In the north-western European countries Norway, Sweden, United Kingdom (UK) and Ireland, variability in the forms, amounts and timing of phosphorus (P) loss from agricultural land is related to national differences in climate, soil, hydrological conditions and agricultural production. The dissolved form of P constitutes 9-93% of the total phosphorus (TP) in water, subsurface drainage can contribute 12-60% and surface erosion 40-88% of TP transfer. TP export in small agricultural streams is generally in the range 0.3-6 kg ha )1 year )1 , with the highest losses in Norway and UK. All four countries are complying with the EU Water Framework Directive and developing a range of measures based on P source with transport controls over P losses. A decreasing trend in TP losses has been detected in agricultural streams following the introduction of measures to reduce erosion in Norway. Average P concentrations in Swedish streams have shown a reduction of nearly 2% per year since 1993 as a result of measures introduced in southern Sweden. However, in two large rivers in agricultural regions of Sweden, the concentrations of suspended solids (SS) and TP were shown to increase by 0.4% and 0.7% per year, respectively, over the period 1975-2004, possibly as a result of climate change. It is too early to detect trends in agricultural contributions to P in surface waters as a result of catchment-sensitive farming (CSF) in the UK and Ireland.

The accumulation, management, and transfer of P in intensive farming systems has increased P export from agricultural watersheds and accelerated eutrophication of surface waters. Even though much research on P has been done in the last 20 years, there are still too few answers to the many questions now being asked regarding agricultural production and environmental quality. To address these concerns, four areas of research are suggested: (i) Soil P testing for environmental risk assessment-What losses are acceptable and can these losses be determined by plot-scale or watershed-scale studies? Threshold P levels in soil and water should be established in combination with an assessment of site vulnerability to P loss. (ii) Pathways of P transport-An analysis of the relative importance of different flow pathways is needed at a watershed scale. (iii) Best Management Practice (BMP) development and implementation-Overall, BMPs must attempt to bring P inputs and outputs into closer balance and should be targeted first to critical source areas within a watershed. Alternative management recommendations, uses, and market demand for manures must be developed. (iv) Strategic initiatives to manage P-To initiate lasting changes, research should focus on consumersupported programs that encourage farmer performance and stewardship to achieve agreed-upon environmental goals. P UOSPHORIJS is an essential element for plant and animal growth and its input has long been recognized as necessary to maintain profitable crop and animal production. Phosphorus inputs can also accelerate the eutrophication of fresh waters around the world (Carpenter et al., 1998-,Foy and Bailey-Watts, 1998). Eutrophication has been identified as the main problem in surface waters having impaired water quality, restricting water use for fisheries, recreation, industry, and drinking (CEC, 1992; USEPA, 1996). Even though much research has been done on P in the last 20 years, many questions are still unanswered regarding agricultural production and its impact on environmental quality (Johnston et al., 1997). Clearly, better information is urgently needed to help resolve production and water-quality issues and provide defensible tools and recommendations for P-based nutrient-management planning, particularly where changes in farm management may be required. From research presented at the recent Organization of Economic Cooperation and Development Conference "Practical and innovative measures for the control of agricultural phosphorus losses to water," Belfast, Northern Ireland, 6 to 19 June 1998 and our own insights, four main research areas are identified as needed to meet these challenges: soil P

Abstract. The phosphorus (P) sorption and desorption dynamics of eleven major agricultural grassland soil types in Ireland were examined using laboratory techniques, so that soils vulnerable to P loss might be identified. Desorption of P from soil using the iron‐oxide paper strip test (Pfeo), water extractable P (Pw) and calcium chloride extractable P (Pcacl2) depended on soil P status in all soils. However, soil types with high organic matter levels (OM), namely peat soils (%OM >30), had lower Pfeo and Pw but higher Pcacl2 values compared to mineral soils at similar soil test P levels. Phosphorus sorption capacity remaining (PSCr) was measured using a single addition of P to soils and used to calculate total P sorption capacities (PSCt) and degree of P saturation (DPS). Phosphorus sorption capacities correlated negatively with % OM in soils indicating that OM may inhibit P sorption from solution to soil. High organic matter soils exhibited low P sorption capacities and poor P reserves (total P, oxalate extractable P) compared to mineral soils. Low P sorption capacities (PSCt) in peat soils were attributed to OM, which blocked or eliminated sorption sites with organic acids, therefore, P remained in the soil solution phase (Pcacl2). In this work, peat and high organic matter soils exhibited P sorption and desorption characteristics which suggest that these soils may not be suitable for heavy applications of manure or fertilizer P owing to their low capacities for P sorption and storage.

No abstract

Phosphorus inputs must be estimated accurately to optimize the economic return to farmers and minimize P loss from soils to surface waters. Currently, P recommendations are based on the diagnosis of field crop responses by chemically extracted soil P. However, the inability of chemical extraction to characterize plant‐available P limits the reliability of these recommendations. Major sources of P mobilized by plant roots include P ions in solution and those from soil constituents, which replenish and buffer solution. A mechanistic evaluation of soil P supply should therefore be based on the description of P ion transfer between soil constituents and solution. Sorption, desorption, electro‐ultrafiltration (EUF), and isotopic exchange studies show that an adequate modeling of this quantity [Q(CP,t)] of P ions must account for both the concentration of P ions in soil solution (Cp) and time (t). In one long‐term field experiment, the Q(CP,t) description was not affected by crop rotation and mineral fertilization histories; therefore, Q(CP,t) changes are fully explained by CP changes. In two field experiments, CP changes were linearly correlated with the cumulative P budget, inputs, and outputs over years. In three field experiments, the soil type effect on the relative maize (Zea mays L.) response curve was taken into account using the ability of soil P to replenish solution P for 1 d. The residual variance of this diagnosis is halved compared to Olsen's extraction. Although more information is necessary, accuracy is improved when soil testing is based on mobility of P ions.