Hydrological modelling of a re-wetted peatland on the basis of a limited dataset for water management

The valley mire ‘Mittleres Trebeltal’ in Mecklenburg-West Pomerania (northeast Germany) has recently been re-wetted in the course of an EU-funded conservation project. On the basis of geological cross-sections, a hydrological model applying the Boussinesq-equation has been developed to investigate the hydrological dynamics of the study area and to evaluate various management strategies. Modelled water levels corresponded well to measured ones. The optimised values of two of the main model parameters—the hydraulic conductivity and the proportions of vegetation types—agreed with measurement values, and, respectively, the real vegetation cover, while the third one, the specific yield, should pragmatically be used as a calibration parameter. Due to a proportion of around 30% of tall fen vegetation, the modelled evapotranspiration clearly exceeded grass evapotranspiration resulting in low groundwater levels in some parts of the fen during summer. Generally, the water levels are controlled by topographic features (i.e. former peat cuttings) and weather conditions during the summer months. When analyzing different management scenarios, even an optimal water supply (ditch water levels near the ground surface) could not raise the groundwater table to ecologically sound levels during summer. However, during winter, these ditch water levels enhanced flooding. The groundwater fluxes towards the river, and, respectively, towards the fen were low. The minor importance of these fluxes for the water balance was caused by very low hydraulic gradients. Therefore, it can be concluded that the risk of nutrient export from the fen into the river is very limited and compound transformations take place within the mire system itself. Hydrological analysis as a prerequisite for the assessment of nutrient and pollutant transport and transformation processes can be carried out—as a first approach—with simple models on the basis of a limited dataset.