Satellite Counts Say Farm Fires Are Falling. After 3 PM, the Picture Changes.
On 13 October, the agriculture and environment ministers sat down at Krishi Bhawan to review the paddy residue season, and the headline numbers were good. Farm fires in Haryana were down 53.3 per cent between 15 September and 12 October compared with last year, Uttar Pradesh was down about 32 per cent, and Punjab had fallen too. Then the same meeting issued the instruction that should reset how every risk team reads those numbers: stop trusting the satellite count on its own, because fires lit after mid-afternoon are the ones it does not see.
The count improved where the system has run longest
The long trend is real, and it is worth stating plainly before the caveats. Punjab and Haryana together recorded about a 90 per cent reduction in fire incidents in the 2025 paddy season against 2022, according to a ministry reply in the Rajya Sabha. Eight years of machinery subsidies, hiring centres and enforcement have moved the two states that defined this problem.
The IARI consortium that tracks residue burning day by day counted 287 incidents across seven states between 15 September and 6 October in its CREAMS bulletin. By 11 October, the seven-state total in the ministry's review stood at 572. The distribution inside that total is the part that matters:
| State | Farm-fire incidents, 15 Sep–11 Oct 2026 |
|---|---|
| Madhya Pradesh | 277 |
| Uttar Pradesh | 121 |
| Punjab | 118 |
| Rajasthan | 45 |
| Haryana | 7 |
| Uttarakhand | 3 |
| Delhi | 1 |
| Total | 572 |
Source: Agriculture Ministry statement, 13 October 2026. Chart: Farmioc.
Haryana, the state with the deepest machinery base, recorded seven incidents in nearly four weeks. Amritsar district alone recorded 60, the highest of any district in the seven states. One national total, two completely different risk pictures — the same lesson this monsoon taught about rainfall averages.
The blind spot is now official
The environment minister's direction at the review was specific: district collectors are to run ground assessments, and patrolling and flying squads are to concentrate after 3 pm, the window when satellites do not pass over the region. Fire data is to be linked across the air-quality commission and the pollution control board, and states are to track harvest schedules and weather alongside it. This is the government stating, on the record, that its own headline metric has a known hole in it — and that burning has been migrating into the hole.
The ground layer is already built. The Commission for Air Quality Management has flying squads in 34 districts — 18 in Punjab, 16 in Haryana — from 1 October to 30 November, reporting daily to its cell at SAS Nagar and coordinating with the Parali Protection Force and state pollution boards. The five Punjab districts named for closest scrutiny at the review — Sangrur, Tarn Taran, Amritsar, Patiala, Gurdaspur — overlap heavily with that deployment.
For anyone who consumes fire counts as data, the implication is uncomfortable and simple. A satellite series is a floor, not a total. Year-on-year comparisons built on it are still useful — the same blind spot existed last year — but an ESG disclosure, a portfolio air-quality screen or a sourcing audit that treats the detected count as the burned area is overstating its own precision. The honest series this season is satellite detections plus ground reports, and the two will not match.
The hotspot map moved, and the old map is stale
The second surprise in the review was geographic. Madhya Pradesh went from 55 incidents to 277 incidents in the same window, a rise of 452 per cent, and Rajasthan tripled from 20 to 60. In the CREAMS count to 6 October, Madhya Pradesh already held 118 of the 287 incidents — and on 6 October alone it accounted for 24 of 33 detections nationally.
Both states were directed to identify affected districts and villages and produce special action plans immediately. Read that as a market signal, not just an administrative one. Residue-management machinery, hiring-centre capacity and straw off-take infrastructure were all built where the fires used to be. The demand is now surfacing in central India, where the Crop Residue Management scheme's money has barely gone: of the ₹4,266.47 crore released between 2018-19 and 2025-26, Punjab received ₹2,056.50 crore and Haryana ₹1,171.71 crore, while Madhya Pradesh received ₹59 crore. An equipment supplier or a straw-aggregation business planning territory on the 2022 hotspot map is planning for a fire season that has already moved.
Machines were bought. A large share is standing idle.
The review also surfaced the number that should worry lenders who financed this machinery. More than 3.54 lakh machines have been supplied under the scheme, including 26,109 in the current financial year — yet a central survey found roughly 40 per cent of older machines in Punjab and 48 per cent in Haryana idle or non-functional. The agriculture minister directed states to repair the existing fleet alongside distributing new machines, and to activate the hiring centres so smallholders can actually rent what the subsidy bought.
That gap between purchase and utilisation is where the scheme's economics leak. The subsidy structure itself is generous — 50 per cent of machinery cost for individual farmers, 80 per cent for custom hiring centres, and 65 per cent, up to ₹1.50 crore, for paddy straw supply-chain projects. For 2026-27, ₹592.57 crore has been allocated and ₹430.87 crore released, and the Centre flagged slow utilisation, asking Punjab for its pending utilisation certificate. Targets set in June — more than 46,000 machines, 910 new hiring centres and 141 supply-chain projects against an estimated 2.762 crore tonnes of paddy stubble — will be judged against a fleet that is, on the government's own survey, half-parked in the two lead states. Asset quality in agri-equipment loan books follows utilisation, not delivery challans.
The exit the policy is actually betting on
Enforcement fills the satellite gap this season. The structural bet is different: give the straw a buyer. The review pushed direct supply chains from farmers to thermal power plants co-firing straw pellets, compressed biogas plants and second-generation ethanol units, and the agriculture minister raised a system under which farmers could gain from practices including carbon credits. In drought-hit areas, states were told to store residue as livestock fodder instead of letting it burn — a reminder that in a deficient-monsoon year, straw is also feed.
The agronomy lever runs in parallel. Farmers get 20–25 days between paddy harvest and wheat sowing, which is why burning persists: it is the fastest way to clear a field inside that window. Short-duration varieties — Pusa 2090 and PR126 in place of the long-duration Pusa-44 — widen the window itself. Every day added there is worth more than a squad added later, because it removes the reason for the fire rather than policing it.
What to do with a count you cannot fully trust
Treat satellite fire data as a detection series, not a census. Any ESG report, credit note or sourcing audit that cites 2026 fire counts should carry the after-3-pm caveat the ministry itself has now stated, and should triangulate with ground reporting where exposure is concentrated.
Re-map residue risk to central India. Madhya Pradesh and Rajasthan are where incidents are growing and where machinery and off-take capacity are thinnest. Portfolio and territory maps drawn on Punjab–Haryana history are a season out of date.
Underwrite utilisation, not procurement. With 40–48 per cent of the older subsidised fleet idle in the lead states, the credit question on CRM machinery is hours worked and hiring-centre throughput — data that exists at district level for anyone who asks for it.
Watch the straw supply chain as an emerging input market. Pellet co-firing, biogas and 2G ethanol demand, backed by 65 per cent supply-chain subsidies, is converting a disposal problem into a priced commodity in the districts where collection gets organised first.
Farmioc DataStore organises satellite and ground data — supply, demand, yield and harvest — by crop and region for the financial sector and agribusiness. The residue season is a clean test of why both layers matter: the satellite sees the pattern, the ground sees the fires it missed, and a decision made on either one alone is a decision made half-blind.
---
Farmioc is India's agriculture data analytics and agri trading company. Explore DataStore for supply, demand and yield analytics, or PriceHub for commodity market intelligence across 200+ crops.