Resource information
The influence of Climate Change on plant development as well as on carbon and nitrogen cycling in soils is an important research topic for Global Change impact assessment at the regional scale. These changes affect the availability and quality of ground and surface waters and accordingly the future productivity of agriculturally used landscapes. The integrated assessment of these changes requires a robust prediction of the potential future characteristics of soil temperature and moisture based on scale-appropriate, process-oriented models. Hence, we present the Soil Heat Transfer Module (SHTM) used as a component of the mesoscale decision support system DANUBIA, which is developed by the multi-disciplinary research project GLOWA-Danube (www.glowa-danube.de). DANUBIA is applied on the Upper Danube catchment to assess the future changes in water availability, quality and use based on Global Change scenarios. In order to cover the temporal and spatial resolution (1h, 1×1km) of the main model as well as the desired investigation period of 50 years, SHTM combines a variable time step conductive heat transfer algorithm with an analytical lower boundary condition to react to long-term climate change with minimal model drift. Changes in soil moisture and soil freezing are explicitly taken into account. The ground heat flux at the soil surface is computed by iterative closure of the energy balance including radiative, latent and sensible energy fluxes. Validation of the heat transfer scheme shows that the variable time step solution improves computational efficiency while imposing only minimal RMSE and phase shift errors. Furthermore, the analytical lower boundary stabilizes the long term heat balance and induces a rather small potential model drift in the order of 0.001Wm⁻². Then the results of the full land surface model including SHTM are compared to measurements at 25 agrometeorological sites. Without site-specific parameterisation other than land cover type, we show that the model performs well (RMSE about 2K) in reproducing daily and annual top soil temperature dynamics over long simulation periods. The analysis of systematic model errors reveals that about 75% of the RMSE is attributable to uncertainties in meteorological input, canopy parameters and snow processes.
