Climate and environment

Bangladesh's warning and evacuation systems have helped reduce disaster mortality. They do not remove exposure to floods, salinity, heat or air pollution, and annual death counts also depend on the hazards that occur. The Water Development Board recorded 25.7 percent of the country flooded in 2023 [BWDB 2023]. Adaptation requires sustained maintenance, water management, land-use decisions and public-health protection alongside emergency preparedness. This chapter separates those tasks and identifies what the available measurements can support.

Floods, water stress, heat and air pollution

EM-DAT logs between 2 and 12 natural disaster events a year across 2000 to 2026 [EMDAT 2026]. The annual death count peaked at 5,635 in 2007, the year of cyclone Sidr, ran above 900 a year on average through the 2000s, and fell to 283 in 2020, 176 in 2022, 20 in 2023 and 124 in 2024 [EMDAT 2026]. The exposure behind those deaths did not fall: persons affected ran at 36.89 million in 2004, 22.93 million in 2007, 8.05 million in 2020 and 48.78 million in 2024, and recorded damage reached 4.60 billion USD in 2004 and 2.49 billion USD in 2020, the year cyclone Amphan made landfall [EMDAT 2026].

The international indices: rank 31 of 174 on the Germanwatch Climate Risk Index for 1993 to 2022 with a score of 12.36, after rank 7 with a score of 28.33 for 2000 to 2019 in the 2021 edition, whose method and window differ [Germanwatch CRI 2025] [Germanwatch CRI 2021], and 6.8 of 10, rank 15 of 163, on the Children's Climate Risk Index, the exposure pillar 7.2 and the child vulnerability pillar 6.5 [UNICEF CCRI 2021].

Flooding is the base condition of the delta. The Water Development Board record shows a third of the land flooding in bad years across seven decades: 50,500 square kilometre, 34 percent of the country, in 1955, and 48,698 square kilometre, 33 percent, in 2021, with 39,991 square kilometre, 27.1 percent, in 2022 and 37,925 square kilometre, 25.7 percent, in 2023, the fifth consecutive year above a quarter of the national territory after 31 percent in 2019 and 40 percent in 2020 [BWDB 2023]. The satellite record: FAO mapping recorded about 28,700 square kilometre flooded and 20.7 million persons exposed in the second half of July 2024, and 747 square kilometre and 0.48 million persons exposed across Sylhet division's four districts in the first half of April 2026 [FAO 2026].

The structural number: 8.38 percent of the population lives below 5 metres elevation, and the average annual share affected by droughts, floods and extreme temperatures was 4.58 percent over 1990 to 2009 [WB WDI 2026].

The slow hazards accumulate rather than strike. Groundwater first: withdrawals run at 34.2 percent of internal renewable resources, and renewable freshwater per head fell from 647 cubic metres in 2017 to 620 in 2022, under the scarcity line [WB WDI 2026]. The Census 2022 drinking water tables show why the stock is drawn down: the unweighted district average has 90.7 percent of households drinking from tube wells and 5.3 percent from piped supply, so the aquifer is the national drinking water utility, and in the salinity belt the utility is failing, with Bagerhat households reporting 22.0 percent drinking surface water and 16.7 percent rainwater against 53.1 percent tube wells, and Satkhira 7.7 percent surface and 4.0 percent rain [BBS Census 2022, district drinking water tables].

The soil side of the salinity belt has a measured reading: a March 2024 in-situ survey of 162 GPS-tagged sites across the Satkhira, Khulna and Bagerhat belt found mean topsoil electrical conductivity of 1.52 millisiemens per centimetre, 29.6 percent of sites above the 2.0 agronomic threshold, 24.1 percent slightly, 4.3 percent moderately and 1.2 percent highly saline [Zenodo 2024, DOI 10.5281/zenodo.14560019]. Extent in hectares intruded has no measured national series in the available data, so any figure for it could not be confirmed; the Water Development Board and Soil Resource Development Institute surveys would settle it.

River erosion is the same kind of problem without any series: the development programme's largest single water allocation, 1,000 crore BDT, is Meghna bank protection at Ramgati and Kamalnagar in Lakshmipur, and Padma bank protection in Kushtia holds the next line at 400 crore BDT [Planning Commission ADP 2025], and the national erosion displacement statistics could not be confirmed; the Water Development Board would settle it.

Heat and air are the urban face of the slow hazards. The April to May 2024 pre monsoon window set the grid record in four of the twelve city reanalysis grids, Khulna and Jessore resolving to the same cell: Khulna's mean daily maximum of 46.4 degrees Celsius and Rajshahi's 45.6 are the hottest April to May readings in their 44 year series, while Dhaka's 42.7 ranks fourth, behind 1999 at 43.1, 1992 at 43.0 and 1995 at 42.9 [NASA POWER 2024]. The country mean temperature anomaly in the Berkeley Earth series averaged minus 0.24 degrees Celsius over 1990 to 1999 and plus 0.47 over 2020 to 2025, with 2023 the warmest year in the record at plus 0.74 [Berkeley Earth 2026].

The air record needs consistent vintages. The series used in chapter 30 places mean annual PM2.5 exposure at 69.3 micrograms per cubic metre in 2023, against a peak of 79.3 in 2018 [WB WDI 2026]. The earlier 42.4 reading came from a different pull and is not joined to this series. The World Health Organization attributes 76,062 deaths to ambient air pollution in 2019 and 117,681 to household air pollution, the larger number standing on the 28.4 percent clean fuel access of 2023 [WHO GHO 2019, ambient and household air pollution attributable deaths] [WB WDI 2026].

Chapter 09 carries the health financing side of that ledger.

On the mitigation side Bangladesh is a rounding error in the global stock and a case study in energy led growth. Total emissions including land use were 265.26 million tonne of CO2 equivalent in 2022: energy accounted for 128.16 million tonne, a sector total that includes electricity and heat at 60.07 million tonne and transport at 15.27 million tonne, agriculture for 90.39 million tonne, waste for 23.41 million tonne, and land use change and forestry was a net 17.59 million tonne [Climate Watch 2024]. Fossil CO2 alone rose in four consecutive years: 95.73 million tonne in 2020, 98.53 in 2021, 102.15 in 2022, 104.53 in 2023 and 108.32 in 2024, 0.28 percent of the global total, about 0.62 tonne per person [GCP 2024, co2].

The driver is the generation mix: coal's share of electricity generation rose from 1.7 percent in 2015 to 10.6 percent in 2023 and 21.5 percent in 2025, the mechanism chapter 10 documents [Ember 2025]. The forest base is thin and static: 14.49 percent of land area in the World Bank series, and 23,217 square kilometre of Forest Department land of which the mangrove, 801,700 hectares in FY23, is 34.5 percent [WB WDI 2026] [Forest Department 2023].

The money is the gap between the plan and the budget. The FY26 Annual Development Programme allocates 8,489.86 crore BDT to the Ministry of Water Resources across 78 projects, 1,380.48 crore BDT to the Ministry of Environment, Forest and Climate Change across 30 projects, and 2,188.48 crore BDT to the Ministry of Disaster Management and Relief across 10 projects, a combined 12,058.82 crore BDT, 5.24 percent of the 230,000.00 crore BDT programme the National Economic Council approved in May 2025, the three printed shares of 3.69, 0.60 and 0.95 percent summing to the same figure [Planning Commission ADP 2025, ministry wise allocation table].

The revised programme's project table for the same three ministries carries 68,795.5 crore BDT of committed project cost against 13,889.7 crore BDT of revised year allocation, a cost stock about five times the year's money, and the Water Development Board alone holds 10,497.0 crore BDT of allocation across 84 projects against 45,147.8 crore BDT of committed cost [Planning Commission ADP 2025, revised FY2025-26 project table]. External climate finance has no series in the available data, so any inflow figure could not be confirmed; the Economic Relations Division records would settle it.

Why adaptation requires more than emergency response

The mortality collapse is the one climate mechanism Bangladesh engineered on purpose, with emergency management instruments. The Cyclone Preparedness Programme, the shelter network, the warning chain and the evacuation drills that ran ahead of every landfall after 1991 killed far fewer people per event than the 1970 storm; Sidr in 2007 was a storm of comparable intensity to the 1991 cyclone and killed a fraction of the people, and the EM-DAT year totals record the regime change [EMDAT 2026]. An evacuation costs almost nothing per life saved next to a polder, so the state bought warnings and shelters early and often.

The same logic does not transfer to the slow hazards, because there is no evacuation from an aquifer going saline, a city heating or an airshed turning toxic; the instruments those hazards require are water pricing, land use control, enforcement and clean energy, and each sits in the political economy of chapters 04, 10, 12 and 14, not in the disaster ministry's drills.

The water mechanism is the delta's own trap. The polders and embankments that protect land from tidal flooding also cut the sediment supply that built and fed the coastal plain, so the protected land subsides and drains poorly while the rivers, confined between embankments, rise; erosion protection on one bank pushes the erosion to the next. The FY26 project list is a catalogue of the mechanism: bank protection at Ramgati and Kushtia, drainage for Chattogram city, Tista irrigation rehabilitation [Planning Commission ADP 2025]. The groundwater mechanism compounds it: with nine tenths of households on tube wells and irrigation drawing on the same aquifer, dry season withdrawal lowers water tables, which on the coast pulls salt inward, which pushes households onto surface and rainwater, the Bagerhat and Satkhira shares above.

The health consequence is measured: illness prevalence ran at 74.65 percent among rural individuals whose households were waterlogged in the past year against 59.83 percent among the unexposed, a risk ratio of 1.25 in the survey of 1,266 individuals across disaster-prone rural households [PLOS Climate 2024], and the financing side of that burden is chapter 09's ledger. The heat and air mechanism is urbanisation without enforcement: kilns, construction dust, exhaust and transboundary smoke hold the airshed above the guideline everywhere, while cooling demand loads onto the coal heavy power system chapter 10 documents [Climate Watch 2024] [Ember 2025].

The finance mechanism explains why the response is smaller than the exposure. The 5.24 percent FY26 share sits inside a budget whose revenue to GDP ratio is chapter 04's central problem and whose execution lapses allocations in the fourth quarter. The Delta Plan 2100, the General Economics Division's 2018 hundred year framework, named the problem correctly and put forest cover at 17 percent of the land, but its investment and financing targets could not be confirmed from the cited sources; the plan document would settle it [GED Plans 2018]. The external leg is a cliff: the grant and concessional windows that LDC status opens, the Least Developed Countries Fund among them, thin out at graduation, which chapter 02 dates to November 2026, and the loss and damage fund agreed at COP28 arrives as an operating entity whose replenishment scale could not be confirmed from the cited sources; the UNFCCC decisions would settle it.

Bangladesh's NDC update carries mitigation and adaptation cost estimates that could not be confirmed from the cited sources; the UNFCCC submission would settle it.

Priorities through 2036

The named projections that frame the FY27 to FY36 window the chapter 15 scenarios assume are the World Bank's. The 2022 Country Climate and Development Report is the registry source for the finding that climate and air impacts subtract from GDP and adaptation pays; its specific GDP loss range and adaptation financing requirement for Bangladesh could not be confirmed from the cited sources, the published document being the source that would settle it, as is its Groundswell based internal climate migration projection for Bangladesh [WB CCDR 2022]. What the verified record supports is the direction of all three channels: affected counts run into the tens of millions in bad flood years [EMDAT 2026], freshwater per head falls about 5.5 cubic metres a year on the 2017 to 2022 trend [WB WDI 2026], and the 2024 pre monsoon set grid records in a third of the city grids with the country anomaly series near its peak [NASA POWER 2024] [Berkeley Earth 2026].

Three transmission channels run from those facts into the macro scenarios of chapter 15. The first is agriculture: the sector that grew at 3.37 percent in FY23 and stayed inside a 3.05 to 3.54 percent band across FY18 to FY23 in chapter 01's record [BBS NA 2024] now holds 44.7 percent of employment, chapter 08's post flood reversal [ILO 2025]; this chapter's arithmetic assumes the sector's growth ceiling tightens as salinity and heat eat into the yield gains that took cereal yields to 5,408 kilogram per hectare in 2024 [WB WDI 2026], and the stall scenario of chapter 15 assumes the agriculture employment share stays above 45 percent into the late 2030s.

The second is the fiscal channel: every adaptation taka competes with the interest bill and the capacity payments in chapter 04's budget, and the reform scenario assumes the state protects a rising climate allocation share while the stall scenario assumes the water budget holds flat in nominal taka and shrinks in real terms. The third is the urban channel: flood and erosion displacement lands in Dhaka's informal settlements, the pressure chapter 14 discusses into land and housing, and this chapter's arithmetic assumes the slow hazards add to the in migration flow the 2020s already delivered.

The main decisions. The water decision, led by the Ministry of Water Resources with the Water Development Board, is whether the programme shifts from reactive bank protection toward sediment management, drainage and coastal polder rehabilitation; the indicator to monitor is the water ADP's maintenance and rehabilitation share [Planning Commission ADP 2025]. The air decision, led by the Department of Environment, is whether the brick kiln conversion, vehicle standards and construction dust rules are enforced against the measured exposure; the indicator to monitor is the PM2.5 series leaving the 40s [WB WDI 2026].

The cooking and energy decision, led by the power and energy divisions with chapter 10's fuel arithmetic, is whether clean fuel access moves off 28.4 percent, because that single number carries the larger of the two air death counts [WB WDI 2026]. The finance decision, led by the Economic Relations Division with the environment ministry, is how much of the post graduation concessional envelope is reserved for climate; the indicator to monitor is the ERD commitment series' climate share. The loss and damage decision, led by the UNFCCC board with the environment ministry as claimant, is whether the fund moves from establishment to disbursement at a scale a single cyclone year's 2.49 billion USD of recorded damage makes meaningful [EMDAT 2026].

Risks and opportunities

Risks. First, a polder and embankment failure: a cyclone landfall at high tide against a run down polder system in the southwest would flood a protected basin rather than a field, and the indicators to monitor are the Water Development Board's damage reports and the EM-DAT damage series against its 4.60 billion USD single year record of 2004 [EMDAT 2026] [BWDB 2023]. Second, a water scarcity regime in the southwest and the barind: if freshwater per head keeps falling about 5.5 cubic metres a year and dry season withdrawals hold, the aquifer breach visible in Bagerhat's drinking water shares spreads up the coast, and the indicators to monitor are the census drinking water tables, the per capita freshwater series and the soil survey's exceedance share [WB WDI 2026] [BBS Census 2022, district drinking water tables] [Zenodo 2024, DOI 10.5281/zenodo.14560019].

Third, an adaptation finance squeeze: graduation closes LDC windows as the slow hazard bill compounds, and if external climate commitments do not replace them, the 5.24 percent allocation share becomes the ceiling rather than the floor, and the indicator to monitor is the ERD climate commitment share against the programme's committed cost of 68,796 crore BDT [Planning Commission ADP 2025].

Upside. First, the adaptation brand is an asset: a country that cut disaster deaths an order of magnitude is the credible claimant for every climate window, and the indicator to monitor is the ERD climate series and the loss and damage fund's first disbursements. Second, the clean cooking pivot is the cheapest adaptation on the table: it attacks the 117,681 household air pollution deaths, the pressure on a 14.49 percent forest base and the fuel import bill of chapter 10 at once, and the indicator to monitor is the clean fuel access series off its 28.4 percent [WHO GHO 2019] [WB WDI 2026].

Third, forced institutional upgrade: the Delta Plan's governance architecture, if its investment targets are published and budgeted, is the vehicle that would move water management from scheme by scheme patronage to basin scale programming, and the indicator to monitor is the first Delta Plan financing report with a budget line [GED Plans 2018].

Testing the policy case

Water, environment and disaster-management agencies should compare maintenance and adaptation delivery with exposure-adjusted losses. Lower annual disaster deaths alone would not establish reduced risk. Continued damage under comparable hazards would challenge the effectiveness of the selected protection measures and justify revisiting design or maintenance.

What to watch

The thresholds below are author-proposed monitoring markers, not official targets or estimated policy effects. Interpret them alongside the source dates and definitions.

1. Flood affected area. Latest cited value 37,925 square kilometre, 25.7 percent of the country, in 2023, the fifth consecutive year above a quarter of the territory [BWDB 2023]. Threshold: a year above 35 percent with damage above the 4.60 billion USD single year record of 2004 marks the embankment failure regime; a run of years below 20 percent marks the sediment and drainage programme working.
2. Renewable freshwater per head. Latest cited value 620 cubic metres in 2022, from 647 in 2017 [WB WDI 2026]. Threshold: a fall below 550 cubic metres would be consistent with the scarcity regime the census drinking water tables already show on the coast; stabilisation above 600 requires withdrawal pricing or recharge programmes absent from the budget.
3. Mean PM2.5 exposure. Latest cited value 69.3 micrograms per cubic metre in 2023, using the same World Bank series as chapter 30 [WB WDI 2026]. Proposed assessment: a sustained decline from the 2023 exposure level would indicate improvement; compare like-for-like estimates and monitoring data. A value below 35 micrograms per cubic metre is an ambitious marker, not evidence that a particular intervention caused the change.
4. Climate relevant development allocation. Latest cited value 12,058.82 crore BDT in the FY26 development programme, 5.24 percent of the 230,000.00 crore BDT the National Economic Council approved, against 68,795.5 crore BDT of committed project cost in the revised programme's project table [Planning Commission ADP 2025]. Threshold: a sustained share above 8 percent marks the adaptation financing turn; a share below 4 percent would be consistent with the stall scenario's assumption that the water budget erodes in real terms.
5. Total emissions including land use. Latest cited value 265.26 million tonne of CO2 equivalent in 2022, with fossil CO2 at 108.32 million tonne in 2024 after four consecutive annual increases [Climate Watch 2024] [GCP 2024, co2]. Threshold: fossil CO2 passing 125 million tonne by around 2030, which the 2020 to 2024 trend of about 3.1 million tonne a year projects absent the fuel switch chapter 10 discusses, locks the energy import path and the airshed together for the decade.

Sources used

[EMDAT 2026] EM-DAT Emergency Events Database, UCLouvain, Bangladesh natural disaster series via bdpolicy.db, series: emdat_disaster_events, emdat_disaster_deaths, emdat_disaster_affected, emdat_disaster_damage.

[Germanwatch CRI 2025] Germanwatch Climate Risk Index 2025, Table 5, Climate Risk Index for 1993 to 2022, Bangladesh rank 31 and score 12.355940 among 174 ranked countries, , accessed 6 September 2026.

[Germanwatch CRI 2021] Germanwatch Climate Risk Index 2021, Bangladesh rank 7 and score 28.33 for 2000 to 2019, via bdpolicy parquet germanwatch_cri_bd.

[UNICEF CCRI 2021] UNICEF Children's Climate Risk Index 2021, Bangladesh score, rank and pillars, via bdpolicy parquet unicef_ccri_bd.

[BWDB 2023] Bangladesh Water Development Board historical flood record via bdpolicy parquet, series: flood affected area and share of total area, 1954 to 2023.

[FAO 2026] FAO flood impact mapping for Bangladesh via bdpolicy parquet, flooded area and population exposed by district and period, 2024 monsoon and April 2026 pre monsoon events.

[WB WDI 2026] World Bank World Development Indicators via bdpolicy parquet wb_environment_bd_deep, series: EN.POP.EL5M.ZS, EN.CLC.MDAT.ZS, ER.H2O.FWTL.ZS, ER.H2O.INTR.PC, EN.ATM.PM25.MC.M3, EN.ATM.PM25.MC.ZS, EG.CFT.ACCS.ZS, AG.LND.FRST.ZS, AG.YLD.CREL.KG.

[BBS Census 2022] Bangladesh Bureau of Statistics, Population and Housing Census 2022, district drinking water source tables via bdpolicy parquet.

[Berkeley Earth 2026] Berkeley Earth country mean temperature anomaly, monthly, 1940 to March 2026, Bangladesh, via the climate/owid_temp_anomaly_bd parquet, Our World in Data mirror, decadal and annual means by chapter arithmetic.

[Zenodo 2024] Bangladesh coastal soil salinity in-situ survey, 162 GPS-tagged sites, Satkhira, Khulna and Bagerhat belt, March 2024, MIT/WECG Laboratory sampling, SRDI salinity classes, DOI 10.5281/zenodo.14560019, via bdpolicy.db series coastal_soil_salinity_*.

[PLOS Climate 2024] PLOS Climate waterlogging and morbidity survey, 1,266 individuals across rural disaster-prone households, article pclm.0000605, via bdpolicy.db series climate_waterlogging_ and climate_survey_.

[WHO GHO 2019] World Health Organization Global Health Observatory via bdpolicy parquet who_gho_bd, series: ambient air pollution attributable deaths (AIR_41) and household air pollution attributable deaths (AIR_11), 2019.

[NASA POWER 2024] NASA POWER reanalysis, T2M_MAX monthly for twelve Bangladesh city grids, 1981 to 2024, via the climate/nasa_power_bd parquet, April to May means and record ranks by chapter arithmetic.

[Climate Watch 2024] Climate Watch historical emissions for Bangladesh, WRI, sector GHG totals including and excluding LULUCF, 2022, via bdpolicy parquet climatewatch_bd_ghg.

[GCP 2024] Global Carbon Project Global Carbon Budget country series, Our World in Data mirror, via bdpolicy parquet climate/owid_co2_bd, series: co2, co2_per_capita, share_global_co2, population, 2020 to 2024.

[Ember 2025] Ember yearly electricity data for Bangladesh, coal, gas and renewables share of electricity generation 2015 to 2025, via the climate/ember_yearly_bd parquet.

[Forest Department 2023] Bangladesh Forest Department, forest land by type FY2022-23, total area and mangrove share, via bdpolicy parquet.

[Planning Commission ADP 2025] Annual Development Programme 2025-2026, Programming Division, Bangladesh Planning Commission, June 2025, ministry wise allocation table, in crore BDT, with every share taken on the 230,000.00 crore programme approved by the National Economic Council on 18 May 2025, , accessed 6 September 2026. The same tag covers the revised FY2025-26 project table held in the available data as bdpolicy.db adp_projects, ministry and agency allocations, project counts and committed costs in lakh taka converted to crore, a separate and later vintage named as the revised programme wherever it is used.

[GED Plans 2018] General Economics Division, Bangladesh Delta Plan 2100, quoted for forest cover share and framework, investment targets not held in the available data.

[WB CCDR 2022] World Bank Bangladesh Country Climate and Development Report 2022, resolving source for GDP impact, adaptation financing and internal climate migration projections; figures not held in the available data.

[BBS NA 2024] Bangladesh Bureau of Statistics national accounts via bdpolicy.db, series: bbs_gdp_growth_agriculture_pct, referenced from chapter 01.

[ILO 2025] ILOSTAT modelled estimates via bdpolicy parquet, series: ilo_emp_temp_sex_eco_nb_a, referenced from chapter 08.