The story · 1981 to 2025
Five things NASA data tell us about South Asia
Each finding is explained in simple words first. The exact numbers are there too, for anyone who wants them.
South Asia is getting warmer, everywhere
Since 1981, the whole region has warmed by about +1.4 °C. That is a steady climb of about a third of a degree every 10 years.
All 234 squares on our map got warmer, and not one got cooler. 2024 was the hottest year on record.
Very hot months, more than 1 °C above normal, used to come about 2 times a year in the 1980s. Now it is about 8 months out of 12.
Gases from burning coal, oil and gas trap heat, like a blanket around the Earth.
More dangerous heatwaves for workers and farmers, and drier land that catches fire more easily.
Solid = clear change · faded = no clear change
For scientists: the exact method and numbersOpenClose
Spring is heating up fastest, especially in the mountains
March to May is already the hottest, driest time of year, just before the monsoon. That is exactly when the warming is fastest.
Mountains warm fastest of all: the Western Himalaya about 4 times faster than the Sri Lanka central highlands, where the sea keeps temperatures steady.
In the mountains, snow melts earlier and the dark ground underneath soaks up more sunlight.
Longer fire seasons, heat damage to wheat just before harvest, and glaciers melting faster.
°C every 10 years. Faded bars: no clear change.
For scientists: the exact method and numbersOpenClose
The monsoon rain is moving, not disappearing
Add up all the monsoon rain over South Asia and the total has hardly changed. But where it falls has changed: the dry northwest (the Indus plain) gets more, while the Ganges-Brahmaputra-Meghna delta and nearby wet regions in the east get less.
This is the big idea of our project: one warming world, but opposite changes in different places.
We checked this with a second, independent rain record. It agrees for the Ganges-Brahmaputra-Meghna delta, Indus plain, Northeast India and Myanmar hills. For some mountain regions the two records disagree, so we don't claim a rain trend there.
Air pollution can weaken the monsoon winds in the east, and warmer seas push more moisture to the northwest. Scientists are still working out the exact mix.
Bigger floods on the crowded plains of Pakistan, and less water for rice farms and rivers in the east.
Solid = clear change · faded = no clear change
% change in monsoon rain every 10 years. Faded bars: no clear change.
For scientists: the exact method and numbersOpenClose
Disasters follow the weather
We lined up the worst disaster years with the weather in the same season. Hot springs came with more forest fires in Central India. Rainy monsoons came with more landslides in the Western Himalaya and more floods on the Indus plain.
Just as useful: in some places the weather does not explain disasters. The Bengal delta keeps flooding even though its own rain is falling, because its floods come from rivers upstream.
Heat dries forests into fuel. Heavy rain soaks steep slopes until they slide, and floods flat land.
Where disasters follow the weather, weather records can warn us early. Where they don't, other causes need attention.
Top: disasters each year; the dark bars are the worst years. Bottom: the weather that season, with those same worst years shaded and marked with big dots. When the big dots sit near the top, bad weather and bad disaster years came together.
For scientists: the exact method and numbersOpenClose
Where to prepare in 2025
In 2025, monsoon rain on the Indus plain was higher than in 87% of years since 1981, and in the Western Himalaya higher than in 96%. Both look like past disaster years.
And in 2025, both regions did suffer widely reported floods and landslides.
We only compare places where the weather and disasters are clearly linked, so the check means something.
Authorities can prepare early: check risky slopes, stock boats and medicine, and warn people. It is not a forecast.
A dot to the right of the line means this year looks like the worst years.
For scientists: the exact method and numbersOpenClose
Our take
The warming is certain, and it is everywhere. What changes from place to place is how it shows up: drying here, heavier rain there, fiercer fire seasons somewhere else. That is why local, season-by-season trends matter more than one big average.
What we can't say (yet)
- • Each map square is 200–280 km wide, so one valley or city can be different.
- • The landslide record covers only 11 years and comes from news reports.
- • Satellites miss fires under clouds and smoke.
- • A link is not a cause: land use, roads and people also play a part.
Words explained
- Trend
- The long-term direction something is moving in, ignoring the ups and downs from one year to the next.
- Clear change (significant)
- A change that is very unlikely to be luck. It does not mean big, only real.
- No clear change
- The ups and downs are too big to tell. It does not prove nothing is changing.
- Every 10 years
- How much something changes per decade. +0.3 °C every 10 years is about 1 °C every 33 years.
- Warmer than normal
- Compared with the average of 1951–1980, a common starting point for climate records.
- Link (correlation)
- Two things that rise and fall together. It is a clue, not proof that one causes the other.
- Hot season / Rainy season
- March–May, the hot dry weeks before the rains / June–September, the monsoon.
- Map square
- The data divides the map into squares about 200–280 km wide; each has one value per year.