When Upper Assam Flooded Differently

Floods are a familiar part of life in Assam. This is exactly why the July 2026 flood in Upper Assam deserves a closer look.
This was not Assam’s largest recent flood. In 2022, according to the Assam State Disaster Management Authority, about 88.5 lakh people were affected, and 181 people lost their lives to flooding. In 2020, approximately 57.89 lakh people across 30 districts were affected, with 125 flood deaths recorded. (ASDMA)

At the height of the July 2026 episode, the scale was much smaller statewide. On July 24, around 7.21 lakh people were affected across 11 districts. Yet nearly 94 per cent of them were concentrated in just three Upper Assam districts—Sivasagar, Charaideo and Jorhat. Sivasagar alone accounted for more than 3.75 lakh affected people. (Akashvani News)
By August 3, the immediate flood situation had improved considerably, with about 1.28 lakh people still affected across seven districts, although the reported flood death toll had risen to 87. (The Times of India)
These numbers highlight the first key difference.
What made the July 2026 flood unusual was not its extent, but its sudden, intense, and focused nature in Upper Assam.
This difference is sometimes overlooked when people try to explain what happened.
Extreme rainfall has been blamed. So have deforestation, coal and stone mining, siltation, blocked drainage, roads, embankments, construction in the hills and hydroelectric projects. Satellite images have circulated as apparent proof of particular causes.
Some of these concerns are legitimate. The difficulty begins when a plausible mechanism is converted into demonstrated causation.
So far, the evidence points to a more complex and scientifically helpful explanation:
An exceptional meteorological event appears to have generated an exceptional flood pulse, which then travelled through a river and floodplain system whose vulnerability may itself have been altered.

The question is therefore not simply whether nature or humans caused the flood.
It is how much of what happened was contained in the rain itself, and how much may have been amplified by the landscape through which that rain became a flood. 
 

To understand the flood, we need to look upstream

The investigation should not begin in Sivasagar town.
The Dikhow and Desang do not originate there. Their behaviour reflects rainfall across much larger catchments extending into the Nagaland hills.

Between July 18 and 20, parts of those hills received extraordinary rainfall. According to IMD data cited in a contemporary analysis, Wokha received around 239 mm over the three days, Mokokchung 167 mm, Mangkolemba 211 mm and Longleng 152.2 mm. Sonari in Charaideo reportedly received about 190 mm of rain overnight. (India Today NE)

The meteorological context is just as important as the amount of rain that fell.
The rainfall fell on catchments that had already received substantial pre-monsoon rain. Reports drawing on meteorological observations described the soils and hills as already saturated following heavy rainfall in April and May. (India Today NE)
A catchment that is already soaked will react differently than one that is dry.

Rainfall on relatively dry soil can infiltrate, be temporarily retained, or gradually move towards streams. Once the soil approaches saturation, a much larger fraction of additional rainfall can become rapid surface and near-surface runoff.
So the relevant relationship is not simply:heavy rain = flood.

It is closer to:rainfall intensity × duration × spatial concentration × antecedent saturation × catchment characteristics = runoff response. 
 

Several of those variables aligned in July.

The event was initially described in some quarters as a “cloudburst” in Nagaland’s Mon district. The terminology does not appear to be technically correct. A scientist at the Meteorological Centre in Kohima told The Indian Express that Mon received around 137 mm over eight to nine hours. The IMD definition of a cloudburst requires more than 100 mm of rainfall in 1 hour over a small area. (The Indian Express)

This difference might seem minor, but it shows a bigger issue in how we talk about disasters. Sometimes, dramatic words take the place of careful measurement.
The rain need not have been a cloudburst to be exceptional.

The Climate Context

Climate change belongs in the picture, but it should not become another single-cause explanation. The IPCC concludes that human-induced warming is intensifying the global water cycle and increasing the frequency and intensity of heavy rainfall in many regions. A warmer atmosphere can hold roughly 7 per cent more moisture for every 1°C of warming, creating conditions in which heavy rainfall can become more intense. (IPCC AR6)

That does not mean climate change “caused” the July 2026 Upper Assam flood. It means such extreme rainfall now occurs against a changing climatic background. Whether climate change made this particular rainfall event more likely or more intense would require a formal attribution study.

And rainfall is only the beginning of a flood. How much water eventually reaches homes and fields also depends on how wet the ground already is, how the catchment converts rain into runoff, the condition of rivers and floodplains, downstream water levels and human alteration of the landscape. Climate change may therefore influence the meteorological trigger; it does not, by itself, explain the hydrological outcome.

The rivers provide independent evidence

There is an important reason not to explain the Sivasagar catastrophe primarily through urban drainage failure, a local embankment problem, or a single excavation site.
The anomaly was already visible upstream.

On July 20, the Desang at Desangpani was reported at 106.36 metres—0.10 metres above its previous highest flood level of 106.26 metres. The same Central Water Commission-linked report showed the Dikhow at Sivasagar in an extreme flood situation, with levels rising. (news.webindia123.com)

Elsewhere along the system, the Dikhow at Bihubar exceeded its previous highest flood level. At the same time, the Desang at Nanglamoraghat subsequently reached approximately 96.52 metres, exceeding its previous record of about 96.49 metres set in 1998.

What matters is not just a few extra centimetres, but the pattern: several gauges on different rivers were all showing extreme or record levels at the same time.
This makes it much less likely that the flood was caused solely by local factors in Sivasagar town.
Something exceptional was already coming out of the catchments.
This strongly suggests an unusual rainfall and runoff event.
But it does not end the investigation.

Has the river itself changed?

The historical Dikhow record raises a second question.
High-water records that have endured for long periods have been approached or exceeded repeatedly in recent years. That pattern is worth investigating.
It does not, by itself, prove climate change.
It does not prove mining.
It does not prove deforestation.
It does not prove siltation.
Instead, it gives us a clue that needs explanation.
Here, an elementary hydrological distinction becomes surprisingly important.
A river gauge measures water level. It does not directly tell us how much water is passing through the channel.
If sedimentation, narrowing or morphological alteration makes a river channel less efficient, the same discharge can produce a higher water level. Conversely, an unchanged channel carrying genuinely unprecedented discharge will also produce a record level.
These are two very different reasons for seeing the same reading on the river gauge.

The decisive question is therefore not merely:
Is the Dikhow getting higher?Does the same discharge now produce a higher river level
 

 than it did twenty years ago?

If the answer is yes, something about the river channel or floodplain hydraulics has changed.

And there is another, equally important question:
Does comparable catchment rainfall now produce a larger or faster discharge peak?
If the answer is yes, the catchment itself is responding differently.
Historical discharge records, rainfall data, river cross-sections and rating curves—the relationship between discharge and gauge height—could distinguish these possibilities.
Right now, there is not enough public evidence to answer this for sure.

A useful piece of negative evidence

One might assume that record-breaking river levels mean progressively larger parts of Sivasagar are being flooded.
The satellite record does not show such a simple progression.
The National Remote Sensing Centre and ISRO have produced a Flood Hazard Zonation Atlas of Assam using multi-sensor satellite observations covering 1998 to 2023. The purpose is precisely to use historical inundation patterns to identify flood-prone areas and support risk assessment. (nrsc.gov.in)

The historical record contains numerous major flood years rather than a straightforward march towards ever-increasing inundation.

This is a helpful reminder not to assume that Sivasagar floods more each year just because the environment is getting worse.
There is, however, an important methodological limitation. NRSC itself notes that satellite overpasses do not always coincide with the flood peak; some low-lying rainwater accumulation may be included, while flash or minor flooding may be missed. Actual flooded area can therefore exceed the area captured in individual satellite observations.

The defensible conclusion is consequently modest:
Available satellite observations do not demonstrate a simple monotonic progression towards ever-greater annual inundation.
The peculiarity of 2026 may instead lie in where the water went, how quickly it arrived, how high particular tributaries rose and what lay in its path.

A vast flood slowly occupying familiar agricultural floodplain is one type of disaster.
A rapidly rising flood entering towns, roads, railway infrastructure and settlements regarded as comparatively safe is another.
Area alone does not measure destructive intensity.

Was the Brahmaputra blocking the exit?

Another explanation deserves careful consideration.
When a tributary approaches a high-elevation receiving river, its drainage capacity can be reduced. The smaller hydraulic gradient may produce a backwater effect, slowing drainage and increasing or prolonging upstream flooding.
The Brahmaputra was indeed elevated during the critical period.
At 6 am on July 20, the Brahmaputra at Neamatighat was reported at 85.88 metres, around 0.34 metres above its danger level of 85.54 metres. (news.webindia123.com)

This makes a backwater contribution hydrologically plausible.
But the evidence does not establish that it was the principal engine of the flood.
At the same time the Brahmaputra was above the danger level, the Dikhow and Desang were experiencing extreme, record-breaking conditions upstream.

The safest interpretation is therefore:

The elevated Brahmaputra probably constrained tributary drainage and amplified inundation, but its quantitative contribution has not yet been established.

This is exactly the kind of detail that should be kept in mind when figuring out what caused the flood.
Several mechanisms can operate simultaneously.

And then there is the dam question

Water was released from a dam during the wider Assam flood episode.
That fact matters.

But geography and chronology matter more.

The operational Doyang Hydroelectric Project in Wokha is a 75 MW NEEPCO installation. CEA records identify Doyang as an operating project on the Doyang river system. (Central Electricity Authority)

By July 23, reports confirmed that water from the Doyang project was being released as reservoir levels rose, prompting Assam authorities to prepare for further flooding downstream. (The Print)

The Doyang system, however, drains towards the Dhansiri and Golaghat, not down the Dikhow into Sivasagar.
The chronology is equally revealing.

Severe flooding in Sivasagar, Charaideo and Jorhat was already well established by July 20–22. More than three lakh people were affected by July 21–22, rivers were above danger levels, and large-scale evacuations were underway before the documented Doyang release became a downstream concern. (The Indian Express)

Thus, two statements that initially appear contradictory can both be true:
Water was released from a dam during the Assam floods.

And:
The documented Doyang release does not explain the preceding Dikhow–Desang flood in Sivasagar and Charaideo.
Confusing those two statements creates a causal relationship that neither geography nor chronology supports.

What about the Dikhu project?

The proposed 186 MW Dikhu Hydroelectric Project raises a different question.
Official Central Electricity Authority documentation classifies the project as in that had received concurrence but wasyet to be taken up for construction. The CEA record describes ongoing problems with land acquisition, project redesign, and the possible reduction in installed capacity, ultimately noting that a recast DPR and fresh concurrence would be required. (Central Electricity Authority)

That documentation does not support the existence of an operational impounding reservoir capable of making a conventional flood release in July 2026.
The appropriate formulation is therefore not that one can prove that no project-related environmental effect of any kind exists.

It is much narrower:
No evidence has been identified of an operational Dikhu reservoir capable of producing the alleged July 2026 dam-release flood.

Construction, excavation or other development activity visible in satellite images may raise separate environmental questions.
It is not evidence of reservoir release.

Mining deserves more serious attention.

One of those separate questions concerns mining and riverbed extraction.
Unlike some explanations assembled after the disaster, concern about extraction along the Dikhow predates July 2026.
Reporting based on Assam Water Resources Department records says officials had warned more than four years earlier that indiscriminate sand and stone excavation along the Dikhow could alter the river’s natural course and increase downstream risk.
(The Times of India)

This is important.
A warning made before a disaster has a different evidentiary status from an accusation invented afterwards.
Riverbed extraction can alter channel depth, width, bank stability, sediment transport and local flow distribution. Hill mining and quarrying may also increase erosion and sediment delivery.
But mechanism is still not attribution.

The correct sequence is not:
Mining occurred → flooding occurred → mining caused the flood.
It is:
Where did extraction occur?
→ How did channel geometry change?
→ Did sediment delivery change?
→ Did the discharge–water-level relationship change?
→ Did downstream inundation behaviour change?

Those are measurable questions.
Until these questions are answered, mining should be seen as a possible factor that contributed to the flood, not as the main cause.

Recent claims that illegal mining and tree felling intensified the flood should therefore be treated as hypotheses requiring hydrological testing, not as established findings merely because the mechanisms are plausible. (The Times of India)

The same applies to deforestation

Forest loss can alter infiltration, erosion, soil stability and runoff.
But saying that vegetation has been lost somewhere in Nagaland does not demonstrate that such loss caused the Dikhow or Desang to behave as they did in July.
Administrative boundaries are hydrologically unhelpful.
A hectare of forest lost in another watershed is irrelevant to the Dikhow flood simply because it lies within Nagaland.
The correct unit of analysis is the contributing watershed.
The questions are therefore spatial:
Where exactly has vegetation been removed?
Is it within the Dikhow or Desang sub-catchment?
Was it on steep slopes or near headwater channels?
What replaced it?
Did measured runoff or sediment yield subsequently change?

Contemporary expert analysis suggests that deforestation, wetland loss, road construction, and excavation could have amplified the flood. Those mechanisms are credible, but their individual contributions in July 2026 have not been quantified. (India Today NE)

This difference should stay at the centre of the discussion.

The mud may tell us something the water cannot

One feature of the aftermath deserves much greater attention: sediment.
After floodwaters receded, some severely affected areas were reportedly left under several feet of sediment, making access difficult even after the water had receded. (The Indian Express)

This detail could provide important clues.
A flood transports not only water but also the physical history of what happened upstream.
An unusually sediment-rich flood might reflect landslides, bank collapse, disturbed slopes, road cutting, quarrying, mining or other erosion.

Sediment quantity, particle characteristics, and mineralogical composition can help identify the origin of that material.
The 2026 event should therefore eventually be reconstructed not merely as a water pulse, but potentially as a sediment pulse.
This might tell us more about human impact than just looking at river height.

 

But once again, one must be precise.
Heavy post-flood sediment deposition demonstrates that large quantities of sediment moved through the system.
It does not, by itself, demonstrate that pre-existing sedimentation had already reduced channel capacity before the flood.
Those are different hypotheses.

What should happen now?

The best approach is not to ignore environmental worries or blame someone before we have measured what is happening.
Upper Assam needs a more integrated basin-observation system.
Rainfall gauges and telemetry should be dense enough across the Nagaland–Upper Assam catchments to detect intense upstream rainfall in real time. River stage and, crucially, discharge should be monitored together. Rating curves and channel cross-sections should be periodically updated rather than assumed to remain constant.
Sediment needs to be measured systematically.

Where major sediment deposition occurs after floods, its amount and likely source should be investigated. Repeated cross-sectional surveys could reveal whether river channels are losing conveyance capacity.
Riverbed extraction should be regulated as a hydrological intervention, not merely as an administrative permit.
Likewise, restoration programmes should be geographically intelligent. Planting trees anywhere within a district does not equate to restoring hydrological function in a high-runoff sub-catchment. Wetlands, abandoned channels and natural flood-storage areas need similar recognition.

Urban drainage requires the same systems approach.
Sivasagar’s drains, palaeochannels, roads, embankments, floodplain depressions and neighbouring rivers are not independent entities. A drainage network adequate for ordinary monsoon rain may fail when intense local rainfall coincides with a tributary already near record level.

Reservoir operations should also be transparent.
For major flood periods, reservoir level, inflow, turbine discharge, spillway discharge and gate-opening times should be publicly available in near real time.

This would help us spot real dam-related effects and quickly rule out false claims.

The flood has already performed the experiment

Ultimately, remarkably few kinds of data could settle much of the present argument.

First, reconstruct catchment-weighted rainfall for major historical Dikhow and Desang floods.
Second, match that rainfall to river discharge, not just water level.
Third, recover historical rating curves and cross-sections.
Fourth, map changes in forests, wetlands, mining, roads, and channels within the actual contributing watersheds.
Then ask two deceptively simple questions.

Does comparable rainfall now produce greater or faster discharge?

If yes, the catchment response has changed.
And:
Does comparable discharge now produce a higher river level or greater inundation?
If yes, channel or floodplain hydraulics have changed.
If neither relationship has materially shifted, then much more of the extraordinary July 2026 response was contained in the meteorological event itself.
This is how we move from guessing to actually understanding what caused the flood.

What, then, was different about 2026?

The statewide comparison made at the outset guards against exaggeration: the July 2026 episode was much smaller in population reach than Assam’s major recent flood years. Its significance lies elsewhere—in the concentration, speed and hydrological behaviour of the Upper Assam event.
The event was extraordinarily concentrated in Upper Assam. The Dikhow and Desang responded rapidly. Multiple river gauges reached extreme or record conditions. Places perceived locally as relatively safe were inundated. Roads, railway infrastructure, villages and towns were affected in a geographically compressed disaster.
That makes the scientifically stronger description:
one of the most hydrologically unusual Upper Assam flood events in the modern observational period.
And perhaps that is why the desire for a single explanation has been so powerful.

A dam did it.
Mining did it.
Deforestation did it.
Climate change did it.
The Brahmaputra did it.
Nature did it.
Hydrology rarely gives us such clear-cut answers.
The emerging evidence instead suggests a chain:
exceptional hill rainfall
→ already wet catchments
→ rapid runoff
→ exceptional tributary levels
→ interaction with an elevated Brahmaputra
→ water entering a landscape containing altered channels, extraction sites, roads, embankments, settlements and constrained drainage.

Observations strongly support the early links in that chain.
We still do not know exactly how much later human changes to the landscape affected the flood.
Being careful and scientific does not mean doing nothing.
Instead, it means we should focus on measuring and understanding what happened.
If mining altered the Dikhow, demonstrate it.

If forest or wetland loss accelerated runoff, measure it.
If the river’s carrying capacity has declined, publish the cross-sections and rating curves.
If a dam aggravated flooding, release the inflow and outflow hydrographs.
If urban drainage trapped floodwater, identify where and why.

The flood itself has already shown us what can happen.
The real question is whether we are ready to learn from what happened.

Perhaps the most useful question left by July 2026 is therefore not:“Who caused the flood?”

It is this: When an exceptional amount of rain fell on Upper Assam’s headwaters, what happened to that water between the hills and the Brahmaputra—and has that journey changed?

If we can answer that, many debates can turn into clear evidence. 

 

¹Associate Professor,Department of Geography, Mangaldai College (Autonomous), ORCID ID: https://orcid.org/0000-0001-9986-9790

Principal sources

Assam State Disaster Management Authority, Annual Activity Report 2020–21 to 2023–24. (ASDMA)

Intergovernmental Panel on Climate Change (IPCC), Sixth Assessment Report (AR6), Working Group I: chapters on the global water cycle and weather/climate extremes. (IPCC)

National Remote Sensing Centre/ISRO, Flood Hazard Zonation Atlas of Assam Using Multi-Sensor Satellite Data, 1998–2023. (nrsc.gov.in)

Central Electricity Authority, State Profile on Hydro Development for Nagaland, including status of Doyang and Dikhu hydroelectric projects. (Central Electricity Authority)

India Meteorological Department observations as reported in contemporary analyses of the July 18–20 rainfall event. (India Today NE)

Central Water Commission observations reported during the July 2026 flood, including river levels at Desangpani, Dikhow and Neamatighat. (news.webindia123.com)

 

(The views expressed in this article are those of the author and do not necessarily reflect the views of India Today NE or its affiliates.)