The combination of two or more meteorological variables that may not be extreme events alone can lead to the occurrence of compound climate extreme events. The simultaneous occurrence of extreme temperature and precipitation values has been investigated in this study, considering the correlation between these two meteorological variables for selected stations in Iran. Bivariate frequency analysis of temperature and precipitation can be used to define a decile based index. Therefore, we used precipitation- temperature decile index (PTDI) as a criterion to detect the type of drought condition in which past and future events will be placed. Falat Markazi and Sarakhs basins, showed the maximum difference of extreme dry/hot condition in near future than historical period in every season. Almost all the basins will experience more severe extreme dry/hot condition suing three climate change scenarios in near future than far future. Probability of a compound drought involving two variables has then been calculated using the non-stationary method in accordance with climate changes. Finally, internal variations of different models were eva‎luated in the projection of compound drought. Tail dependence between these two variables were first calculated for this purpose using observational data for precipitation and temperature between 1961 and 2021 using various copula functions. Stations where the compound drought is caused by the concurrent occurrence of extreme values of precipitation and temperature were then identified based on the goodness of fit test of tail dependent copula functions. The proportion of stations with compound drought behavior is highest in the autumn season (45%), compared to other seasons. Findings indicate a negative correlation between these two variables particularly in the spring and winter. In these two seasons, the average return period across all stations is 31 and 36 years, which raises the probability relative to the situation in which these two variables are independent of one another by about three times. Using the 100-member large ensemble model cesm2, the non-stationary analysis was used to determine the time dependent correlation and bivariate probability of occurrence of their extreme values for the future period. The findings demonstrate the change of the correlation between these two variables' extreme values from 1961 to 2100. In most stations, the tail dependence between the two variables of temperature and precipitation decreased. Three seasons of autumn, winter, and spring all experienced an increase in the return period of compound drought occurrence as a result of this issue, but summer season experienced a decrease in this return period, from 62 years in the historical period to 57 years in the future period. Additionally, the range of changes of the return period and the most critical return period of compound drought have been projected spatiotemporally due to internal variations brought on by the different initial conditions of each of the models. Compared to other seasons, the variations in spring were less pronounced. Additionally, these internal variations, which are a significant source of uncertainty, change over time and present greater interval in future. The north and west of the country have the lowest vriability, according to the spatial pattern generated by the eva‎luation of internal variations. According to the findings, the probability of a compound drought in future, compared to the historical period, is almost two times higher under the critical model. In addition, the critical model predicted an increase in the probability of compound drought occurrence relative to the average of the models in the range of 18 to 45% in various seasons on average in the studied stations.

آزاده احمدي

رامتين معيني , محمدحسين گل محمدي