Sensitivity of temperature to physical parameterization schemes of<scp>RegCM4</scp>over the<scp>CORDEX</scp>‐Southeast Asia region

<jats:title><jats:styled-content style="fixed-case">ABSTRACT</jats:styled-content></jats:title><jats:p>This study examines the simulated temperature over Southeast Asia (<jats:styled-content style="fixed-case">SEA</jats:styled-content>) using the Regional Climate Model version 4.3 (<jats:styled-content style="fixed-case">RegCM4</jats:styled-content>.3), and its sensitivity to selected cumulus and ocean surface flux schemes. Model simulations were conducted for the<jats:styled-content style="fixed-case">SEA</jats:styled-content>domain at 36 km spatial resolution for the period of 1989–2008, as part of the Southeast Asia Regional Climate Downscaling/Coordinated Regional Climate Downscaling Experiment‐Southeast Asia (<jats:styled-content style="fixed-case">SEACLID</jats:styled-content>/<jats:styled-content style="fixed-case">CORDEX</jats:styled-content>‐Southeast Asia) project. A total of 18 sensitivity experiments were conducted with a combination of six cumulus parameterization schemes and three ocean surface flux schemes. The model's skill in representing mean, maximum and minimum temperatures is evaluated against observed gridded data sets. Results indicate a predominant cold bias in all simulations, particularly over mainland<jats:styled-content style="fixed-case">SEA</jats:styled-content>(Indochina) during the season of December to February. Nevertheless, the seasonal correlation is highest over this region. The cold bias of the model is also evident in the temperature distributions, such that there are more cold months than observed, which may be associated with the underestimation of the daily maximum temperature. A few simulations also reveal a warm bias over some areas in the Maritime Continent. Further examination shows that both radiative and surface fluxes influence the simulated temperature, which may also have effects that partially offset each other in some areas. Comparison of the sensitivity experiments reveals differences in model performance, and underlines the importance in choosing the appropriate configuration for<jats:styled-content style="fixed-case">RegCM4</jats:styled-content>.3 before it is used to downscale climate projections, particularly for the<jats:styled-content style="fixed-case">SEA</jats:styled-content>region. This study also shows a strong influence of the choice of cumulus scheme on temperature. Based on performance metrics for temperature among the schemes tested, the Massachusetts Institute of Technology (MIT) Emanuel cumulus scheme and the Biosphere‐Atmosphere Transfer Scheme version 1e (BATS1e) ocean surface flux scheme can be used in future simulations for the region.</jats:p>