Climate extremes have large impacts on human societies and natural ecosystems. Projection of changes in climate extremes is very important for long-term planning. The current study investigated future changes in extreme precipitation events over Iran based on 18 CMIP5 models for the period 2006-2100. National gridded data from the Asfazari database were used to evaluate climate model simulation. Results indicate that models with higher spatial resolution (CCSM4 and MRI-CGCM3) perform better than those with lower resolution in capturing the spatial features of extreme precipitation events. Bias correction was applied to the models and the projected changes were assessed with the nonparametric modified Mann-Kendal trend test and Sen slope estimator at a 95% confidence level. Annual total precipitation (PRPCTOT) and rainy days (RD) were projected to decrease but the intensity and frequency of precipitation extremes were predicted to increase significantly. The projected decreases were larger in northwestern parts than other regions, with PRPCTOT decreasing by 18 to 22 mm decade-1 and RD by 4 to 4.8 d decade-1. Although there were discrepancies in rates between the models, extreme precipitation events over Iran were generally projected to increase. An increase in consecutive dry days (CDD) was predicted for most regions by the end of the 21st century under RCP8.5, with the largest increase of 5 to 6.8 d decade-1 found for northwestern Iran. In eastern areas of Iran, where precipitation occurs extremely rarely, the number of days with daily precipitation exceeding 10 mm (R10) or even 20 mm (R20) were projected to increase significantly. In conclusion, these changes suggest an increased risk of flash floods in Iran from increased extreme precipitation under the RCP8.5 emission scenario.
