The relative constancy of hydrophysical and hydrochemical parameters in springs sets them apart from other ecosystems. As groundwater flow is unidirectional and continual in springs, plant species occurrence is expected to be controlled mainly by hydrochemical properties. In central Germany and northwest Czech Republic we investigated the vegetation of springs and corresponding hydrological parameters to reveal (1) driving factors of species composition, (2) niche characteristics of potential indicator species, and (3) the response time to hydrological changes.

(1) NMDS, Mantel tests, and path analyses disclosed that plant species composition is essentially determined by springwater chemistry, more precisely by an acidity gradient with pH-value, Al and Mg as major factors. Catchment traits exert a minor influence on the vegetation of springs, but they are part of the functional chain. This chain is driven by atmospheric depositions, which have strongly influenced forested catchments in the last decades, especially on siliceous bedrock with poor buffering capacity. The acidification of ground , spring- and headwaters reflects this circumstance.

(2) Aiming for the monitoring of springwater acidification and recovery we identified indicator species by calculating the probability of species occurrence with respect to pH-value, using generalized additive models. To test the hypothesis that niche widths and response optima are independent of observational scale we compared the results of niche modeling among two different scales: an entire-spring and a raster-based within-spring sampling approach. Whereas response optima did not differ significantly between scales, systematically narrower niche widths (pH amplitudes) were observed on the within-spring scale. This suggests a fine-scale habitat selection of stenoecious species, which is only captured by the within-spring sampling approach.

(3) To unravel the driving factors of vegetation dynamics we applied a NMDS to interannual differences of hydrological and vegetational data. Those were calculated from repeated measurements between 1989 and 2006, separately for time intervals of different length. Interannual species turnover was found indeed, but its hypothesized sensitivity to changes in the acidity status of the springwater could only be verified for the long-term intervals of at least one decade. We suppose that the vegetation dynamics of forest springs is subject to species inertia (delayed response, persistence patterns) and can be affected by a modified acidity status only in the long run. But, extreme events (drought), mechanical disturbance (wallowing animals) or changes in insolation (logging) can mark short-term tipping points in species occurrence.