Storm-runoff processes in a mainly waterlogged low mountain range
catchment in the national park Hunsrück-Hochwald, SW-Germany: Assessing
the impact of wetland restoration using multiple linear regression
analysis and ANCOVA.
Abstract
This study investigates potential effects of wetland restoration on
storm flow dynamics in a mainly waterlogged low mountain range catchment
located in SW-Germany. Here, wetland drainage networks are being sealed,
aiming to achieve rising soil water tables and reestablished peat
vegetation. With the help of hydrograph separation, multiple linear
regression (MLR) and covariance analysis (ANCOVA), runoff-governing
storm properties and sealing influences were analyzed. Results show,
that not only natural storm parameters (precipitation sum, rainfall
intensity, antecedent precipitation and temperature) exert influence on
storm-runoff, but sealings also led to significantly altered processes:
On the one hand, storm-runoff coefficients increased in sealed
catchments, resulting most likely from more saturated soils, providing a
smaller infiltration capacity. This is a desired effect of rewetting but
coincidently a downside regarding storm flood prevention. On the other
hand, lag times, meaning the timespan between rainfall occurrence and
the hydrograph starting to rise, were noticeably prolonged. This effect
potentially can be beneficial when it comes to storm flood prevention.
Overall, statistical models including sealings showed more satisfactory
results describing stormflow variance compared to models without
sealings. Therefore, sealings do exert – statistically proven – an
effect on storm runoff. The heterogeneity of the results, representing a
dense gauge network spread over an investigation area of roughly 7.5 km²
shows, that a high-resolution sampling, both spatially and temporally,
is vital. That is since runoff processes in waterlogged low mountain
range catchments are still poorly understood.