RegCM4 in climate simulation over CORDEX‐MENA/Arab domain: selection of suitable domain, convection and land‐surface schemes

ABSTRACT In order to find out the optimal setting for downscaling the Coupled Model Inter Comparison Project Phase 5 (CMIP5) database, this study describes the most suitable domain, convection and land‐surface schemes within Regional Climate Model version 4 (RegCM4) over the Middle East and North Africa (MENA) domain. The activity is carried out in compliance with COordinated Regional climate Downscaling EXperiment (CORDEX) initiative, where domain suitability is assessed through seven simulations for the period 2001–2005. After the selection of domain, eight simulations are carried out to find better performing convection and land‐surface schemes within RegCM4 for the same time period of 5 years. These experiments are conducted using ERA‐Interim reanalysis datasets; however, the selection of optimal settings is also validated using other reanalysis datasets. Finally, using the same settings, data of five CMIP5 general circulation models (GCMs) are also downscaled to assess the applicability of RegCM4. Statistical measures, such as correlation coefficient, bias, root mean square difference and standard deviation, are taken into consideration to compare the simulated data against the gridded Climatic Research Unit (CRU) and Tropical Rainfall Measuring Mission (TRMM) datasets. RegCM4‐simulated rainfall and temperature show better spatial distributions and magnitudes against observations, as compared to ERA‐Interim data, which may be considered as an added value of the model. In analysing rainfall and temperature data for the 11 sub‐domains (each an 8° × 8° box), the region encompassing the area 7°S–45°N and 27°W–76°E, named CORDEX‐MENA/Arab domain, is found to be the most cost effective and suitable domain. Further analysis over this domain shows Biosphere and Atmosphere Transfer Scheme (BATS) to be the better performing land‐surface scheme, whilst Grell with Fritsch–Chappell closure (GFC) comes up as the most suitable option for convection scheme in the model. Of five CMIP5 datasets downscaled using RegCM4, rainfall and temperature patterns of four are similar to the reanalysis and observations. Therefore, it is recommended that RegCM4 may be employed with the aforementioned settings in downscaling the CMIP5 multi‐model database for climate change impact and adaptation studies for the region in future.