CORDEX Ensemble-based drought projections for Sindh Province of Pakistan under climate change

Climate change, the most crucial challenge of modern times, has affected the whole globe, including the countries that play a little role in producing greenhouse gas (GHG) emissions. Pakistan, the fifth most populous country contributes less than 1% to global GHG emissions but stands eighth on climate vulnerability index (Eckstein et al. 2021). Climate change is a major contributor to the increased frequency and intensity of droughts, the most expensive natural disaster. It is estimated that Pakistan's gross domestic product (GDP) may be slashed by 18–20% by climate change per year by 2050 (World Bank 2022). With a history of droughts, mainly in Southern parts, Pakistan faces a major drought cycle every 16–20 years and a lesser impact drought every 3–4 years (NDMA 2018). Sindh Province of Pakistan is particularly vulnerable to droughts; 65% of its landmass receives less than 100 mm of average rainfall annually. The province receives 80% of its annual rainfall in 3 months of monsoon while the remaining 9 months are extremely dry. During most recent events in 2014, 2015, and 2016, respectively, caused 326, 398, and 476 fatalities of children under the age of five (Durrani 2018). Research projects a significant increase in the number of individuals affected and the extent of financial damages (Buhr et al. 2018). Developing countries like Pakistan are more vulnerable to climate-driven disasters due to several factors including capacity issues (Aziz & Yucel 2021).

Assessment of climate change impacts is a well-focused research area and researchers have made use of general circulation models (GCMs) and regional climate models (RCMs) to project future climate trends. Majority of the researchers have conducted studies using GCM data and eventually worked on larger scales like continents, countries, and basins as the resolution of GCMs allowed (Burke & Brown 2008; Lu et al. 2019; Aadhar & Mishra 2020; Cook et al. 2020; Zhai et al. 2020; Su et al. 2021; Supharatid & Nafung 2021). The body of research can be categorized into several key areas. First is focused on the identification of regions that are becoming increasingly susceptible to droughts due to altered precipitation patterns and rising temperatures. In short, these studies have explored the future droughts and assessed the likelihood as well as severity of prolonged dry conditions (Cook et al. 2020; Spinoni et al. 2020; Zhai et al. 2020). The second category of research delves into the complex mechanisms that cause and govern droughts, including precipitation, temperature, oceanic influences, and other atmospheric conditions. In this category, several drought indices have been used to investigate the impacts of these factors as well as the accuracy of these drought indices (Vicente-Serrano et al. 2010; Jiang et al. 2015; Zhang et al. 2020). This category of research focuses on the explanation of drought development, progression, termination, and the factors that contribute to these traits. In the third category, hydrological modeling has been used for a more meticulous examination of the impact of droughts on water resources. This category takes into account changes in streamflow, groundwater, and reservoir levels (Fleig et al. 2006; Corzo Perez et al. 2011; Van Loon et al. 2012). Research in this category is imperative for long-term management and planning. In the fourth category, the socioeconomic and ecological implications of droughts have been investigated. This involves evaluating the consequences of droughts for agriculture, water supply, energy production, and vulnerable communities (Edalat & Stephen 2019; Crausbay et al. 2020). Climate models, including GCMs, RCMs, and the coordinated regional downscaling experiment (CORDEX), have achieved many advancements in recent years that enhance the utility of these models in drought assessment and management (Giorgi & Gutowski 2015). GCMs now offer relatively higher resolution than their previous versions. Hence, more accurate representations of precipitation patterns, temperature changes, and atmospheric circulation are possible to draw. These improvements make GCMs helpful in assessing long-term drought trends and their global-scale implications with more accuracy (Raju & Kumar 2020). RCMs, on the other hand, have witnessed extensive progress in downscaling global climate projections to regional or local levels. The enhanced computational power and improved parameterizations have enabled RCMs to provide higher resolution data successfully capturing localized climatic realities which GCMs are unable to catch. This ability of RCMs results in more precise drought assessments at a regional scale (Chokkavarapu & Mandla 2019). Moreover, the CORDEX has enabled international collaboration to enhance regional climate simulations. These advancements in GCMs, RCMs, and CORDEX have led to more accurate and detailed assessment of drought at respective scales. These models have effectively been used in various sectors, including agriculture, water resources, and disaster risk reduction. In a nutshell, a richer array of data and tools are available now to explore the dynamics of drought for science-based policy formulation and planning at multiple scales.

The studies conducted so far on drought assessment and characterization in Sindh Province of Pakistan have predominantly centered on historical drought events. These investigations have primarily focused on comprehending drought patterns during the pre-1960 and post-1960 periods and have used standardized precipitation index (SPI) and standardized precipitation evapotranspiration index (SPEI) to evaluate the severity and impact of these droughts (Khan & Gadiwala 2013; Adnan et al. 2015; Ullah et al. 2020). However, it is crucial to highlight that, to the best of our knowledge, there is a significant research gap in the context of future drought projections in the region. Addressing this gap is of paramount importance due to the ongoing and projected shifts in global climate patterns. With the increasing uncertainty surrounding future climate and its influence on drought events, it is imperative to explore future drought scenarios using advanced climate model data. While both global climate models (GCMs) and RCMs provide valuable insights, the CORDEX data emerge as especially pertinent for such studies covering small areas. As stated before, CORDEX data with its finer spatial resolution are suitable for projecting future drought patterns in Sindh Province. By utilizing finer resolution models, more detailed and accurate projections can be drawn for a thorough understanding of climate change impacts on drought occurrences and severity in a region like Sindh Province. There is an urgent need for research leveraging CORDEX data to fill the void in our understanding of future drought scenarios. These research questions specifically need to be answered in order to provide critical insights for the management of water resources, agriculture, drought-induced migration, and other socioeconomic impacts: (1) How will climate change impact the droughts in the Sindh Province of Pakistan throughout the 21st century under different representative concentration pathways (RCPs)? (2) What are the trends in the SPEI under RCP 2.6 and RCP 8.5 scenarios? and (3) Which areas within Sindh Province are expected to experience the most significant changes in drought conditions? In this regard, we hypothesized that (1) the frequency and severity of droughts in Sindh Province will increase significantly under the RCP 8.5 scenario compared to the RCP 2.6 scenario due to higher GHG emissions and resultant climate change impacts; (2) the SPEI will show a decreasing trend over the 21st century with more pronounced declines under RCP 8.5, indicating worsening drought conditions; and (3) specific areas within Sindh Province will exhibit worse drought conditions, leading to greater challenges for local water resources.

Therefore, this study is aimed at projecting future droughts in Sindh Province using CORDEX data under RCP 2.6 and RCP 8.5. The objectives are to: (1) calculate SPEI using CORDEX ensemble for 21st century; (2) detect and visualize drought trends for near-future, mid-future, and far-future periods. The analysis will help investigate the potential impacts of climate change on drought occurrences. The results of the study will provide valuable information for policymakers and stakeholders in the region to adopt effective drought management strategies.

Sindh Province, located in southern Pakistan, is home to 42.4 million people and has a total area of 140,914 km², centered on 26.1°N, 68.5°E. The region has a semi-arid to arid climate and receives annual rainfall of less than 200 mm in the southern parts. The province is highly dependent on monsoon rains for agriculture and other livelihood activities. The majority of the rainfall occurs during the monsoon season (July–September). The summer season (May–September) is characterized by harsh weather conditions wherein temperature can reach up to 50°C (Muslehuddin & Faisal 2006).

The province lies in the lower Indus Basin, which is a major river system in South Asia. The Indus River System is the lifeline of the region as it serves as the primary source of water for irrigation, industry, and domestic uses. The Indus River provides about 90% of the total surface water in the region. Groundwater in the province is found in the form of shallow to deep aquifers which are largely brackish, hence, unfit for irrigation. In some places, it is used for irrigation by supplementation with canal water (Shahab et al. 2019). Several factors affect the hydrology of Sindh Province including topography, climate, land use, and water management practices. In the wake of climate change and droughts, it is imperative to adopt optimum water management for sustainable development in the province.

The province experiences frequent droughts giving rise to various economic, environmental, and social implications. The southern parts of Sindh including the Thar region are particularly characterized by severe impacts of droughts, leading to damages including human lives. Therefore, detailed futuristic studies of droughts in the province are crucial for understanding this complex disaster and its interaction with climate. Such research can provide insights into the potential risks of droughts and their impact on the region's water resources, socioeconomics, and ecosystems leading to the formulation of informed strategies for drought management and adaptation.

To improve the accuracy of climate model projections, bias correction techniques are often used to adjust the systematic biases that exist in the models. Despite the potential for controversy, using bias correction is advised for hydro-climate analysis (Berg et al. 2012). One such technique is distribution mapping which involves fitting parametric probability distribution functions (PDFs) to the observed and modeled data and then matching their cumulative distribution functions (CDFs). The distribution mapping method assumes that the CDF of the modeled data is similar to the CDF of the observed data, but shifted due to the bias. The resulting bias-corrected data can be used to improve the accuracy of drought projections and other climate impact assessments (McGinnis et al. 2015). In this study, we employed distribution mapping to bias-correct CORDEX data using the Climate Model Data for Hydrologic Modeling (CMhyd) tool. The bias-corrected data were used for further analysis (S.1 may be referred in Supplementary File for additional figure).

In the present study, SPEI was calculated at 6-, 12-, and 24-month timescales as the aim is to grasp the long-term impacts of drought over a time span of 120 years (1981–2099). Seasonal and short-term droughts are assessed by calculating the drought index at smaller timescales like 1 and 3 months. Whereas, drought indices are calculated at larger timescales like 6, 12, and 24 months for capturing longer impacts related to agricultural droughts, long-term water availability, groundwater, reservoirs situation, and associated socioeconomic aspects, etc. Long-term assessment is useful for monitoring prolonged drought events and their cumulative effects on water resources, ecosystems, and water supply systems. Above all, longer timescales help in assessing the influence of climate variability and climate change on drought patterns. Hence, they are particularly useful for water resources planning and management (Vicente-Serrano et al. 2010).

Based on the SPEI, a drought event is defined as a time period that is characterized by severity, length, and frequency, and during which the index values are constantly −0.5 or below. Run theory, a probabilistic technique, is widely used to derive drought characteristics like duration and severity. According to run theory, drought severity is an event when the values of SPEI are below the threshold, whereas duration is the time when the value of SPEI remained continuously below the threshold and such durations of droughts are identified as events. Drought frequency is the number of times the SPEI falls below the threshold (Yevjevich 1969).

The seasonal trends in the period 1981–2010 do not show any significant difference from the yearly trends except for minor changes in the magnitude of trends. However, the overall trends did not shift from negative to positive or vice versa. Keeping in view the similarity between yearly and seasonal trends, the future analysis was carried out only yearly basis.

Under RCP 8.5, the drought is expected to increase with a strong trend continuously growing toward the end of the century. The areas like Hyderabad, Padidan, and surroundings where the SPEI trends were positive or weak negative, during the historical period, now exhibited strong negative SPEI trends, indicating the projection of strong droughts in the future. This is in accordance with the report of the Intergovernmental Panel on Climate Change that predicts that Hyderabad will be the warmest city in Pakistan by the year 2100. The report also foresees increased droughts in Pakistan (IPCC 2022).

The impacts of climate change on droughts in Sindh Province, Pakistan, are part of a global trend of increasing climate variability and extremes. Rising global temperatures are causing shifts in precipitation patterns, leading to more frequent and severe droughts worldwide. The findings of this study, particularly the projected worsening of drought conditions under the RCP 8.5 scenario, align with global projections indicating that higher GHG emissions will intensify drought intensity and frequency. By focusing on Sindh Province, a region highly vulnerable to climate change, the urgent need for adaptive strategies is underscored, that are applicable both locally and in other drought-prone areas globally. This broader perspective enhances the global relevance of this research, offering valuable insights for policymakers and stakeholders aiming to abate the adverse effects of climate change on water resources and agricultural sustainability. A recent study revealed an increasing trend of mean annual temperature in Sindh Province, particularly at stations Nawabshah, Sukkur, Umerkot, Larkana, Badin, and Dadu at a rate of ∼0.03, 0.034, 0.037, 0.03, 0.028, and 0.045 °C/year, respectively (Qureshi et al. 2023). Large-scale changes in wind speed, relative humidity, and geopotential height anomalies are also considered potential drivers of droughts in the region (Ullah et al. 2022). Consequently, increasing drought trends have been reported in the past, particularly after 1960 when droughts were found to be more severe and longer (Ahmed et al. 2023). In addition to climate change, drought development has also been impacted by anthropogenic activity (Magnan et al. 2021). The findings of our study align with the previous studies as all the stations except Hyderabad and Padidan showed a decreasing SPEI trend and moderate to severe droughts were detected from 1998 to 2002 (Adnan et al. 2015). In addition to climate change, several anthropogenic activities led to the development of droughts. Population growth has led to increased pressure on the resource base and resulting competition for water. The need for water resources is progressively growing in terms of quantity, quality, and guaranteed water supply rate (Rosa et al. 2020). The overuse of water resources, a decrease in plant cover, and a drop in groundwater levels are only a few of the ecological and environmental issues that will weaken the resistance to drought and hasten its onset. Furthermore, excessive GHG emissions brought on by quick social and economic growth encourage the onset of drought. Thus, as a result of climatic and anthropogenic influences, the drought situation in the Province is becoming worse (Alamgir et al. 2016).

The study forecasts drought occurrence in Sindh Province under both scenarios, RCP 2.6 and RCP 8.5. The trend analysis was carried out for three time periods, i.e., near future (2011–2040), mid-future (2041–2070), and far future (2071–2099). In the context of climate scenarios, the adoption of RCP 2.6, characterized by stringent mitigation measures, presented a persistent occurrence of drought events spanning the entire century. However, these droughts exhibited a comparatively lower degree of severity when juxtaposed with the more alarming RCP 8.5 scenario, representing a business-as-usual trajectory. Conversely, under the RCP 8.5 scenario, a consistent upward trajectory in the frequency of droughts was observed, starting from the near future and extending into the far future. It is anticipated that, under RCP 8.5, drought occurrences will manifest with greater frequency. Specifically, the province is expected to experience a higher incidence of both severe and moderate droughts in this scenario, in stark contrast to the outcomes projected under RCP 2.6. Furthermore, extreme drought events were forecasted for locations such as Paddidan and Hyderabad in both scenarios. Notably, trend analyses revealed adverse SPEI trends in both scenarios, with a notably more pronounced trend observed in the RCP 8.5 scenario. Examination of the average durations of drought events revealed variations across different locations. Under the RCP 2.6 scenario, the average duration in months was calculated to be 12 at Badin, 7.42 at Chhor, 20.91 at Hyderabad, 16.08 at Karachi, 16.79 at MohenjoDaro, 21.15 at Nawabshah, 11.69 at Padidan, and 12.24 at Rohri. On the other hand, under the RCP 8.5 scenario, considerably longer drought durations were projected, with average durations (in months) of 24.57 at Badin, 27.43 at Chhor, 24.35 at Hyderabad, 19.47 at Karachi, 25.56 at MohenjoDaro, 29.71 at Nawabshah, 18.45 at Padidan, and 17.79 at Rohri. This analysis underscores the expectation of significantly prolonged drought events under the more extreme RCP 8.5 scenario compared to the relatively milder RCP 2.6 scenario.

In this situation, Sindh, as a province that is particularly vulnerable to droughts due to its arid climate, needs to chalk out a comprehensive adaptation plan. Mitigation efforts to reduce GHG emissions require significant financial resources, technical expertise (Princiotta 2021), and policy interventions at a national and international level, which may not be feasible for a province. Sindh Province which contributes slightly more than 15% to Pakistan's total GHG emissions (Government of Pakistan 2021), can do very little on the mitigation front. The mitigation efforts have to be steered at the global level as climate change is not a local problem and most of its drivers act at the global scale. Moreover, there exist apprehensions about the goals set under RCP 2.6, though still possible to achieve technically, the pace of progress is not at par. On the other hand, adaptation measures are more localized and can be tailored to the specific needs of the region. Additionally, adaptation measures can provide immediate benefits to the affected communities, such as improved access to water and food security. Therefore, investing in adaptation measures can help Sindh cope with the increasing frequency and intensity of droughts and ensure the well-being of its population.

A comprehensive adaptation plan given the agro-climatic and economic realities of the province should be centered on key components such as (i) early warning systems and drought preparedness, (ii) water conservation, (iii) water augmentation, and (iv) crop adjustments. Development of early warning systems as well as emergency response plans can also help in reducing the impact of drought on affected communities. The water conservation component includes measures to improve water management practices, groundwater recharge, and efficient irrigation systems. Water augmentation may include water storage like rainwater harvesting, water diversion, treating and reusing wastewater or other non-potable sources of water for irrigation or other non-drinking purposes, and desalinating seawater or brackish water to make it potable. Additionally, promoting drought-resistant crops and technologies, such as genetically modified crops that require less water, can help mitigate the impact of drought. However, it is important to note that the success of these measures largely depends on the level of community participation, government support, and financial resources allocated toward their implementation. Meanwhile, stronger advocacy efforts are needed to call for climate justice for Pakistan in general and Sindh Province in particular. The climate problem has mainly been caused by the industrially developed world and the countries like Pakistan are facing the consequences of an action they never committed. Climate finance should be channeled keeping in view the needs of the countries with little contribution to GHGs (Colenbrander et al. 2018) and the ‘Polluter Pays Principle’ should be adopted at the global level in this regard, in letter and spirit.

To objectively examine the temporal evolution, geographical distribution, and gridded trend features of drought in Sindh Province under RCP 2.6 and RCP 8.5, the SPEI was used as a drought indicator. Run theory was used to extract the length and severity of drought occurrences, and the marginal distribution functions of drought length and severity were fitted. The analysis of the historical versus future projected SPEI under both RCPs indicated increased droughts in Sindh Province as per all four CORDEX models used in this study. Consequently, Sindh is likely to face water management challenges in the future.

Under RCP 2.6, which assumes strong mitigation measures, there was a persistent presence of droughts throughout the century, but with less intensity compared to RCP 8.5. In contrast, under RCP 8.5, which represents a business-as-usual scenario, there was a continuous increase in droughts from the near future to the far future. Southern Sindh is expected to face more severe drought conditions compared to Northern Sindh.

Droughts, in general, are expected to be more frequent under RCP 8.5. The province will face severe and moderate droughts under this scenario compared to RCP 2.6. Extreme droughts were forecasted at Paddidan and Hyderabad under both scenarios. Trend analysis showed negative SPEI trends under both scenarios, and these were stronger under RCP 8.5.

The average durations of droughts in months under RCP 2.6 were found to be 12 at Badin, 7.42 at Chhor, 20.91 at Hyderabad, 16.08 at Karachi, 16.79 at MohenjoDaro, 21.15 at Nawabshah, 11.69 at Padidan, and 12.24 at Rohri. Under RCP 8.5, average durations in months were calculated to be 24.57 at Badin, 27.43 at Chhor, 24.35 at Hyderabad, 19.47 at Karachi, 25.56 at MohenjoDaro, 29.71 at Nawabshah, 18.45 at Padidan, and 17.79 at Rohri. This projects much longer droughts under RCP 8.5 compared to RCP 2.6.

The results necessitate the implementation of effective adaptation strategies in Sindh Province. An adaptation approach, given its localized nature, is more feasible for Sindh Province. Some possible strategies could include water conservation and augmentation measures, the development of drought-resistant crops, and improving water storage and distribution systems.

The negative SPEI trend implies worsening drought conditions over time. This worsening trend highlights the escalating impact of climate change on regional water resources, posing substantial risks to agriculture, ecosystems, and human livelihoods. It underscores the urgent need for effective adaptation strategies to address the growing challenges posed by climate-induced droughts.

This research was supported by the National Key R&D Program of China (2021YFC3000203).

Z. H., H. Y., and Z. W. conceptualized the whole article. Z. H. and H. Y. developed the methodology; M. A., M. F., M. G. and M. I. A. rendered support in the literature search; Z. H., M. I. A., M. U., and R. A. visualized the data; Z. H. drafted the original data; Z. W., H. Y., R. A., and M. U. contributed to critical review; Z. H. edited the article; H.Y. arranged the funding.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.