Comparison of contrasts in rainfall and drought characteristics in the Chambal basin in Madhya Pradesh and Rajasthan

Droughts resulting from insufficient rainfall over extended time periods cause significant hydrological imbalances which generally lead to water scarcity, reduction of river flows, and depletion of groundwater, crop stress, reduced crop yields, and crop damages (Gaiha et al. 2015). Many researchers have evaluated drought characteristics based on different criteria used by them to define a drought (Hoyt 1938; Blumenstock 1942; Condra 1944; Ramdas & Malik 1948; Thornthwaite 1948; Ramdas 1960; Herbst et al. 1966; Yevjevich 1967; Warrick 1975; World Meteorological Organization 1975; Gadgil & Yadamani 1987). The India Meteorological Department (IMD), 1971 criterion based on the percentage departure of rainfall from its long-term mean, is widely adopted in India (Narain et al. 2006; Namdev & Pandey 2017; Karinki & Sahoo 2019; Akhtar et al. 2021) and also used here. Different variables indicate the impact of drought including groundwater level, soil moisture, streamflow, reservoir storage, and lake water levels (Triana et al. 2020). The meteorological drought propagates into hydrological drought as groundwater and surface water levels reduce further (Van Loon 2015), and the hydrological drought thereafter propagates agricultural drought, directly affecting the crop growth and its survival (Mishra & Singh 2010; Thomas et al. 2014).

Indicator-based approaches have been widely used for the evaluation of drought characteristics Palmer Drought Severity Index (Alley 1984), Self-calibrated Palmer Drought Severity Index (Wells et al. 2004), Standardized Precipitation Evapotranspiration Index (Vicente-Serrano et al. 2010), Effective Drought Index (Byun & Wilhite 1996), China Z-index, and Deciles (Gibbs & Maher 1967), and these have their own advantages and limitations. The Standardized Precipitation Index (SPI) (McKee et al. 1993; McKee et al. 1995) owing to its utility and flexibility is widely accepted and more conveniently used to predict different types of drought at various time scales in all climatic regimes. The SPI is a probabilistic indicator that describes the degree to which the cumulative precipitation of a specific period departs from the average state (Guhathakurta et al. 2017). The SPI is easy to calculate and convenient to apply, as it is based on a single variable, namely the long-term monthly rainfall (Guhathakurta et al. 2017). The SPI provides real-time monitoring of the intensity and spatial extension of droughts, at different time scales of 3, 6, 12, and 24 months. Several studies have also highlighted the strength of the SPI (Keyantash & Dracup 2002; Guhathakurta et al. 2017) based on comparing it with other indexes based on various performance evaluation criteria including robustness, tractability, transparency, sophistication, extendibility, and dimensionality (Thomas et al. 2015). Moreover, the WMO in 2009 recommended the SPI as the main meteorological drought index that countries should use to monitor and follow drought conditions (Hayes 2011).

Li et al. (2012) studied the drought evolvement using the SPI in the Huaihe river basin and concluded that the drought intensity increased in the 21st century, whereas the drought-affected areas decreased. Basamma et al. (2017) carried out the assessment of drought using the SPI in the Mewar district of Rajasthan (RJ). Sánchez et al. (2016) used the SPI at 1- and 3-month time scales to represent soil moisture drought and crop stresses. Rahman et al. (2018) studied the spatial rainfall variability and used the SPI for drought assessment in the Khyber Pakhtunkhwa province of Pakistan. Saini et al. (2020) emphasized the capability of SPI for predicting drought and found that the north-western, south-eastern, and north-eastern parts of RJ had higher drought severity, as compared to lower severity in the central regions. Won & Kim (2020) investigated the impact of climate change on future droughts using the SPI and Evaporative Demand Drought Index (EDDI) over Korea to devise a drought severity–duration–frequency (SDF) curve. Sharafati et al. (2020) compared the SPI with the Standardized Precipitation Evapotranspiration Index (SPEI) to understand the importance of evapotranspiration on meteorological droughts and suggested high correlations between the SPI and the SPEI, indicating the efficiency of SPI for predicting drought based on a single parameter. Das et al. (2020) studied the drought scenario in the Luni River basin using the SPI and investigated drought trends using the Mann–Kendall (MK) test. The comprehensive analysis revealed the impact of climate change on drought.

Water scarcity is one of the major issues facing the Chambal basin which is located in a semi-arid region in India. It is believed that many of the districts in MP bordering RJ have seen increased occurrences of droughts and dry spells in recent times. As the droughts are initiated by rainfall deficiency, the evaluation of the rainfall characteristics in MP and RJ will shed some information. Therefore, a comparison of the rainfall characteristics as well as the drought characteristics will help to understand the similarities and contrasts of both these phenomena in the Chambal basin in MP and RJ which is the main objective of this paper. The evaluation of drought characteristics will pave the way for planning for drought mitigation strategies for the reduction of drought impacts.

River Chambal located in the central and western regions of India flows through MP, RJ, and Uttar Pradesh (UP) in India and is the main tributary of Yamuna River, which ultimately joins the Ganga River system. The Chambal basin lies between 22°27′ and 27°20′ North latitudes and 73°20′ and 79°20′ East longitudes. The location map of the study area is shown in Figure 1. The basin lies in the semi-arid zone of north-western India. The average annual rainfall in the Chambal basin is 760 mm. It is Perennial River with a length of 960 km with an average annual discharge of 14,380.416 Mm³/year. The study area has been limited up to the confluence of Banas River with Chambal River. The study is limited to the comparison of drought characteristics in the Chambal basin falling in MP and RJ only as is clearly reflected in the title. Sixteen districts of MP are falling in the Chambal basin which includes Sheopur, Shivpuri, Guna, Vidisha, Bhopal, Rajgarh, Shajapur, Sehore, Dewas, Indore, Dhar, Ujjain, Agar, Ratlam, Mandsaur, and Neemuch, whereas 15 districts of RJ are falling in the Chambal basin which includes Ajmer, Jaipur, Dausa, Karauli, Tonk, Sawai Madhopur, Bundi, Baran, Kota, Bhilwara, Rajsamad, Udaipur, Chittaurgarh, Pratapgarh, and Jhalawar. Moreover, out of the total basin area of 143,219 km² up to its confluence with River Yamuna, only 1,101 km² falls in UP, whereas the major portion of the basin is located in MP (76,854 km²) and RJ (65,264 km²). The land-use and land-cover (LULC) map of the study area is given in Figure 2.

The daily rainfall data for various districts and falling in Madhya Pradesh (MP) and RJ regions located in the Chambal basin were obtained from the Water Resource Departments of the respective states. The length of the data used in the analysis is for a period of 34 years from 1985 to 2018. No data gaps were observed in the datasets. The consistency of rainfall records has been checked using the double-mass curve analysis, in which the cumulative annual or seasonal rainfall of the station of interest is compared with the cumulative annual or seasonal rainfall of the reference station. The rainfall of the reference station is considered as the mean rainfall at the adjacent stations. The spatial analysis has been carried out using the inverse distance weighted interpolation technique in the Spatial Analyst Tool of ArcGIS10.1 software. This method is based on the assumption that things that are close to one another are more alike than those that are farther apart (http://www.gisresources.com/types-interpolation-methods_3/) and estimates the values of unsampled points by averaging the values of surrounding sampled points. The LULC map of the study area has been obtained from the Oak Ridge National Laboratory Distributed Active Archive Center (ORNL DAAC) (Roy et al. 2012, 2015). The next section deals with the methodology adopted to compare the rainfall and drought characteristics in various districts of MP and RJ falling in the Chambal basin. The methodology adopted for the identification of rainfall characteristics involves the computation of annual and seasonal rainfalls, coefficient of variation in annual rainfall, 1-day maximum rainfall, number of rainy days, and trend analysis using the non-parametric MK test, whereas the identification of drought characteristics involves the computation of the departure of annual/seasonal rainfall, probability analysis of annual/seasonal rainfall, SPI-based evaluation of drought frequency, and drought severity.

Mann presented a non-parametric test of randomness with respect to time, which is a particular application of Kendall's test for correlation and is commonly known as the ‘MK’ test (Abdullahi et al. 2015). The MK test is a statistical test widely used for the analysis of trends in climatology and hydrologic time series (Alhaji et al. 2018). There are two advantages of using this test. First, this is a non-parametric test and does not require the data to be distributed normally. Second, the test has low sensitivity to sudden interruptions due to unequal time series (Panda & Sahu 2019).

The statistic represents the number of positive differences minus the number of negative differences for all the differences considered (Longobardi & Villani 2010).

The presence of a statistically significant trend is evaluated using the Z value. A positive value of Z indicates an upward trend, and its negative value indicates a downward trend. The Z values were tested at a 0.05 level of significance.

In the end, Q is computed by a two-sided test at 100 (1−α)% confidence interval, and then a true slope can be obtained by the non-parametric test. A positive value of Q indicates an upward or increasing trend, and a negative value of Q gives a downward or decreasing trend in the time series.

As per the IMD classification (Appa Rao 1986), an area is considered to be affected by drought if the seasonal rainfall during the year is less than 75% of its normal. However, depending on the climatic conditions, crop type, and crop practices in the study area, a rainfall anomaly of −20% may generally lead to water stress for crop growth thereby affecting crop health and crop yields. Therefore, those years with seasonal departure less or equal to −20% have been considered as a drought year. The drought severity has thereafter been classified into four categories as given in Table 1 (Pandey et al. 2008, 2010; Thomas et al. 2016; Surendran et al. 2019; Wable et al. 2019). The departure analysis of seasonal rainfall has been carried out for the various rain gauge stations for the identification of drought years and drought severity in the identified drought years.

Higher values of the RDI are indicative of higher drought proneness of the area and based on this information, and the priority areas for undertaking mitigation measures can be prioritized.

Drought indicators have been used for the evaluation of drought characteristics, and various indicators are in use for the evaluation of the various types of droughts. The SPI is one of the most widely used indicators for the assessment of drought characteristics (McKee et al. 1993; Heim 2000; Wilhite 2000; Cheval 2015). The SPI can be used to describe the characteristics of both short- and long-term droughts based on different time scales, and it has the benefits of statistical consistency also. The SPI is the best-suited index for drought risk analysis due to its intrinsic probabilistic nature (Guttman 1999). Due to its standardization, the SPI can be effectively used to compare drought conditions among regions with different climatic conditions and for time periods (Bonaccorso et al. 2003). Hayes et al. (1999) showed that the SPI can be used operationally to detect the start of the drought, its spatial extension, and temporal progression.

However, the three-parameter gamma distribution instead of the two-parameter gamma distribution is considered to produce more robust values of the SPI. If m is the number of zeros in a precipitation time series, states, then q can be estimated by m/N and tables of incomplete gamma function can be used to determine cumulative probability H(x). The cumulative probability is then transformed to the standard normal random variable Z with mean zero and variance one, which is considered to be the value of the SPI.

For 0.5<

For 0.5<

where c₀=2.515517, c₁=0.802853, and c₂=0.010328, and d₁=1.432788, d₂=0.189269, and d₃=0.001308.

The drought severity has been evaluated in the study area using the classification given by Hayes et al. (1999) as given in Table 2. A drought event occurs during the period when the SPI is continuously negative and reaches the intensity of −1.0 or less, and it ends when the SPI becomes positive. The SPI can be employed to compute the frequency, duration, and intensity of droughts. The summation of the SPI values for all the months within a drought event gives the magnitude of drought. The ratio of drought magnitude to its duration gives the drought intensity. A 3-month SPI (3 m SPI) has been used as a short-term or seasonal drought index and a 6-month SPI (12 m SPI) for intermediate-term drought index, whereas 12-month is a long-term drought index.

The SPI for a month/year in the period of record is dependent upon the timescale. The SPI can be used for all stations having more than 30 years of rainfall data (Hayes et al. 1999).

The contrasts and similarities in the rainfall pattern of the regions in the Chambal basin falling in MP and RJ were carried out using the statistical approach. The average annual rainfall in the Chambal basin is 812.67 mm which varies between 938.23 mm in the MP region of the basin and 687.11 mm in the RJ region. The average coefficient of variation of annual rainfall is 28% for the MP region and 30% for the RJ region of the basin. The coefficient of variation ranges between 24 and 38% in the MP region and between 24 and 35% in the RJ region, which indicates the very high rainfall variability for both regions. The RJ region receives much lower rainfall but with much higher variability as compared to the MP region, except a few districts of MP which include Shivpuri, Agar, and Rajgarh. The spatial variation of average annual rainfall in the study area is given in Figure 3(a), and the spatial variation of coefficient of variation is given in Figure 3(b).

Even though the variability in the annual rainfall in the MP region is high (28%), this region receives sufficient rainfall to meet the various demands of the competing users. The lack of water scarcity in this part of the basin is mainly due to a lack of sufficient water storage infrastructure and over-dependence on groundwater for meeting the crop water demands. Off-late floods have also been recorded here. However, the rainfall pattern and its distribution in the RJ region are quite different. The lower annual rainfall coupled with higher variability is the major cause of the water shortages. As a consequence of these, the occurrences of dry spells and droughts are higher in the RJ region in the Chambal basin.

The MK and Sen's slope estimator have been used for determining the trend. The investigation of trends has been carried out for seasonal rainfall (Jun–Sept) and annual rainfall. Similarly, the trend has also been evaluated for the 1-day maximum rainfall and the number of rainy days to detect any climate change signals in the rainfall pattern. The brief details of the results are presented in the following subsections.

70% of the seasonal rainfall occurs in the two principal monsoon months of July and August. A statistical summary of the rainfall pattern of the Chambal basin in the MP and RJ regions is given in Table 3. Over the longer term, significant trends in seasonal rainfall have not been generally observed in the MP region of the Chambal basin except at Dewas, which exhibits a significant increasing trend at a 0.1% level of significance. The negative trend was observed at Shajapur (−4.53), Vidisha (−0.97), Mandsaur (−1.75), and Guna (−1.28). The magnitude of trend in seasonal rainfall in the MP region varies between −4.53 mm/year in Shajapur and 7.49 mm/year in Dewas. However, for the RJ region, significant increasing trends in seasonal rainfall have been observed at Udaipur, Chittaurgarh, Rajsamad, and Bundi district at a 0.05% level of significance and also in Tonk and Dausa at a 0.1% level of significance, while a negative trend has been observed in Jhalawar (−0.41) and Kota (−0.22). The magnitude of trend in seasonal rainfall in the RJ region varies between −0.41 mm/year in Jhalawar and 9.21 mm/year in Udaipur. Similar results have been observed for the annual rainfall trends in the MP and RJ regions of the Chambal basin. The spatial variability in rainfall quantum, intensity, and duration are responsible for water scarcity and the drought-like situation in some parts of the basin, while other regions in the vicinity remain unaffected. The seasonal rainfall pattern in Shajapur in the MP region and Udaipur in the RJ region is given in Figure 4(a) and 4(b), respectively.

The IMD defines a rainy day as a day with a rainfall of 2.5 mm or more. If the rainfall on any day is greater than or equal to 2.5 mm, it has been considered as a rainy day and all the rainy days in each year have been accumulated to compute the number of rainy days. The 1-day maximum rainfall corresponds to the maximum rainfall that has occurred on a rainy day within a year. The time series of the annual 1-day maximum rainfall was extracted from the station rainfall data for comparisons. The 1-day maximum rainfall gives an idea about the extreme event magnitude in the basin over a longer term. The higher magnitude of 1-day maximum rainfall signifies the increase in the magnitude of flash floods and urban drainage issues, whereas a decrease in the 1-day maximum rainfall may result in lesser soil erosion and higher groundwater recharge prospects. The result of the MK test and Sen's slope estimated for the 1-day maximum rainfall and the number of rainy days has been summarized in Table 4. The decreasing trend in the 1-day maximum rainfall has been recorded at 7 out of 16 districts in MP and at 6 out of 15 districts in RJ. However, the magnitude of decreasing trend is not significant except at Shajapur in MP (the significant decreasing trend at a 0.1% level of significance), while Sheopur in MP and Udaipur and Bundi in RJ showed an increasing trend at a 0.1% level of significance. The magnitude of trend in the 1-day maximum rainfall varies between −1.26 mm/year in Shajapur and 0.81 mm/year in Shivpuri in the MP region, while it varies between −0.53 mm/year in Jaipur and 1.91 mm/year in Udaipur in the RJ region of the Chambal basin.

The decrease in the magnitude of 1-day rainfall may lead to reduced peak flows in the various tributaries of the Chambal river system. This may be a cause of concern to water resources managers in providing adequate quantum of water to satisfy the increasing water demands of a burgeoning population as a scenario of very high rainfall variability along with a reduction in the annual rainfall and 1-day maximum rainfall which may cause further stress in the already water-scarce region. The plot showing the 1-day maximum rainfall at Shajapur in MP and Jaipur in RJ is given in Figure 5(a) and 5(b), respectively.

The number of rainy days is another indicator that gives information on the temporal distribution of rainfall. On average, there are 48 rainy days in the MP region, while there are 39 rainy days in the RJ region of the Chambal basin. The annual number of rainy days varies between 37 days in Shivpuri and 62 days in Sheopur in the MP region, whereas it varies between 30 days in Ajmer and 48 days in Baran in the RJ region of the Chambal basin. A significant increase at a 0.05% level of significance has been recorded in the annual number of rainy days at Vidisha in the MP region, whereas, in the RJ region, most of the districts depicted an increase in the annual number of rainy days except at Jhalawar which had a decreasing trend. The magnitude of the positive trend in RJ was observed as Udaipur (+0.27) at a 0.1% level of significance, Chittaurgarh (+0.25) at a 0.05% level of significance, and Rajsamad (+0.33) at a 0.01% level of significance. Figure 6(a) and 6(b) shows the annual number of rainy days in the various districts of the MP and RJ regions, respectively.

The drought years have been identified based on the departure analysis of seasonal rainfall. A negative value of rainfall departure indicates drought condition, and a positive value indicates the wet condition. The drought years have been identified for various districts in the study area. The drought years in the Chambal basin falling in the MP region are given in Table 5. Dhar, Sehore, Bhopal, Neemuch, and Sheopur faced a minimum of 7 drought years, followed by Indore, Dewas, and Ratlam which had 8 drought years, whereas 10 drought years occurred at Ujjain, Agar, Rajgarh, Vidisha, and Guna. Shajapur faced a maximum of 12 drought years during the period of analysis followed by Shivpuri with 11 drought years. Generally, most of the districts in the MP region of the basin experienced 7–10 drought events, and the drought frequency varies between 1 in 3 years and 1 in 5 years. 1985, 1987, 1989, 1992, 2000, 2002, 2005, 2007, 2009, 2010, and 2017 have been the predominant drought years in which the major portion of the region in MP falling in the Chambal basin was under drought. The departure of the seasonal rainfall at Ujjain and Shajapur in MP is given in Figure 7(a) and 7(b), respectively.

The drought years in the Chambal basin falling in RJ are given in Table 6. Pratapgarh was under drought for 5 years and Chittaurgarh faced 6 drought years, whereas Rajsamad, Baran, Bundi, Tonk, and Jaipur faced 8 drought years. 9 drought years occurred at Jhalawar followed by 12 drought years at Udaipur, Bhilwara, and Ajmer during the period of analysis. Generally, most of the districts in the RJ region of the basin experienced 5–10 drought events, and the drought frequency in the RJ region varies between 1 in 3 years and 1 in 4 years. 1985, 1987, 1989, 1993, 1998, 1999, 2000, 2002, 2005, 2006, 2007, 2009, 2010, and 2017 have been the predominant drought years in which the major portion of the RJ region falling in the Chambal basin was under drought. The departure of the seasonal rainfall at Bhilwara and Ajmer districts in RJ is given in Figure 8(a) and 8(b), respectively. The maximum seasonal departure of −98.16% occurred at Shivpuri district in the MP region during 1997, whereas the maximum seasonal departure of −82.9% occurred at Tonk district in the RJ region in 2002. The Chambal basin falling in the RJ region faces frequent drought with higher severity as compared to the MP region.

The drought-prone districts are identified based on the probability analysis of seasonal rainfall. The probability analysis of seasonal rainfall indicates that 7 out of 16 districts in the MP region falling in the Chambal basin are prone to drought except for Indore, Dewas, Sehore, Bhopal, Agar, Mandsaur, Guna, Sheopur, and Neemuch and constitute about 44% of the basin area in MP. Figure 9(a) and 9(b) shows the probability of exceedance of seasonal rainfall in Ratlam district of MP and Udaipur district of RJ, respectively, in the Chambal basin. The probability distribution of seasonal rainfall in the Chambal basin is given in Table 7. In the RJ part of the basin, except for Pratapgarh, Chittaurgarh, Rajsamad, Kota, Bundi, Ajmer, and Jaipur all the remaining districts are drought-prone and constitute about 54% of the basin area in RJ.

The drought severity evaluated based on the seasonal rainfall departure has been classified into four classes, namely mild drought, moderate drought, severe drought, and extreme drought based on their severity ranges. The drought severity and drought frequency for the study area falling in MP and RJ regions is given in Tables 8 and 9, respectively.

The comparison of the drought frequencies between the regions falling in MP and RJ clearly indicates that the frequency of drought is higher in the RJ region and generally varies between 1 in 3 years and 1 in 4 years, whereas the drought frequency in the MP region varies between 1 in 3 years and 1 in 5 years. Most of the basin area in RJ has higher drought frequencies as compared to those in the MP region. This makes it amply clear that the scenario of drought is worse in RJ as compared to the MP region in the Chambal basin.

The RDI has been used to prioritize the drought-prone areas so that the drought relief and mitigation efforts can be focussed in these districts which are at higher priority. The ranking system is designed such that higher weights have been assigned for droughts of higher severities. A weight of 1 is assigned for mild drought, 2 for moderate drought, 3 for severe drought, and 4 for extreme drought, which is subsequently summed up for the complete time series and normalized. The RDI for the various districts falling in MP and RJ is given in Table 10.

The RDI varies between 0.38 at Bhopal and 0.91 at Shivpuri in the MP region of the Chambal basin, whereas it varies between 0.47 in Pratapgarh and 0.91 at Udaipur in the RJ region of the basin. In the RJ region, Udaipur has the highest priority (0.91) followed by Sawai Madhopur, Tonk, Dausa, Rajsamad, Kota, Bhilwara, and Karauli district where the RDI is more than 0.70. Moreover, 10 out of 15 districts have very high RDI values. In the MP region, there are only 3 out of 16 districts with an RDI greater than 0.7. Shivpuri has the highest priority followed by Shajapur, Ratlam, and Ujjain. The comparison of the RDI in the MP and RJ regions of the Chambal basin indicates that the districts falling in the RJ region are more prone to drought as compared to the districts in MP.

The meteorological drought characteristics have been evaluated based on the SPI. The SPI can also be used to assess the various types of drought based on multiple-scale SPIs. However, 3 m-SPI has been used for the evaluation of meteorological drought characteristics in the Chambal basin. The temporal variation of the 3 m-SPI at Shivpuri district in MP is given in Figure 10. The 3 m-SPI values less than −1.0 are considered as a drought event. It can be observed that 1986, 1987, 1991, 1992, 1997, 2002, 2004, 2006, and 2013 have been the drought years with varying degrees of severity. The extreme droughts (3 m-SPI< −2.0) have been observed in 1997 and 2002, whereas severe soil moisture droughts (−2.0<3 m-SPI<−1.5) have been observed in 1986, 1987, 1992, 2004, 2006, and 2013. The number of drought events, drought severity, and drought frequency evaluated from the time series of 3 m-SPI is given in Table 11.

In the MP region of the Chambal basin, a maximum of 9 extreme drought events occurred in the Mandsaur district followed by 8 drought events in Agar, Dewas, Dhar, and Sheopur districts. The number of extreme drought events varied from 3 events at Guna and Shajapur to 9 events at Mandsaur. Similarly, the number of severe drought events varied between 2 events at Shivpuri and 14 events at Ratlam followed by 12 and 11 events at Indore, Guna, Rajgarh, Sheopur, and Ujjain district, respectively. The number of moderate drought events varied between 14 events at Shivpuri and 44 events at Sheopur followed by 41 events at Guna and 31 at Shajapur. Twenty-four moderate drought events occurred at Indore followed by 23 events at Rajgarh, and 22 events at Sehore, Ujjain, and Vidisha. The maximum drought duration of 63 months was observed at Sheopur followed by 55 months at Guna. The maximum drought severity varies between −42.29 at Bhopal and −88.81 at Sheopur followed by −75.12 at Guna and −62.68 at Neemuch. The drought intensity, which is a ratio of the drought severity and the number of drought events, gives an overall indication of the drought scenario at various districts. In the MP region, the drought intensity is maximum at Shivpuri (−1.87) followed by Agar (−1.6), Dhar (−1.59), and Bhopal (−1.57).

Similar assessments for the evaluation of drought characteristics using the SPI have been carried out for the various districts in the RJ region in the Chambal basin. The temporal variation of 3 m-SPI at Sawai Madhopur is given in Figure 11. It can be observed that 1985, 1987, 1989, 1991, 2000, 2002, 2005, 2006, 2007, 2009, 2015, and 2017 have been drought years with varying degrees of severity. The extreme droughts (3 m-SPI<−2.0) occurred in 1989, 1991, and 2002, whereas severe droughts (−2.0< 3 m-SPI<−1.5) have been observed during 1987, 1988, 2000, 2003, 2006, 2010, 2015, and 2017. The number of drought events, combined drought severity, and the drought intensity have been evaluated from the time series of 3 m-SPI and are given in Table 12.

A maximum of 10 extreme drought events occurred in Jaipur and Tonk followed by 9 events at Bundi, Jhalawar, and Pratapgarh. The number of extreme drought events varied between 2 events at Rajsamad and 8 events at Chittaurgarh followed by 7 events in Dausa, Karauli, and Udaipur. Similarly, the number of severe drought events varied between 6 events at Ajmer, Baran, and Jhalawar and 16 events at Rajsamad followed by 14 events at Udaipur, 12 events at Tonk, and 9 events at Bhilwara, Chittaurgarh, Dausa, Jaipur, and Sawai Madhopur. The number of moderate drought events varied between 16 events at Ajmer and 34 events at Jaipur followed by 33 events at Karauli and Sawai Madhopur, and 30 events at Kota. The drought duration was maximum at Jaipur (52 months) followed by 47 months at Karauli and Sawai Madhopur, and 45 months at Udaipur. The drought severity varied between −42.61 at Ajmer and −79.95 at Jaipur followed by −69.41 at Udaipur. The drought intensity was maximum at Ajmer (−1.64) followed by Chittaurgarh and Tonk (−1.63).

The 3 m-SPI-based evaluation of the meteorological drought characteristics indicates that the Chambal basin faces regular droughts based on the analysis of the long-term data of 34 years spanning from 1985 to 2018. Droughts are experienced both in the RJ region and the MP region. An attempt has been made to understand the contrasts by a comparison of drought characteristics for the basin falling in MP and RJ. Even though both regions face drought, the drought severity and the drought duration are comparatively higher in the area falling in RJ. Five districts in RJ have a total drought severity of more than −65.0, namely Jaipur (−79.95), Karauli (−68.99), Sawai Madhopur (−66.46), Tonk (−68.25), and Udaipur (−69.41). However, the scenario is quite different in the basin area falling in MP, wherein the total drought severity of more than −65.0 was observed at only Sheopur (−88.81), and Guna (–75.12) which are located close to border RJ. Indore (−61.8), Neemuch (−62.68), Ratlam (−60.46), and Ujjain (−61.01) had a total drought severity greater than −60.00. Most of the districts fell under the severity range of −40.0 to −60.00. Therefore, higher drought severities are observed in the RJ region as compared to the MP region of the Chambal basin.

The drought intensity varied between −1.37 and −1.64 in the RJ region and between −1.37 and −1.87 in the MP region of the Chambal basin. As such, there is not much difference in the drought intensity in MP and RJ regions. Also, drought frequency generally varies between 1 in 3 years and 1 in 4 years in the RJ region, whereas the drought frequency varies between 1 in 3 years and 1 in 5 years in the MP region of the basin. Therefore, the basin area falling in RJ has a higher frequency of drought occurrence, with the repetition of drought every 3–4 years. The analysis shows that out of the 15 districts falling in the RJ region, 8 districts are drought-prone, i.e., 54% area of the Chambal basin falling in the RJ region is drought-prone, whereas out of the 16 districts falling in the MP region, 7 districts are prone to drought, i.e., 44% area of the Chambal basin falling in MP. Therefore, it can be seen that the basin area falling in the RJ region is more vulnerable to drought as compared to the basin area falling in MP.

The Chambal basin falling in MP and RJ is one of the highly water-stressed regions in India. The regular occurrences of drought and degradation of the natural resources have led to desertification in many areas of RJ and are also believed to be fast encroaching into MP. The comprehensive investigations of the changes in the rainfall pattern and its distribution along with the evaluation of drought characteristics helped to have an insight into the two contrasting basin areas falling in MP and RJ. Owing to the higher annual rainfall in the MP region (938.23 mm) as compared to the RJ region (687.11 mm), the water availability scenario is favorable in the MP region as compared to the RJ region. However, the very high variability of rainfall (28–30%) in both MP and RJ has led to regular water scarcity and droughts. A decreasing trend of 1-day maximum rainfall has been recorded in 7 out of 16 districts in MP and 6 out of 15 districts in RJ.

1985, 1987, 1989, 1992, 1993, 1998, 1991, 2000, 2002, 2005, 2007, 2009, 2010, and 2017 were some of the predominant drought years spread over the complete basin area. The drought frequency varies between 1 in 3 years and 1 in 4 years for the basin area falling in RJ, whereas it varies between 1 in 3 years and 1 in 5 years in the MP region. The prioritization-based evaluation of districts in both regions emphasizes more concern with respect to the district that is located in the RJ region of the basin as compared to MP, and therefore, these districts should be taken up for drought mitigation activities on a priority basis. The drought severity was much higher in the RJ region as compared to the MP region of the basin. It can be concluded that even though drought occurs in both MP and RJ regions, the drought scenario is more severe in the RJ region owing to lower seasonal rainfall and higher rainfall variability. Moreover, the soil and land-use/land-cover characteristics are also not favorable from the point of view of soil moisture retention leading to desertification. The adverse effect of drought is felt more in the basin areas falling in RJ as compared to MP. The present work is based on the evaluation of meteorological drought only based on a single drought index SPI which is based on precipitation only. The study can be taken further by the application of more drought indicators for the evaluation of soil moisture, hydrological, and agricultural drought characteristics.

Climate change-related impacts add another dimension to the challenges due to drought in the basin. The increase in the average maximum and minimum temperatures may lead to much higher occurrences of heat waves which may not augur well for the crops being grown in an already water-stressed region with limiting soil conditions. Future research should include the investigation of the changes in the climate and the drought scenario that may be projected for future time horizons using global circulation model outputs. A real-time drought monitoring system and a decision support system for drought may prove to be an effective adaptation tool for the mitigation of droughts in the near term and into the future.

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