Tracing Biogeochemical Processes in Small Catchments Using Non-linear Methods

Since the 1980s a variety of biogeochemical catchment studies have been set up to investigate the cycling of water and solutes. Groundwater and streams have been sampled to investigate the dominant processes of solute turnover in the subsoil and to monitor their long-term changes. Usually a variety of processes interact partly in a highly non-linear manner. Consequently, identifying the dominant processes is not an easy task. In this study, a non-linear variant of the principal component analysis was used to identify the dominant processes in groundwater and streamwater of two forested catchments in the East Bavarian–West Bohemian crystalline basement. The catchments are approximately 60 km apart, but exhibit similar bedrock, soils, climate, land use, and atmospheric deposition history. Both have been monitored since the end of the 1980s until today, that is, during a period of dramatic decrease of atmospheric deposition of sulfur and accompanying base cations. Time series of component scores at different sites were investigated. Non-linear long-term trends were determined using a low-pass filter based on a Lomb–Scargle spectrum analysis.