Controls on the phosphorus content of fine stream bed sediments in agricultural headwater catchments at the landscape-scale

There have been no landscape-scale assessments which quantify the relative importance of the organic and mineral properties of BS (bed sediment) and associated catchment characteristics (geology, land cover and mean topsoil phosphorus (P) content) for BSP concentration. Mid infra red diffuse reflectance spectrometry was applied to estimate the quantities of organic matter, dithionite extractable aluminium- (Ald) and iron (Fed), kaolinite, dioctahedral clay and mica (D&M) minerals in 1052 snapshot samples of fine (<150 μm) BS in small to medium-sized (5–55 km2) agricultural headwater catchments across a large area (15 400 km2) of central England. Analyses included estimates of BS specific surface area, cerium (Ce) concentrations (enriched in P-bearing apatite and P-fertilsers), and catchment average topsoil P content.Simple linear regression demonstrated that the proportion of variance in BSP explained by specific components of BS across all catchments declined in the following order: Ald > Fed > topsoil P = kaolinite = residual iron> organic matter = Ce> D&M > mineral SSA. No single component accounted for more than 36% of the variance in BSP. Multiple regression – including a classification of bedrock lithology and proportions of arable and grassland by area – accounted for 61.9% of the variance in BSP. The proportion of arable and grassland by area in each catchment was also a statistically significant predictor of BSP. Across this large region – with widely differing geology and soils – Fed in BS is more strongly associated with kaolinite than D&M minerals because iron-oxyhydroxides and kaolinite form contemporaneously during pedogenesis. The SSA of fine bed sediments is largely determined by catchment area, fitted accurately using a power function.

► A multiple regression model based on a variety of bed sediment properties accounts for 61.9% of the variance in sediment phosphorus concentration over a landscape of 15 000 km2.
► Total bed sediment phosphorus was strongly correlated with extractable iron and aluminium oxides plus kaolinite.
► The cerium content of bed sediment – from both natural apatite and fertilisers/manure – was also correlated with sediment phosphorus.
► The areal proportions of arable and grassland in each catchment was a significant predictor of bed sediment phosphorus.