Estimating seepage intensities from groundwater level time series by inverse modelling: A sensitivity analysis on wet meadow scenarios

Abstract

Summary Mesotrophic wet meadows with an upward seepage of fresh, alkaline groundwater are famous for their high species richness. However, due to the lack of seepage data on an appropriate spatial scale, no quantitative relationships have been established as yet between seepage and the occurrence of seepage-dependent plant communities. Since there is no established method to directly measure upward seepage in the field, we investigated the possibility of inferring the seepage intensity by using measurable hydrological quantities such as ground and surface-water levels. To this end, we designed 16 representative plots of virtual hydrological situations, using known sets of geohydrological parameters. Then we applied the integrated soil–water–atmosphere–plant model SWAP to generate ‘measured’ time series of ground and surface-water levels for these plots. Finally, using the SCEM-UA optimisation algorithm, we calibrated parameters that affect seepage onto these time series. We analysed how the accuracy and uncertainty of calibrated seepage fluxes depend on the measurement interval of input data and on the accuracy and uncertainty of inferred local geohydrological parameters and boundary conditions. Our analysis shows that it is possible to make reliable estimates of seepage intensities from data provided by easy to place piezometers and water level gauges. For application on real datasets, the analysis gives insight into the limitations of both the approach and the data requirements. For example, data supplied every fortnight was found to be just as valuable for seepage estimation as modern high frequency measurements. When setting up a monitoring programme, our method can help to decide what and when to measure. Furthermore, our method can be used establish quantitative relationships between seepage and plant communities on an appropriate spatial scale.