The flood and drought events over arid and semi-arid regions under climate change: Northern Fars province

Abstract

The melted ices, reduced snow covers, increased air temperature, change in frequency and duration of droughts and change in precipitation patterns are evidences of the climate change (Ehsani et al., 2017; Huber and Knutti, 2011; IPCC, 2014). Different alternatives of driven factors (population and economic growth, technology efficiency and type of used energy) are considered in the 21st century to project the greenhouse gasses in the Representative Concentration Pathways (RCPs). These assumptions lead to the three RCPs, including a stringent mitigation scenario of RCP2.6, an intermediate scenario of RCP4.5 and, a very high greenhouse gasses emission scenario of RCP8.5 whose radiative forcing is +2.6, +4.5, +6 and +8.5 W/m2, respectively in the year 2100 relative to pre-industrial values. The global warming is +1, +1.8, +2.2 and +3.7 ℃, respectively at the end of the 21st century (2081-2100) relative to 1986-2005 (IPCC, 2014). The study of the temporal variability of the observed precipitation and temperature in southern Iran showed that the precipitation and temperature trends are negative and positive, respectively (Tabari and Talaee 2011a, 2011b; Kousari and Asadi Zarch 2011; Tabari et al. 2011; Kousari et al. 2011). There are large dams such as Dorudzan and Molasadra and natural lakes of kaftar, Arzhan and Parishan over the study area which may adversely affected due to climate change in future. Therefore, the objective of this study is to find how the climate variables of precipitation and temperature, drought duration and intensity of extreme storms will change under climate change.

The six climatic stations are selected based on the longest common daily records of precipitation and temperature data (1978-2005) to investigate the climate change in the study area. In this study, the daily outputs of the 22 CMIP5 models are applied for climate change study. The LARS-WG is used to downscale each GCM-predicted data in 6 stations. The observed daily data during the 28-year period of 1978-2005 are inputted to the LARS-WG to calibrate it for each station. Then, to verify the LARS-WG at each station, the 28-year daily precipitation, minimum and maximum temperature are generated and statistical characteristics of the generated data (statistical distributions and mean values) are compared with observed ones at significance level of 0.01 using Kolmogorov–Smirnov (K-S) and Student’s t-test (Semenov and Barrow, 2002). The verification results imply that the LARS-WG reproduces the daily data well in 6 stations. The change factors are calculated using the given GCM data and relevant grid box to downscale that GCM output for each climatic station. the change factors are calculated using a transient approach in which 5-year future time windows shifted forward by 1-year.

The downscaling results in Abadeh station indicate that the mean annual temperature will increase from 14.3 ℃ during the period 1978-2005 to 16.1, 16.7 and 17.7 ℃ under the RCP2.6, RCP4.5 and RCP8.5, respectively, leading to the warming by +1.8, +2.4 and +3.4 ℃, respectively. The downscaling results show that the mean annual precipitation will decrease from 452 mm during the observation period (1978-2005) to 386, 353 and 357 mm under the RCP2.6, RCP4.5 and RCP8.5, respectively, leading to precipitation reduction by 14.6%, 22% and 21%, respectively. The intensity of extreme storms with different return period are estimated using the maximum annual series during the observation period and three RCPs. The goodness of fit on statistical distributions of Exponential (1P and 2P), Gamma (2P and 3P), General Gamma (3P and 4P), Log-Gamma (2P), General Extreme Value, Normal, Log-Normal (2P and 3P), Pearson (2P and 3P) and Log-Pearson (3) shows that the General Extreme Value is the best-fitted statistical distribution on maximum annual series. Then, the intensity of 2, 5, 10 and 20-year return period storms are determined using the fitted statistical distribution. The results indicate that the 2-year return period storms will be intensified, the 5-year storms will not be changed, but the intensity of the 10-year and 20-year storms will decrease in future due to climate change. The 12-month SPI analysis indicates that the wet and normal periods will slightly increase while droughts will decrease in the period 2040s under the three RCPs. The SPI analysis in the second period (2070s) show that the wet periods will not change, the normal periods slightly increase, but drought conditions may be intensified especially in RCP2.6 and RCP8.5.