Evaluation of a composite drought index to identify seasonal drought and its associated atmospheric dynamics in Northern Punjab, Pakistan

Highlights

• Rainfed areas are highly vulnerable to drought.

• CDI performs better to monitor Agro-meteorological droughts in rain-fed areas.

• CDI identified four severe to extreme droughts for both Rabi and Kharif seasons.

• Associated with large-scale circulations revealed the severe and extreme droughts.

Abstract

Drought is one of the most devastating climate extremes in terms of its spatial extent and intensity. Rainfed areas are extremely vulnerable to drought, but effective monitoring may lessen the impact of such events. This study developed a composite drought index (CDI) for monitoring and assessing seasonal droughts in rainfed areas of the Potwar Plateau of Pakistan, using remotely sensed and observed meteorological datasets. We identified four severe-to-extreme drought periods in the Rabi season (wheat; 2000–01, 2001–02, 2009–10, and 2011–12) and four such events in the Kharif season (maize; 2000–2002 and 2009). An intense agro-meteorological drought was experienced in 2000, which reduced the wheat and maize yields to −54.6% and −29.9%, respectively. Our analysis revealed that these conditions could be explained by the vertically integrated moisture flux divergence (MFD), moisture transport, and total precipitable water (TPW) anomalies. For example, the presence of a strong MFD anomaly over the study area was responsible for preventing moisture transport from the Arabian Sea and Bay of Bengal, resulting in dry conditions. The index developed here can effectively monitor seasonal droughts in rainfed areas, which may help inform strategies to lessen the impact of such events.

Introduction

Pakistan is an agricultural country, however, the majority of the country consists of arid to semi-arid areas (Adnan et al., 2017; Haider and Adnan, 2014), with insufficient rainfall for growing crops and other vegetation. Although the agriculture sector has satisfactorily ensured the food security of Pakistan's growing population, it is vulnerable to climate change (Safdar et al., 2014; Ahmed and Schmitz, 2011). Two thirds of Pakistan's land area has an annual rainfall of under 250 mm, whereas only 8% of the geographical area receives more than 500 mm annual rainfall (Khan and Khan, 2015; Adnan and Khan, 2009; Alam, 2000). Areas with irrigation systems can cope with the scarcity of rainfall, but the rainfed and arid regions are prone to frequent droughts.

In the recent times, Asia has experienced an increasing frequency of droughts (Leng et al., 2015; Ganguli and Reddy, 2014). Droughts are an unavoidable recurring natural hazard, especially in rainfed agricultural areas. They have a relatively strong spatial impact in arid regions due to the spatial variability of rainfall in these areas (Ward and Makhija, 2018; Miyan, 2015). Droughts are responsible for more agricultural losses than any other natural disaster (Wilhite, 2000). They leave agricultural lands devastated, with large losses in crop yields having a serious impact on the country's gross domestic product, public health, environment, and economic development (Ain et al., 2019; Zhang et al., 2015; Hao and &Agha Kouchak, 2014; Zargar et al., 2011; Kao and Govindaraju, 2010; Wu et al., 2001; Le Houérou, 1996). Although the causes of drought are still uncertain, the shortage of irrigation water and the overgrazing of pastures by livestock, especially during periods of low rainfall, have contributed its effects (White et al., 2003).

Droughts are generally defined as prolonged periods with well below normal rainfall in a region leading to an extreme shortage of water (Gonzalez‐Hidalgo et al., 2009). Four types of droughts are generally recognized: meteorological, agricultural, hydrological, and socio-economic. (Guo et al., 2019; Huang et al., 2016; Mehran et al., 2015; Eklund and Seaquist, 2015; Wilhite, 2000; Agnew and Warren, 1996; Palmer, 1965).

A tool or index is required to monitor any type of drought (Hayes et al., 2011; Keyantash and Dracup, 2002). Drought indices simplify the complex relationships among meteorological parameters and provide a useful tool for the community and stakeholders. These indices can be used to reassess historical droughts and obtain an insight into the probability of drought recurrence, while also being an aid for planning, forecasting, and the design of certain agricultural applications. Hundreds of drought indices have been developed. Some are univariate, such as the standardized precipitation index (SPI), while others are bivariate, such as the standardized precipitation evapotranspiration index. There are also multivariate drought indices such as the Palmer drought severity index and composite drought index (CDI) (Kim et al., 2018; Waseem et al., 2015; Zargar et al., 2011). In recent years, remotely sensed data and drought indices have been widely used for drought monitoring and prediction (Abuzar et al., 2019; Zhong et al., 2019; Dalezios et al., 2018).

The onset and magnitude of drought in a region is still a paradox for researchers due to the insufficient number of ground meteorological observatories. However, satellite-based remotely sensed data has partially resolved the problem by providing a fast and economical way of acquiring information. This data is easily available and helps to identify drought onset, its temporal and spatial extent, and magnitude (Enenkel et al., 2016; Tadesse et al., 2014). These satellite-based remote sensing (RS) datasets have been widely used for crop and drought monitoring from the mid to late 1980s (Pan et al., 2017). The Moderate-Resolution Imaging Spectroradiometer (MODIS) provides high-resolution data on a near real-time basis for the assessment of drought and its propagation and intensity over an area where in-situ data is sparse (Khan et al., 2018; Klisch and Atzberger, 2016; Akhtar, 2014; Ghauri and Khan, 2013; Caccamo et al., 2011). The RS technology along with meteorological observations can effectively be used for agro-metrological drought monitoring with a CDI. Multivariate indices and CDIs are now in common use over almost all of the globe. For example, a CDI was developed and used for agricultural drought in Europe as well as drought vulnerability assessment in the USA (Baptista, 2014; Sepulcre-Canto et al., 2012). In Morocco, a CDI is considered to be a state-of-the-art tool for drought monitoring (Bijabar et al., 2018). Waseem et al. (2015) has developed and used a CDI to identify the driest year and driest month for the sub-basin of the Han River, South Korea.

In Pakistan, drought monitoring is generally conducted using the SPI, although Adnan et al. (2015) have evaluated the performance of almost 15 indices. The SPI alone may not accurately describe the drought conditions (especially their onset) in arid and rainfed regions, which tend to have a naturally drier climate. Moreover, the SPI on a one-month scale does not generate reliable results in comparison with the ground truth (Leasor et al., 2020). It is therefore imperative to combine meteorological observations with satellite-based drought products to monitor agro-meteorological drought effectively. Due to the high spatial and temporal variability of precipitation, seasonal droughts occur frequently in Pakistan. The performance of a CDI has been tested to determine the occurrence of agro-meteorological drought. The role of atmospheric dynamics during droughts has also been identified. This paper is organized as follows: Section 2 provides the data and methods, Section 3 presents and discusses the results, and finally in Section 4 the conclusions of the study are presented.

The Potwar Plateau of Pakistan (PPP) consists of the four districts of northern Punjab (Attock, Chakwal, Jhelum, and Rawalpindi), and the capital territory of Islamabad (Rashid and Rasul, 2011). The region lies between two rivers, the Indus and Jhelum, with an elevation varying from 300 to 600 m, and encompasses an area of about 13,000 km². The Geographical location of PPP is 32.5°–34°N and 72°–74°E (Fig. 1). Due to the high elevation, low temperature, and plentiful annual rainfall (1000–2000 mm), this region has a semi-arid to humid climate classification (Adnan et al., 2009, 2020). Two weather systems, the southwest summer monsoon (July to September) and eastward propagating western disturbances (December to March) prevail over the region, bringing rainfall during the Kharif and Rabi seasons, respectively (Hussain and Lee, 2009; Latif and Syed, 2016; Ahmed et al., 2020; Adnan et al., 2020). The mean annual temperature is between 13 °C and 26 °C (Adnan et al., 2017). Two major crops, maize and wheat, are grown in Kharif (May to September) and Rabi (October to April), respectively. The crop yield of this region contributes about 10% of the country's agronomy (Ashraf, 2016). Only the southwest summer monsoon precipitation relieves the moisture stress and fulfills the water demands of both seasonal crops (Adnan et al., 2018). The seasonal variability in rainfall can affect the agricultural sector, which contributes 24% of the GDP of the country (Latif and Syed, 2016).

The study area experiences a high spatial and temporal variability of rainfall. Agricultural activity is highly dependent upon rainfall. A deficiency of seasonal rainfall is one of the major causes of drought in the region. Due to the complex terrain, this region contains no canals as an irrigation system and the groundwater is very shallow. There are frequent seasonal droughts that impact on the crop yield in the region. Therefore, this area needs to be investigated using a CDI to monitor drought.