Deluge: Floods, Urban Flooding, Cloudbursts and GLOFs

Water disasters are India's most frequent and most human-made category of calamity: floods, urban flooding, cloudbursts and glacial lake outburst floods. The monsoon delivers the water, but encroached floodplains, concretised cities, mismanaged dams and a warming cryosphere convert rain into ruin. Over 40 million hectares, 12 per cent of India's land area, is flood-prone, floods affect about 7.5 million hectares every year, and a single cloudburst over a Himalayan valley can now trigger a glacial lake outburst downstream.

This article works through the water chain in order: river floods, city floods, cloudbursts, and GLOFs. Each section defines the hazard completely, maps India's vulnerability, and shows what the state has built, from embankments to sponge cities to satellite-monitored glacial lakes.

Floods: the overflow onto dry land

A flood is the overflow of water onto normally dry land, caused by heavy rainfall, river overflow, dam failure or storm surges. NDMA's working definition is tighter: a flood is an overflow of water that submerges land and causes destruction. Floods are the most frequent disaster India faces, and unlike earthquakes they recur on a predictable calendar, which makes the failure to manage them a policy choice rather than bad luck.

The scale is staggering. India has 329 million hectares of geographical area, of which over 40 million hectares (12 per cent) is classified as flood-prone. On average, floods affect about 7.5 million hectares annually and some 30 million people. The most exposed basins are the Ganga system (with the Yamuna, Gandak and Ghaghara), the Brahmaputra with its enormous sediment load, the Kosi (the Sorrow of Bihar, infamous for changing course), the Mahanadi, the Godavari and the Damodar (the Sorrow of Bengal).

Floods take four main forms:

Natural causes include intense monsoon rain that exceeds absorption limits, cyclones, melting snow and glacial lakes, and flat alluvial topography with poor drainage, as in the Brahmaputra valley. The human causes are the ones UPSC rewards: floodplain encroachment (illegal construction on riverbeds, as along the Mithi in Mumbai), dam mismanagement (sudden releases overwhelmed Kerala in 2018; the CAG found Emergency Action Plans were absent for about 99 per cent of large dams), deforestation of catchments (which intensified the Kerala floods of 2018-19), mining and siltation that raise riverbeds, and embankments that expert committees, from the G.R. Garg Committee (1951) to the National Flood Commission (1980), have shown can worsen flooding by blocking natural drainage and silt deposition.

NDMA's flood guidelines divide the response into structural and non-structural measures. Structural measures are embankments and levees, watershed management with check dams and afforestation, anti-erosion works, and flood shelters on raised platforms. The non-structural half is where modern flood management lives: flood forecasting and early warning through the CWC's network of about 592 monitoring stations and IMD's Flash Flood Guidance System, which gives localised alerts for nearly 30,000 watersheds, hours ahead of the event; floodplain zoning legislation to keep construction out of the river's way; wetland conservation; and basin-wise management plans. The FloodWatch India app now puts 7-day forecasts from 592 stations in the public's pocket.

• Flood Management Programme
• Sponge City Mission
• Build Back Better and CDRI

Reservoir operations and dam inflow forecasting

Every large dam faces the same dilemma: keep the reservoir full for irrigation, power and drinking water, or keep it partly empty as a flood cushion to absorb monsoon inflows. Getting this balance wrong is how dams turn from flood controllers into flood makers, when sudden emergency releases coincide with heavy downstream rainfall. The technical answer is dam inflow forecasting, the prediction of how much water will enter a reservoir over the coming hours and days, so that releases can be planned gradually instead of dumped in panic.

The Central Water Commission (CWC), India's apex technical body for water resources, runs the country's flood forecasting network and issues inflow forecasts for major reservoirs. The system combines upstream rain gauges and river gauges, satellite rainfall estimates and rainfall-runoff models to project inflows, and the forecasts flow to dam authorities, state irrigation departments and disaster management authorities. When the chain works, gate operations follow a rule curve rather than a crisis; when it breaks, through data gaps, poor inter-state coordination or ignored forecasts, downstream towns pay for the release. For mains, inflow forecasting is the institutional answer to the recurring charge of dam-induced floods.

Urban flooding: a man-made disaster in natural clothing

Urban flooding is the inundation of land or property in cities or other built environments, caused by rainfall or coastal storm surges overwhelming the capacity of drainage systems such as storm sewers. As Gilbert F. White wrote, floods are acts of God, but flood losses are largely acts of man, and nowhere is that truer than in Indian cities: Bengaluru (2024), Delhi (2023), Mumbai (2020) and Chennai (2015) have all gone under in the last decade.

Meteorological triggers are heavy monsoon rain, cyclonic depressions and cloudbursts, but the city itself manufactures the disaster. Unplanned land use disrupts natural drainage (Hyderabad, 2020); encroachment on floodplains and lakes removes the city's sponges (Chennai, 2015; Bengaluru's Bellandur-Varthur lakes, 2022); concretisation makes rain run off instead of infiltrating; solid waste chokes storm drains (Mumbai, 2005); the urban heat island intensifies convective downpours; and high tides block stormwater discharge in coastal cities (Mumbai, 2021). Climate change then raises the frequency of the extreme rainfall that these weakened cities cannot handle.

The impacts go beyond wet roads: drowning and electrocution (as in Delhi's Old Rajendra Nagar in 2024), damage to metros and power lines, business closures, dengue and malaria from stagnant water, and sewage-contaminated runoff poisoning urban water bodies. NDMA's urban flood guidelines answer with five pillars: early warning and communication (a national hydrometeorological network and wider Doppler radar coverage), scientific drainage design based on catchment boundaries and contour data, vulnerability analysis with dedicated Urban Flood Cells in city administrations, emergency preparedness through Emergency Operations Centres and an Incident Response System, and capacity building with training and documentation of best practices.

The way forward is a shift from grey infrastructure to blue-green resilience. Water-sensitive urban planning must put flood-risk assessment into master plans; sensor-based smart drainage and Airborne Laser Terrain Mapping can give real-time alerts; reviving urban lakes and wetlands restores natural buffers (Biome Environmental Trust's recharge wells cut Bengaluru's flood risk); and sponge cities absorb and reuse rainwater through permeable pavements, rain gardens and green corridors, with Chennai's Sponge Park initiative (2024-25) in the Kovalam basin as India's first successful adaptation and Digital Twins of cities being piloted.

Cloudbursts: concentrated atmospheric fury

A cloudburst is an enormous amount of precipitation in a short period, sometimes with hail and thunder, capable of creating flood conditions. When rainfall of more than about 10 cm per hour is recorded over an area of roughly 20 to 30 sq km, it is classified as a cloudburst event. Cloudbursts represent the concentrated fury of atmospheric instability over mountainous regions. Recent events include Mandi in Himachal Pradesh (2025) and Pithoragarh in Uttarakhand (2024).

The mechanism is orographic and violent. Strong upward air currents trap raindrops inside the cloud, letting them grow until the currents weaken and release the accumulated moisture all at once. When moist monsoon air is forced up Himalayan slopes, it cools rapidly and condenses; narrow valleys concentrate the cloud over a small zone, and steep terrain accelerates the deluge. Global warming loads the atmosphere with more moisture, black-carbon soot from Himalayan forest fires and vehicles provides extra condensation nuclei (IIT Kanpur research), and disturbed monsoon patterns make the bursts less predictable.

A cloudburst's damage is instant and cascading: flash floods, landslides and mudslides destroy homes, roads and bridges, isolate communities, and damage crops and livestock. Critically, cloudbursts in the high Himalayas are now increasingly triggering glacial lake outburst floods: the Sikkim disaster of 2023 began at South Lhonak Lake when a lateral moraine collapsed into the lake. NDMA-aligned measures include Doppler radars and the completed Himalayan Radar Network (2025), a dense grid of local weather stations every 5 to 10 km in high-risk zones, community drills, and keeping construction out of natural drainage paths, with self-draining gabion retaining walls preferred over rigid concrete that bursts under pressure.

The 2022 Amarnath tragedy shows what a cloudburst does to a pilgrimage. In July 2022 a cloudburst near the holy cave shrine in Jammu and Kashmir triggered flash floods that swept through tent colonies and langars along the yatra route, killing pilgrims and underlining how exposed high-altitude religious gatherings are to sudden weather. Earlier cloudburst-linked disruptions on the route had already flagged the risk, but the combination of lakhs of yatris, narrow valleys and near-zero warning time makes the Amarnath Yatra one of India's hardest pilgrimage disaster-management challenges.

The management response is now a template for pilgrimage DM: mandatory registration and daily caps on yatris, weather-based go or no-go advisories from IMD, restrictions on halting near streams and in landslide-prone stretches, pre-positioned NDRF, SDRF and medical teams along both the Baltal and Pahalgam routes, and RFID tracking of pilgrims. The mains lesson is transferable: wherever faith concentrates crowds in hazard-prone terrain, from the Char Dham to the Kumbh, disaster management must be built into the pilgrimage calendar, not bolted on after a tragedy.

GLOFs: Himalayan time bombs

A glacial lake outburst flood (GLOF) is the sudden release of water from a glacial lake when its natural dam, usually made of loose moraine debris or ice, fails. These are high-energy floods: a wall of water, ice and debris that can travel tens of kilometres downstream, destroying everything in its path. The 2013 Kedarnath disaster was partly triggered by such a lake breach after intense rainfall.

The trend data is alarming. Glacial lakes in India expanded by 33.7 per cent between 2011 and 2024, 67 high-risk lakes grew by over 40 per cent in surface area, and the National Remote Sensing Centre mapped 28,043 glacial lakes across the Indus, Ganga and Brahmaputra basins in 2023. The Wadia Institute puts the Himalayan glacier retreat rate at 15 to 20 metres per year; the IPCC's Sixth Assessment Report warns Himalayan glaciers could lose 30 to 50 per cent of their mass by 2100; and ICIMOD projects 70 to 80 per cent of Hindu Kush Himalaya glaciers could vanish by 2100 under high emissions.

GLOFs are classified by what holds the water back:

The defining Indian case is the Sikkim GLOF of 3-4 October 2023. About 14.7 million cubic metres of frozen lateral moraine collapsed into South Lhonak Lake, generating a wave about 20 metres high that breached the moraine dam and sent a flood down the Teesta. The flood destroyed the 1,200 MW Teesta Stage-III dam, killed 55 people (74 missing), washed away bridges and roads, and the project proponent has estimated about Rs 4,189 crore to revive the plant. Studies found no evidence of a cloudburst, and downstream hydropower projects, tunnels and towns multiplied the damage, proving that the GLOF risk is as much about what sits downstream as about the lake itself.

The state's answer is the National GLOF Risk Mitigation Project (NGRMP), approved in 2024 at Rs 150 crore for Arunachal Pradesh, Himachal Pradesh, Sikkim and Uttarakhand, funded largely from the National Disaster Mitigation Fund. It covers hazard and risk assessment, monitoring and early warning, site-specific structural and non-structural mitigation, and community awareness. The Central Water Commission monitors 902 glacial lakes with a risk-indexing framework (and 2,485 lakes larger than 10 hectares by remote sensing from June to October); a 2025 NDMA-led scientific assessment of 189 high-risk lakes classified 56 as very high risk; NDMA's October 2020 guidelines lay down the SOP for averting GLOF and landslide-lake outburst flood threats; and the Wadia Institute of Himalayan Geology maintains detailed lake inventories for Uttarakhand and Himachal Pradesh.

The remaining agenda is clear: real-time early warning on every high-risk lake (no end-to-end GLOF early warning system is commissioned yet), no-construction zones near glacial lakes and in downstream flood paths, mandatory GLOF studies for all Himalayan dams (ordered after Teesta-III), an integrated glacier-monitoring authority linking GSI, IMD and ISRO, and panchayat-level preparedness so that the last mile is not the weakest link.

• CAP-based alert system
• Mission Mausam (2024-26)
• Global cooperation

The February 2021 Chamoli disaster belongs in any GLOF discussion as a cautionary twin. A massive rock-ice avalanche detached in the Ronti Gad catchment of Uttarakhand's Chamoli district, and the resulting debris flow thundered down the Rishiganga and Dhauliganga valleys, destroying the Rishiganga hydroelectric project, severely damaging the Tapovan-Vishnugad project and killing scores of workers. Early reports suspected a glacial lake outburst, but subsequent scientific assessment pointed to a hanging glacier collapse compounded by rockfall, a cascading hazard rather than a classic moraine-dam failure.

The lesson reshaped Himalayan risk thinking: monitoring cannot stop at mapped glacial lakes but must cover hanging glaciers, unstable slopes and the valleys below them, and hydropower projects in such terrain need hazard assessments that account for cascading events, not single triggers. Chamoli is now the standard mains example of why the Himalayas demand multi-hazard assessment rather than single-hazard checklists.

Drought: assessment and management

Drought is the disaster that arrives without drama. Unlike a flood or cyclone, it has no single moment of impact; it is a creeping disaster, a slow-onset deficiency of precipitation that deepens over weeks and months into crop failure, drinking-water scarcity, livestock distress and distress migration. Because its onset is gradual and its geography diffuse, drought rarely commands the urgency of sudden disasters, yet it affects more people over larger areas and for longer durations than almost any other Indian hazard. For UPSC, drought is the test of whether an aspirant understands disasters beyond the spectacular.

Four types of drought are distinguished, and the distinction drives the response. Meteorological drought is a prolonged deficiency of rainfall compared to the long-period average, the trigger condition. Agricultural drought follows when soil moisture falls short of what crops need at critical growth stages, linking rainfall failure to food production. Hydrological drought is the depletion of surface and groundwater, visible in falling reservoir levels and drying wells, and it typically lags the meteorological drought by months. Socio-economic drought occurs when water scarcity disrupts supply systems and livelihoods, the point at which a natural anomaly becomes a human crisis. A good mains answer moves through this sequence rather than treating drought as a single event.

The assessment playbook is the Manual for Drought Management, issued by the Department of Agriculture and Farmers Welfare, which prescribes how drought is to be declared, monitored and relieved. The Manual lays down declaration criteria built on multiple indicators rather than rainfall alone, defines the roles of central and state agencies, and details the relief measures to be activated, from employment generation to cattle care. Its core institutional message is that drought declaration must be evidence-based and timely: late declaration delays relief, and relief delayed is relief denied.

The evidence increasingly comes from space. NADAMS, the National Agricultural Drought Assessment and Monitoring System, is a satellite-based system that assesses agricultural drought using vegetation indices, rainfall data and soil information, with assessments issued at regular intervals during the cropping seasons. Run with the support of the National Remote Sensing Centre, NADAMS lets administrators see crop stress developing in near real time instead of discovering it through crop-cutting experiments months later. It is the drought counterpart of flood forecasting: an early-warning system for a slow disaster.

Assessment rests on a dashboard of indicators, because no single number captures drought. Rainfall deviation and the length of dry spells measure the meteorological shortfall; crop sown area compared to normal reveals farmers' own verdict on the season; vegetation indices from satellites show crop vigour; soil moisture indicates agricultural stress; and groundwater levels and reservoir storage track the hydrological lag. States combine these into drought declarations at district or sub-district level, triggering the relief machinery.

Relief follows a standard repertoire: gratuitous assistance to the most vulnerable, drinking-water supply through tankers and emergency schemes, cattle camps with fodder, and relief employment on public works. The employment pillar links directly to MGNREGA, the Mahatma Gandhi National Rural Employment Guarantee Act, under which additional days of wage employment are provided in drought-notified areas beyond the normal guarantee, converting relief into durable assets like ponds and check dams. Mitigation looks further ahead: watershed development, rainwater harvesting and groundwater recharge to drought-proof landscapes; contingency crop plans prepared by agricultural research institutions that tell farmers what to sow when the monsoon fails or arrives late; crop insurance under the Pradhan Mantri Fasal Bima Yojana to transfer part of the risk; and drought-tolerant crop varieties. The mains synthesis: drought management has moved from relief-centric to a cycle of assessment, early warning, relief and mitigation, with MGNREGA as the bridge between relief and resilience.

The deeper reform is to treat drought as a managed risk rather than a recurring surprise. That means drought-proofing vulnerable districts through assured irrigation where feasible, diversified cropping away from water-guzzling monocultures, fodder banks for livestock, and drinking-water security plans that do not depend on emergency tankers every summer. It also means protecting the assessment system itself: timely data from states, honest reporting of sown areas, and keeping relief politics from distorting declarations. A drought policy that only activates after crops have withered is relief; a policy that keeps the sown area, the soil moisture and the groundwater under watch all year is management.

A related 2020 shock: the desert locust invasion

In 2019-20 India faced its worst desert locust invasion in decades, a biological disaster riding on climatic triggers. Unusual rains in the Arabian Peninsula and the Horn of Africa enabled massive breeding, and swarms moved through Pakistan into Rajasthan, Gujarat, Punjab and beyond, threatening standing crops across lakhs of hectares. A single swarm can contain millions of insects and consume vegetation equivalent to the food needs of thousands of people in a day, making locusts a fast-moving food-security emergency.

Control was a coordinated operation led by the Locust Warning Organisation under the Ministry of Agriculture, working with state governments: ground surveillance teams tracked swarm movements, vehicle-mounted and aerial spraying applied insecticides, drones were deployed for difficult terrain, and India coordinated with Pakistan and the UN Food and Agriculture Organization on cross-border swarm intelligence. The episode's lesson for disaster management is that some disasters arrive from across borders on the wind, and preparedness means surveillance networks and rapid-response capacity, not just stockpiles.

Cloudburst, GLOF and flood: how they differ

Mains answers gain marks from precise distinctions. The three are often confused:

Where India's flood damage concentrates

Flood proneness is national, but flood damage is regional. Of India's roughly 40 million hectares of flood-prone land, about 33% of the country's flood damage occurs in Uttar Pradesh and Uttarakhand, about 27% in Bihar and about 15% in Punjab and Haryana. The middle and lower courses of north Indian rivers such as the Ganga, Brahmaputra, Kosi, Damodar and Mahanadi flood repeatedly because their gradients flatten almost to nothing, while peninsular rivers flood differently: mature channels on hard rock beds have shallow basins that spill quickly.

What triggers a GLOF: six mechanisms

A glacial lake outburst flood (GLOF) is the sudden discharge of a large volume of water stored in a glacial lake. The triggers matter for early warning because each leaves a different signature for monitoring agencies to watch.

UPSC and this topic: PYQ weightage

Water disasters are among the most frequently asked DM topics in the mains. The actual questions:

• 2016 (12.5 marks)
• 2022 (10 marks)
• 2016 (12.5 marks)
• 2024 (15 marks)

Key Terms

• National GLOF Risk Mitigation Project (NGRMP): The National GLOF Risk Mitigation Project, approved in 2024 with an outlay of Rs 150 crore, targets glacial lake outburst flood risk in four Himalayan states: Himachal Pradesh, Uttarakhand, Sikkim and Arunachal Pradesh. The central share comes from the National Disaster Mitigation Fund. Example: the project funds lake-lowering, early warning and mitigation works at high-risk glacial lakes
• flood forecasting and early warning: Flood forecasting and early warning is the system of monitoring rainfall and river levels, modelling inflows and issuing advance alerts to vulnerable areas. The Central Water Commission is India's nodal flood-forecasting agency. Example: CWC's flood forecasts give downstream districts lead time to evacuate before rivers cross danger levels
• glacial lake outburst flood (GLOF): A glacial lake outburst flood is the sudden, catastrophic release of water from a glacial lake when its moraine or ice dam fails, sending a wall of water and debris downstream. Retreating Himalayan glaciers are creating more such lakes. Example: the October 2023 South Lhonak GLOF in Sikkim destroyed the Teesta-III dam

• Chamoli 2021 rock-ice avalanche: The February 2021 Chamoli disaster in Uttarakhand began with a massive rock-ice avalanche in the Ronti Gad catchment, whose debris flow destroyed the Rishiganga project and damaged Tapovan-Vishnugad on the Dhauliganga. Initially suspected to be a glacial lake outburst, it was assessed as a cascading rock-ice failure. Example: Chamoli is now the standard case for multi-hazard assessment in the Himalayas.

• Central Water Commission (CWC): The Central Water Commission is India's apex technical organisation for water resources, which runs the national flood forecasting network and issues inflow forecasts for major reservoirs. Its forecasts let dam authorities plan releases gradually instead of in panic. Example: CWC inflow forecasts feed the rule curves that govern gate operations during the monsoon.

• Manual for Drought Management: The Manual for Drought Management, issued by the Department of Agriculture and Farmers Welfare, prescribes how drought is declared, monitored and relieved in India. It lays down multi-indicator declaration criteria, agency roles and relief measures from employment to cattle care. Example: States invoke the Manual's criteria when notifying drought-affected districts to trigger relief.

• MGNREGA in drought relief: The Mahatma Gandhi National Rural Employment Guarantee Act provides additional days of wage employment in drought-notified areas beyond the normal guarantee, converting drought relief into durable assets like farm ponds and check dams. Example: Drought relief works under MGNREGA simultaneously give income support and build water-harvesting structures that reduce future drought vulnerability.

• Locust Warning Organisation: The Locust Warning Organisation, under the Ministry of Agriculture, is India's nodal agency for locust surveillance and control, coordinating ground teams, spraying operations and cross-border swarm intelligence. Example: During the 2020 invasion it coordinated vehicle-mounted spraying, drones and FAO-linked tracking of swarms.

• Flood Management Programme: The Flood Management Programme is the centrally sponsored scheme funding flood control works such as embankments, drainage and anti-erosion measures in states. It has supported critical protection works along major rivers. Example: embankment strengthening on the Kosi under the programme protects Bihar's flood-prone districts
• Riverine (fluvial) floods: Riverine or fluvial floods are floods caused by rivers overflowing their banks after prolonged rainfall or snowmelt in the catchment. They are slow to build and can inundate vast plains. Example: the annual Brahmaputra floods that submerge Assam's plains are riverine floods

• Amarnath 2022 cloudburst: The July 2022 cloudburst near the Amarnath cave shrine in Jammu and Kashmir triggered flash floods that swept through pilgrim camps along the yatra route, killing yatris. It exposed the vulnerability of high-altitude religious gatherings to sudden weather. Example: The response template now includes registration caps, IMD-based advisories, restricted halting zones and pre-positioned response teams.

• Mission Mausam (2024-26): Mission Mausam (2024-26) is the Ministry of Earth Sciences' mission to make India weather-ready and climate-smart through advanced observation, high-resolution modelling and last-mile dissemination. Example: the mission is upgrading Doppler radars and developing hyperlocal forecasts

Practice questions