ABSTRACT
This paper presents a model of catchment-scale solute transport applied to the case of nitrate in a small agricultural watershed (Alton Pancras: <10 km2) in the River Piddle catchment, Dorset, UK. The solute-transport model calculates stream-nitrate response to catchment-scale nitrate loading on an annual time-step assuming: an initial baseline concentration, Cb ; a nitrate load to concentration conversion factor, α; some dispersion characteristic of the catchment (hydrogeological) system, Pe ; a time delay between initial catchment nitrate loading and stream concentration response, ta ; and the Mean (catchment) Travel Time (MTT). Historical land-use and management data are used to estimate the net annual loading of nitrate to the catchment between 1930 and 2007, and parameters are based on model fits to observed annual average stream-nitrate concentrations between 1981 and 2004.
A simple graphical translation of the catchment nitrate loading and river concentration data suggests an MTT of 37 years. The Mean Absolute Error (MAE) reached a minimum at Pe → ∞, and beyond this the MAE rises to a stable plateau of 0.26 mg/l. Estimates of α and Cb converge with increasing Pe . For this particular catchment the value of Pe suggests catchment-scale dispersion may be ignored, taking Pe → ∞, and thus allowing model simplification. This allows catchment nitrate loading and stream response to be related by a simple linear model.
Alternative catchment nitrate-loading scenarios are considered, assuming that fertiliser inputs between 1930 and 2007 were cut by 25%, 50%, 75% and 100%. Our models show two points of interest: fertiliser inputs are only partially responsible for the stream concentration rises between 1970 and the present, but the future peak in around 2017 will be almost 50% attributable to fertiliser inputs. It is noted that rises in stream-nitrate concentration observed to date result from a combination of grassland ploughing, increasing animal inputs and fertiliser application, rather than the latter. Hence, policies that rely solely on fertiliser management address only around a third of the total inputs.
The results demonstrate that, in groundwater-dominated catchments, MTTs are of the order of several decades, even in the smallest of watersheds. Therefore diffuse pollution strategies implemented now will not have a measurable impact on the river, in this case, for almost 40 years. Further, in this particular catchment, stream-nitrate concentrations will continue to rise due to past land use and management, peaking just before 2020.
