Predicting critical source areas of sediment in headwater catchments

Mitigation of diffuse nutrient and sediment delivery to streams requires successful identification andmanagement of critical source areas within catchments. Approaches to predicting high risk areas forsediment loss have typically relied on structural drivers of connectivity and risk, with little considera-tion given to process driven water quality responses. To assess the applicability of structural metrics topredict critical source areas, geochemical tracing of land use sources was conducted in three headwateragricultural catchments in Co. Down and Co. Louth, Ireland, within a Monte Carlo framework. Outputswere applied to the inverse optimisation of a connectivity model, based on LiDAR DEM data, to assess theefficacy of land use risk weightings to predict sediment source contributions over the 18 month studyperiod in the Louth Upper, Louth Lower and Down catchments. Results of the study indicated sedimentproportions over the study period varied from 6 to 10%, 84 to 87%, 4%, and 2 to 3% for the Down Catch-ment, 79 to 85%, 9 to 17%, 1 to 3% and 2 to 3% in the Louth Upper and 2 to 3%, 79 to 85%, 10 to 17%and 2 to 3% in the Louth Lower for arable, channel bank, grassland, and woodland sources, respectively.Optimised land use risk weightings for each sampling period showed that at the larger catchment scale,no variation in median land use weightings were required to predict land use contributions. However,for the two smaller study catchments, variation in median risk weightings was considerable, which mayindicate the importance of functional connectivity processes at this spatial scale. In all instances, arableland consistently generated the highest risk of sediment loss across all catchments and sampling times.This study documents some of the first data on sediment provenance in Ireland and indicates the needfor cautious consideration of land use as a tool to predict critical source areas at the headwater scale