Explore
Flagship studies, sector briefs, and recurring reports, by topic.
Long-form data narratives that walk through an argument.
Recurring advisor series, from weekly reads to annual reviews.
The daily policy prescription, generated each morning.
Seven sector deep-dives: banking, trade, energy, agriculture, and more.
Ask the corpus: answers grounded in published work, with citations.
Grounding verified
Sunamganj / Sylhet flash flood pre-monsoon
The hazard in scope is the pre-monsoon flash flood in the Sunamganj and Sylhet haor, the seasonal wetland basins of the northeast. This is a fast-onset, event-class hazard operated as a nowcast: the gap between an upstream rainfall signal and floodwater arriving in the haor is measured in hours and a few days, not weeks. The single most exposed asset is the boro paddy crop, which is harvested in this same pre-monsoon window. When the water arrives before the harvest, a year of food and household income is lost in a single event, and the loss falls on a population whose only embankments are the seasonal submersible dykes that protect the standing crop.
The lead responsible body is the Ministry of Disaster Management and Relief (MoDMR), per the GovTwin entity registry. The current operational state for this hazard is not wired to a live indicator (current_state is null). That is itself the first finding: a nowcast hazard with no standing real-time trigger means the warning-to-action chain is improvised every season rather than pre-agreed. The policy task is not to predict the flood better in the abstract. It is to convert whatever warning lead time exists into financed, pre-positioned action before the water arrives.
First, before the next pre-monsoon season, MoDMR issues the trigger protocol circular (action 1). This is the cheapest, fastest step and it unlocks everything else, because it forces the trigger source, the budget authority, and the named officers to be decided on paper rather than in the crisis. Second, in parallel, wire the real-time feed (action 2) so the protocol has a signal to fire on. Third, secure the pre-financed budget line with the Ministry of Finance (action 3) so that when the trigger fires, money moves the same day. Pre-positioning (action 4) is then executed within the season, and the after-season review (action 5) closes the loop and sets the agenda for year two.
The binding constraint is fiscal and procedural, not technical. Pre-financing early action requires the Ministry of Finance to release money on a forecast, before any damage is confirmed, which cuts against standard post-event relief practice and audit habits. Without that pre-agreement, the protocol degrades into another warning that no one is funded to act on. The second constraint is institutional coordination: the warning function and the response function sit in different bodies, and a nowcast hazard fails at the handoff between them unless one circular names the handoff explicitly.
The haor flash flood cannot be stopped, but the days of warning before it can be bought back as harvested crop and evacuated households if MoDMR converts a vague seasonal alert into a financed, pre-agreed early-action protocol. The decisive move is the protocol circular plus a pre-financed budget line that releases on the trigger, so that money and crews move before the water, not after the loss.
Where this hazard bites hardest, ranked by population in the 100-year floodplain, population-weighted with WorldPop 2020. Source: multi-hazard exposure atlas (GloFAS, FABDEM, MODIS, ACAG).
| District | Division | population in the 100-year floodplain |
|---|---|---|
| Jhalokati | Barisal | 98.3% |
| Pirojpur | Barisal | 97.5% |
| Lakshmipur | Chittagong | 96.9% |
| Gopalganj | Dhaka | 96.4% |
| Barguna | Barisal | 96.1% |
| Barisal | Barisal | 96% |
| Madaripur | Dhaka | 94.8% |
| Shariatpur | Dhaka | 94.1% |
The figures and responsible bodies cited in this prescription are drawn from the platform's own data and the GovTwin registry listed below.
Drafted by an Opus writer grounded in the facts above. Where the prescription cites a figure, it is drawn from those facts. The diagnosis derives from the BDPolicyLab crisis taxonomy; the responsible body and budget from the GovTwin registry. Recommended actions are the think tank's policy judgment.