A High-Resolution Regional Climate Model Physics Ensemble for Northern Sub-Saharan Africa

Abstract

While climate information from General Circulation Models (GCMs) are usually too coarse for climate impact modelers or decision makers from various disciplines (e.g., hydrology, agriculture), Regional Climate Models (RCMs) provide feasible solutions for downscaling GCM output to finer spatiotemporal scales. However, it is well known that the model performance depends largely on the choice of the physical parameterization schemes, but optimal configurations may vary e.g., from region to region. Besides land-surface processes, the most crucial processes to be parameterized in RCMs include radiation (RA), cumulus convection (CU), cloud microphysics (MP), and planetary boundary layer (PBL), partly with complex interactions. Before conducting long-term climate simulations, it is therefore indispensable to identify a suitable combination of physics parameterization schemes for these processes. Using the European Centre for Medium-Range Weather Forecasts (ECMWF) reanalysis product ERA-Interim as lateral boundary conditions, we derived an ensemble of 16 physics parameterization runs for a larger domain in Northern sub-Saharan Africa (NSSA), northwards of the equator, using two different CU-, MP-, PBL-, and RA schemes, respectively, using the Weather Research and Forecasting (WRF) model for the period 2006–2010 in a horizontal resolution of approximately 9 km. Based on different evaluation strategies including traditional (Taylor diagram, probability densities) and more innovative validation metrics (ensemble structure-amplitude-location (eSAL) analysis, Copula functions) and by means of different observation data for precipitation (P) and temperature (T), the impact of different physics combinations on the representation skill of P and T has been analyzed and discussed in the context of subsequent impact modeling. With the specific experimental setup, we found that the selection of the CU scheme has resulted in the highest impact with respect to the representation of P and T, followed by the RA parameterization scheme. Both, PBL and MP schemes showed much less impact. We conclude that a multi-facet evaluation can finally lead to better choices about good physics scheme combinations.

Added value of an atmospheric circulation pattern‐based statistical downscaling approach for daily precipitation distributions in complex terrain

“…Our findings, therefore, support the suggestion that more than one metric should be provided in the evaluation process (e.g. Chai & Draxler, 2014; Laux et al, 2021a).…”

Evaluation and projection of precipitation in Pakistan using the Coupled Model Intercomparison Project Phase 6 model simulations

“…where CRU and GCM represent the precipitation product from CRU and CMIP6 GCM datasets, respectively, n is the number of observations of the respective time period under consideration, and σ is the standard deviation. Many studies (Guo et al, 2017;Laux et al, 2021;Shiru and Chung, 2021;Oduro et al, 2021) asserted that climate models' ability to capture the daily precipitation variability varies from one model to another depending on its initialization and spatiotemporal resolution. Therefore, to explore the temporal skill of all GCMs relative to the CRU data (You et al, 2017a), we employed interannual variability skill (IVS) for precipitation data accumulated on monthly, annual, and seasonal scales as an alternative to interannual standard deviation (Zhang et al, 2018).…”

Multivariate Bias‐Correction of High‐Resolution Regional Climate Change Simulations for West Africa: Performance and Climate Change Implications

“…Moreover, we quantify the uncertainties from the boundary conditions, that is, from the different driving GCMs (Figure ) to highlight if and where substantial differences exist. It is found that apart from WRF‐M, all RCMs show a relatively low sensitivity to the driving GCMs, which can be attributed to the effect of internal model physics (Gnitou et al., 2021; Laux, Dieng, et al., 2021).…”