Where the Salt Is

Executive Summary. Bangladesh's climate file is not primarily an international advocacy problem. It is an internal triage problem: deciding which coastal land is saved by adaptation, which is given up in a planned way, and where the displaced are settled in advance. No single agency owns that decision today. Approximately 15.9 million Bangladeshis live below five metres of elevation on land that already absorbs the leading edge of sea-level rise, salinity intrusion, and cyclone landfall, while the country contributes less than half a percent of global greenhouse gas emissions. Current adaptation spending is around USD 1 billion per year, against CEGIS-tracked annual flood-and-cyclone damage that averages roughly USD 2 billion over the long run. The Tarique Rahman government's 180-day plan does not name climate adaptation as a priority. The FY27 budget decides whether this stays a residual line in the development budget or becomes a national-security file with an institution and a financing rule behind it.

Late January 2026 in a union of Shyamnagar upazila in Satkhira, on the seaward edge of the Sundarbans, the boro sowing window is closing. The scene that follows is a representative illustration, not an interview with a named individual; the person and household details are illustrative, but every quantity is drawn from a verifiable source. When the soil salinity in a coastal boro farmer's half-prepared field passes roughly 6 deciSiemens per metre, most conventional boro paddy varieties suffer heavy yield depression in the seedling and tillering stages. The Bangladesh Rice Research Institute has released a series of salt-tolerant varieties precisely for these soils: BRRI dhan47, BRRI dhan67, BRRI dhan73, BRRI dhan78. By the institute's own station data, these varieties hold materially higher yields than the conventional varieties at high salinity levels.

The problem is not the variety; it is delivery. The upazila agriculture office issues a November circular announcing that salt-tolerant seed will be available; in many unions the seed reaches the union parishad weeks after the optimum transplanting date, when it is no longer usable for that season. The farmers who collect it late keep it for the following year, because using it then would mean missing the entire boro cycle.

The arithmetic spine of this scene is not illustrative but geographic: the salt has come, the seed exists, and the state has not connected the two in time. As the salinity gradient advances, every late seed shipment moves a field one season closer to being uneconomical for rice. The visible end of that failure is the household's younger members leaving for garments and tourism work, which shows up in the statistics as internal migration.

The number that matters and the number that does not

Bangladesh's contribution to global greenhouse gas emissions is less than 0.5 percent of the global total (NDC 3.0, 2022 basis). Per-capita greenhouse gas emissions are around 1.48 tonnes of CO2 equivalent, roughly a quarter of the world average and among the lowest of any sizeable country. Whatever we do or do not do on the mitigation side, our marginal impact on the global temperature trajectory rounds to zero. This is not a moral statement; it is an arithmetic statement. The country's climate problem is not what we put into the atmosphere. It is what the atmosphere is doing to a delta with 1,265 people per square kilometre.

The vulnerability numbers are large. Approximately 15.9 million Bangladeshis live below five metres of elevation, measured from MERIT DEM v1.0.3 elevation and WorldPop 2020 population over the whole country. The five-metre LECZ covers approximately 26,745 square kilometres of land, about 18 percent of the country's total land area. The under-one-metre band covers 4,079 square kilometres. The ten-metre band covers 63,726 square kilometres.

The sea is rising in the Bay of Bengal; the IPCC AR6 puts the global mean rate at approximately 3.7 millimetres per year over 2006 to 2018, and the local relative rise in the northern Bay of Bengal is higher once regional tide-gauge and delta-subsidence patterns are added. Bangladesh Meteorological Department long-run observations show clear trends in surface temperature and in the seasonal balance of rainfall: mean temperature rising, and the split between monsoon and dry-season rainfall growing more extreme. On average a large share of the country's land area is flood-affected each year, and major cyclones make landfall every few years.

The direction of salinity intrusion is verifiable. World Bank river-salinity modelling projects that by 2050 the 5-ppt salinity line moves inland by approximately 24 kilometres on the southwest coast (Khulna-Satkhira) and approximately 8 kilometres on the southeast coast (Bhola-Patuakhali). The consequence is drinking-water scarcity in coastal towns, a shortage of dry-season irrigation water, and significant change in coastal aquatic ecosystems.

The displacement numbers are real. The International Organization for Migration's first nationwide estimate, released in December 2024, put cumulative internal displacement from natural disasters in Bangladesh at approximately 4.96 million people (4,955,527), about 85 percent of them living in rural unions, with the largest shares in the Chittagong, Dhaka, and Rajshahi divisions. Most of the moves are short-distance. The visible end of the displacement is the households whose teenage children have moved to Mirpur garments factories or to Cox's Bazar tourism work. The invisible end is the field that goes uneconomical, the village that loses its primary school because enrollment collapses, the union parishad that loses its tax base.

Two facts have to be held together. First, our mitigation effort is marginal because our emissions are marginal. Second, our adaptation problem is one of the largest in the world because the geography is unforgiving and the density is high. The policy implication of holding both facts together is that the climate file is not primarily an international advocacy file. It is an internal triage file.

What internal triage looks like

There are three lines on the triage map and the country has not built any of them at scale.

The first is in-place adaptation: the technical and infrastructural work that keeps a coastal field productive for the next twenty years rather than the next five. This is salt-tolerant rice variety distribution that actually arrives at the union parishad before the sowing window closes. It is rationalised polder management with the Water Development Board's khals and sluice gates maintained on a schedule rather than rebuilt after each cyclone breach. It is brackish-water aquaculture conversion subsidies that let farmers shift from monsoon-rice plus dry-season-rice to a shrimp-cum-bagda cycle when the salinity gradient passes the conventional crop threshold. It is rainwater-harvesting infrastructure that gives coastal households a dry-season drinking-water source that is not the deepening tube-well chasing receding fresh-water lenses. None of these is technically novel. All of them require institutional execution that has not happened.

The second is managed retreat: the deliberate, planned, compensated relocation of households and infrastructure from the polders and shoreline strips that the adaptive infrastructure cannot save on a 30-year horizon. The current pattern is distress-driven outmigration in which the family abandons the field and the village without compensation, the male wage-earner finds informal-sector work in Khulna or Dhaka, and the household ends up in peri-urban tenancy with no land title and no economic anchor. The alternative is structured: land-for-land or land-for-cash conversions in receiving areas (Khulna metropolitan area, Mongla economic zone, peri-Chittagong, the Faridpur belt), with portable land titles, transfer of agricultural extension and stipend programmes to the receiving area, and explicit decommissioning of the polder embankment maintenance budget once the area is depopulated. This is politically difficult because it admits that some land will be returned to the sea. It is economically inevitable because the alternative is the same outcome reached through repeated reconstruction cycles that the budget cannot sustain.

The third is upstream investment that pre-positions the receiving infrastructure: the housing stock, school capacity, health facility capacity, drainage and water supply that the climate-displaced households will need in the receiving areas. The current Khulna metropolitan plan assumes a population growth trajectory that significantly under-projects the climate-displacement inflow. The current peri-Chittagong land-use plan has the same gap. The Greater Faridpur belt has been identified in successive long-term plans as a strategic receiving area for inland migration; the infrastructure to make it absorptive at the projected scale has not been built. The political economy of receiving-area investment is hard because the visible voters are the existing residents, not the future climate migrants, and infrastructure pre-positioned for future demand is easy to defer through one budget cycle after another.

The institutional gap

The state machinery to deliver internal triage exists in pieces and is not assembled. The Ministry of Disaster Management and Relief responds after the cyclone hits and the embankment breaches. The Bangladesh Water Development Board maintains the polders and the embankments. The Department of Agricultural Extension distributes (or fails to distribute) the salt-tolerant rice seed. The Ministry of Environment, Forest and Climate Change holds the international climate-finance file and produces the NDC documents and the national adaptation plans. The Bangladesh Climate Change Resilience Fund and the Bangladesh Climate Change Trust Fund hold pools of domestic and pooled donor money. The Local Government Engineering Department and the city corporations of Khulna, Chittagong, Mongla, and Cox's Bazar handle receiving-area infrastructure.

None of these institutions is responsible for the triage map. There is no single agency whose mandate is to look at a specific at-risk union, decide whether it can be saved with adaptation or has to be retreated from, and execute the relevant programme. The result is that each institution operates within its narrow mandate and the strategic question goes unanswered. The gap is visible in the official accounts: MoEFCC-tracked adaptation spending runs at around USD 1 billion per year, much of it donor-funded, while CEGIS-framed flood-and-cyclone damage averages roughly USD 2 billion per year over the long run. The distance between what is spent and what the damage costs is the part that nobody is assigned to close.

The central fix: an authority that owns the triage map

The institutional reform required is not a new-agency slogan but a specific accountability structure. A consolidated Climate Adaptation and Resettlement Authority should be created under the head of government (not merely a wing of MoEFCC, since its work spans the Water Development Board, the Department of Agricultural Extension, the Ministry of Disaster Management and Relief, LGED, and the coastal city corporations), with a distinct budget line marked in the FY27 budget. Its function is to build and annually update a triage map at union and upazila granularity: which pockets get in-place adaptation, which get managed retreat, which get receiving-area investment, then to commission the relevant programme through the line agency that holds the asset.

To do that work and not become another coordination committee, the Authority needs four things specified up front, not left to a later rule.

Decision rights. A statutory mandate, set by act of parliament rather than executive circular, to direct programming and condition the climate-relevant budget releases of MoEFCC, the Water Development Board, the Department of Agricultural Extension, the Ministry of Disaster Management and Relief, LGED, and the coastal city corporations against the triage map. Without budget-conditioning power, the map is advisory and the line agencies revert to their narrow mandates.

A financing rule. On the FY26 nominal base of about USD 510 billion (IMF World Economic Outlook), the additional internal-triage spend that closes the gap to a fully resourced programme is small as a share of GDP: an extra USD 1.5 to 2.0 billion per year from domestic resources is roughly 0.3 to 0.4 percent of GDP. Because the existing flow is already mostly domestic, the realistic planning assumption is that the gap is closed primarily from domestic fiscal space, with international finance treated as upside rather than the base case.

Staffing for execution, not administration. A lean technical core (geospatial triage, water and agronomy, resettlement economics, public finance) with authority to second engineers and extension staff from the line agencies into time-bound delivery teams, rather than a parallel field bureaucracy.

A measurable success signal. The first test is operational, not financial: salt-tolerant seed and the corresponding extension package reaching the union parishad before the boro transplanting window in every designated in-place-adaptation upazila, reported by upazila each season. If the Authority cannot fix the seed-delivery failure that already has a known technical answer, it will not deliver managed retreat, which has none.

The precedent for failure is concrete. The post-2008 Bangladesh Climate Change Strategy and Action Plan governance structure was supposed to do something like this and did not, because it had thematic areas and pooled funds but no agency with decision rights over the line ministries. The post-2024 reform window, which the Tarique Rahman BNP government (sworn in on 17 February) inherited from the Yunus interim period, is the opening to build the authority with the mandate and the budget envelope the earlier structure lacked.

The Tarique Rahman government's published 180-day priority plan does not include climate adaptation among its four named priorities (law and order, essential-goods prices, electricity and gas, and railway connectivity). The omission is the relevant political fact. The scientific case has been made for two decades; the institutional case has been made for one; the budget case is being made now in the first BNP-era fiscal cycle. Whether the FY27 budget tabled later this year treats the climate file as a long-arc national-security question or as a residual category of the development budget is the test. The Sundarbans, on which several million coastal residents depend directly or indirectly for their livelihoods, sits in the middle of both the sea-level-rise and the salinity trajectories.

What it costs and what it buys

Implemented seriously over a fifteen-year horizon, the additional annual cost of internal triage is on the order of USD 3 to 4 billion per year in 2026 prices, against the roughly USD 1 billion current adaptation-spending base. On the IMF-projected 2026 nominal GDP base of about USD 510 billion, USD 3 to 4 billion is roughly 0.6 to 0.8 percent of GDP. The composition is about one-third in-place adaptation, one-third managed-retreat-and-compensation, and one-third receiving-area infrastructure. The National Adaptation Plan (2023 to 2050) prices the full long-run programme at about USD 230 billion over the period.

At current realisable pledge and disbursement velocity, the international climate-finance contribution to this number is on the order of USD 1.0 to 1.5 billion per year. The Loss and Damage Fund agreed at COP28 and the new collective quantified goal negotiations may raise this eventually, but realised disbursement has historically been a fraction of the headline pledge. The gap that has to be closed from domestic resources is therefore about USD 1.5 to 2.0 billion per year, roughly 0.3 to 0.4 percent of GDP, the kind of fiscal space that opens once the tax base is broadened, without new external borrowing.

What the spending buys is two things. First, the prevention of the unmanaged collapse scenario in which the receiving areas are overwhelmed by distress migration on a timetable they cannot absorb, the urban informal sector swells past what formal-sector job growth can absorb, and the political-economy stress of large unintegrated populations in peri-urban Dhaka, Chittagong, and Khulna creates governance instability beyond the climate file itself. Second, the preservation of agricultural productive capacity in the parts of the coastal belt that can be kept productive with adaptation, which preserves the rural-economy base and the food-security buffer the country has depended on for the entire post-independence period.

The strongest objection is the political and moral risk of managed retreat: if the state announces that an area will be given up, it can trigger a collapse in land values, forced eviction, and an uneven burden on the poorest households. The objection is serious, and it is exactly why the success signals make advance, portable compensation the first condition. The conclusion would change if global mitigation accelerated enough to bend the sea-level and salinity trajectories sharply down, or if adaptation technology (more tolerant varieties, affordable desalination) improved fast enough that land considered marginal today stayed productive. Neither is safe to assume on current evidence. The alternative to internal triage is not preservation of the status quo; it is the same outcome reached chaotically, in one shock rather than gradually, with spending concentrated in receiving cities rather than distributed across the planning horizon. Tropical-cyclone damage alone is estimated by the IMF at 1 to 2 percent of GDP, and the recurring distress-migration and reconstruction cycle already spends the money, just chaotically and after the fact.

The series, and the coastal field

Six reforms have been described in this series. Tax mobilisation, the foundation that pays for everything else. Banking recapitalisation, the financial-system precondition for productive credit. LDC-graduation cushioning, the diplomatic file that protects the export base. Energy transition, the industrial-policy file that determines whether the manufacturing base can move up the value chain. Education quality, the human-capital file that determines whether the demographic dividend translates into productivity. Climate adaptation, the geographic file that determines whether the parts of the country we cannot relocate are kept liveable.

Each of these reforms has a binding constraint that is not analytical: the reforms have been studied; the data is available; the policy texts have been drafted. Each has a binding constraint that is political: the reforms require asking concentrated interest groups to pay for diffuse public goods, and the political economy of asking has not been resolved by any government in the last twenty years.

The BNP government has roughly twelve months from its February inauguration to commit publicly to the package and roughly five years to execute the first half of it visibly enough that the second half becomes politically self-sustaining. The reforms are interdependent: without tax revenue, the banks cannot be recapitalised; without recapitalised banks, the energy transition cannot be financed; without the energy transition, the post-LDC industrial upgrading cannot happen; without the industrial upgrading, the education investments cannot translate into the labour market; without the demographic dividend playing out productively, the climate file cannot be financed at the scale required. Each reform's success is each other reform's precondition.

A field in a coastal union is the smallest unit on the policy map. The seed that did not reach it in time is one administrative failure. The salt that has come is one geographic fact. The country that has the analytical capacity to know exactly what to do and has not done it is the country that the reform stack in this series is trying to repair. The work is not impossible. The work has not been started.

Sources

• Low-elevation coastal zone land area and population by elevation band, derived from MERIT DEM v1.0.3 and WorldPop 2020 via Google Earth Engine. Reproducible: scripts/derive_lecz_gee.py.
• Bangladesh Meteorological Department, temperature and rainfall trend series. bmd.gov.bd
• Intergovernmental Panel on Climate Change, AR6 Working Group I (global mean sea-level rise approximately 3.7 mm/year over 2006 to 2018; regional Bay of Bengal estimates). ipcc.ch/report/ar6/wg1
• River salinity and climate change (2050 inland shift of the 5-ppt isoline): Dasgupta, Kamal, Khan, Choudhury, Nishat, "River Salinity and Climate Change: Evidence from Coastal Bangladesh", World Bank Policy Research Working Paper 6817, 2014. documents1.worldbank.org/curated/en/522091468209055387
• International Organization for Migration, first nationwide estimate of disaster-induced internal displacement in Bangladesh (4,955,527 people), launched December 2024. bangladesh.iom.int
• CEGIS, annual flood and cyclone economic damage assessment. cegisbd.com
• Government of Bangladesh, Nationally Determined Contribution (NDC 3.0), 2025: greenhouse gas emissions less than 0.5 percent of global total and per-capita 1.48 tCO2e (2022 basis); annual climate damages around USD 3 billion (around 1 percent of GDP). Department of Environment. unfccc.int.pdf)
• Government of Bangladesh, National Adaptation Plan 2023-2050: total adaptation financing requirement about USD 230 billion. moef.gov.bd
• International Monetary Fund, Article IV / staff reports: projected tropical-cyclone damage on the order of 1 to 2 percent of GDP. imf.org
• International Monetary Fund, World Economic Outlook (April 2026), nominal GDP (NGDPD) Bangladesh 2026 about USD 510.7 billion. imf.org/external/datamapper/NGDPD/BGD
• BRRI salt-tolerant rice varieties (BRRI dhan47/67/73/78) research station performance. brri.gov.bd
• COP28 Loss and Damage Fund, operational status. unfccc.int
• MoEFCC, Bangladesh Climate Change Resilience Fund and Climate Change Trust Fund disbursement reports. moef.gov.bd