Comparison of LARS_WG and RegCM4 models in simulation and post-processing of annual temperature and rain fall data in Great Khorasan

In this study, rainfall and temperature simulated annual in north-east of Iran (Great Khorasan) In the period 1987-2011. Basedon the results, in LARS model during verification period (2007-2013), average annual rainfall raw bias is equal to 53.63 millimeterand post- processed is -11.25. Shortly, in annual period, in 84% of studied stations, post-processing techniques have been effectiveand bias error rate has decreased heavily in more stations. Based on the results, in Reg-CM4 model during verificationperiod (2006-2011), average annual rainfall raw bias is equal to 85.3 millimeter and post- processed is 61.04. Shortly, in annualperiod, in 75% of studied stations, post-processing techniques havebeen effective and MA technique is more effective. So absolutebias error after processing an average annual rainfall of LARSequal to 6/13 and RegCM4 model is 61 .average annualtemperature raw bias is equal to 0.096 millimeter and post- processed is -0.432. This is more than bias without post-processing hence, post-processing is not effective in all stationsand only in 46% of stations are good. Simulation of 2--meter temperature data in climate stations with RegCM modeland MA technique show high effectiveness. Average annual temperature rawbias in RegCM4 model is equal to -2.78 millimeter andpost- processed is -0.05. In all stations modeled annual temperaturewith Observed data has difference less than 0.1. Also inrainfall simulation, LARS_WG model is more better than RegCMmodel and in simulation of annual temperature data, dynamic RegCM4model show more clear Objectivity of researched area rather thanLARS_WG model