Climate change and Indian floods: what the science says

After every major flood in India, the same question comes up within hours: was this climate change? It is a fair question, and it deserves a more careful answer than "yes" or "no". Climate change and floods in India are linked, but the link runs through several steps, and the evidence is stronger for some steps than for others. This article follows those steps one at a time, using the IPCC's Sixth Assessment Report (AR6), the Ministry of Earth Sciences (MoES) assessment of climate change over the Indian region, and peer-reviewed studies.

Does climate change cause floods?

Climate change does not cause a flood on its own, but it changes the conditions in which floods happen. A flood needs water to arrive faster than a channel, a drain or a floodplain can carry it away. Climate change acts on the first part of that sentence: it can change how much rain falls in a short time, how fast snow and ice melt, and how high the sea stands at the coast. It does not act on the second part, the capacity of channels and drains, which depends on land use, embankments and maintenance.

That is why scientists separate three questions:

1. Has heavy rainfall become more frequent or more intense?
2. Have river floods, meaning high flows in rivers, followed?
3. Is flood damage rising, and how much of that is climate and how much is people and property in harm's way?

The answers are "yes, in many places", "in some regions, less clearly" and "mostly exposure, with climate adding to it". The rest of this article explains each.

The physical starting point: a warmer atmosphere holds more water

Warmer air can hold more water vapour. AR6 states that at the global scale, the intensification of heavy precipitation follows the increase in the maximum moisture the atmosphere can hold as it warms, about 7 % per 1 °C of global warming (high confidence) [1]. This is the Clausius–Clapeyron relation. It is a thermodynamic guide, not a forecast for any one storm: regional increases also depend on changes in atmospheric circulation and storm dynamics [1].

The arithmetic compounds with each degree. The chart below is an illustrative calculation, 1.07 raised to the power of the warming, not a projection for any place.

AR6 adds that the frequency of heavy rain events rises non-linearly with warming and more for rarer events: 10-year and 50-year events become approximately double and triple as frequent, respectively, at 4 °C of warming [1]. This matters for flood design, because rarer floods drive the design of bridges, embankments and spillways. For the meaning of a return period, see what a "100-year flood" really means.

What has already changed in India's rainfall?

India's monsoon is a puzzle for this question. The MoES assessment finds that all-India summer monsoon mean rainfall decreased over 1951–2015, notably over the Indo-Gangetic Plains and the Western Ghats, with rising anthropogenic aerosols appearing to play a role (medium confidence) [2]. Yet at the same time, the frequency of localized heavy rain occurrences over India increased during 1951–2015 (high confidence) [2].

Two studies show how both can be true:

• Goswami and colleagues used daily rainfall over central India and found significant rising trends in the frequency and magnitude of extreme rain events, and a significant decreasing trend in moderate events, over 1951–2000. The seasonal mean showed no significant trend because the contribution from increasing heavy events was offset by decreasing moderate events [3].
• Roxy and colleagues found a threefold increase in widespread extreme rain events over central India during 1950–2015. They linked it to growing variability of the low-level monsoon westerlies over the Arabian Sea, which drives surges of moisture supply across the central subcontinent [4]. AR6 reports that this increase has been attributed to the combined effect of warming of the western Indian Ocean (Arabian Sea) and the intensification of irrigation water management over India [1].

In other words, the monsoon has not simply become wetter. Seasonal totals have been flat or lower, while the rain has become more concentrated into heavier spells. A river does not respond to the season's total; it responds to the heavy spells.

How much of this is climate change, and how much is something else?

AR6 is careful here. It concludes that human influence has contributed to the intensification of heavy precipitation in three continents with more abundant observations, including Asia (high confidence), but that at the scale of individual AR6 regions the evidence of human influence on extreme precipitation is limited, though new evidence is emerging [1]. For India, the MoES assessment says the rise in localized heavy rain is partly attributed to greenhouse-gas warming and partly to aerosols, land-use change, atmospheric stability and urbanisation [2]. So the honest statement is: climate change is a significant contributor, mixed with other regional factors that scientists are still separating.

Our own work on rainfall extremes under climate scenarios points the same way: projected heavy rain events above 40 mm a day increase across most of India, even under a low-emission scenario, though models differ markedly by season and region [5].

Have Indian river floods increased?

River floods are harder to read than rainfall. AR6 says significant trends in peak streamflow have been observed in some regions over past decades (high confidence), but changes in river floods are more uncertain than changes in pluvial floods (floods caused by rain exceeding local drainage capacity), because "complex hydrological processes and forcings, including land cover change and human water management" are involved [1]. A more intense downpour does not always produce a bigger river flood. AR6 adds that water regulation and management have in general increased resilience to flooding, masking the effect of an increase in extreme precipitation on flood probability in some regions, so more extreme rain does not always mean more river floods [1].

For India, the MoES assessment concludes that an increased frequency of localized heavy rainfall on sub-daily and daily timescales has enhanced flood risk (high confidence), and that urban floods are on the rise [2]. Using the Dartmouth Flood Observatory database of severe floods, it reports a statistically significant increase of about one severe flood event per decade over India during 1985–2019 [2]. That database is compiled from reported events, so changes in reporting can influence it. It is a signal, not a measurement of river peaks.

The gauged record is more cautious. Ali, Modi and Mishra found that a majority of river basins in the Indian sub-continent saw a rise in extreme precipitation during 1966–2005, but that these trends were mostly insignificant at the 5 % level [7]. Part of the difficulty is that flood peaks are rare events and long, continuous records are scarce. For how flows are derived from levels, see rating curves and their uncertainty.

At the global scale, Gudmundsson and colleagues analysed river-flow records from 7,250 observatories for 1971–2010. They found that trends computed from model simulations match the observations only if radiative forcing that accounts for anthropogenic climate change is included, and that simulated effects of water and land management do not suffice to reproduce the observed pattern [8]. That is a global result, not an India-specific one.

A worked comparison: a flood where rainfall and management combined

The 2018 Kerala flood shows why attribution is not simple. Mishra and colleagues found that Kerala received 53 % above-normal rainfall for the monsoon season up to 21 August 2018, and that 1-, 2- and 3-day extreme rainfall in August 2018 had return periods of 75, 200 and 100 years. Six of seven major reservoirs were above 90 % of full capacity on 8 August, before the extreme rain, and that rainfall over the catchments of three major reservoirs had a return period of more than 500 years for some durations between 1 and 15 days [9]. They concluded that above-normal seasonal rainfall, high reservoir storage and unprecedented extreme rainfall together worsened the flooding [9]. (This paper is a preprint from an open journal review process and was not taken forward to publication; read its figures as a study under discussion, not a final result.)

The MoES assessment states that it is difficult to attribute a specific extreme monsoon event, such as the Kerala flood of 2018, to anthropogenic climate change, yet it is robustly anticipated that a warming atmosphere will result in more severe weather [2]. Both statements hold together: climate change loads the dice, but a particular flood also depends on how full the reservoirs were and how the water was managed.

What about the Himalaya?

Hill and mountain catchments add snow, ice and steep slopes. AR6 notes that where snow cover is seasonal, snowmelt is the main cause of extreme river flooding over large areas, and that changes in such regions are uncertain because warming increases precipitation but also shortens the period of snowfall accumulation [1]. For South Asia, AR6 Chapter 12 assesses with medium confidence that monsoon floods will be more intense [10].

In our analysis of the 7–11 July 2023 floods in Himachal Pradesh, using station data and the ERA5, SMAP-L4 and FLDAS-NOAH reanalysis products, we found that the compounding of extremely heavy rainfall with near-saturation antecedent soil moisture and snowmelt was the leading factor in inflating and sustaining the flood peak [6]. A flood in the hills is rarely a one-factor event: rain, wet soil and melt coincided. See also river monitoring in hill and Himalayan catchments and snow and glacier melt.

What do projections say for the rest of this century?

Projections point mostly one way for flood hazard, with uncertainty in the size.

• Heavy rain: MoES notes that the frequency of extreme precipitation events may increase all over India, and more prominently over central and southern parts [2]. A study of subdaily rainfall found that a rise of 1.5 °C (2.0 °C) in global mean temperature is projected to cause a 20 % (25 %) increase in 3-hourly precipitation maxima at the 100-year return period, and that 3-hourly maxima rise almost twice as fast as 24-hourly 100-year maxima [11]. Short, intense bursts are what overwhelm urban drains; see why cities flood.
• River floods: MoES reports that flood propensity is projected to increase over the major Himalayan river basins, such as the Indus, Ganga and Brahmaputra (high confidence) [2]. Ali, Modi and Mishra project a significant increase in the frequency of multi-day extreme precipitation and flood events under both RCP 2.6 and RCP 8.5, and found that the projected increase in flood frequency can be reduced to half under the low-emission scenario compared with the high one [7].
• Streamflow: in our projections for 128 Indian river basins, using four hydrological models and 12 climate models, the indicators showed an increase in high-flow frequency and duration, meaning greater flood risk, while low-flow outcomes were more uncertain [12].
• Compound events: in a copula-based analysis of IMD gridded rainfall, we found shorter return periods for floods and droughts at most grid points, but no increase in the co-occurrence of floods and droughts [13].

The coast: sea level and cyclones

Coastal flooding adds two more factors. The MoES assessment reports that global mean sea level has risen by 1.7 mm a year since 1901 and the rate has accelerated to 3.3 mm a year since 1993, and that the north Indian Ocean rose at 3.3 mm a year over 1993–2015 [2]. For cyclones, it reports a significant reduction in the annual frequency of tropical cyclones over the North Indian Ocean from 1951 to 2018, but a significant rise in post-monsoon very severe cyclonic storms over 2000–2018 (high confidence). Models project a rise in cyclone intensity (medium confidence) and in cyclone rainfall intensity (medium to high confidence) [2]. Fewer storms therefore do not mean weaker ones.

Climate is not the only driver of flood damage

Flood damage depends on hazard, exposure and vulnerability. More people, homes and roads on floodplains raise losses even if the rain does not change. Urbanisation seals ground and shortens the time rain takes to reach a drain or stream; see time of concentration. The MoES assessment adds that human influences such as land-use changes, irrigation and mismanagement of dams and reservoirs can aggravate the magnitude and frequency of floods [2]. AR6 likewise names land cover change and human water management among the factors that make river-flood changes uncertain [1].

The consequence for planners is practical: you cannot read climate change from a single flood, and you cannot dismiss it either. Hazard and exposure are changing at the same time.

What does this mean for flood risk and design?

Traditional design assumes the past is a good guide to the future, so a 100-year flood estimated from an old record stays the 100-year flood. If heavy rain is intensifying, that assumption weakens. Practical responses include:

• Keep records long, continuous and quality-checked, so that trends can be separated from noise. A gauge that is moved, or a rating that is not updated, can hide or fake a trend.
• Test designs against a plausible increase in rainfall intensity. The scaling in the chart above gives a first idea of the size of the sensitivity.
• Treat floodplain maps as living documents. See flood hazard maps and floodplain zoning.
• Invest in monitoring and warning, because rivers will be tested against a changing climate whatever the cause. How fast a river rises is the basis of timely warnings; see flash floods vs river floods.
• Know where floods happen. See flood-prone areas of India.

Key points

• A warmer atmosphere can hold about 7 % more moisture per 1 °C, and AR6 says heavy precipitation intensifies along this rate at the global scale (high confidence) [1].
• In India, mean monsoon rainfall fell over 1951–2015 while localized heavy rain became more frequent (high confidence) [2]; widespread extreme events over central India rose threefold from 1950 to 2015 [4].
• River-flood trends are less certain than rainfall trends because soil, land cover, reservoirs and embankments intervene [1].
• Individual events, such as Kerala 2018, cannot simply be labelled as climate change; rainfall, reservoir levels and exposure all mattered [2][9].
• Projections indicate rising flood hazard in the Indus, Ganga and Brahmaputra basins and a rising share of short, intense rain, but the size of the change varies by model and scenario [2][7][11].
• Flood damage rises with exposure as well as climate, so long records, monitoring and warning systems matter in any case.

Where ClimMaTech fits

At ClimMaTech we work on the measurement side of this question: reliable river level, rainfall and flow records are what allow trends to be seen and warnings to be issued. Our flood warning page describes how monitoring stations, a cloud platform and alert devices fit together for communities and departments in flood-prone catchments.

References

1. S. I. Seneviratne, X. Zhang, M. Adnan, et al., "Weather and Climate Extreme Events in a Changing Climate", Chapter 11 in Climate Change 2021: The Physical Science Basis, Contribution of Working Group I to the Sixth Assessment Report of the IPCC, Cambridge University Press, 2021. ipcc.ch
2. R. Krishnan, J. Sanjay, C. Gnanaseelan, M. Mujumdar, A. Kulkarni, S. Chakraborty (eds.), Assessment of Climate Change over the Indian Region: A Report of the Ministry of Earth Sciences (MoES), Government of India, Springer, 2020 (chapters 3, 6, 8 and 9). doi:10.1007/978-981-15-4327-2
3. B. N. Goswami, V. Venugopal, D. Sengupta, M. S. Madhusoodanan, P. K. Xavier, Increasing trend of extreme rain events over India in a warming environment, Science 314(5804), 1442–1445, 2006. doi:10.1126/science.1132027
4. M. K. Roxy, S. Ghosh, A. Pathak, R. Athulya, M. Mujumdar, R. Murtugudde, P. Terray, M. Rajeevan, A threefold rise in widespread extreme rain events over central India, Nature Communications 8, 708, 2017. doi:10.1038/s41467-017-00744-9
5. V. Gupta, V. Singh, M. K. Jain, Assessment of precipitation extremes in India during the 21st century under SSP1-1.9 mitigation scenarios of CMIP6 GCMs, Journal of Hydrology 590, 125422, 2020. doi:10.1016/j.jhydrol.2020.125422
6. V. Gupta, B. Syed, A. Pathania, S. Raaj, A. Nanda, S. Awasthi, D. P. Shukla, Hydrometeorological analysis of July-2023 floods in Himachal Pradesh, India, Natural Hazards 120(8), 7549–7574, 2024. doi:10.1007/s11069-024-06520-5
7. H. Ali, P. Modi, V. Mishra, Increased flood risk in Indian sub-continent under the warming climate, Weather and Climate Extremes 25, 100212, 2019. doi:10.1016/j.wace.2019.100212
8. L. Gudmundsson, J. Boulange, H. X. Do, et al., Globally observed trends in mean and extreme river flow attributed to climate change, Science 371(6534), 1159–1162, 2021. doi:10.1126/science.aba3996
9. V. Mishra, S. Aaadhar, H. Shah, R. Kumar, D. R. Pattanaik, A. D. Tiwari, The Kerala flood of 2018: combined impact of extreme rainfall and reservoir storage, preprint (discussion paper), Hydrology and Earth System Sciences Discussions, 2018. doi:10.5194/hess-2018-480
10. "Climate Change Information for Regional Impact and for Risk Assessment" (Asia, section 12.4), Chapter 12 in Climate Change 2021: The Physical Science Basis, IPCC AR6 WGI, Cambridge University Press, 2021. ipcc.ch
11. H. Ali, V. Mishra, Increase in subdaily precipitation extremes in India under 1.5 and 2.0 °C warming worlds, Geophysical Research Letters 45(14), 6972–6982, 2018. doi:10.1029/2018GL078689
12. V. Sharma, S. Barbhuiya, V. Gupta, Projections of future streamflow for India informed by CMIP6 global climate models, Hydrological Processes 40(3), e70476, 2026. doi:10.1002/hyp.70476
13. S. C. Kumar, V. Manikanta, M. K. Jain, V. Gupta, Deciphering the intra annual dynamics of evolving compound flood-drought interaction in Indian hydrology: insights from a copula-based approach, Theoretical and Applied Climatology 156(4), article 199, 2025. doi:10.1007/s00704-025-05427-w