The southwest (June to September) and northeast (October to December) monsoons are crucial for India’s economy. More than 50 per cent of the net sown area is rainfed, which accounts for approximately 40 per cent of the total production (Government of India 2023). Nearly 55 per cent of India’s population is engaged in agricultural and allied sectors, and well-distributed rainfall during the kharif and rabi seasons is regarded as the backbone of the sector (Government of India 2023). Beyond its impact on agriculture, these monsoons play a pivotal role in water resource management and realising India’s clean energy aspirations, particularly through the production of hydro-generated energy, which is important for it to achieve its ambitious net-zero targets.

However, monsoonal variability triggers contrasting wet and dry conditions, leading to hydro-meteorological disasters. As per an analysis of climatic disasters between 1971 and 2020 by the Council on Energy, Environment and Water (CEEW), around 75 per cent of India’s districts are prone to severe hydro-meteorological disasters such as floods, droughts, and cyclones. Moreover, nearly 40 per cent of these districts show a swapping pattern, alternating between drought and floods, which leads to compounding risk. Such swapping trends can be primarily attributed to evolving rainfall patterns (Mohanty 2020). Consequently, an estimated 80 per cent of the population now resides in regions bearing the brunt of climate-induced vulnerabilities (Mohanty and Wadhawan 2021).

Current climate change trends suggest that global warming records could be breached sooner than anticipated, leading to heightened climate variability and more frequent and intense extreme weather events, which will impact an even larger population (WMO 2023). This has a direct bearing on India’s monsoons, as evidenced in 2023, when unpredictable rainfall patterns triggered severe floods in some cities due to unprecedented heavy downpours, while core monsoon states faced notable rainfall deficits. For instance, Chandigarh received nearly half of its total annual rainfall in just 50 hours, while Maharashtra and Kerala encountered deficits of around 46 per cent and 60 per cent in June, respectively (IMD 2023).

In light of this rapidly evolving climate risk landscape, the development of effective climate adaptation policies at the sub-national level hinges on access to an exhaustive spectrum of climate information, spanning across diverse and granular spatial and temporal scales. In this issue brief, we try to decode changing patterns with respect to one of the most critical climatic variables for India – rainfall. We undertake this assessment at the tehsil level to quantify and understand trends in changes in rainfall during the southwest monsoon (between June and September) and the northeast monsoon (between October and December). We analyse the trends over the past 40 years (1982–2022), with a particular focus on the quantification of changes over the past decade, between 2012 and 2022. Our primary aim with this assessment is to offer fundamental insights into shifting rainfall patterns in recent years. These insights are intended to serve as valuable resources for policymakers and stakeholders across various sectors, including water management, agriculture, disaster preparedness, energy, and urban planning, among others, to help them make climateinformed decisions.

1.1 Monsoon variability and climate change​

The monsoons inherently exhibit substantial spatial and temporal variability (Guhathakurta et al. 2020). The significant impact of this variability across sectors has made the study of its dynamics crucial in the context of climate change and the underlying atmospheric and oceanic processes. This has led to comprehensive research efforts spanning the climate science and policy domains (Gadgil 2003; Pai et al. 2014; Rajeevan, Bhate, and Jaswal 2008; Roxy et al. 2017; Guhathakurta and Rajeevan 2008; Goswami et al. 2006; Kumar et al. 2006). In the past two decades, researchers have leveraged both observational trend analyses and machinelearning-driven atmosphere–ocean coupled general circulation models to dissect climate fluctuations and assess potential long-term adverse effects. This analytical approach has provided valuable insights into the intricate dynamics of the monsoon system and its implications.

In its report titled “Assessment of Climate Change over the Indian Region”, the MoES delved into the long-term historical and projected changes in both the southwest and northeast monsoons across India (Krishnan et al. 2020). The study examined climatic trends between 1901 and 2015 and the projections under RCP4.5 and RCP8.5 till the end of the twenty-first century. Based on past climatic trends, the study reported that there was no discernible long-term trend in southwest monsoon rainfall across India as a whole. However, it highlighted pronounced spatial variability. Significant declining trends in rainfall were identified over Kerala, the Western Ghats, parts of central India, and select northeastern states. Conversely, regions such as Gujarat, the Konkan coast, Goa, Jammu and Kashmir, and the east coast experienced a significant increase in rainfall. Intriguingly, the study also found that an increase in the rate of heavy rain events offset a decrease in the rate of moderate-intensity rainfall events – that is, in regions such as central India, it was observed that the frequency of heavy rainfall events (rainfall more than 100 mm/day) increased as the frequency of moderate-intensity rainfall events decreased (rainfall between 5 and 100 mm/day). Moreover, the report underscored the fact that climate change influences not only the southwest monsoon, but also the northeast monsoon, as it was observed that the variability in northeast monsoon rainfall had escalated during the period 1959 to 2016.

Analysing climate change projections using the CMIP5 multi-model ensemble, CORDEX-SA, and NASA Earth Exchange Global Daily Downscaled Projections (NEXGDDP) simulations at a high-resolution 50-km grid, the assessment indicated a wetter future for the southwest monsoon in India. It predicts an increase of about 6 per cent under RCP4.5 and 8 per cent under RCP8.5 up to 2050 (near future) over India’s core monsoon region, while by the end of the twenty-first century (far future), the projected increase in rainfall is expected to be approximately 10 per cent and 14 per cent as per RCP4.5 and RCP8.5, respectively. With regard to the northeast monsoon, the simulations indicated a wetter northeast monsoon in the range of 10 per cent to 50 per cent under both RCP4.5 and RCP8.5 scenarios in the near and far future across the country.

Building upon the MoES report, the IMD employed 25- km resolution grid data to assess southwest monsoon rainfall across 29 states and union territories. It concluded that during the period 1989 to 2018, while no changes were found across the majority of the country, noteworthy declines in southwest monsoon rainfall were observed in the states of the Indo-Gangetic plains and the northeastern states (IMD 2020).

In general, a comprehensive analysis of the existing literature underscores the prevailing consensus that the majority of India is poised for intensified monsoon seasons characterised by increased rainfall. However, capturing short-term and spatial variabilities at a granular scale has been a challenge. The prevailing modelled assessments predominantly concentrate on prolonged long-term trends at a coarse resolution, thereby limiting their precision in encapsulating immediate effects and within-season variations. The studies have not been able to account for intricate within-season nuances, spanning across months and days, and even the variations within a district. This gap also extends to India’s current definition of a ‘normal’ monsoon, which is calculated as the nationwide average of total rainfall received throughout all monsoon months and does not incorporate spatial variabilities or wet and dry extremes.

Furthermore, the inherent complexities of the Indian monsoon systems introduce a further layer of uncertainty, particularly when contrasted with more predictable variables such as temperature. Studies highlight a lack of proficiency in accurately simulating the monsoons, which in turn amplifies uncertainties in comprehending future shifts (Chaturvedi et al. 2012; Saha et al. 2014; Sharmila et al. 2015; Krishnan et al. 2016). The wide inter-model spread in simulated precipitation changes over South Asia adds to the ambiguity in assessing regional hydroclimatic response (Kripalani et al. 2007; Annamalai, Hamilton, and Sperber 2007; Turner and Slingo 2009; Sabade, Kulkarni, and Kripalani 2011; Fan et al. 2010; Hasson, Lucarini, and Pascale 2013; Saha et al. 2014).

Therefore, there is an urgent need to further quantify these changes in monsoon rainfall over the recent decade at a finer spatial resolution. Consequently, in this issue brief, we embark on a comprehensive evaluation of these changes, which have been overlooked in prior assessments. This includes an examination of alterations in monthly patterns, a region-specific threshold-based assessment concerning wet extremes, and a meticulous tehsil-level analysis. Our goal is to present this information comprehensively and cohesively. To this end, we selected the following objectives for our study.

1.2 Objectives

We undertook a tehsil-level rainfall variability assessment using IMDAA’s latest fine-resolution reanalysis data available at a 12-km spatial grid. This marks, to the best of our knowledge, the first tehsil-level assessment of rainfall variability in India. Our main goal is to provide fundamental insights into monsoon variability and its extremes over the past decade, between 2012 and 2022. By focusing on key monsoon seasons, we aim to assist policymakers and stakeholders from different sectors in making climate-informed decisions. We analysed inter-annual changes (year to year) and intra-annual variations (within months and extreme wet and dry anomalies) using well-established statistical criteria and indices recommended by the WMO and IMD. The main objectives of our study are as follows:

• analysing shifts in India’s rainfall patterns during the southwest and northeast monsoon seasons;

• investigating changes in rainfall patterns across India at the tehsil scale during these monsoon seasons to quantify year-to-year variability and monthly anomalies; and

• Identify dry and extreme rainfall hotspots in the tehsils, including heavy rainfall events and dry days, and highlight tehsils with high percentages of total rainfall occurring within short periods.

The risks associated with climate change exhibit nonlinear characteristics, wherein the impacts can rapidly escalate despite the shifts in average conditions being gradual (King et al. 2015). Altered rainfall patterns give rise to both chronic and acute risks for ecosystems. Chronic risks emerge through gradual shifts such as reduced mean rainfall, while acute risks stem from sudden, intense events such as extreme rainfall and prolonged dry spells. These risks are interconnected, and their mitigation requires coordinated efforts across sectors and stakeholders (Pachauri et al. 2014). Notably, abrupt changes in precipitation during the Indian monsoon on decadal and centennial scales are evident from high-resolution climate proxy records spanning thousands of years (Berkelhammer et al. 2012; Sanyal and Sinha 2010).

In India, there are traditionally four primary seasons (refer to Box 1). Here, we focus on two pivotal seasons for agricultural and water resource management activities – the summer or southwest monsoon season (referred to as JJAS, since it occurs between June and September) and the post-monsoon or northeast monsoon season (referred to as OND, since it occurs between October and December). By utilising indices that capture both chronic and acute risks linked to rainfall variability, our analysis aimed to provide comprehensive information on shifts in the rainfall patterns at the tehsil level in India.

3.1 Changing rainfall patterns during the southwest monsoon

As depicted in Figure 5, a trend analysis of the all-India JJAS demonstrates an overall intensification when averaged across the country, exhibiting an increasing trend at an approximate rate of 3 mm/year (as indicated by Sen’s slope). Conversely, our investigation at the district and tehsil level uncovered a landscape characterised by substantial spatial variability within this trend.

To dig deeper into this pattern and uncover districts and tehsils that have shown significant year-to-year variations over the past four decades, we followed a two-step approach. First, we categorised each year based on the IMD criteria, dividing them into deficient, normal, or excessive rainfall years at the district and tehsil level. First, we categorised each year based on the IMD criteria, dividing them into deficient, normal, or excessive rainfall years at the tehsil level. At the same time, we also conducted a country-level study to determine the rainfall received by India as a whole and categorised it as normal, below-normal, or abovenormal rainfall.

Figure 5 All-India averaged JJAS rainfall shows a slightly increasing trend at the rate of 3 mm/year (not statistically significant)

Source: Authors’ analysis

It’s important to note that the IMD uses different criteria for rainfall categorisation at the regional level and the national level. For regional categorisation, we used specific thresholds outlined in Table 2. However, at the national level, rainfall is considered ‘normal’ when it falls within 10 per cent of its long-period average (LPA) or ranges from 90 per cent to 110 per cent of the LPA. Rainfall is deemed ‘below normal’ if it’s less than 90 per cent of the LPA and ‘above normal’ if it’s more than 110 per cent of the LPA (IMD 2019).

Our analysis reveals that nearly 30 and 38 per cent of India’s districts witnessed a high number of deficient and excessive rainfall years (more than 10 years of 40), respectively in the past 40 years. Moreover, 23 per cent of these witnessed both a high number of deficient as well as excessive rainfall years such as New Delhi, Bengaluru, Nilgiris, Jaipur, Kachchh, and Indore. Further, at the tehsil level, nearly 24 per cent of the tehsils in India have experienced a high number of both excessive and deficient rainfall years.5 Intriguingly, during this period, from 1982 to 2022, based on current criteria, we identified 29 years classified as ‘normal’ monsoon years, alongside 9 years classified as ‘abovenormal’ and 3 years as ‘below-normal’ at the country level. The tehsils demonstrating such pronounced year-to-year variability are primarily concentrated in the Marathwada region of Maharashtra, along with tehsils of Rajasthan and Gujarat.

While these trends highlight a concerning pattern of erratic monsoon behaviour from one year to another, they emphasise that looking at monsoons at a coarse resolution such as a national scale doesn’t adequately account for the substantial spatial variabilities observed across the country.

We analysed changes in the past decade at the tehsil level, which revealed that there is an overall increasing trend for JJAS rainfall; 55 per cent of tehsils showed an increase, while 11 per cent showed a decrease, compared to the baseline of 1982–2011 by more than 10 per cent, as seen in Figure ES1. The latter is of particular concern as a significant number of these tehsils are concentrated in the Indo-Gangetic plains and Himalayan states, where agricultural activities in the kharif season rely heavily on JJAS rainfall and are also highly vulnerable to hydro-meteorological disasters such as floods and droughts.

Figure 6 24% of tehsils have witnessed both a high number of deficient as well as excessive rainfall years in last four decades.

Source: Authors’ analysis

Notably, the degree of positive change surpassed that of negative change in most regions experiencing changing trends. For instance, the most substantial decrease was observed in the tehsils of Assam and Meghalaya, which have traditionally been high monsoon rainfall areas, with reductions exceeding 30 per cent (greater than 1200 mm/year compared to their LPA) in tehsils such as Pachim Nalbari Circle, Boitamari Circle, and Barnagar Circle of Assam among others. Conversely, the most notable increase was seen in the tehsils of traditionally drier areas, such as Rajasthan, Gujarat, the Konkan region, central Maharashtra, and parts of Tamil Nadu, where the increase in JJAS rainfall exceeded 30 per cent when compared to the baseline of 1981–2011.

Although the analysis of changes in JJAS rainfall as a whole offers a broad overview of variations, a closer examination on a monthly basis shows that there is significant month-to-month variation across several tehsils, as seen in Figure 7. Noteworthy among these changing patterns is the continuous decline in rainfall in tehsils situated within the Indo-Gangetic plain states, alongside substantial month-to-month fluctuations in eastern Maharashtra.

Source: Authors’ compilation

Figure 7 Out of the total number of tehsils that witnessed an overall decrease in JJAS rainfall, 68% of tehsils consistently experienced reduced rainfall in all JJAS months, while 87% showed a decline during the June and July.

Source: Authors’ analysis

Notably, the months of June and July are important for agricultural activities in India, because this is the period of sowing and establishment of crops. Our analysis showed that of the tehsils witnessing an overall decrease in JJAS rainfall, approximately 68 per cent of tehsils experienced reduced rainfall across all four months, while around 87 per cent experienced decreasing rainfall specifically in June and July, which is a critical phase of the monsoon for the timely sowing of Kharif crops. Moreover, in June, when the monsoon sets in over Kerala and advances to the western coast of India, some parts of the state witnessed a decrease in rainfall.

Our analysis revealed that nearly 55 per cent of the tehsils in the country are witnessing an increase in JJAS rainfall. However, it’s crucial to assess its seasonal distribution. In our previous report on extreme hydrometeorological disasters in India, we found a nearly three-fold increase in flood frequency nationwide, impacting nearly 34 more districts since 2005 (Mohanty 2020). Therefore, we investigated whether regions with heightened JJAS rainfall also experience more frequent and intense heavy rainfall events, using a daily variability analysis (i.e., heavy rainfall days, moderate rainfall days, and dry days). Furthermore, we evaluated the contribution of these events to seasonal rainfall.

Source: Authors’ compilation

Figure 8 64% of Indian tehsils saw an increase in the frequency of heavy rainfall days during JJAS

Source: Authors’ analysis

We categorised wet extremes based on tehsil-specific thresholds corresponding to the 90th and 99th percentile of the baseline. For simplification purposes, we refer to them as heavy and very heavy rainfall days. We found that nearly 64 per cent of Indian tehsils experienced an increase in the frequency of heavy rainfall days by 1-15 days per year in the past decade. In addition, a few tehsils in Tamil Nadu, Gujarat and western coast of India also saw an increase in the frequency of very heavy rainfall days during JJAS by 1-5 days per year. While the majority of the country did not see any significant change in very heavy rainfall day frequency, these events, exceeding the 99th percentile threshold, typically occur rarely – about one or two days per monsoon season.

Significant increases in wet extremes have been particularly evident in the tehsils of Maharashtra, Gujarat, Rajasthan, western Madhya Pradesh, and peninsular India. Interestingly, we also noted that tehsils in the top four states with the highest gross state domestic product – namely, Maharashtra, Tamil Nadu, Gujarat, and Karnataka – experienced an increase in the frequency of heavy rainfall days (Forbes India 2023).

Although we observed an increase in the frequency of wet rainfall extremes, our analysis does not reveal significant trends in the intensity of these occurrences. However, a statistically significant finding does emerge – the proportion of total seasonal rainfall attributed to wet extremes is on the rise in India during JJAS. This trend is of particular concern as it is related to the distribution of rainfall within a season. This could be one of the reasons behind the recent occurrences of flash floods, such as in Delhi, Uttarakhand, Himachal Pradesh (in 2023), and Bangalore (in 2022).

Source: Authors’ compilation

Figure 9 India is getting a higher proportion of its JJAS rainfall from short-lived wet extremes.

Source: Authors’ analysis

Source: Authors’ compilation

Regarding changes in the occurrence of dry days and moderate rainfall days, our analysis indicates a similar pattern to changing overall monsoon patterns. For example, the majority of the tehsils in the country which saw an increase in JJAS rainfall also saw a decrease in frequency of dry days and an increase in frequency of moderate rainfall days. However, an important observation emerges: the frequency of dry days is notably on the rise in specific tehsils along the Indo-Gangetic plains and the northeast. Interestingly, we identified contrasting trends of more dry days coupled with fewer moderate rainfall days in the affected areas. This distinct pattern is discernible in approximately nine per cent of all tehsils across India.​

Figure 10 Increase in the number of dry days is offset by the decrease in moderate rainfall days in 9% of tehsils during JJAS rainfall.

3.2 Changing rainfall patterns during the northeast monsoon​

While the southwest monsoon greatly impacts India’s economy, delivering the majority of its annual rainfall from June to September, the northeast monsoon from October to November holds significance for the cultivation of rabi crops across the country. This monsoon is particularly crucial for peninsular India and states such as Andhra Pradesh and Tamil Nadu, which receive up to 60 per cent of their annual rainfall during this period. The northeast monsoon significantly affects the cultivation of rice and maize in Tamil Nadu and Andhra Pradesh (Krishna Kumar et al. 2004). The rainfall during the northeast monsoon exhibits high variability both spatially and temporally. The coefficient of variation (inter-annual) for the northeast monsoon rainfall averaged over five subdivisions is 25 per cent, exceeding that of the southwest monsoon rainfall averaged over the entire country (10 per cent). The peninsula in the south is renowned for its substantial year-to-year variability (Rajeevan et al. 2022).

As seen in Figure 11, we observe a similar phenomenon with the northeast monsoon as we did with the southwest monsoon. Over the past four decades, there has been a slight increasing trend in the rainfall during the northeast monsoon across the country. This trend reflects a gradual increase at a rate of 0.7 mm/day, according to Sen’s slope estimator, though it’s worth noting that this trend is not significant.

Figure 11 The northeast monsoon over India shows a slightly increasing trend, at the rate of 0.7 mm/year (not statistically significant)

Source: Authors’ analysis

Over the past decade, there have been significant changes in the northeast monsoon patterns across the regions it impacts. As depicted in Figure ES1, regarding alterations in OND rainfall within the core states it affects, our analysis revealed that the OND rainfall increased in nearly 80 per cent of Tamil Nadu’s tehsils, 44 per cent of Telangana’s tehsils, and 39 per cent of Andhra Pradesh’s tehsils in the past decade compared to the climate baseline by more than 10 per cent.

Furthermore, even though the northeast monsoon mainly affects peninsular India and the OND rainfall contribution constitutes only a small portion of the annual rainfall in the Indian mainland, we observed changing patterns in these tehsils as well. For example, the tehsils showing an increase in OND rainfall in the past decade compared to the climate baseline are in Odisha, south West Bengal, and the Konkan region. Conversely, a decrease is observed in tehsils of Uttarakhand, Himachal Pradesh, Haryana, and northeastern states such as Meghalaya and Assam.

The OND season, also known as the retreating monsoon season or post-monsoon season, witnesses a shift in the rainfall zone to the southern regions of India, Sri Lanka, and the adjacent sea. As the southwest monsoon recedes from the northern parts of India, the mean sea level pressure and upper tropospheric wind circulation patterns undergo a swift transition from the summer monsoon type to the winter type. By October, the ITCZ or monsoon convergence zone, previously situated over northern India, starts shifting southwards (Rajeevan et al. 2022).​

Source: Authors’ compilation

Figure 12 48% of tehsils in India saw an increase in October rainfall in the past decade.

Source: Authors’ analysis

Our analysis of monthly OND rainfall changes reveals a remarkably variable trend characterised by significant month-to-month variations. Notably, our findings indicate that around 48 per cent of the tehsils have been witnessing an increase in October rainfall during the past decade by more than 10 per cent, which could be due to the delayed withdrawal of the southwest monsoon over the mainland, as evident in 2023 (IMD 2023).

While examining wet and dry rainfall extremes, we observed minor shifts occurring in India’s coastlines, as the frequency of heavy rainfall days was found to be increasing in the range of 1-5 days per day in the past decade. However, the majority of the country experiences a lower intensity of heavy rainfall events during this period compared to the southwest monsoon period and hence, the extent of the increase is relatively modest in comparison and could be attributed to the post monsoon cyclonic activity in the region. Our analysis revealed no notable or substantial changes in the frequency and intensity of very heavy rainfall events, dry days, or moderate rainfall days.