Investigating short-term dynamics and long-term trends of SO4 in the runoff of a forested catchment using artificial neural networks

The impact of long-lasting non-point emissions on groundwater and streamwat er in remote watersheds has been studied at numerous sites. In spite of sub stantially decreasing emissions in the last decade, recovery has not yet be en observed in all cases. This trend might be masked by the considerable sh ort-term variability of the chemical hydrographs. In this study, artificial neural networks are applied to investigate the SO4 dynamics in the runoff of a small forested catchment susceptible to SO4 deposition. Empirical mode ls are fitted to the short-term dynamics at a time step of one day. About 7 5% of the variance of the SOS data is explained by the instantaneous discha rge, short-term history of discharge and the moving average of SOS concentr ation in throughfall. In contrast, neither air temperature as an indicator for biological activity nor a snowmelt indicator based on the temperature s um increase the performance of the model. The model is used to investigate long-term trends in sub-regions of the phase space spanned by the identifie d input variables. According to the model, decreasing emissions have a sign ificant effect on runoff SO4 concentration only during the first severe sto rms at the end of the vegetation period. This suggests to focus on these ev ents as indicators for recovery of the topsoil layers. (C) 2001 Elsevier Sc ience B.V. All rights reserved.