چكيده لاتين :
Regional climate models such as RegCM3 are capable to simulating different climate processes. Modeling the climate can detect differences between real climate and model-simulated climate in the area under study. Variety of simulations have been done for sensitivity investigation of the model to the physical computational processes and schemes such as radiation, convection, land use, resolution, domain size and center of domain. In this work, sensitivity of the RegCM3 has been measured to the different convective schemes including Grell, modified Kuo and Emanuel. Grell scheme itself has been divided into two different schemes of Arakawa-Schubert (AS) and Fritch-Chapel (FC). Winters (started from Decembers up to February) of 1997 and 2000 were the period of study. Model operation in producing the monthly amount of precipitation was calculated by comparing model output with observed precipitation of 151 synoptic stations of Iran and CRU reanalysis data. According to the results of this research, the regional appropriate schemes with their errors are: Kuo with -%24 error for the southeast, Kuo with %16.5 error for the northeast, Emanuel with %85 error for the central part of Iran, Kuo with %20 error for southwest, Emanuel with -%10 error for west, Grell-FC and Kuo with -%80 error for the northeast and Kuo with -%16 error for the southwest of Caspian Sea. Simulationʹs errors for the central and the northwestern parts of Iran have been rejected statistically because of high amount of errors. But according to mean error analysis of two years simulations, Grell schemes with approximately % 20 errors have good results for winter precipitation simulation of Iran.
