Extreme rainfall over the hills of Nagaland and vulnerable conditions on the Assam plain both played a part. Whether the rain was a cloudburst, how it compares with the historical record and how much damage was made on the ground are questions that different bodies have answered differently.
In the third week of July 2026, floodwater reached Sivasagar and Charaideo in Upper Assam, districts that the North-Eastern Space Applications Centre (NESAC) describes as historically not flood-prone. By 10 August the death toll stood at about 100, including 13 in Nagaland, according to Mongabay India. The phrase that spread quickly was that Nagaland had triggered the floods. What can be said with confidence, and what cannot, depends on which body is speaking.
What was observed
A meteorological report on the flash-flood situation, whose issuing agency was not identified in the press coverage, recorded extreme rainfall over the Nagaland hill catchments on 18 to 20 July. Nagaland’s average was 85 millimetres, 181 per cent above normal. Mokokchung recorded 238.9 millimetres, 493 per cent above normal, and Zunheboto and Kiphire were 419 and 427 per cent above. On the Assam side, Charaideo recorded 170 millimetres (436 per cent above normal), Golaghat 147.9, Jorhat 105.2 and Sivasagar 93.2.
NESAC traced a landslide near Mon town at about 06:30 on 19 July, on slopes of 30 to 45 degrees made of weathered lateritic material in which rainfall raised pore-water pressure and cut shear strength. It concluded that landslides and rainfall in the upper catchments produced the floods downstream on the Dikhow, Disang and Jhanji rivers. NESAC cites more than 400 millimetres at Mokokchung over the period, well above the 238.9 millimetres in the meteorological report; the coverage does not explain the difference, which may reflect different stations, windows or datasets.
What the authorities said
Chief Minister Himanta Biswa Sarma attributed the disaster to heavy rainfall and a cloudburst in Nagaland, and announced interim relief of ₹15,000 per affected family. A member of the Assam Legislative Assembly, Debabrata Saikia, blamed unchecked sand, stone and coal mining, including open-cast coal mining in Nagaland, and riverbank extraction. These are political statements and claims, not measurements.
What the meteorological agencies reported
The India Meteorological Department documented continuous heavy to very heavy rainfall across Nagaland, Arunachal Pradesh and eastern Assam in July. It ruled out a cloudburst, according to Mongabay India and Border Lens. The distinction has a technical basis: a cloudburst is conventionally an intense burst of 100 millimetres or more of rain in a single hour over a small area. The totals reported here, accumulated over two to three days, describe prolonged extreme rainfall. Hourly data would settle the question, and none has been published in the coverage reviewed.
What scientists concluded
World Weather Attribution published an analysis on 14 August 2026 covering Upper Assam, Nagaland and Arunachal Pradesh from late June to early August. Its team found that the return periods of both the three-day and 30-day rainfall events in the study region were under two years, that observed rainfall showed no trend, and that climate models indicated 1.4°C of human-induced warming had not materially altered the likelihood or intensity of such heavy rain. It cautioned that uncertainty is high because station density in the region is low.
The analysis emphasised exposure and vulnerability: high population density in the Brahmaputra valley, land-use change including unplanned urbanisation and deforestation, and rigid embankments that can worsen flooding by blocking drainage. The Brahmaputra basin researcher Mirza Zulfiqur Rahman told Mongabay that persistent wetland degradation and infrastructure projects have hampered natural drainage.
A report from Assam University, Silchar, with contributors from the International Water Management Institute and other institutions, dated August 2026, reaches a wider conclusion: compound causation. It records that Assam as a whole received 594.4 millimetres between 1 June and 29 July, 29 per cent below the normal 839.6 millimetres, so the flood came despite a statewide deficit. It attributes the event to concentrated hill rainfall after about 72 hours of sustained rain had saturated catchment soils. South-bank tributaries, the Dikhow, Disang, Jhanji and Dhansiri, exceeded danger levels and breached embankments before the Brahmaputra itself peaked, and a release of about 209 cubic metres a second from the Doyang reservoir in Nagaland by 24 July compounded flooding along the Dhansiri in Golaghat.
The same report documents embankment breaches on the Bhogdoi in Jorhat on 20 July and on the Jhanji at Jamuguri, embankments dating from the 1950s to the 1970s, and the shrinkage of the Panidihing wetland from 18.9 square kilometres in 2001 to 5.9 in 2021.
Reconciling extreme rain with an unremarkable return period
The two findings are not contradictory, although they answer different questions. The meteorological figures measure how far specific stations exceeded their normal for three days. The attribution analysis assessed rainfall over a large region, and found such totals to be common. A storm can be extreme for a small steep catchment while remaining ordinary when averaged over a wide area. It is an editorial inference, not a finding, that the hill catchments of Nagaland may have received rain that was locally exceptional but regionally unremarkable. Testing it requires return-period analysis for the catchments themselves.
What remains uncertain
Several questions are open. How much did mining and deforestation in the Naga hills contribute to runoff and sediment? Reports say a high Brahmaputra stage reduced the drainage capacity of its tributaries; the size of that effect is not quantified in the sources reviewed. Public estimates of damage conflict as well: Mongabay reports 119,000 hectares of cropland submerged, the attribution analysis more than 45,000, and the Assam University report about 21,000 damaged, while its count of about 2.12 lakh people affected at the late-July peak sits far below the roughly 10 lakh reported on 26 July elsewhere. The figures probably refer to different dates and definitions, and the university authors call all 2026 impact figures provisional.
Describing the floods as triggered from Nagaland is defensible as a statement about where the rain fell. As a claim about cause, it leaves out the plain, the river stage and the drainage on which the depth and duration of flooding depended.
Venkata Reddy Konatham
