Thirteen Percent Down, a Hundred Stories Up: Monsoon 2026's Patchwork of Extremes

The 2026 southwest monsoon is over, and it goes into the record as 13% below normal — a below-normal season, but not a drought year, and the kind of number that fits in a headline and settles nothing. Because −13% is an average, and an average is a story about a country that does not exist. Resolve the same four months to 724 districts and the season runs from −91% to +104%. Fourteen states contain both a district that lost most of its rain and a district that got half as much again as it should have. Uttar Pradesh finished the season at +8% while holding both. Below are the two maps that show it, and the four-act story of how a catastrophic June turned into a merely disappointing year.

First, the Picture Everybody Publishes

Seasonal rainfall departure by state, 01 Jun – 29 Sep 2026, against IMD's published long-period normal · IMD departure classes and IMD's standard rainfall colours · boundaries are state polygons, not IMD meteorological subdivisions

At state resolution the season reads as a familiar geography of disappointment. The northeast carries the worst of it — Meghalaya −59.6%, Manipur −44.7%, Arunachal Pradesh −43.6% — which is remarkable in a region whose problem is normally too much rain, not too little. The other deep red is Punjab at −41.7%, and the southern peninsula is uniformly short: Karnataka −30.0%, Kerala −27.1%, Andhra Pradesh −23.2%, Tamil Nadu −18.3%.

Against that, only two patches of blue matter. Odisha finished +30.5%, the standout surplus of the season, and the Ganga plain held up — Uttar Pradesh +8.2%, Uttarakhand +7.6%, West Bengal +2.9%. Ladakh's +122% looks dramatic and means very little: its seasonal normal is around 22 mm, so a few extra showers swing the percentage wildly. Percentages on tiny normals are the most over-read number in Indian rainfall reporting.

Even here the national figure is already misleading. Fifteen of thirty-seven states and union territories finished Deficient or worse. The country as a whole did not.

Now Zoom In, and the Average Falls Apart

The same season, the same data, resolved to 724 districts · IMD departure classes and colours · 01 Jun – 29 Sep 2026 · 31 districts have no published IMD figure and are hatched

This is the map that should be published alongside the number. The smooth state blocks dissolve into a mosaic. A corridor of blue runs inland from the north Bay of Bengal coast — Odisha, Jharkhand, eastern and central Uttar Pradesh, Bundelkhand — and almost everything west and south of it is red. Deficit and surplus sit side by side across district lines, sometimes within sight of each other.

IMD class	Departure	Districts	Share
Large Deficient	−60% and below	35	4.8%
Deficient	−20% to −59%	280	38.7%
Normal	−19% to +19%	307	42.4%
Excess	+20% to +59%	87	12.0%
Large Excess	+60% and above	15	2.1%

724 districts with a published IMD figure, 01 Jun – 29 Sep 2026.

Fewer than half the districts in India — 42% — actually had a normal season. Nearly 44% finished Deficient or Large Deficient, and 14% finished in surplus. The single national figure of −13% describes the arithmetic of that distribution; it describes the lived experience of almost nobody.

The extremes are worth naming. Ramanagara in Karnataka finished −91.5%, with 39 mm of rain across four monsoon months. Churachandpur in Manipur −79.8%, West Jaintia Hills in Meghalaya −79.3%, Mansa in Punjab −72.0%. At the other end, Gautam Buddha Nagar in Uttar Pradesh took +103.7%, with Sambhal, Mahoba, Hamirpur, Kanpur Nagar and Banda all between +77% and +97% — the Bundelkhand and central-UP belt doubled its season. Two districts sit higher still — Bageshwar at +163% and Tirunelveli at +338% — but both have small seasonal normals, and those percentages rest on modest absolute totals.

The Season in Four Acts

The other thing the seasonal total hides is that this was not one season. It was a disaster followed by a rescue.

Month	All-India departure	What drove it
June	−32.9%	A slow, stalling advance and almost no organised systems
July	+3.6%	Bay of Bengal lows and depressions begin forming and tracking inland
August	−12.8%	A weaker, more intermittent system train
September	−7.2%	Late systems, including the Deep Depression of 23–25 September
Season	−12.7%	Rounds to the published −13%

All-India observed departures, VayuMet Monsoon Monitor; September to 29 Sep.

A −33% June is drought territory. Had the rest of the season continued at that rate, 2026 would be remembered very differently. What changed in July was the Bay of Bengal. The surplus corridor on the district map is not a coincidence of geography — it is a track. Low pressure areas and depressions forming over the north Bay and moving west-northwest across Odisha, Chhattisgarh, Madhya Pradesh and up the Ganga plain lay rain down in a band, and that band is precisely where the blue sits.

The recovery was not a general strengthening of the monsoon — the southern peninsula and the northwest stayed dry throughout. It was narrow, system-driven rainfall along a repeated track. Districts under that track did well; districts a few hundred kilometres either side did not. That is how a country ends up with a −13% average and a 195-point spread.

What This Season Should Change

Intraseasonal variability did the work, not the seasonal mean. A forecast that called 2026 "slightly below normal" would have been right and useless. Sowing windows, reservoir draw-down and insurance triggers turned on a June that failed and a July that rescued it — not on a number that only existed once the season was over.

The unit has to be the district, not the state. A state average of +8% and a district reality of −68% cannot guide the same decision, and the state average is the one that gets quoted. And in a season carried by a few Bay systems on a repeated track, the difference between a good year and a bad one is largely which side of the track you farm.

Comments