Assessing the Impact of Climate Change on the Drought Status in Tabriz Station During Future Periods Using LARS-WG

Drought is one of the most widespread and devastating natural hazards, which is compounded by climate change. Indicators are widely used to provide an overview of drought conditions. In this study, the impacts of climate change on drought status in Tabriz station during future periods were investigated using Deciles Index (DI) and the Standardized Precipitation Index (SPI). First, the daily output data of HadGEM2 model under RCP2.6, RCP4.5 and RCP8.5 scenarios were downscaled by LARS-WG version 6 and the ability of the model was confirmed to simulate the past climate (1987-2016) in Tabriz. Then, the precipitation was simulated for Future periods of 2021-2040, 2041-2060, and 2061-2080. Using simulated precipitation data, drought status in Tabriz was assessed using two drought indices on an annual scale. The results show that in most of the studied years, the number of droughts decreased in all three future periods compared with the base period and the number of wet years increased. The results of drought monitoring and its prediction for future periods can be used in natural resource management as well as water resource management planning. In recent years, weather and climate researchers have identified climate change as the most important concern due to increasing greenhouse gas emissions and global warming. Drought is one of the most important and most common disasters affected by climate change that is slowly and progressively causing environmental, agricultural and economic damage in both dry and humid climates around the world (Li et al., 2013). Since drought affects different segments of society such as water resources, agriculture, industry, economy, health, etc., monitoring and evaluation of this factor in the present and in the future is necessary in order to provide proper planning in different parts of society. Climatologists are currently simulating climate variables using general atmospheric circulation models (Barrow and Yu, 2005). The main purpose of these models is to calculate three-dimensional climate indices in specific grids. The outputs of these models have low spatial accuracy. Therefore, if their output directly enters hydrological models, it increases uncertainty. Downscaling methods are used today to increase the spatial accuracy of these data. Downscaling methods are divided into two categories: dynamic and statistical (Beecham et al., 2014). Statistical methods are commonly used in climatic studies. In this study, the output of HadGEM2 model under RCP2.6, RCP4.5 and RCP8.5 scenarios were downscaled by statistical method and LARS-WG model. The daily climatic variables such as minimum temperature, maximum temperature, precipitation and sunshine for Tabriz station were produced for the next three periods of 2021-2040, 2041-2060 and 2061-2080. Then, using simulated rainfall data, the drought status of Tabriz station was evaluated using two decile indices (DI) and Standardized precipitation index (SPI). The study area in this research is Tabriz Synoptic Station which is geographically located in northwest of Iran. The data used include observed and simulated data. The observed data were related to 1987- 2016. The simulated data included HadGEM2 model that was downscaled by LARS-WG under RCP2.6, RCP4.5 and RCP8.5 scenarios. LARS-WG as one of the most popular models for generating stochastic data was used to produce daily minimum and maximum temperatures, precipitation and radiation for present and future climatic conditions. This model is more applicable than others due to repeated computation, less data input, simplicity and performance (Kilsby et al., 2007). After primary data analysis, daily precipitation series were generated and then LARS-WG was implemented. Subsequent to analyzing the input data and the initial statistical studies, the base state scenarios were implemented for the observed data and the precipitation data were simulated. Model validation based on observed and simulated precipitation values showed high agreement of the model with the observed data. Then precipitation data were tested for normality of distribution. The results indicated that the precipitation data followed the normal distribution. After ensuring that the model was capable to simulate precipitation data series for Tabriz station, it was run for three periods of 2021-2040, 2060-2041 and 2061-2080 using HadGEM2 output under RCP2.6, RCP4.5 and RCP8.5. Then annual drought was calculated using SPI and DI indices. Investigation of drought status using SPI index revealed that in most of the years, the number of droughts decreased compared with the base period (1987-2016) and the number of wet periods increased. Evaluation of the DI index also showed that in all future periods the number of extreme, severe and mild droughts decreased, in compare with the base period, but the number of moderate droughts increased. According to this index, the percentage of normal years would increase significantly in all three future periods, but the percentage of wet years would show a significant decrease.