A critical re-evaluation of controls on spatial and seasonal variations in nitrate concentrations in river waters throughout the River Derwent catchment in North Yorkshire, UK.

Since mean nitrate concentration along single river channels increases significantly with percent arable land use upstream of sampling points and autumn/early winter flushes in nitrate concentration are widespread, it is generally concluded that farmers contribute most of the nitrate. For the River Derwent in North Yorkshire, the correlation between nitrate concentration and percent arable land use is much poorer when tributary data are included in the equation, because of greater variations in dilution by water draining upland areas and in other N input sources. For the whole river system therefore, percent upland moorland/rough grazing land cover is an appreciably better predictor than percent arable land use for nitrate concentration. Upland land use encompasses the higher precipitation and runoff in such areas, and the subsequent greater dilution downstream of both arable land runoff and effluent from treatment works, as well as an inverse correlation to percent arable land use. This is strongly supported by the observation that, for the Derwent, Meteorological Office rainfall data alone proved even better than percent moorland rough grazing for predicting nitrate concentration. The dilution effect is therefore substantial but highly seasonal; lower runoff and dilution in summer offset the lower leaching losses from arable land, and higher dilution and runoff in winter offset greater nitrate leaching losses from arable soils. Because of this, coupled to improved efficiency of nitrogen fertilizer use, seasonality trends in nitrate concentrations that were pronounced a decade ago now have all but disappeared in the catchment.