Particulate phosphorus and suspended solids losses from small agricultural catchments: links to stream and catchment characteristics.

Abstract

Excessive phosphorus (P) inputs from agriculture are well established as a contributor to freshwater eutrophication. Decreasing these inputs is an important step in improving the ecological state of impaired waters. Particulate P (PP) is a significant contributor to diffuse P inputs in agricultural catchments. Identifying the main correlates for PP losses is an important step in reducing these inputs. However, there are few studies of long term temporal and spatial dynamics of PP in agricultural streams. Here, we investigate the relative importance of hydrology, catchment characteristics and geochemistry on PP concentrations and fluxes in agricultural headwaters. We evaluate long-term monitoring data from eleven small (<35 km2 ) Swedish catchments with at least seven years of measured flow and flow proportional water quality sampling. Using parametric and non-parametric regression together with principal components analysis (PCA), we identify in-stream and catchment variables relevant for predicting PP concentrations, e.g., suspended solids concentrations (SS), soil texture and average catchment soil P content, measured as ammonium lactate/acetic acid extractable P (P-AL). We show that PP is primarily correlated to SS concentrations, which in turn are correlated to average clay content and land use. However, the SS:PP relationships differ between catchments. No correlation between PP concentrations in the stream and soil P content was found. An increasing clay content decreases the slope of the relationship between SS and PP, i.e., in catchments with higher clay content, less PP is transported per unit SS. The PP/SS ratio increased significantly (p < 0.05) over time in four catchments, despite limited changes in SS or PP concentrations. Our study highlights the importance of long time series since the enrichment of P on SS in the streams is only detected when using long term monitoring data.