Land surface models are used for a better understanding of hydrological processes and energy fluxes of the system soil-vegetation-atmosphere. However, the prediction of the spatial and temporal heterogeneity of states and fluxes with land surface models on small scales and high resolutions is still a challenge in hydrological modelling. This work focuses on the comparison of soil moisture measured by a sensor network with a high spatiotemporal resolution and modelled with ParFlow-CLM using different levels of model complexity and spatial resolution.
Developed at the Lawrence Livermore National Laboratory, ParFlow is designed to simulate fully saturated as well as variably saturated flow fully coupled with overland flow on large scales and for high spatial resolutions. On this account, ParFlow is created to run in parallel on high performance computers. The Community Land Model (CLM) is embedded as a module in ParFlow which substitutes the soil column of CLM to improve the representation of groundwater and overland flow. Like other land surface models CLM describes complex processes using simplifying assumptions and empirical approaches (e. g. neglecting lateral exchange processes). CLM was originally designed for lateral resolutions of 500 m x 500 m up to several kilometers. In our project the integrated model ParFlow-CLM is currently applied with lateral resolutions of 10 m x 10 m, 2 m x 2 m and 1 m x 1 m to the 27 ha grassland TERENO test site at Rollesbroich located in the Eifel (Germany). So far the effects of high resolutions and different grid scales on ParFlow-CLM simulations are not well examined. Within the scope of this work the impact of different lateral and vertical model resolutions as well as soil layer complexities of ParFlow-CLM are investigated and quantified for the Rollesbroich model to find the impact on the spatiotemporal soil moisture patterns. Amongst others (i. e. lysimeter devices, eddy covariance towers, discharge measurements etc.) the Rollesbroich study site is equipped with a wireless sensor network (SoilNet ) measuring soil water content and temperature delivering long term as well as temporally and spatially high resolution, which allows a detailed model evaluation. It is expected that an analysis and quantification of the effects of model complexity and resolution will improve the understanding of structural model uncertainties and identify possible scaling discrepancies of the model simulations on sub catchment scale. This will lead to more accurate simulation results for planned high resolution ParFlow-CLM Data Assimilation studies for Rollesbroich and other similar study sites