Vol. 78: 189–204, 2019 CLIMATE RESEARCH Published online August 22 https://doi.org/10.3354/cr01568 Clim Res OPENPEN ACCESSCCESS Historical analysis (1981−2017) of drought severity and magnitude over a predominantly arid region of Pakistan Saadia Hina1,2,3,*, Farhan Saleem2,4 1State Key Laboratory of Numerical Modelling for Atmospheric Sciences and Geophysical Fluid Dynamics, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, PR China 2College of Earth and Planetary Sciences, University of Chinese Academy of Sciences, Beijing 100049, PR China 3Department of Environmental Sciences and Engineering, Government College University Faisalabad, Faisalabad 38000, Pakistan 4International Center for Climate and Environment Sciences, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, PR China ABSTRACT: Droughts are one of the multifaceted meteorological disasters affecting agriculture, livestock and water resources worldwide. Pakistan has a semiarid climate system with a high degree of interannual rainfall variability. This study evaluates the applicability and comparison of 3 drought indices (deciles index [DI], standardized precipitation index [SPI] and reconnaissance drought index [RDI]) in Pakistan. Monthly rainfall and temperature data (1981−2017) from 30 weather stations were used to analyze the current status of drought occurrence in terms of severity and magnitude. A nonparametric Mann-Kendall test and Sen’s slope estimates were applied on drought indices to determine the statistical significance and magnitude of the trend. The DI cap- tured the dry episodes in the region well, as Baluchistan and Sindh provinces have been seen to be more susceptible to droughts. The indices of SPI and RDI were well correlated at 3, 6, 9 and 12 mo timescales. Province-level analysis revealed the highest number of drought years during the 3 mo timescale and the lowest number of drought years during the 12 mo timescale. Overall, a lin- early increasing trend of SPI and RDI (towards wetness) was observed, whereas the province-level analysis showed a statistically significant trend at the 95% confidence level for Khyber Pakhtunkhwa, Punjab and Sindh in the long-term drought analysis. Moreover, analyses of histori- cal drought years and their intensity have been investigated and compared with a recent long drought episode (1999− 2002). The analysis of historical drought events highlights the challenging nature of drought management in Pakistan. The outcomes of this study would help water resource managers to investigate drought response measures for drought preparedness in the country. KEY WORDS: Drought indices · Severity · Trend analysis · Climate change · Pakistan 1. INTRODUCTION ported worldwide (Gocic & Trajkovic 2014, Duffy et al. 2015, Zhao & Dai 2015, Hui-Mean et al. 2018). Drought is one of the most damaging hydrometeo- Droughts are directly linked to water availability, so rological hazards that are recurrent in nature (Mishra any shift in drought characteristics due to a warming & Singh 2010, Guha-Sapir et al. 2012, Samaniego et climate would affect water demand and food security al. 2013). An increase in the severity and magnitude (Nam et al. 2015, Touma et al. 2015). The phenome- of drought episodes due to warming has been re - non of drought mostly persists over arid to semiarid © The authors 2019. Open Access under Creative Commons by *Corresponding author: [email protected] Attribution Licence. Use, distribution and reproduction are un - restricted. Authors and original publication must be credited. Publisher: Inter-Research · www.int-res.com 190 Clim Res 78: 189–204, 2019 environments where economies are strongly de - characterization (Xie et al. 2013, Ahmed et al. 2016, pendent on agriculture (Kazmi et al. 2015, Ahmed et Haroon et al. 2016, Adnan et al. 2018). In recent al. 2018a). Interannual rainfall variability makes an decades, climate change has severely affected socio- arid region more susceptible to drought because of economic and environmental conditions in Pakistan below-average rainfall (Adnan et al. 2017). Such pro- and surrounding areas of Southwest Asia (Abbas et longed drought conditions may in turn cause severe al. 2018a,b). desertification, increased frequency of aerosol load- The current study is designed to monitor drought- ing and more intense dust storm events affecting prone areas of Pakistan. This paper identifies and human health, ecosystems and regional climate, as compares the severity and magnitude of drought ob served in Southwest Asia (Kaskaoutis et al. 2012, in different provinces of Pakistan by calculating Rashki et al. 2012, 2013a,b, 2015, 2017). The timely drought indices (DI, SPI and RDI) at 3, 6, 9 and 12 mo information obtained through various drought in - timescales. The time series analysis of SPI and RDI dices about the start and areal extent of droughts is helps to determine historical drought episodes. This beneficial for making appropriate contingency plans study will help policy makers to formulate contin- (Morid et al. 2006). gency plans for climate change-induced drought- Some studies on drought assessment focus on a sin- prone areas in the region. gle index (Du et al. 2013, Gocic & Trajkovic 2014, Adnan et al. 2015). Climatic conditions vary region- ally in the world; thus, to have a complete picture of 2. MATERIALS AND METHODS drought including its spatiotemporal extent and severity, it is effective to calculate and use various 2.1. Study area drought indices for drought assessment and monitor- ing (Mendicino et al. 2008). To quantify droughts, Pakistan is a predominantly arid region lying many studies have been carried out in different within the geographic coordinates of 23.38°−30.25° N regions of the world, and almost 50 indices have latitude and 61.78°−74.30° E longitude, with a total been developed based on different variables. Despite land area of 796 096 km2. Fig. 1 shows weather sta- the high number of drought indices available, most tions spatially distributed in different provinces of provide only a general description of droughts with Pakistan as well as Pakistan’s borders with China to no information about drought risk related to pre- and the north, the Arabian Sea to the south, India to the post-drought episodes (Kim et al. 2011, Huang et al. east and Iran and Afghanistan to the west. Agricul- 2015, Li et al. 2015, Ahmed et al. 2016). tural practices contribute 25% of the total gross do- Various meteorological organizations generally mestic product of Pakistan, whereas the projected use some of the well-known indicators to monitor cultivated land area percentage is about 27.87%. droughts worldwide such as the standardized precip- Based on its geography, there is great variability in itation index (SPI), standardized precipitation evapo- transpiration index and Palmer drought severity index in the USA (Palmer 1968, McKee et al. 1993); reconnaissance drought index (RDI) in Europe; deciles index (DI) in Australia (Gibbs & Maher 1967); and China Z index in China (Wu et al. 2001) as endorsed by the World Meteorological Organization (Hayes et al. 2011, Svoboda et al. 2012). Pakistan is predominantly categorized as an arid country (low precipitation and high temperatures) with an agricultural-based economy (Adnan et al. 2015, 2017). According to the land utilization survey, 34.15 Mha of land area is agricultural, while 23.60 Mha of land area is uncultivable; 25% of the culti- vated land is rainfed, which plays a vital role in the country’s economy (Adnan & Khan 2009, Kazmi et al. 2015). The economy’s heavy reliance on agriculture has made Pakistan more susceptible to drought risks. Fig. 1. Study area, showing provinces and weather stations Many studies in Pakistan have focused on drought in Pakistan Hina & Saleem: Drought assessment in Pakistan 191 Table 1. Weather stations in Pakistan purpose, monthly rainfall (mm) and maximum and minimum temperature (°C) data covering the histori- Station Station Elevation Latitude Longitude cal period of 37 yr (1981−2017) were obtained from no. name (MSL) (° N) (° E) the Pakistan Meteo ro logical Department. A list of weather stations along with their latitudes and longi- 1 Nokkundi 682 28.81 62.76 2 Dalbadin 848 28.98 64.46 tudes is presented in Table 1. Three drought indices 3 Panjgur 968 26.75 64.00 (DI, SPI and RDI) were calculated using monthly cli- 4 Lasbella 87 26.23 66.62 mate data of 30 weather stations spatially distributed 5 Khuzdar 1231 27.50 66.38 at the province level in the region. 6 Kalat 2015 29.17 66.44 7 Sibbi 133 29.55 67.88 8 Quetta 1626 30.2 66.58 9 Zhob 1405 31.21 69.28 2.3. Drought indices 10 Barkhan 1097 29.92 69.74 11 Jacobabad 55 28.59 68.47 To have a clear idea about the severity and magni- 12 Badin 09 24.61 68.62 13 Hyderabad 28 25.23 68.25 tude of drought episodes in the region, 3 types of 14 Chhor 05 25.67 69.30 drought-related indices were used: DI, SPI and RDI. 15 Padidan 46 26.91 67.98 A brief description of the indices follows. 16 Khanpur 88 28.67 70.66 17 Bahawalpur 110 29.35 71.93 18 Bahawalnagar 161 29.96 73.21 19 Multan 122 30.49 71.33 2.3.1. DI 20 Lahore 214 31.52 74.40 21 Sargodha 187 32.14 72.76 One of the simplest drought index calculation meth- 22 Sialkot 255 32.52 74.65 ods is DI. According to this method, long-term monthly 23 Jhelum 287 32.96 73.76 24 Mianwali 210 32.71 71.55 rainfall data of an extended length are ranked in 25 D.I.Khan 171 31.93 70.85 descending order to make a cumulative frequency 26 Peshawar 327 34.11 71.61 distribution (Gibbs & Maher 1967). Tigkas et al. 27 Balakot 995 34.79 73.45 (2015) grouped DI into 5 classes, shown in Table 2.