Long-term rainfall variability of Indian river basins in the context of global warming and climatic indices

Asymmetric global warming has emerged as a defining challenge that fundamentally alters the hydrological cycle, disturbing the spatio-temporal distribution of rainfall and hence threatening water resources worldwide. This disruption has cascading effects on agricultural productivity, water availability, and disaster risk, especially in monsoon-dominated regions such as India. The Intergovernmental Panel on Climate Change (IPCC 2021) projects a significant increase in South Asian monsoon variability and a 7% intensification of extreme rainfall with each 1 °C rise in global temperature. These changes heighten the risks of floods and droughts, intensifying the challenges of water resource management in a warming climate, particularly in regions like India where monsoons play a crucial role in agriculture, hydrology, and overall socio-economic stability. Several studies have highlighted a broad spectrum of impacts driven by global warming, with regional variations in rainfall trends (Arnell 1999; Milly et al. 2005; Lau & Wu 2007; Bates et al. 2008, and many more). For instance, significant decreases in annual and seasonal rainfall have been reported in parts of central and north India (Duhan & Pandey 2013; Pingale et al. 2014; Mathew et al. 2021). Praveen et al. (2020) found a significant decreasing trend in seven subdivisions between 1901 and 2015, and Sahoo & Yadav (2022) documented that in the north Indian homogeneous zone. Some regions in the south and west have experienced upward trends in precipitation during specific seasons. Nageswararao et al. (2019) discovered an increase in northeast monsoon rainfall over the southern peninsula between 1959 and 2016, while Chowdari et al. (2023) documented that in post-monsoon rainfall in 11 districts of Karnataka. Earlier, Singh & Ranade (2010) discussed the effects of climatic changes on the wet and dry spell characteristics, suggesting a potential shift in monsoon dynamics.

Studies on river basins reveal a complex and varied impact of changing rainfall patterns over time, reflecting the nuanced effects of climate change and regional variability. Ranade et al. (2008) observed no significant long-term trends but highlighted a diminishing trend in wet-season rainfall across some major basins in Central India. In a broader analysis, Kumar & Jain (2011) reported a decreasing trend in rainfall across 15 river basins over 135 years. Contrastingly, Jain et al. (2017) found an increasing trend in the number of rainy days in the Cauvery, Brahmani, and Baitarani basins between 1951 and 2012. Deka et al. (2013) documented a significant decline in monsoon and post-monsoon rainfall in northeast India's Brahmaputra and Barak basins between 1901 and 2010, while Taxak et al. (2014) identified an 8.4% overall decline in annual rainfall within the Wainganga basin from 1901 to 2012. Additional studies reveal decreasing trends in the Satluj basin during 1984–2010 (Kumar et al. 2015), the Ganga basin during 1901–2000 (Gajbhiye et al. 2016), and the Kosi basin during 1901–2000 (Srivastava et al. 2021). However, peninsular rivers present a contrasting picture, showing increasing rainfall variability (Varikoden et al. 2020). This heterogeneous and often contradictory nature of rainfall trends underscores the critical need for long-term, multiscale studies that examine regional and basin-specific dynamics to better understand and manage the evolving challenges posed by climate variability and its impact on water resources.

Spectral analysis has played a significant role in deciphering Indian summer monsoon rainfall (ISMR) variability, revealing dominant oscillations across various time scales. Early work by Murakami (1977) identified prominent 5- and 15-day peaks in the ISMR spectrum, laying a foundation for subsequent studies. Rangarajan (1994) further highlighted the dominance of the first two harmonics in explaining homogeneous ISMR variability. Expanding on these findings, Vijayakumar & Kulkarni (1995) revealed significant periodic variations in monsoon rainfall across 29 subdivisions, with cycles spanning 2–23 years and 3–3.9 years. Using advanced multichannel singular spectrum analysis (SSA), Krishnamurthy & Shukla (2007) identified 45- and 20-day oscillations, along with three seasonally persistent components, emphasizing the multi-scale nature of ISMR variability. Joshi & Pandey (2011) delved deeper, uncovering rainfall cycles lasting 10–20, 20–30, 30–40, and 40–50 days across four Indian subregions. These periodicities are intricately tied to interannual variability, influenced by changes in mean atmospheric circulation and modulated by the coupled ocean-atmosphere system, including factors such as sea surface temperatures, snow cover, and unpredictable intra-seasonal dynamics (Krishnamurthy & Shukla 2000).

Global climatic indices such as the El Niño Southern Oscillation (ENSO), Indian Ocean Dipole (IOD), and Pacific Decadal Oscillation (PDO) play a pivotal role in significantly modulating rainfall variability. ENSO, for example, generally exhibits an inverse relationship with ISMR although this link has weakened in recent decades due to the interaction with other factors such as sea surface temperature (SST) anomalies in the Indian Ocean, the IOD, and the PDO (Krishnamurthy & Krishnamurthy 2014; Kucharski & Abid 2017; Kurths et al. 2019). Notably, the warm phase of the PDO correlates with reduced monsoon rainfall, with dry conditions intensifying when coupled with El Niño events, while cold PDO phases and La Niña years are linked to wetter monsoons (Krishnan & Sugi 2003).

Beyond the Pacific, the Atlantic Multidecadal Oscillation (AMO) has emerged as another key player, with its warm phase associated with enhanced monsoon activity over India (Luo et al. 2018). Similarly, the Antarctic Oscillation (AAO) positively impacts Indian rainfall by intensifying cross-equatorial flows during its positive phase (Huang et al. 2021). The Arctic Oscillation (AO) also influences the North Indian monsoon, modulating the Rossby wave structure and interacting with Eurasian snow cover and land-sea thermal gradients (Coumou et al. 2018). Meanwhile, the North Atlantic Oscillation (NAO) has a strong correlation with peak monsoon rainfall, and tropical north and south Atlantic warming (TNA and TSA) affects the monsoon by altering the structure and intensity of the Intertropical Convergence Zone (Kuchaarski et al. 2008; Xavier et al. 2018). These multifaceted interactions underscore the complexity of global teleconnections in shaping Indian monsoon variability.

Despite these advances, existing research often focuses on specific regions, short-term periods, or isolated climatic indices, resulting in fragmented insights into the long-term impacts of global warming on rainfall variability across India. A comprehensive, long-term analysis, based on consistent fixed rain gauge station observations, is still lacking, leaving critical gaps in understanding rainfall patterns and their implications for water resource management at a national scale. While existing studies have established important links between global climatic indices – such as ENSO and IOD – and rainfall variability, these analyses predominantly target ISMR trends or specific subregions, overlooking the nuanced impacts at the river basin level. Studies have demonstrated that global climatic indices strongly affect the magnitude and frequency of annual and seasonal peak flows (Gurrapu et al. 2016, 2023). However, efforts to directly link these indices to peak flow events often neglect their intermediate relationships with basin-specific rainfall patterns, introducing uncertainties and reducing the reliability of predictions. Research, investigating the influence of global teleconnection patterns on rainfall across India's diverse river basins over an extended period, is scarce, creating challenges for effective hydrological and flood management planning. This gap limits the development of holistic strategies for water resource management, hydrological planning, and climate adaptation.

Intra-seasonal and inter-annual monsoon rainfall fluctuations are profoundly influenced by monsoon circulation intricately tied to global temperature distribution that drives summer rainfall across the Asia-Indo-Pacific region. The core of the monsoon circulation is located over subtropical Asia (the Tibet-Himalaya-Middle East sector), characterized by the warmest-thickest troposphere, lowest pressure, and largest zonally oriented upper tropospheric anticyclonic circulation. The air converges into the monsoon trough and adjacent areas near the Tibetan Plateau, rises and accumulates at upper levels over subtropical Asia, disperses through anticyclonic circulation, subsides over subtropical and polar highs, and eventually returns from the lower layers of highs and converges into the monsoon regime, completing the circulation cycle (Ranade & Singh 2014). Our previous studies have shown that during boreal summer, a warmer and thicker troposphere over the eastern hemispheric north subtropics strengthens monsoon circulation, often leading to above-average rainfall across extensive regions of India (Ranade & Singh 2019, 2021). Meteorological analyses consistently highlight an east-west oscillation in the Tibetan anticyclone core during the summer monsoon. These patterns raise critical questions about how global temperature variations may modulate monsoon circulation and, in turn, influence the spatio-temporal distribution of rainfall across India.

By integrating the above perspectives, this paper aims to address the gaps in understanding long-term rainfall trends by conducting a comprehensive analysis of the longest available (1813–2020) instrumental basin-scale rainfall series developed from 316 fixed rain gauge station observations. The study investigates evolving rainfall patterns, their relationship with global teleconnection indices, and the broader implications of global warming. The key objectives of the study are as follows:

• 1. To uncover long-term trends and spatio-temporal variability in annual and monsoon rainfall of 11 major and nine independent minor basins, as well as the West Coast Drainage System (WCDS) in India using the longest instrumental rainfall series.

• 2. To investigate the influence of global climatic indices on rainfall variability at the river basin scale and analyze the evolving relationships between them under the influence of global warming.

• 3. To understand the impact of global tropospheric warming on rainfall distribution and variability of each river basin.

Thus, the paper explores the broader implications of integrated long-term basin-scale rainfall trends, recent shifts and evolving connections with global climatic indices and global warming contributing to improved water resource management and climate adaptation strategies. It is expected that the study results represent critical insights into region-specific dynamics and their implications for hydrological systems vital for fostering sustainable water resource management, improving disaster resilience, and shaping targeted climate-informed policies and adaptation strategies in a rapidly warming world.

The longest instrumental area-averaged monthly rainfall dataset of 11 major and 36 minor river basins is available earliest from 1813 to 2000 (Sontakke & Singh 1996; Sontakke et al. 2008). In this dataset, the river basins are classified into major and minor basins based on the classification by Rao (1975) and ‘National Atlas for Thematic Mapping Organization (NATMO 1996). Twenty-seven of the 36 minor basins are sub-basins of five major basins, such as the Indus (3), the Ganga (13), the Brahmaputra (3), the Godavari (5) and the Krishna (3). The Sabarmati, the Mahi, the Narmada, the Tapi, the Mahanadi and the Cauvery are the major basins without any distinct minor basin. It also includes nine independent minor basins, the WCDS (combined catchment area of 25 small rivers originating in the Sahayadri Range), and the country as a whole. The data span from 1813 to 2000 and are prepared in two phases. In the first phase, the dataset for the period 1901–2000 was developed using the simple arithmetic mean of all available gauges in each basin using data from 316 fixed well-spread rain gauge stations obtained from the India Meteorological Department (IMD) (Mooley & Parthasarathy 1984). In the second phase, the dataset was extended backwards from 1900 to 1813 with the starting year varying for each basin. A theoretically vindicated numerical method was used on the limited available observations (Sontakke & Singh 1996). All reconstructions were made considering 1901–2000 as the reference period.

The study also used the longest available datasets for 10 different global climatic indices from five major areas surrounding the Indian subcontinent, as given in Table 1. The longest data series is for NAO (197 years), succeeded by PDO (167 years), AMO (165 years), and AO (164 years), while the shortest records are for TNA and TSA (73 years). The monthly temperature and geopotential height dataset of standard isobaric levels (1,000–250 hPa) across the globe at 2.5° resolution for the period 1949–2020 from the National Centers for Environmental Prediction – National Center for Atmospheric Research (NCEP-NCAR) reanalysis (Kalnay et al. 1996) also used to quantify the tropospheric warming over the northern subtropical belt.

The longest instrumental area-averaged monthly rainfall datasets were updated for the period 2001–2020 using the ratio method suggested by Rainbird (1967) and approved by the World Meteorological Organization (WMO), as per the following steps.

• (i) For updating the basin-scale series, each station value in a particular basin boundary was extracted from the corresponding location value of the grid from the 1 × 1° gridded rainfall dataset for the period 1951–2020.

• (ii) The area-averaged new dataset for the period 1951–2020 is prepared for all river basins and All India by the simple arithmetic mean of selected grid values within the basin boundary.

• (iii) The ratio between new and old rainfall values for each month was calculated for the common base period between the two datasets, i.e. 1951 and 2000.

• (iv) The new dataset from the years 2001–2020 is then updated by dividing individual month rainfall by the average of the ratio for the corresponding month calculated from the base period.

Area-averaged annual and monsoon (June through September: JJAS) rainfall series for all river basins, WCDS and All India has been developed and analyzed. Table 2 documents the basic statistical parameters calculated for the period 1901–2000 for annual and monsoon rainfall, respectively. Normally, (1901–2000) the mean annual rainfall of all major river basins varies from 742.8 (±269.1 mm) over Sabarmati to 2,528.5 (±284.4 mm) over WCDS. The coefficient of variation of the annual rainfall varies from 9.6% (Brahmaputra) to 36.2% (Sabarmati). The year-wise highest rainfall varied between 1,116.4 mm (Krishna) and 3,161.6 mm (Brahmaputra). All India gets 1,165.9(±106.2 mm) rainfall annually with the highest rainfall of 1,435.3 mm during the year 1917 and the lowest rainfall of 895.7 mm during the year 1918. For independent minor basins, the statistics is given in Table 2. The mean (1901–2000) monsoon rainfall of the major river basins varies from 581.4 (±101.9 mm) over Krishna to 1,706.6 (±181.1 mm) over Brahmaputra basin. Normally, WCDS receives 1,957.0 (±273.3 mm) rainfall in a year. The coefficient of variation of the monsoon rainfall varies from 12% (Ganga) to 37.3% (Sabarmati). The highest monsoon rainfall across the basins varied between 831.3 mm (Krishna) and 2,291.4 mm (Brahmaputra), while that of the lowest from 194.9 mm (Sabarmati) to 1,359.3 mm (Brahmaputra). Normally, All India gets 906.5 (±88.3 mm) rainfall during monsoon season with 9.7% of coefficient of variation (CV). During the period the highest rainfall was 1,088.8 mm during the year 1961 and the lowest was 661.3 mm during the year 1918. For independent minor basins, the statistics are given in Table 2.

The average distribution of rainfall throughout the river basins shows that the annual/monsoon rainfall of them is greatly influenced by their geographical locations, altitude, topography, proximity to the sea, the influence of global climatic patterns, onset and duration of southwest and northeast monsoon, extreme rain events and local microclimate. It is evident that basins nearer the coasts such as Mahanadi, Damodar, Suvarnarekha, Kasai, and so on, typically experience heavier rainfall because of oceanic factors like the abundance of moisture, interplay between ocean currents and wind patterns that influences the monsoon activity. The northeast monsoon and oceanic impact are responsible for high rainfall amounts in basins such as Cauvery and Pallar and Ponnaiyar. The rainfall amount in Indus, Ganga, Narmada, and so on is influenced by topography. A combination of oceanic and geographic factors results in extremely high rainfall in some basins such as Brahmaputra, Surma and WCDS.

In order to quantify statistically significant changes in the recent 20 years (2001–2020) compared to the last century (1901–2000), Students' t-test is used. River basins, such as Ganga, Brahmaputra, Cauvery, Surma, Brahmani, and Pallar & Ponniyar, have experienced a significant decrease in rainfall in recent decades. These basins are critical for India's water supply and food production, and a sustained decrease in rainfall could have severe socio-economic impacts. For instance, the Ganga basin saw a decrease of −6.3% (−6.6%) in annual (monsoon) rainfall. Similarly, the Brahmaputra and Cauvery basins saw a decrease of −8.4% (−7.6%) and −13.6% (−18.7%), respectively. Independent minor basins also saw a significant decrease in annual and monsoonal rainfall (Table 3). In contrast, certain basins entered a wet phase in recent years but were not statistically significant. Only the Surma basin experienced a significant increase in annual and monsoon rainfall by 3.9 and 7.9%, respectively. However, this localized wet phase contrasts sharply with the broader dry trends in other parts of the country, underscoring the spatial variability of monsoonal rainfall. As a result of the majority of basins experiencing negative percentage change in rainfall in recent years, the national average of annual and monsoon rainfall shows a significant decrease of −1.1% compared to the previous century.

Trend values are utilized to measure significant shifts in climatic time series, providing valuable insights for future process evolution, risk analysis, infrastructure planning, and agricultural and water resource management.

Visual examination reveals that the consistently observed significant peaks can be grouped into three categories: (i) waves with a periodicity of less than 10 years (short-term fluctuations); (ii) waves with a periodicity of 10–30 years (decadal variations); and (iii) waves with a periodicity of more than 30 years (long-term variability). The combined percentage of variance explained by these three wavelength bands is calculated. For annual rainfall, the combined variance of short-term variations across the basins varies from 54.3% (Godavari) to 79.4% (Ganga), the decadal variability from 14.4% (Ganga and Mahanadi) to 19% (Krishna and Brahmaputra), and the long-term variability ranges from 4.8% (Sabarmati) to 21.1% (Godavari). For the monsoon rainfall, the combined variance across the basins varies as follows: short-term variations 53.9% (Godavari) to 79.5% (Ganga); decadal changes 10.3% (WCDS) to 25% (Godavari); and long-term variability 5% (Sabarmati) to 22% (Godavari). For the country as a whole, the combined variances of short-term variations of annual and monsoon rainfall are 79 and 80%, respectively, those of decadal fluctuations are 10.6 and 9.4%, respectively, and for long-term variations, they are 10.1 and 10.2%, respectively. The findings clearly indicate that short-term variability (cycles less than 10 years) is the dominant factor in rainfall variability across India, and can lead to unpredictable swings between droughts and heavy rainfall, which makes it challenging to manage water resources effectively. While long-term climate trends (such as global warming) may influence rainfall patterns, their impact is often overshadowed by more immediate, shorter-term weather cycles. As a result, strategies to cope with water availability and extreme weather events must prioritize short-term variability over decadal or long-term trends.

The effectiveness of extrapolating rainfall time series using harmonic analysis is dependent on the degree of regularity in the series. The variance contribution of many oscillations found through harmonic analysis shows comparable values. It shows that the extrapolation of seasonal/annual rainfall time series using simple classical harmonic analysis is not feasible. More sophisticated techniques, such as appropriate filtering of the dataset to retain predictable components and noise removal using SSA or variable/non-integer harmonic analysis that can capture realistic periodicities, are suggested for future studies.

Figure 9(b) illustrates the interannual fluctuations of 10 climatic indices according to the existing records. Most of these indices do not exhibit a significant long-term trend over an extended period, except for dipole mode index (DMI), AAO, TSA, and TNA, which display a statistically significant increasing long-term trend. The increasing trend in the DMI can be attributed to the warming of the western Indian Ocean, the change in frequency and intensity of ENSO events and ongoing changes in global circulation patterns due to warming (Cai et al. 2014). Warming of the tropical Atlantic and North Atlantic could lead to an increase in TSA and TNA indices (Watanabe & Kimoto 2000; Knight et al. 2006).

In recent years, various oscillations have shown increasing or decreasing epochal patterns. The strength of the robust, enduring concurrent relationship/teleconnection between the longest monthly and monsoonal rainfall of All India with each of the 10 climatic indices is represented by correlation coefficients (CC) in Table 4. Climatic indices like Niño 3.4, Southern Oscillation Index (SOI), and AO consistently show significant correlations with rainfall across all months during the monsoon season. Niño 3.4 has the strongest negative correlation with ISMR (CC = −0.60). El Niño events (warmer sea surface temperatures in the central and eastern Pacific) are known to suppress monsoon rainfall due to weakened convection over India. This effect is especially pronounced in September when monsoon withdrawal begins. The strongest monthly negative correlation in September (CC = −0.45) indicates that ENSO's effect intensifies towards the end of the monsoon season. SOI with the atmospheric counterpart of Niño 3.4 displays a significant positive correlation, also peaking in September (CC = 0.40), and has the second-highest correlation with ISMR (CC = 0.48). Positive SOI events enhance the monsoon's strength by promoting favourable moisture transport and convection over the Indian subcontinent.

The AO impacts mid-latitude atmospheric circulation, including the monsoon. It has a significant negative correlation with rainfall, strongest in September (CC = −0.31). This could be due to changes in the subtropical jet stream and its interaction with the monsoon trough, which influences monsoon dynamics. The PDO influences long-term climate variability in the Pacific and can modulate ENSO effects. PDO shows negative correlations with rainfall only in June (CC = −0.30) and September (CC = −0.28) indicating that negative phases of the PDO (cooler Pacific waters) suppress the early and late monsoon rains by altering the atmospheric circulation, reducing moisture inflow into the Indian subcontinent. NAO and DMI exhibit significant but weaker correlations with rainfall in June (NAO: CC = 0.27) and September (DMI: CC = −0.17). Their effects on the Indian monsoon may be more localized or secondary compared to other indices. AMO, TSA, and TNA do not display significant correlations with overall India-level rainfall but have notable correlations at the basin scale.

The analysis of ISMR and climatic indices shows that correlations between them are stronger in June and September, but weaker during July and August, which are critical months for monsoon rains. This suggests that the monsoon is becoming less predictable in its peak period, possibly due to the complex interactions between multiple indices leading to nonlinear responses in monsoon rainfall.

Only a few indices show significant correlations consistently throughout the monsoon season, e.g. Niño 3.4 has a persistent significant negative correlation with rainfall of the Indus and Ganga basins from June through September. SOI is positively correlated with rainfall in the Brahmaputra basin throughout the monsoon season and with the Ganga basin, except in August. AO is positively correlated with the Ganga basin's rainfall throughout the monsoon season. PDO shows a strong negative correlation with rainfall in the Indus basin, except in August. For most other basins, no single climatic index consistently correlates with monthly rainfall for more than two months in a season. Additionally, most indices show stronger correlations with basin rainfall in June and September, and comparatively weaker in July and August. The results are in line with those of All India. They suggest that these relationships are influenced by multiple factors, including the phase, timing, and interactions between the indices as well as broader climatic trends like Arctic warming and complicate the traditional connections.

Figure 11(b) illustrates the 31-term running correlation between monsoon rainfall across 11 major basins and selected climatic indices that have shown significant long-term correlations. For basin-level rainfall, the correlation analysis revealed that several climatic indices have recently lost their association with rainfall. Niño 3.4 and SOI, however, continue to exhibit significant correlations with the Indus, Ganga, Brahmaputra, Godavari, Krishna, Tapi, and Cauvery basins, while the Sabarmati, Mahi, Narmada, and WCDS basins have lost these connections. Additionally, AO has maintained significant correlations with the Godavari, Narmada, and Tapi basins, but its association with other major basins, including the Indus, Ganga, Krishna, and others, has diminished. Many climatic indices have recently lost their strong associations with rainfall. This weakening may reflect local and regional variations in how these basins respond to broader climatic signals, possibly influenced by land use changes, irrigation practices, or changes in atmospheric circulation patterns.

Niño 3.4 and SOI, while traditionally linked to monsoon rainfall, are also affected by interactions with other indices such as the PDO, MJO, and the IOD. The intensity and timing of ENSO events, along with their interaction with other indices, result in varying impacts on seasonal rainfall. Further analysis shows the interconnectivity between many climatic indices. For instance, the PDO negatively correlates with SOI and AO (CC = −0.31 and −0.35, respectively) and positively with Niño3.4 (CC = +0.48). Similarly, Niño3.4 has a strong positive correlation with the DMI (CC = +0.50) and a negative correlation with AO (CC = −0.33). DMI also exhibits positive correlations with AO and AAO (CC = +0.23 and +0.24, respectively). TSA and TNA are strongly positively correlated with the AMO (CC = +0.31 and +0.86, respectively) and AAO (CC = +0.29 and +0.34, respectively). TSA and TNA also share a positive correlation (CC = +0.36). The interrelationships between various climatic indices show that the global atmospheric and oceanic phenomena do not operate in isolation, further complicating their individual relationships with monsoon rainfall. The timing, intensity, and phase of one index can alter the behaviour of others, impacting the seasonal distribution of rainfall in complex ways.

We found that global climatic indices are the weak representative of monsoon rainfall performance, especially during July and August. This study highlights the rainfall peak during July and August is primarily driven by the location, size, intensity, and reach of the well-established Asia-Pacific monsoon circulation. Its location, form, size, and intensity are primarily impacted by the tropospheric temperature/thickness contrast between the Tibetan anticyclone and eight deep highs worldwide.

The study provides comprehensive insights into the intricate patterns of interannual rainfall variability by analyzing the climatological and fluctuation features, recent years' changes in annual and monsoonal rainfall of 11 major basins, nine independent minor basins, WCDS, and All India. The research highlights the multifaceted interplay of internal and external factors shaping the high spatio-temporal variability in monsoon rainfall. Using reconstructed rainfall data dating back to 1813, the study underscores the substantial variability in normal annual rainfall, which ranges from 487.7 mm (Luni basin) to 2,519.5 mm (Surma basin), and monsoon rainfall, varies from 255.3 mm (Vaigai basin) to 1,706.6 mm (Brahmaputra basin). Long-term trends reveal distinct epochs of wet and dry periods, with multiple transitions between these phases from 1813 to 2020. The most recent dry epoch (2000–2020) continues to affect major river basins, particularly the Ganga, Brahmaputra, and Cauvery basins. In contrast, certain basins, such as Krishna and Sabarmati, have witnessed a wet phase in recent decades. Statistical analysis of the last two decades (2001–2020) indicates significant declines in annual and monsoon rainfall in key basins, particularly in the Ganga (−6.3, −6.6%) and Brahmaputra, (−8.4, −7.6%), signalling a shift in regional rainfall patterns. Collectively, these basins contribute a significant portion of the country's rainfall, making these changes critical for water resource management. On the other hand, the Surma basin stands out for its increased rainfall in both annual (+3.9%) and monsoon (+7.9%) terms, indicating regional disparities in rainfall trends. However, the overall All India rainfall has decreased marginally by −1.1%, a change not statistically significant.

Distinct epochs are identified over time scales of 15, 31, 51, and 101 years. Specific basins, such as the Brahmaputra and Brahmani, have experienced consistent declines in monsoon rainfall, while Pennar and WCDS have seen persistent increases. Although recent decades have shown a decreasing trend in All India rainfall, particularly in the shorter time windows, the 101-year cycle indicates a modest but statistically significant increase of 1.3% in monsoon rainfall. This suggests that while fluctuations dominate at shorter time scales, there is overall long-term stability in the monsoon rainfall. These findings are crucial for researchers and policy-makers focused on sustainable water resource management. The study demonstrates that short-term fluctuations are observed as the primary driver of interannual variability, accounting for more than 75% of rainfall variability across basins, followed by decadal shifts (15%) and long-term variability (10%), as reflected in the spectral analysis.

Furthermore, the study explores the relationship of year-to-year and multidecadal variability of ISMR with global climatic indices, revealing complex and evolving interconnections. The Niño 3.4 index has a strong negative correlation (−0.60) with ISMR, while other indices such as SOI, AO, PDO, and DMI also influence rainfall patterns with varying degrees of correlation. Spatial analysis shows that SOI, Niño3.4, and AO are the most influential indices across all river basins, with significant monthly correlations during the monsoon season. Notably, Niño 3.4 is consistently linked to rainfall in the Indus and Ganga basins, while SOI has a strong positive correlation with the Brahmaputra and Ganga basins, AO has a correlation with the Ganga basin and PDO to the Indus basin. Interestingly, many climatic indices exhibit strong correlations with rainfall in June and September, but weaker correlations during July and August, the peak of the monsoon season. It could be due to the overlapping influences of various phenomena such as Niño 3.4, AO, PDO, and DMI, that disturb the linear relationship. The indices directly influence the transition and intensity of monsoon winds, so the correlations may be more pronounced at the onset and retreat phases of the monsoon.

The study also emphasizes the dynamic nature of correlation with ISMR over time. For instance, the AO has shown a weakening relationship with ISMR since the 1980s, preferably due to Arctic warming affecting key basins like the Indus, Ganga, Krishna, Cauvery, and WCDS while correlations with basins like Godavari and Narmada have strengthened. These findings align with previous research, indicating multiple switchovers in the associations between these indices and ISMR over time (Adarsh & Janga Reddy 2019). The findings underscore the intricate nature of global teleconnections and their varying impacts on regional rainfall patterns. While Niño 3.4 and SOI remain key predictors of monsoon rainfall variability, the influence of other indices like PDO and AO is less consistent, and their correlations have shifted over time. These evolving relationships have significant implications for hydrological modelling and infrastructure planning. As historical correlations are often used to predict annual and seasonal peak flows and water availability in river basins, the study emphasizes the importance of revisiting these predictive models in light of shifting climatic influences to reduce uncertainties in future projections and ensure sustainable management of water resources. This also highlights the need for adaptive and region-specific approaches in water resource management and climate risk planning, as climatic indices can have differential impacts across basins and time scales.

The research also highlights the connection between monsoon rainfall patterns and global tropospheric warming, particularly in relation to the Tibetan anticyclone and the Afro-Asian highlands. Northern and eastern basins, such as Indus and Brahmaputra, exhibit strong correlations with the warm, thick troposphere over eastern and central subtropical Afro-Asian highlands. Central Indian basins, like the Godavari and Krishna, show significant relationships with warming patterns over the northern hemisphere and southern tropical and subtropical regions, reflecting the global nature of monsoon circulation. In contrast, southern basins, such as the Cauvery, are primarily influenced by equatorial climate dynamics and the strength of the southwesterly monsoon flow plays a larger role than large-scale monsoon circulations. These findings underscore the intricate relationship between global tropospheric warming and regional rainfall patterns, highlighting the need for a nuanced understanding of monsoon dynamics in the context of global climate variability. As global temperatures continue to rise, understanding these interactions will be crucial for predicting monsoon performance and managing water resources effectively.

The considerable rising or falling tendency in monsoon rainfall across India has some rationale. However, neither of them can conclusively claim that they are uniform and random. This study demonstrates the importance of understanding both the short- and long-term variability of monsoon rainfall across 20 river basins of India. The research findings indicate that short-term fluctuations are the primary drivers of rainfall variability in India, overshadowing decadal and long-term trends. This makes predicting future rainfall patterns based on long-term cycles alone unreliable. The study highlights significant declines in monsoonal rainfall in key basins such as the Ganga, Brahmaputra, and Cauvery, alongside modest increases in others. The basin-specific trends emphasize the need for localized water management strategies to address the diverse impacts of changing rainfall patterns.

The intricate web of interactions between climatic indices collectively influences monsoon rainfall patterns reflecting regional differences and evolving nature. The shifting correlations over time underscore the dynamic nature of these relationships, driven by a multitude of factors. It suggested the predictive models to account for these shifting correlations to ensure accurate water management projections. Additionally, the study underlines the influence of global tropospheric warming on rainfall patterns, highlighting the critical need to consider the broader implications of climate change on India's monsoon dynamics. The monsoon circulation, though vigorous, exhibits a distorted 3D structure and variable interaction with global climatic indices and uneven global warming. This leads to heterogeneous rainfall distribution over space and time, with unpredictable and inconsistent extreme rain events. As India continues to face the challenges of global temperature rise and climate change, these findings underscore the urgency of developing adaptive, region-specific strategies for managing river basins, assessing groundwater recharge potential, and aiding local agricultural planning for sustainable development and ensuring water security for future generations.

The author is grateful to the Director, National Institute of Hydrology, Roorkee for the necessary facilities to pursue the study and to the India Meteorological Department for providing the necessary station and gridded rainfall datasets.

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

The authors declare there is no conflict.