Withdrawal began on 19 September with cumulative rainfall still about 15% below normal, and two days later than its normal date, so this was not a monsoon that left early. A Bay of Bengal system will bring heavy rain to the east this week. It misses most of the deficit, and arrives late for the crop already in the ground.
India's southwest monsoon began withdrawing from West Rajasthan on 19 September, two days later than its normal date of 17 September. Cumulative rainfall for the season so far stands at about 15% below normal. Withdrawal has only begun, and the season is not over.
That timing matters, because the intuitive explanation for a short monsoon is that it left too soon. This one did not. Withdrawal onset is not the same thing as season duration, and the problem here was rainfall amount, timing and distribution rather than an early exit.
Now a Bay of Bengal system is arriving. As of IMD's monsoon activity bulletin of 21 September, it is a well marked low pressure area over the Bay, "very likely to become depression and deep depression during the next 48 hours." IMD's bulletin of 20 September had it reaching the north Andhra Pradesh and south Odisha coasts around the morning of 23 September. The intensification is a forecast, not yet an observation. Heavy to very heavy rain is forecast for Odisha and coastal Andhra Pradesh on 23 and 24 September, with widespread rain across Gangetic West Bengal, Jharkhand and Bihar through 26 September.
It will be a substantial rain event. It will not be a monsoon rescue, and the reason is geography before it is anything else.
| Region | Departure from normal | In the depression's path? |
|---|---|---|
| South Peninsula | −28% | Partly, mainly coastal Andhra Pradesh |
| East and Northeast India | −25% | Partly, especially Jharkhand, Bihar and Gangetic West Bengal |
| Northwest India | −10% | No |
| Central India | −6% | Yes, especially Odisha |
| All India | −15% |
Regional departures are as of 13 September; the all-India figure held at −15% through 19 September.
Read that table against the forecast track and the mismatch is hard to miss. In IMD's four-region scheme Odisha sits in Central India, not the east, and Central India carries the smallest deficit of the four at −6%. That is where the heaviest rain is forecast.
The two regions that actually need it do less well. East and Northeast India, down 25%, catches the inland remnant across Jharkhand, Bihar and Gangetic West Bengal. The South Peninsula, worst hit at −28%, is only clipped: north coastal Andhra Pradesh lies in the track, but the interior and the southwest do not, and Kerala is down 23% at the far end of the region from the system.
So the heaviest rain is forecast for the region with the least to make up. Most of the deficit will not be repaired by rain falling on the other side of it.
The benefits are real. Late September rain arrives just before the rabi sowing window, and in the states the system will cross it can recharge reservoirs, lift groundwater and leave a soil-moisture profile that supports wheat, gram and mustard planted from late October. Rabi in eastern India leans heavily on residual soil moisture and stored water, so this is close to ideal timing for that crop.
The same rain does nothing for a groundnut crop in Gujarat or a rain-fed crop in interior Karnataka, because it will not fall there.
The harder limit is not geography, it is time.
Kharif crops were sown in June and July. For many pulses and oilseeds already approaching maturity, moisture stress experienced during flowering and pod filling cannot be reversed, because the number of pods and the weight of the grain are already set. Heavy late rain can instead cause lodging, sprouting, quality losses and harvest delays, and IMD is forecasting isolated extremely heavy rainfall in Odisha and coastal Andhra Pradesh. Some later-planted rice and cotton fields may still benefit, so the effect will vary by crop and by planting date.
So the same weather system is, at the same moment, good for the crop not yet planted and bad for the crop waiting to be cut. That is the whole argument in one sentence, and it is why "monsoon revival" headlines in late September should be read carefully.
India is better insulated this year than the rainfall figures alone suggest. The government's pulses buffer stands at about 4.5 million tonnes, above the 3.5 million tonne norm and well above the 1.8 million tonnes held in May 2025. Duty-free imports of tur and urad run through 31 March 2027, and pulse imports actually fell about 30% in 2025-26.
A buffer caps spikes; it does not create supply. Stocks get drawn down and have to be rebuilt from somewhere. Retail tur and urad were both near Rs 123 a kilogram in September. It is a shock absorber with finite travel, and a second short season would test it.
This season is consistent with El Niño expectations without being a clean demonstration of them, and the monthly sequence shows why. June finished 35% below normal, with the all-India deficit reaching roughly 40% at its worst. Then July delivered a small surplus, 283.3 mm against a 280.5 mm normal, pulling the season deficit back to about 13%. August then came in at 213.3 mm, the seventh lowest August since 2001, and the deficit widened again.
No single seasonal driver produces that shape. A dry June, a normal July and a dry August is the signature of intraseasonal variability: where the monsoon trough sat, how long the break spells lasted, and how many Bay of Bengal systems formed and where they tracked. El Niño loaded the dice toward a dry season and the season came out dry, but the month-to-month pattern was written by other things. We made the same point about Indonesia's fire season, where the regional outcome also turned on more than the Pacific.
The broader lesson is what a seasonal total conceals. A 15% deficit sounds survivable. A 15% deficit made of a very dry June, a normal July and a poor August, concentrated in the South Peninsula and the east, is a worse thing than 15% spread evenly. Timing and distribution do more damage than the headline number shows, and they are the part a seasonal outlook is least able to tell you in advance.
Confirms: IMD's end-of-season figure, due in early October, closes near −15% with the South Peninsula still 25% or more below normal, and the late-September rain shows up in eastern reservoir levels and rabi sowing rather than in kharif output estimates.
Refutes: the all-India deficit narrows by more than about four points between 19 September and the season close, meaning this system delivered far more widely than its track implies. Also refuting: kharif production estimates coming in near normal.
By when: IMD's season-end statement in early October, and the government's first advance estimates of kharif production, normally issued during the autumn.