English · हिन्दी · తెలుగు · தமிழ் · മലയാളം · मराठी · ಕನ್ನಡ

Rainfall, variability and forecasting

Climate of India · Part 3 / 3 · Fact-checked on 8 November 2026 · how we check

Written with AI from the sources listed at the bottom of this page, then fact-checked by a separate AI.

Where rain falls in India and why; the coefficient of variation; droughts and floods; IMD's definitions of heat waves, cold waves and cloudbursts; Köppen's climatic regions of India; the monsoon as a unifying influence; and how IMD forecasts the monsoon against the long period average.

Parts 1 and 2 showed how the monsoon works and how it shapes the seasons. This last part looks at the result on the ground: how much rain different parts of India receive, how dependable that rain is, what happens when it fails or comes in excess, how India's climates can be grouped, and how the India Meteorological Department forecasts the season. These are practical questions. A farmer choosing a sowing date, an engineer planning a city's water and an official handling food stocks all depend on the same rainfall figures, because so much still hangs on the monsoon: how big the harvest is, what villages earn, and how much water is saved for the dry season.

The national average hides huge differences from place to place. Annual rainfall tops 200 cm in three kinds of area: the west coast and the windward slopes of the Western Ghats; the hills of Meghalaya and much of the north-east; and the sub-Himalayan belt of the east, including northern West Bengal. All three receive orographic rainfall where moist monsoon winds run into high ground and are forced to rise. On the southern face of the Khasi hills, Mawsynram and Cherrapunji (Sohra) record some of the highest yearly rainfall anywhere on Earth. A moderate 100 to 200 cm falls over most of the eastern half of the country: the middle and lower Ganga plain, Odisha, Chhattisgarh, the eastern Peninsula and the southern part of Gujarat. These are the areas where rice can usually be grown on rainwater alone.

At the other extreme, less than 50 cm a year falls over western Rajasthan, the Kachchh region of Gujarat and Ladakh. Two things keep western Rajasthan dry: there are no mountains there to push the monsoon air upward, and the Bay of Bengal current reaches it only at the very end of its journey, when little moisture is left. Ladakh is a cold desert, walled off from the monsoon by the Himalaya, where much of the scanty moisture falls as snow. A belt of 50 to 100 cm covers the interior of the Peninsula east of the Western Ghats, where the Deccan lies in the rain-shadow area, together with eastern Rajasthan, Punjab, Haryana and northern Gujarat. The overall picture follows two simple trends: rainfall in the Peninsula falls away from the coasts towards the interior, and in the north it falls as one travels westward across the plains.

The amount of rain matters, but so does how dependable it is. The usual yardstick is the coefficient of variation: the standard deviation of yearly rainfall shown as a percentage of the average rainfall. A low value means the rain is much the same from year to year; a high value means it swings widely. In India the two measures pull in opposite directions. The coefficient stays below roughly 25 per cent in the rain-rich west coast, Western Ghats, north-east and eastern plains, while dry regions such as western Rajasthan, Kachchh and Ladakh have values above 50 per cent. Put simply, a bad monsoon is most likely exactly where people are least able to cope with one. Variability also works within a season, through a late onset, a long break or an early withdrawal.

When the monsoon fails, drought follows. IMD treats an area as having a meteorological drought when its rainfall falls more than 25 per cent short of normal, and a shortfall of more than 50 per cent counts as severe. Agricultural and hydrological drought follow when the shortage reaches soil moisture and crops, and then rivers, reservoirs and groundwater. The areas most prone to drought lie in the low-rainfall, high-variability belts: western Rajasthan and Gujarat, and the rain-shadow interiors of Maharashtra, Karnataka, Telangana and Andhra Pradesh. Poor monsoons across the country, as in 2002 and 2009 and in two weak years running, 2014 and 2015, cut farm output and pushed up food prices, and several of these came in El Niño years. Because a large share of India's farmland still depends on rain rather than irrigation, a failed monsoon reaches well beyond the village.

Too much rain is as damaging as too little. Floods come when heavy monsoon rain falls on catchments that cannot drain it fast enough, and they are worst in the Brahmaputra and Ganga plains, where rivers carry heavy loads of silt, shift their courses and spill over low banks; Assam and Bihar suffer almost every year. Coastal Odisha, Andhra Pradesh and Tamil Nadu flood when cyclones strike, and large cities such as Mumbai and Chennai have flooded when extreme rain fell on built-up land with poor drainage. In 1980 the Rashtriya Barh Ayog (National Flood Commission) put the flood-prone area of India at roughly 40 million hectares. Because the same season can bring floods to one state and drought to another, rainfall statistics are always read region by region, not only as a national total.

IMD also defines extremes of temperature and rain, so that a warning means the same thing everywhere. For the plains, IMD's heat-wave test has two routes: a maximum of 40°C or more that is also at least 4.5°C above normal, or an actual maximum of 45°C or more regardless of the normal; for hill stations the 40°C floor becomes 30°C. A cold wave on the plains is declared when the minimum temperature is 10°C or lower and well below normal, or when it drops to 4°C or lower. Heat waves are most common from March to June over the north-west, central India and the interior of Telangana and Andhra Pradesh, while cold waves strike the north between December and February. A cloudburst is an extreme local downpour of about 100 mm or more in an hour over a small area; it is typical of Himalayan slopes, where it sets off flash floods and landslides.

Climatic regions bring order to this variety. The Köppen climate classification, applied to India, uses monthly temperature and rainfall to sort the country into a few types, each written as a short code of letters. The opening capital letter names the broad group: A means tropical and warm all year, B dry, C warm with a cool season, D cold-wintered, and E the coldest highland climates. Small letters add the season of rain or drought, such as 'w' for a dry winter and 's' for a dry summer. So most of the Peninsula is Aw, a tropical savanna with a long dry season, while the Coromandel coast, which gets much of its rain in the cooler months, is As. The table below lists the main types and where they occur.

For all these differences, the monsoon gives the subcontinent a common rhythm. Almost everywhere the farming year turns on the arrival and departure of the rains, and most of India receives the bulk of its rain in the same few months of the year. The same seasonal reversal of winds controls the whole region, and a good or bad monsoon is felt across the country as a national event, in harvests, prices and the level of water in reservoirs. The Himalaya make this possible: the range keeps out the bitterly cold winter winds of Central Asia and, in summer, traps the rain-bearing monsoon air within the subcontinent. So the monsoon is seen as something that binds the country together, though it is also a source of risk, because its arrival, breaks and departure change from one year to the next.

Forecasting this uncertain system is the task of the India Meteorological Department, set up in 1875 and now part of the Ministry of Earth Sciences, with its headquarters in New Delhi. IMD measures each season's rain against the long period average (LPA) (LPA): the average June-to-September rainfall over the whole country across a fixed run of years. IMD now calculates the LPA over 1971–2020, which gives a figure of about 87 cm. A season is then placed in one of five categories according to how its rain compares with the LPA, from deficient at the bottom to excess at the top, as shown in the table below. The categories matter beyond meteorology: a forecast of a below-normal season, for example, prompts governments to plan seed supplies, contingency crops and food stocks well before the rains arrive.

IMD issues its first long-range forecast for the south-west monsoon in mid-April and updates it around the end of May, adding forecasts for individual months and for broad regions of the country. Since 2021 it has used a multi-model ensemble approach, which combines the output of several dynamical climate models instead of relying on a single statistical formula. The dynamical models came out of the National Monsoon Mission, which the Ministry of Earth Sciences started in 2012 and which built a coupled ocean–atmosphere forecasting model for India at the Indian Institute of Tropical Meteorology, Pune. Forecasts must weigh El Niño, the Indian Ocean Dipole and the other signals described in Part 1, and they carry a stated margin of error, so a 'normal' forecast is a probability rather than a promise. Within the season, IMD also issues forecasts a few weeks ahead and daily warnings for heat, heavy rain and cyclones.

Taken together, the three parts of this chapter move from cause to effect. Part 1 explained the machinery: pressure and winds, the ITCZ, the jet streams, Tibet, El Niño and the Indian Ocean Dipole. Part 2 followed that machinery through four seasons, with their local storms, the two branches of the south-west monsoon, the north-east monsoon and cyclones. This part has shown the outcome: heavy rain where moist winds meet mountains, dry lands in the rain shadows and the far west, rainfall that is least reliable where it is scarcest, the twin risks of drought and flood, the climatic types that sum up the pattern, and the forecasting system that tries to see each season coming. The same monsoon that gives India a shared calendar also leaves its farms and water supply exposed, which is why the subject links geography with the economy and disaster management.

Exam traps

• The coefficient of variation is HIGH where rainfall is LOW (western Rajasthan, Kachchh, Ladakh) and low where rainfall is high. A high value means LESS reliable rain.
• IMD's long period average for the south-west monsoon is now based on 1971–2020 and is about 87 cm; it refers to June–September rainfall over the whole country, not the annual total.
• 'Normal' is 96–104% of the LPA, not 90–110%. Below 90% is deficient; above 110% is excess; 90–95% is below normal and 105–110% above normal.
• IMD's meteorological drought: rainfall more than 25% below normal; more than 50% below normal is severe.
• Heat wave on the plains: at least 40°C and 4.5°C or more above normal, or 45°C regardless of normal; for hill stations the threshold is 30°C, not 40°C.
• In Köppen's scheme the Coromandel coast is 'As' (rain in the cooler months, dry summer); the west coast south of Goa is 'Amw'. Do not swap them.
• The under-50 cm belt is western Rajasthan, Kachchh and Ladakh; the rain-shadow Deccan generally gets 50–100 cm, not under 50 cm.
• IMD was set up in 1875 and is under the Ministry of Earth Sciences; the National Monsoon Mission (2012) is a Ministry of Earth Sciences programme, not an ISRO one.

Recap

• Over 200 cm: west coast and Western Ghats, Meghalaya and the north-east, the eastern sub-Himalaya (orographic rain); Mawsynram and Cherrapunji among the wettest places on Earth.
• Under 50 cm: western Rajasthan, Kachchh, Ladakh (cold desert); 50–100 cm: rain-shadow Deccan, eastern Rajasthan, Punjab, Haryana.
• Rain decreases from the coasts inland and from east to west across the northern plains.
• Coefficient of variation: under about 25% in wet areas, over 50% in dry areas; the driest places have the least reliable rain.
• Meteorological drought: more than 25% below normal (over 50% severe); floods worst in the Brahmaputra and Ganga plains; about 40 million ha flood-prone (Rashtriya Barh Ayog, 1980).
• Heat wave on the plains: 40°C and 4.5°C above normal, or 45°C; hills 30°C. Cloudburst: about 100 mm or more in an hour over a small area.
• Köppen in India: Amw west coast, As Coromandel, Aw most of the Peninsula, BWhw western Rajasthan, Cwg Ganga plain, Dfc Arunachal, E high Himalaya.
• IMD (1875, Ministry of Earth Sciences): LPA about 87 cm (1971–2020); normal = 96–104%; first forecast mid-April, update end of May; multi-model ensemble since 2021; Monsoon Mission 2012.

Practice questions

Sources

• India Meteorological Department, Ministry of Earth Sciences: rainfall statistics of India, long-range forecast bulletins for the south-west monsoon and the long period average (1971–2020), forecast categories, and definitions of heat wave, cold wave, cloudburst and meteorological drought — https://mausam.imd.gov.in
• Indian Institute of Tropical Meteorology, Pune: monsoon variability and the Monsoon Mission coupled forecast model — https://www.tropmet.res.in
• Ministry of Earth Sciences: the National Monsoon Mission (launched 2012) and the multi-model ensemble forecasting system — https://moes.gov.in
• National Disaster Management Authority: guidelines on drought, floods and heat waves, and the flood-prone area assessed by the Rashtriya Barh Ayog (1980) — https://ndma.gov.in