Agriculture, phosphorus and eutrophication: a European perspective

The enrichment of European waters with anthropogenic sources of nutrients (nitrogen (N) and phosphorus (P)) contributing to eutrophication, and the resulting impairment of water use for recreation, industry and drinking, has become a major environmental issue in recent decades. Concern over eutrophication first emerged in Europe’s coastal waters with subsequent commitments by bordering countries to reduce nutrient emissions to the sea (De Walle Foy, 2005). With the introduction of the EU Water Framework Directive (WFD) in December 2000, there is now a legislative framework to implement catchment controls over P inputs to EU waters from all sources, including those from agriculture. According to EEA (2005), the agricultural contribution to total annual P loads in EU waters is often about 50% (range 25–75%) with higher proportional contributions in countries where point-source P inputs have been markedly reduced by wastewater treatment. Controls over transfers of P from agricultural land have therefore been judged a necessary part of the integrated catchment management needed to achieve the good ecological status demanded by Europe by 2015. Until the early 1990s, there had been very little EU research conducted on phosphorus transfer in runoff from agricultural land. As the potential linkages between agricultural intensification, increased P concentrations in land runoff and eutrophication became more widely appreciated (e.g. Tunney et al., 1997; Edwards Kronvang, 2002; Chardon methods which have worked well for P saturated sandy soils (Chardon & Schoumnas) do not work well on the more calcareous soils of the Mediterranean region (Torrent et al.). Each country has developed its own method of soil testing for crop available P and these tests will continue to be used to monitor soil P build-up and risk of P release to runoff. However, integrating the large spatial variability in soil P, within and between fields, to the mobilization and delivery of P within catchments remains a major challenge (Page et al., 2005). This more localized heterogeneity in P transfer adds a further tier of complexity to the regional and national variation discussed above. Mobilization of P in land runoff during storm events occurs as a result of detachment or solubilization of soil P and any amendments or crop residues present at the soil surface (Figure 1). Delivery of P encompasses aspects of transport including the proximity of the field to the watercourse and the presence of connection channels (gateways, roads, tramlines, culverts and drains) or retention barriers (hedges, bunds, ponds and wetlands) in the landscape. Areas with a high risk of both P mobilization and delivery have been termed ‘critical source areas’, and the concept that P transfer can be transport limited rather than source limited is an important one for targeting of control measures. This is illustrated by Strauss et al. who predicted that the majority of P export in an Austrian catchment was confined to a small area of highly vulnerable soils. Six papers explore the mobilization and delivery of P at various scales ranging from laboratory measurements, to small boxes (near zero delivery) to variably sized catchments and to large river basins. Differences in P mobilization reflect inherent differences in climate and soil type and the modification of this inherent vulnerability by farm management practices. Under EU Fp5 funding, Withers et al. describe the development of a simple laboratory test to estimate the intrinsic mobilization by rainfall impact of particulate P (PP) using a wide range of European soils. Butler & Haygarth show how tillage and reseeding of permanent grassland can greatly increase PP transfer within a high rainfall area whilst Uusi-Kämppäet al. recorded increased dissolved P (and to a lesser extent PP) concentrations in ditch runoff when animals were grazing a forested feedlot. The importance of land management in defining critical source areas in catchments is explored further in a study of two headwater catchments by Hodgkinson & Withers. This paper reminds us of the great temporal and spatial heterogeneity of P transfers and the challenges in separating the role of land use from complexities of land management. In larger catchments, where major point source inputs (municipal and industrial wastewater discharges) are more dominant, agriculture’s contribution to eutrophication must be explored through modelling. Two papers describe the application of models to define agriculture’s contribution to total P loads in catchments. Kronvang et al. reviews catchment P export at different scales within Scandinavia (catchments 50 000 km2). These data again illustrate the very large spatial and temporal variability in P export that can occur between catchments as a result of variation in flow, population density and retention of P within open water bodies. As the scale of monitoring increases, controlling factors other than those operating within fields and farms become active. Collins et al. explore the use of modelling to identify different sources of agricultural P transfer in three catchments suffering diffuse pollution. The separation of P mobilization by source – soil, fertilizer/manure and farmyard, and of P delivery between surface and sub-surface pathways, provided a means of identifying what mitigation measures might be most appropriate and where to target them. These modelling studies confirm the general findings from (e.g. Ulén et al.) reviews that estimates of P export from agricultural land are very variable (0.1–6 kg P ha–1), with the majority occurring in runoff/erosion and with variable composition (form of P). Again, the issue of variability and heterogeneity is evident. The variation in the form and bioavailability of transported P and differences in the mode and timing of delivery between point (continuous discharge) and diffuse (episodic discharge) sources is a theme explored by Edwards & Withers. In reviewing the discussions of the COST 832 group working on impacts, they conclude that the timing of P transfers in relation to biological demands is an important factor determining the significance of different sources, and one which is often overlooked when developing control measures. They suggest that there are a number of rural sources of P other than runoff from farmed land (septic tanks, farmyards, road runoff) that can have a major ecological impact in headwater streams. These small rural sources are often lumped together as ‘agriculture’ and hence ignored in source apportionment studies. They conclude that a greater understanding of the ecological impacts of different sources of P is required to ensure field and catchment-based measures will be effective in controlling catchment P export. Four papers investigate the effectiveness of various source and transport controls over P transfer from agricultural land. Schärer et al. compare different options for reducing P transfer in overland flow from permanent grassland soils, which tend to preferentially accumulate P at the soil surface due to lack of incorporation. Their results suggest that reductions in P transfer in the short-term require more drastic action than just omitting P fertilizer. Owen et al. suggest that different buffer features (e.g. riparian grass strips) are not very effective at trapping the finer particles suspended in runoff, that delivers P to watercourses. Using export-coefficient modelling, Johnes et al. advocate a two-tier approach that combines broad regional policies (taking sensitive land out of production and ceilings on fertilizer use) with more targeted management measures in high-risk areas at the farm and catchment scale. Strauss et al. further illustrate the importance of the identification of these high-risk areas (critical sources) in the cost-effectiveness of reducing P export. The precise role of agriculture in eutrophication still remains poorly understood and accurate source apportionment at different scales and relevance to impacts remains a crucial gap in our research portfolio that needs addressing. The papers presented here show how heterogeneous, complex and variable P export from agricultural land can be, and this makes it difficult for us to arrive at simple generalized rules of the type required for modelling. One pattern that emerges, however, is that the largest P loads exported from agricultural land occur in runoff (highly variable composition) and soil erosion (particulate) and are often delivered to the watercourse in excess winter rainfall when biological activity is minimal. In contrast, other rural sources (farmyard and road runoff, septic tank discharges) that are also variable in composition, but delivered more continuously during the year, maybe more ecologically relevant than runoff from traditionally farmed land, even though they may not constitute a large proportion of the total annual P load (e.g. Douglas et al., 2007). Phosphorus transfers in runoff from agricultural land will also be greatly modified by river basin processes before reaching the monitoring outlet and it is perhaps not surprising that Kronvang et al. could not find any relationship between catchment P loads and P surpluses in their review. As shown in the final paper of this issue by Foy, point source discharges still remain the predominant sources of bioavailable P in Europe’s major rivers and are likely to have the greatest ecological impact. In reviewing a number of catchment studies in the UK, Jarvie et al. (2006) raise similar concerns and suggest that estimates of agriculture’s contribution to eutrophication based simply on annual loads is misleading. In reality, each system is different and agriculture may well contribute to specific eutrophication problems related to P in some headwater streams and in some lakes, especially where there is a high-soluble P component being delivered (Reynolds & Davies, 2001). However, links between agriculture and poor water quality must not be automatically assumed just because it is a rural catchment. The role of climate change and the potential impact it will have on modifying P transfers will only exacerbate the complexity of these issues, presenting us with many new science and policy challenges for the future. As editors of the special issue, our over-arching conclusion is that ironically, as we learn more from a wider range of studies in a greater diversity of European landscapes, we also have to concede that a greater complexity is emerging and is causing us to re-evaluate our understanding of Agriculture, Phosphorus and Eutrophication. We hope that this special issue and the broad portfolio of work it provides, can help foster a new and more enlightened understanding. The authors wish to acknowledge the EU funding of COST Action 832, the enthusiasm and dedication of the many scientists involved in the different working groups and the particular support for COST activities given by the Department of the Environment, Food and Rural Affairs (Defra) in the UK. PMH acknowledges Defra funding on projects PE0118 and WQ0109. IGER and SoilCIP acknowledge core funding from the UK Biotechnology and Biological Sciences Research Council (BBSRC). Finally, the guest editors gratefully acknowledge the excellent editorial support from Bryan Davies at Soil Use and Management.