arable land is subdivided into drained and undrained, on the basis of dominant soil association at 1 km2 supplied by SSLRC, and background data on the area of land with artificial drainage. The distinction allows losses to be partitioned into drainflow and surface runoff pathways. Surface runoff/erosion losses are moderated by a factor representing landscape retention. Erosion risk is defined using an SSLRC methodology on the basis of slope, soil texture and crop type (after Palmer (1993)). A digital terrain model derived from Ordnance Survey data is used to define slope characteristics. Soil texture is derived from SSLRC soil association data. Managed additions of fertiliser and manure are estimated from crop areas and livestock numbers (using for example FMA et al. (1998) and Smith et al. (2000)). The exported estimates for all landscape types are modified by hydrologically effective rainfall (HER) which is generated at 1 km2 resolution within the MAGPIE system. In order to facilitate comparison with observed data, human inputs are estimated from population data using the coefficients of Johnes et al. (1996) and added to model predictions.

Application was made to catchments draining into the western part of the Humber estuary. For comparison with observed data, EA harmonised monitoring stations containing extensive time-series of molybdate-reactive P and total P concentration lie within these catchments, along with gauges monitored under the NERC LOIS programme. Results represent annual total P (dissolved and particulate) loss estimates. The spatial distribution of model output indicates high levels of P loss to be predominantly associated with urban centres. Other localised hotspots occur over agricultural land where pig and especially poultry farming is important. The lowest annual exports are predicted to occur in upland areas where intensive land practice is largely absent.

Ongoing MAFF contract PE0105, involving collaboration between the University of Sheffield, University of Reading, SSLRC and ADAS, is developing a new P Indicator tool to model P loss at the river basin, regional and national scale. The new tool builds on the ADAS methodology described here, together with those of the AERC National Export Coefficient Model and the SSLRC P-Expert System.


Johnes, P. J., Moss, B. and Phillips, G., 1996. The determination of total nitrogen and phosphorus concentrations in freshwaters from land use, stock headage and population data: testing of a model for use in conservation and water quality management. Freshwat. Biol., 36: 451-473.

Lord, E. I. and Anthony, S. G., 2000. MAGPIE: A modelling framework for evaluating nitrate losses at national and catchment scales. Soil Use and Management, 16: 167-174.

FMA, MAFF & ADAS, 1998. The British Survey of Fertiliser Practice for 1997. HMSO, London.

Palmer, R. C., 1993. Risk of Soil Erosion in England and Wales. SSLRC (Cranfield University), Silsoe, UK.

Smith, K. A. and Frost, J. P., 2000. Nitrogen excretion by farm livestock with respect to land spreading requirements and controlling nitrogen losses to ground and surface waters. I. cattle and sheep. Biores. Tech., 71, 173-181.

Predicting and mapping diffuse phosphorus loads and concentrations down major stream networks using export coefficient modelling

C. Jordan1, R. V. Smith1 and S. O. McGuckin2

1Agricultural and Environmental Science Division, Department of Agriculture and Rural Development, Newforge Lane, Belfast BT9 5PX

2Department of Agricultural and Environmental Science, Queen’s University of Belfast, UK

Phosphorus (P) enrichment of surface waters leading to eutrophication is considered to be one of the major pollution problems facing developed countries. Diffuse losses of phosphorus from agriculture are often the dominant source of P in many UK catchments (Foy and Withers, 1995). Annual losses of P from catchments can be estimated using export coefficient modelling (McGuckin et al., 1999).

Export coefficients for soluble reactive phosphorus (SRP) and total phosphorus (TP) were derived for a range of land cover classes (Table 1) using P data collected from experimental stream catchments in Northern Ireland (McGuckin et al., 1999). The land cover classification and data used were those from the CORINE map for Ireland 1990 (Jordan et al., 2000). River catchments and sub-catchments were delineated by analysis of a digital elevation map (DEM) on a 50 m grid for Northern Ireland using the hydrological extension functions within the ArcView Geographic Information System (GIS) (ESRI, 1999). If surveyed water features are not available, the DEM allows the location of the stream network to be determined accurately at the 1:50,000 scale provided there are no man-made features which would divert the natural flow of water. Annual stream flows at any point in the network were estimated from the difference in the long-term (1961-90) rainfall and evapo-transpiration values at that point.

A 50 m grid of land cover was created from the vectors of the CORINE land cover map for Northern Ireland. Using an Avenue script, an ArcView accumulation grid of the stream network, weighted by the export coefficient values for the land cover classes at each point, was then created within the ArcView spatial analysis environment. The resulting grid represents the P load accumulated to each point on the stream network due to contributions from the surrounding land cover. As the grid squares used were 50m x 50m = 0.4ha, the accumulated grid can be easily adjusted to give kg P ha-1 for both SRP and TP losses. The ArcView software allows flexible mapping and colour-coding of the resultant values, for example, by dividing the range into, say, 5 equal intervals. The load at any point can be converted to the equivalent P concentration by dividing the loading grid by the flow grid. As with the P loading grid, the ArcView software can colour-code the resultant values so that, for example, only P concentrations above a certain threshold value are highlighted. In this way, the change in P loading or P concentrations can be mapped throughout a major river catchment and lengths of network affected by high P determined. Thus, sections of the network that exceed water quality guidelines may be easily targeted for remedial work.

Table 1. Export coefficients for SRP and TP for some key land cover classes in Northern Ireland.

CORINE land cover class

SRP export

TP export

kg P ha-1 yr-1

kg P ha-1 yr-1

Non-irrigated arable land



Good pasture



Poor pasture



Coniferous forest




ESRI, 1999. Environmental Systems Research Institute Inc., ArcView GIS, V3.2, Redlands, California, USA.

Foy, R.H. and Withers, P.J.A. 1995. The contribution of agricultural phosphorus to eutrophication. Proc. Fert. Soc. 365 : 1-32, Peterborough, England.

Jordan, C., McGuckin, S.O. and Smith, R.V. 2000. Increased predicted losses of phosphorus to surface waters from soils with high Olsen-P concentrations. Soil Use Manage. 16: 27-35.

McGuckin, S.O., Jordan, C. and Smith, R.V. 1999. Deriving phosphorus export coefficients for CORINE land cover classes. Wat. Sci. Tech. 39: 47-53.

Monitoring the transfer of phosphorus from agricultural grasslands to first order streams: parameters for a predictive model

A. Joynes 1, P. M. Haygarth1, K. Beven 2 , J. Freer2, A. Fraser3.

1Institute of Grassland and Environmental Research, North Wyke Research Station, Okehampton, Devon, UK

2Department of Environmental Sciences, Lancaster University, Lancaster, UK

3Soil Survey and Land Research Centre (Cranfield University), North Wyke, Okehampton, Devon, UK


Elevated levels of phosphorus (P) and other nutrients in watercourses have the undesirable effect of contributing to eutrophication and a reduction in the water quality (Foy and Withers, 1995). Phosphorus is of particular importance because it is often the nutrient which limits algal growth (Gibson, 1997). Concentrations of P as low 35 m g L-1 (Vollenweider and Krekes, 1982) have been shown to accelerate algal growth if the other nutrients are already in abundance and environmental conditions are correct. It is therefore important for farmers to minimise the risk of P transport from grasslands to streams by adopting best management practices. To evaluate P transfer in a catchment, it would be useful to have a model, which could be used by land managers and policy makers to predict how agricultural practices may affect the transport of P from grassland to the aquatic environment. This abstract describes field studies in two hydrologically contrasting catchments to provide high-resolution data for such a model. Most previous work on P transfer has been carried out on small-scale artificially drained lysimeters and plots (Haygarth et al., 1998), but the approach described here focuses on two typically managed agricultural catchments draining into natural first order streams.

The approach

Two catchment areas feeding first order streams, managed as agricultural livestock grasslands in Devon were selected to provide continuous data for a period of 12 months. The catchments are in close geographical proximity to each other so as to have similar climate and rainfall. One catchment is characterised by a well drained fine loamy soil, while the other is dominated by a slowly permeable, seasonally waterlogged clay soil (Dystrochrept and Typic Haplaquept respectively, as described by Soil Survey Staff, 1975).

From each catchment, hydrometric data was obtained for rainfall, discharge, turbidity, rain and stream conductivity and temperature, with data being recorded at 15-minute intervals. Discharge was monitored using a flume and pressure transducer installed in a stillwell. Water samplers (EPIC) were used to obtain representative samples during periods of discharge. Samples were flow weighted and controlled by a data logger, which triggered daily sampling when the streams were at base flow, hourly during periods of high discharge and at 30 minute during extremely energetic hydrological events, which may only occur three or four times a year.

The samples collected were analysed for all forms of P, suspended solids, turbidity and nitrate. A programme of systematic soil sampling from sectors with in the catchments will provide soil P status data. Catchment management perturbations were determined using a questionnaire, recorded daily by farm users, which defined the actual time of events. The questionnaire covers every aspect of farm management that may affect the form and magnitude of P transfer, such as fertiliser applications and stock movements.

Data from the logger, laboratory analysis and the catchment management information is then assimilated in time order. These data, along with a theoretical framework, could be used as the basis for the development of a mechanistic model (see Freer et al. this volume).

Progress to date

Although the study is in its early stages with approximately 30 days of data collected, some interesting findings have already started to emerge. The slow draining catchment exhibited discharges with a mean total P concentration 1020 m g L-1, while a mean total P concentration of 55 m g L-1 was observed from the freely draining catchment.


Foy, R. H. and Withers, P. J. A. (1995) The contribution of agricultural phosphorus to eutrophication.

The Fertiliser Society. Proceeding No. 365. 32pp.

Gibson, C. E. (1997). The Dynamics of phosphorus in Freshwater and Marine Environments. In 'Phosphorus loss from soil to water' (H. Tunney, O. T. Carton, P. C. Brookes and A. E. Johnston eds.), pp. 119 - 135. CAB International, Wallingford.

Haygarth, P. M., Hepworth, L. and Jarvis, S. C., 1998. Forms of phosphorus transfer in hydrological pathways from soil under grazed grassland, European Journal of Soil Science, 49, 65-72.

Soil Survey Staff (1975) Soil taxonomy, a basic system of soil classification for making and interpreting soil surveys. Agriculture Handbook, USDA, No. 436.




Quantifying phosphorus pathways within small Danish arable catchments

B. Kronvang, R. Grant and A. R. Laubel

National Environmental Research Institute, Department of Streams and Riparian Areas, Silkeborg, Denmark

Separation of stream discharge into two flow components

A deterministic precipitation/runoff model (NAM-model) has been set up for 5 sandy and 10 loamy catchments in Denmark (Table 1). The NAM-model is calibrated to the measured daily discharge from each catchment. The NAM-model splits daily discharge into two flow components that can be assumed to mimic runoff from the unsaturated zone and saturated zone, respectively. The average annual runoff from the unsaturated and saturated zone has been calculated for the 5 sandy and 10 loamy catchments during the 10-year period of 1989-98 (Table 1).

Table 1: Description of the two classes of catchments (loamy: n=10 and sandy: n=5) regarding average catchment area, proportion of sandy soils, land use and runoff characteristics.


Catchment size

Sandy soil





unsaturated zone


saturated zone

Loamy catchments

7.4 km2

10 %

74 %

13 %

13 %

81 mm

245 mm

Sandy catchments

13.8 km2

96 %

73 %

10 %

17 %

179 mm

185 mm

Phosphorus budget for arable catchments

The phosphorus (P)-cycling in Danish agriculture is monitored annually in six small agricultural catchments using the questionnaire surveys at the field level and measurements of P in soil water, drainage water, groundwater and stream water (Andersen et al., 1999). The agricultural monitoring data collected enable us to establish an overall P-budget for small catchments which reveals the major input and outputs of P from arable soils. The information obtained on average annual P-concentrations in the root zone, subsurface tile drainage water, upper groundwater and deeper groundwater from the intensively monitored 3 sandy and 3 loamy catchments are used to establish an average catchment budget for P-emissions via different hydrological pathways. The aim is twofold: (i) assess the importance of each hydrological pathway at the catchment level and (ii) assess any discrepancies between the emission approach (sum of all emissions) and the actual measured export (immission) of P from agricultural land within the catchments.

Emissions of DRP and PP via different hydrological pathways in the 5 sandy and 10 loamy catchments are shown in Table 2. The estimated DRP-loss is considerably lower than the measured DRP-export from diffuse sources within the 2 types of catchments. The estimated PP-loss is also considerably lower than the measured PP-loss from the catchments. In the latter case PP-loss via soil and bank erosion can possibly explain the missing ‘PP’ in the budget. In the case of DRP fields with a high P-emisson via leaching or surface runoff (Critical Source Areas) must contribute substantially to the budget.

Table 2. Estimation of the emission of dissolved reactive phosphorus (P) and particulate P via different pathways as compared to the measured export from the catchments.


Sandy catchment

Loamy catchment

Sandy catchment

Loamy catchment


Dissolved reactive P

(kg P ha-1)

Particulate P

(kg P ha-1)

Soil water





Upper groundwater





Lower groundwater





Tile drainage water





Sum of pathways to streams





Measured in streams





1DRP-loss from unsaturated zone is calculated based on either DRP-concentration in soil water or upper groundwater


Andersen, H.E., Kronvang, B. and Larsen, S.E. (1999) Agricultural practices and diffuse nitrogen pollution in Denmark: Empirical leaching and catchment models. Water Science and Technology 39(12), 257-264.

On-farm evaluation of phosphorus management strategies in Pennsylvania

J. L. Weld1, D. B. Beegle1, A. N. Sharpley2 and R. L. Parsons3

1Dept. of Agronomy, The Pennsylvania State University, University Park, PA USA

2USDA-ARS Pasture Systems and Watershed Management Research Unit (PSWMRU), University Park, PA USA

2Dept. of Community Development and Applied Economics, University of Vermont, Burlington, VT USA


Currently, a coordinated effort, with Penn State, USDA-Agricultural Research Service (ARS), and other state and federal agencies, is underway exploring environmental issues and potential management approaches related to agricultural phosphorus (P). Additionally, the states in the Mid-Atlantic region of the US, including Pennsylvania (PA), are cooperatively working on P related issues. In support of these efforts, Penn State and the USDA-ARS, PSWMRU have been conducting research on P behaviour, loss mechanisms, and management strategies, including work on the P Index and ARS leadership of the National P Project in the US. As a first step toward practical implementation of P-based management, we evaluated three P management strategies (Table 1) proposed in the USDA/EPA Unified National Nutrient Management Strategy (USDA/EPA, 1999).

Table 1. Summary of phosphorus management strategies used in this project.


Phosphorus Management Approach

Crop Soil Test P

Based on current Penn State soil fertility recommendations

Environmental P Threshold

Based on Mehlich-3 soil test P threshold level; Phase 1=300 ppm P;
Phase 2=200 ppm P

Phosphorus Index

Transport factors: Erosion, runoff, leaching, and connectivity to water

Source Factors: Soil test P, Fertiliser P rate, timing , and application method, and Manure P rate, timing , and application method


This was a two phase project involving 10 commercial nutrient management plan (NMP) writers and 11 cooperating PA farms. Each farm worked with a NMP writer to develop a nitrogen (N)-based NMP and three modified P-based NMP’s. The cooperating farms represented dairy, poultry, swine, and combined animal production enterprises as well as varied levels of animal production intensity. The different nutrient management approaches were compared for impact on NMP development, NMP implementation, and farm economics. Phase 2 used revised environmental P threshold and P Index strategies to re-evaluate farm NMP’s and economics. Feedback on the NMP development process was gathered using farmer and NMP writer surveys as well as debriefing meetings with the writers.


Management and financial impacts of the three P strategies depended factors such as farm location and animal density making the implications farm specific. However, the following generalisations characterise the P management impacts. The crop soil test P strategy was the most restrictive, the environmental P threshold strategy had moderate restrictions, and the P Index was the most flexible. The economic results varied, ranging from slightly positive to very negative impacts. Generally manure export over long distances resulted in the most significant financial impacts. Based on Phase 1, environmental P threshold and P Index approaches were modified for Phase 2 by: lowering the environmental P threshold (Table 1), and in the P Index, reducing the importance of the leaching factor, introducing a new approach for estimating connectivity to the stream, and adding a screening tool. In the surveys, farmers and NMP writers indicated the P Index strategy made the most sense. The writers indicated the P Index was also more likely to be implemented by farmers. The P Index NMP required the most additional time to develop, but providing supporting and reference materials may reduce the development time.


U.S. Department of Agriculture and U.S. Environmental Protection Agency. 1999. Unified national strategy for animal feeding operations. March 9, 1999. Internet URL: http://www.epa.gov/owm/finafost.com.