Assessing the Performance of CMIP5 GCMs in Copula-Based Bivariate Frequncy Analysis of Drought Characteristics in the Southern Part of Karun Catchment

Drought is an extreme event and is a creeping phenomenon as compared with other natural disasters, which has great effects on the environment and human life. During 1997 to 2001, a severe 40-year return period drought affected half of Iran's provinces, with a loss in the agricultural sector estimated at more than US$ 10 billion (National Center for Agricultural Drought Management, http://www.ncadm.ir) and a Gross Domestic Product (GDP) reduction of about 4.4% was reported (Salami et al., 2009). A more severe drought period (2007–2009) devastated the country on a larger scale than the previous drought period. A 20% average reduction of precipitation has been reported for 2008 compared with a 30-year average (Modarres, et al. 2016). It was found that the longest and most severe drought episodes have occurred in the last 15–20 years (1998-2017) (Ghamghami and Irannejad, 2019). A drought is characterized by severity, duration and frequency. These characteristics are not independent of each other, and droughts cause significant economic, social and ecosystem impacts worldwide (IPCC, 2013). Probabilistic analysis of drought events plays an important role for an appropriate planning and management of water resources systems and agriculture, especially in arid or semi-arid regions. In particular, estimation of drought return periods can provide useful information for different water sectors under drought conditions. In this study, the capability of two CMIP5 GCMs in estimating the joint return period of severity and duration of drought using copula have been investigated in the Southern part of the Karun Basin.In this study, three type data have been used. These include monthly precipitation and temperature observed at synoptic stations and gridded data in 1975-2005 were obtained from IRIMO (the Iranian Meteorological Organization) and CRU (http: https://crudata.uea.ac.uk/cru/data) as well as the outputs of two GCM (HadGEM2-ES and IPSL-CM5A-MR) from CMIP5 (http;//cmip-pcmdi.llnl.gov/CMIP5/) in the period of 1975-2005 for historical. Following the Intergovernmental Panel on Climate Change (IPCC, 2013), the first ensemble member (r1i1p1) from two GCMs were selected. RCPs are estimation of radiative forcing (RF), where RCP2.6 and RCP4.5 represents 2.6 and 4.5 W.m-2 and RCP8.5 represents 8.5 W.m-2 at the end of the 21th century (Goswami, 2018). Assuming a drought period as a consecutive number of intervals where SPEI (Vicente-Serrano et al. 2010) values are less than −1, two characteristics are determined, namely: extreme drought length and severity. Hydrological phenomena are often multidimensional and hence require the joint modeling of several random variables. Copulas model have become a popular multivariate modeling tool in many fields where multivariate dependence is of interest and the usual multivariate normality is in question. Among the copula-based drought frequency analysis, Elliptical and Archimedean copulas have been the most popular used equations. In this paper, we focus on copulas based multivariate drought frequency analysis considering drought duration and severity. Return period is defined as ‘‘the average time elapsing between two successive realizations of a prescribed event’’ (Salvadori et al.,2011).