Dynamic Regional Climate Modeling and Sensitivity Experiments for the Northeast of Brazil

This research topic has two main goals. The first is to supply spatially and temporally consistent data for a drought period in the northeast of Brazil, generated by a dynamic mode1ing approach. The second is to elaborate a state-of-the-art regional-scale climate model, which can be applied to simulate scenarios of future climatic developments in the northeast of Brazil. For this, two different types of such models were adapted to the region of interest and were investigated for their ability to reproduce present-day extreme weather conditions. To assess a model’s performance here in terms of the specific drought conditions in the WAVES project area under agricultural aspects, a new validation algorithm has been developed and applied. With this method, errors can be quantified on a statistical basis and can then be accumulated to form general quality measures. Possible sources of inaccuracies were identified as being linked to the representation of hydrological soil processes. On the basis of these results, sensitivity experiments were designed and carried out to investigate the effect of implementing various modified descriptions of the soil hydrology on the accuracy of the model simulations. Finally, a regional climate model version was elaborated which is able to reproduce present-day climatic conditions within a range of errors as characteristic for coarser-scale dynamic modeling approaches, but with a better representation of smaller scales. Remarkable improvements were achieved regarding the representation of drought conditions by the dynamic climate model. It is, however, still a challenging task for the future to further reduce the range of errors for regional-scale features in the simulation results to such an order of magnitude as is known from gridded observations. To reach this objective, further investigations on soil—water and energy balances and near-surface vertical exchange processes are required.