Bihar’s flood risk is changing, not just repeating

Bihar’s flood story is both familiar and evolving. While many communities face recurring inundation, erosion and infrastructure damage, changing flood patterns are pushing the geography of risk into new settlements, exposing more households to river floods, embankment breaches, flash floods and prolonged waterlogging

Every year, Bihar counts its floods. Districts affected. Gram Panchayats affected. People evacuated. Relief camps, boats, damaged houses, families receiving gratuitous relief. These numbers matter; they let the state mobilise quickly. But after three consecutive years of flood figures, a harder question is becoming difficult to avoid: what does Bihar learn from one flood before the next arrives?

The Disaster Management Department’s reports offer a starting point. On October 15, 2024, it recorded 28 affected districts, 979 Gram Panchayats and 52.46 lakh people. On October 15, 2025, the figures stood at 23 districts, 600 Gram Panchayats and 48 lakh people. The September 24, 2026 report shows 15 districts, 576 Gram Panchayats and 50.49 lakh people, a total likely to rise after the recent flooding along the Gandak. These are snapshots from different stages of different floods, so they do not measure severity. Yet the pattern is striking. The reported geography of impact has shifted considerably, while the number of people affected stays in a similar range. What changed because of what Bihar experienced the year before?

Vulnerability in Bihar is neither wholly repetitive nor wholly new.

Some embankments, settlements, roads and households face recurring exposure and accumulating loss. Elsewhere, a river shifts course, a breach redirects water, or a flash flood reaches people never considered at high risk. A single breach is an event; repeated breaches near the same spot signal persistent vulnerability, demanding a different response from a newly affected area. And when a flood behaves unexpectedly, that should reshape the state’s understanding of risk.

Every flood produces information. Breaches are reported, roads cut, crops lost. Engineers inspect embankments; communities know where water entered and where it stagnated. When the water recedes, what happens to all this knowledge? Does a repeatedly flooded location acquire a risk profile? Does a newly affected area prompt a revision of vulnerability maps? Does a known breach or drainage failure shape how infrastructure is designed and maintained before the next monsoon?

Bihar’s machinery for emergency response is extensive. What is harder to see is a learning loop connecting one year’s experience to the next year’s decisions. A response system asks: what must we do now?

A learning system asks: why did this happen here, what did we learn, what must change, and who will make sure it does before it happens again? The first is indispensable during a flood. The second decides whether the state can move from recurring emergency management towards sustained risk reduction.

Once the emergency passes, the questions should change. Why did the breach occur? Had the location already been flagged? Were earlier interventions temporary or permanent? Did the flood contradict existing risk assessments? Most importantly, what must change before the next monsoon?

The answers cannot stay locked in departmental files. Water Resources handles embankments, Disaster Management coordinates response, Rural Works handles roads, Agriculture handles crops, Public Health Engineering handles drinking water. A household experiences none of these separately.

A breached embankment, a flooded road, a damaged handpump, a lost crop, a closed school and an unusable house are one flood. Post-flood learning must see them as one system. Development partners have a role here. Bihar has long worked with bilateral and multilateral agencies, academic institutions and civil society on risk reduction, forecasting, livelihoods and community preparedness. The opportunity now is to deepen that engagement: supporting systematic post-flood assessments, documenting recurring and emerging vulnerabilities, bringing together evidence across sectors, and translating it into preparedness before the next monsoon.

None of this requires another large survey. It requires the reporting system to capture what the administrative record largely misses: what happens inside affected households. Every significant flood location could have a post-flood action record. It could document damage to houses, stored grain, livestock and productive assets; disrupted livelihoods and lost school days; health challenges, especially for women and adolescent girls when water, sanitation and menstrual hygiene facilities fail; and how long displaced families take to resume normal life. It could also record the suspected cause, the flood typology, the infrastructure involved, the corrective action required, the responsible institution, the deadline and the status before the next monsoon.

Status is the important word here. A repaired embankment is not a resolved vulnerability. A family receiving relief has not necessarily recovered. A reopened road does not mean the local economy has resumed. A restored handpump does not prove a household has safe drinking water.

A rebuilt school may stay empty because children’s homes and livelihoods have not recovered. The record must therefore distinguish the event, the action taken, the household consequences, the extent of recovery, and whether the intervention actually reduced the risk of repeated loss. That is a far better measure of governance than another annual total of people affected. The same logic applies to places that flood repeatedly. Bihar could maintain a simple repeat-impact register at ward or Gram Panchayat level, recording how often a location has been affected, by what type of flooding, which infrastructure failed, what was done, and whether it changed the outcome. A location hit once needs emergency response; repeated exposure points to persistent vulnerability that demands prevention. That steers public money towards reducing known risks rather than repairing damage.

Embankments need the same discipline. A breach should trigger a documented technical review: was the cause overtopping, seepage, erosion, animal burrowing, poor maintenance or unusual hydraulics? What was done afterwards? Did it recur? Such reviews should aim not at blame but at ensuring that what the incident revealed is available when planning, maintenance and budgeting begin.

The principle extends further. Drainage congestion, road closures, crop losses and displacement all need memory. Even relief does. A household receiving assistance year after year shows that the relief mechanism works, but should also prompt the question of why the exposure persists. This is no argument against relief, which remains essential. But a good system must distinguish between responding well to a recurring problem and reducing the recurrence itself.

Megh Pyne Abhiyan’s household-level loss assessment of the 2024 floods, supported by the Tata Trusts, shows how such evidence-gathering can begin. It showed that vulnerability is multidimensional and varies across flood typologies. Its value lies in continuing across successive floods rather than ending with a report. Repeated ground evidence shows what recurs, what changes, and where interventions work.

That is how evidence becomes institutional memory. The aim is not another parallel reporting system but a stronger capacity within the existing one to retain, compare and use what it learns. The three-year figures matter again here: the number of people affected stays in roughly the same range despite wide variation in spread. That should prompt an annual review. Which vulnerabilities persisted, and which were addressed? Which interventions worked, and which failed? Which locations flooded again, and which households were displaced again? And what changed because the system learned?

Such questions would move flood governance from a cycle of incident, response and expenditure towards one of incident, evidence, learning, correction and prevention. The annual report can remain the rapid snapshot. Alongside it, Bihar needs a post-flood learning record that follows significant incidents beyond the receding water and separates damage from recovery, and repair from risk reduction.

Bihar will still keep having floods. Rivers will rise, embankments will be tested, settlements will stay exposed. Progress cannot be measured by the absence of the next flood. It should be measured by whether the next flood finds Bihar less vulnerable because it remembered the last one. Not simply how it counts a flood, but how it remembers one.

Flood vulnerability in Bihar is neither entirely repetitive nor entirely new from one year to the next. Some settlements, embankments, drainage systems, roads, agricultural lands and households face recurring exposure, accumulating losses and vulnerabilities over time. At the same time, different flood typologies can alter the geography of impact, bringing new settlements and households into the path of riverine flooding, embankment breaches, flash floods, erosion, waterlogging or combinations of these

The writer is the Managing Trustee of Megh Pyne Abhiyan; Views presented are personal.