Executive Summary. Satellite elevation data shows 41,958 km2 of Bangladesh, roughly 28% of the country, already sits at or below one meter above sea level, and the LECZ curve is nearly flat across the first three meters: raising the threshold from one meter to three meters adds barely 2,300 km2. The thesis follows directly: the first meter of sea-level rise will not add incrementally to flood exposure, it will swamp terrain that is functionally already at sea level. Sea levels along the Bangladesh coast are rising at 4.58 mm per year, and the IPCC's high-emission scenario (SSP5-8.5) projects roughly 0.77 m of rise in the Bay of Bengal by 2100, more once local land subsidence is added. Night-time land surface temperatures have risen 1.5 degrees Celsius since 2012, adding roughly 10% to atmospheric moisture capacity and feeding the rainfall volatility the country already cannot plan around. The interim government should anchor the response in the existing Bangladesh Delta Plan 2100 by raising delta-management spending from the current 0.8% of GDP toward the planned 2.5%, codifying the one-meter LECZ as a binding land-use category, and converting satellite rainfall and night-temperature records into probabilistic advisories with named owners and measurable triggers.

From orbit, Bangladesh does not look like a country under siege. The green of its rice paddies, the silver threads of its rivers, the white of its cloud cover: these are the colors of a nation that has learned to live with water. The satellite data tells a different story. Strip away the visible spectrum and measure elevation, and a single finding governs everything that follows: 41,958 km2, about 28% of Bangladesh, already lies at or below one meter above sea level, and the next two meters of elevation add almost no land. The country is not approaching the waterline. Across its coastal third, it is already at it.

This is Part 4 of the Satellite Bangladesh series. Where previous installments tracked floods, urbanization, and vegetation change, this one examines the slower, more structural threat: coastal exposure, rainfall volatility, and a warming land surface that is reshaping the terms of life for 170 million people.

A country below the waterline

The Low Elevation Coastal Zone, or LECZ, is the standard measure of a country's exposure to sea-level rise: how much land sits below a given elevation threshold. For Bangladesh, the shape of the answer matters more than any single number.

At the one-meter threshold, 41,958 square kilometers of Bangladesh's land area already lies at or below that elevation. That is roughly 28% of the country's total area of approximately 147,570 km2. Raise the threshold to two meters and the number barely changes: 42,840 km2. At three meters, it is 44,265 km2. The curve only begins to climb steeply at higher thresholds: 50,317 km2 at five meters (34% of the country), and 65,310 km2 at ten meters (44%).

The near-flatness of the curve between one and three meters is the critical finding, and it is what separates this analysis from a generic warning about a low-lying delta. The first meter of sea-level rise, which the IPCC's high-emission scenario projects for the Bay of Bengal this century, would not nibble at the margin of an otherwise dry country. It would inundate land that is already functionally at sea level. The difference between one meter and three meters of elevation across 42,000 to 44,000 km2 is, in practical terms, the height of a single storm surge. This is consistent with the long-standing IPCC finding, cited in Bangladesh's own planning literature, that a one-meter rise would submerge on the order of one-fifth of the country's landmass.

WorldPop population estimates for 2020 indicate that approximately 5.26 million people live within the five-meter LECZ. This is a conservative figure. It captures only the population residing below five meters of elevation, not the far larger number who depend economically on coastal zones or who would be displaced by saltwater intrusion into groundwater and agricultural land that extends well beyond the inundation boundary.

The three coastal divisions, Barishal, Chattogram, and Khulna, bear the overwhelming burden of this exposure. Khulna, home to the Sundarbans and the country's most vulnerable coastline, faces the highest composite risk. Barishal, with its low-lying chars and islands, is close behind. Even Chattogram, despite its hillier eastern terrain, has a densely populated coastal strip that sits squarely in the danger zone. The LECZ figures here are national-level, not disaggregated by division, so division-specific vulnerability mapping requires ground-level survey data beyond the scope of this satellite analysis.

The rain is getting wilder

Coastal elevation is a static vulnerability. Rainfall is a dynamic one, and the satellite-era precipitation record from 1985 to 2023 shows it becoming more unpredictable.

Mean annual rainfall across Bangladesh has oscillated between 2,130 mm in 1989, the driest year in the satellite record, and 3,148 mm in 2017, the wettest. The year 2023, at 2,134 mm, was nearly as dry, the second-driest on record. That is a ratio of roughly 1.48 to 1 between the extremes. For a country whose agricultural calendar, drainage infrastructure, and flood defenses were designed around a narrower band of rainfall, this volatility is itself a crisis.

The monsoon component, which delivers the bulk of annual precipitation between June and September, shows its own pattern of instability. Monsoon rainfall ranged from 1,293 mm in 1992 to 1,931 mm in 2004. More recently, the swings have been sharp: 1,355 mm in 2022 followed by 1,864 mm in 2024, a 38% jump in two years.

What makes these numbers dangerous is not any single extreme year but the inability to predict which kind of year is coming. A farmer in Barishal planning boro rice cultivation faces fundamentally different conditions depending on whether the monsoon delivers 1,300 mm or 1,900 mm. Drainage engineers in Dhaka must design for both. Insurance actuaries must price for both. The satellite record says that both are equally plausible in any given year.

The 2017 spike is worth isolating. At 3,148 mm mean rainfall, that year was an outlier by any measure. It coincided with catastrophic flooding in Sylhet and Sunamganj that displaced millions and caused billions of taka in agricultural losses. The maximum recorded rainfall in 2017 reached 7,419 mm at the wettest station, more than double the national mean. That spatial concentration, heavy rainfall piling into already-saturated catchments, is the mechanism that turns high-rainfall years into disaster years.

The land is warming

Behind the rainfall volatility sits a thermal trend that is easier to miss but no less consequential. MODIS land surface temperature data from 2000 to 2024 reveals what is happening to the ground itself.

Daytime land surface temperatures have fluctuated between 26.6 degrees Celsius and 27.8 degrees Celsius over the 25-year record, with no strong directional trend. The surface heats and cools with seasonal and inter-annual variation, but the envelope has remained relatively stable.

The night-time record is different. Night LST started at 20.6 degrees Celsius in 2000 and reached 20.9 degrees Celsius in 2024, but the trajectory is not linear. The coolest year was 2012 at 19.4 degrees Celsius, creating a U-shaped pattern: cooling from 2000 to 2012, then warming sharply from 2012 to 2024, a 1.5-degree rise in twelve years. That recent warming in night-time surface temperature matters for two reasons.

First, warmer nights mean the atmosphere retains more moisture. The Clausius-Clapeyron relation dictates that for every degree of warming, the air can hold roughly 7% more water vapor. A 1.5-degree increase in overnight temperatures since 2012 translates to about 10% more moisture capacity, which feeds directly into more intense rainfall events when convective systems develop.

Second, warmer nights reduce the thermal recovery window for crops, livestock, and human bodies. Rice yields decline measurably once night-time temperatures rise above roughly 22 degrees Celsius: the canonical IRRI field record finds grain yield falling about 10% for each 1-degree increase in growing-season minimum temperature (Peng et al., PNAS, 2004). Heat stress in humans is a function not just of peak daytime temperature but of cumulative thermal load, and nights that do not cool below 21 degrees prevent the physiological recovery that hot-climate populations depend on.

The combination of volatile rainfall and rising night-time temperatures creates a compounding effect that no single dataset captures on its own. More moisture in the atmosphere, plus unpredictable monsoon timing, plus a coastline that sits within a few meters of sea level: this is the structural equation that defines Bangladesh's climate exposure.

What the satellites cannot see

The satellite record measures elevation, precipitation, and temperature with impressive precision. What it cannot measure is the human infrastructure that determines whether physical exposure translates into actual harm.

It cannot see the condition of the embankments that protect Khulna's shrimp farms from tidal surges. It cannot assess whether the polders in Barishal have been maintained or whether their sluice gates function. It cannot tell us how many of the 5.26 million people in the five-meter LECZ have access to cyclone shelters, or whether those shelters have been provisioned for the next event.

It also cannot capture the slow-onset damages that produce no dramatic imagery: the salinization of freshwater aquifers as sea levels creep upward, the declining productivity of coastal soils, the mental health toll on communities that live in permanent uncertainty about whether the next monsoon will be a 1,300 mm year or a 1,900 mm year.

These are the gaps that ground-level survey data, community reporting, and institutional monitoring must fill. The satellite provides the map of exposure. The response requires a different kind of intelligence entirely.

What policy must do

The satellite evidence points to three interventions. Bangladesh already has the umbrella framework: the Bangladesh Delta Plan 2100 (BDP 2100), adopted in 2018 and coordinated by the General Economics Division of the Planning Commission. The gap is funding and operational specificity, not strategy on paper. The plan calls for delta-management spending to rise to about 2.5% of GDP per year; current spending sits near 0.8%. Each recommendation below names an owner, a budget order-of-magnitude, and a signal by which success can be measured.

First, make the one-meter LECZ a binding land-use category, not an advisory line. Owner: the General Economics Division and the Ministry of Housing and Public Works, through the National Building Code and divisional master plans. The fact that 41,958 km2 sits below one meter of elevation is not new to climate science, but it is not yet operationalized in zoning, building codes, or capital-budgeting rules. Every new road, school, and hospital sited within the one-meter LECZ should be required to meet an elevated flood-resilience standard before it can draw public funds. Budget signal: this is a design-standard change, largely cost-neutral at the policy level, raising unit construction cost in the zone on the order of single-digit percentages rather than requiring a new program. Success signal: 100% of new ADP-financed public buildings inside the one-meter LECZ certified to the elevated standard within three fiscal years, reported in the annual BDP 2100 progress review.

Second, fund and mandate a probabilistic seasonal forecast as the BDP 2100 short-term flagship. Owner: the Bangladesh Meteorological Department and the Flood Forecasting and Warning Centre under the Ministry of Water Resources. Current systems forecast around historical means, but the mean is the least useful number here: the gap between a 2,130 mm year and a 3,148 mm year is the gap between drought stress and catastrophic flooding. Seasonal products must publish probability ranges, not point estimates, with named downstream users (district agriculture offices, city drainage authorities) required to act on the bands. Budget signal: instrumentation, computing, and dissemination on the order of tens of millions of US dollars, well inside the roughly 2% of GDP BDP 2100 gap between current and planned delta spending. Success signal: forecast skill (Brier score) for the June-to-September monsoon total improved year on year, and a measurable rise in the share of coastal districts that adjust boro and aman planting on the basis of the seasonal band.

Third, fold night-time land-surface temperature into the existing agro-met advisory, with a hard alert threshold. Owner: the Department of Agricultural Extension and the Bangladesh Meteorological Department, building on the agro-meteorological advisory already issued to farmers. The upward drift in night LST is a slow threat to rice yields and human health that current advisories do not track. The trigger is already known from the field literature: when night temperatures exceed roughly 22 degrees Celsius during the boro flowering window, yield loss begins. Budget signal: this is an add-on to an existing service, not a new institution, in the low single-digit millions of US dollars for data integration and SMS dissemination. Success signal: a night-temperature heat alert reaching boro-growing unions and urban heat-health systems within 24 hours of a threshold breach, with documented uptake measured in the next crop-cutting survey.

Bangladesh cannot move its coastline upward. It cannot control the monsoon. It can measure what is happening, model what is coming, fund the plan it has already written, and assign each task an owner who can be held to a number. The satellite record has done the measuring. The rest is delivery.

Part 5 will examine the human geography that these physical forces are reshaping: spatial inequality, poverty persistence, and the widening fracture between those who benefit from Bangladesh's growth and those who are being left behind.

Sources

• SRTM elevation / LECZ: SRTM 30 m DEM, processed via Google Earth Engine. Internal output: bd_gis/outputs/coastal/lecz_areas.csv. These LECZ areas are an internal satellite-derived series; the one-meter result (about 28% of land area) is directionally consistent with the long-standing IPCC finding, cited in Bangladesh's planning literature, that a one-meter rise would submerge roughly one-fifth of the country.
• WorldPop population 2020: WorldPop 100 m resolution, clipped to 5 m elevation zone. https://www.worldpop.org/
• CHIRPS rainfall: Climate Hazards Group, 5.5 km, 1985-2023. https://www.chc.ucsb.edu/data/chirps. Internal outputs: bd_gis/outputs/climate/rainfall_timeseries.csv, monsoon_rainfall.csv.
• MODIS LST: NASA Terra MOD11A2, 1 km, 2000-2024. https://lpdaac.usgs.gov/products/mod11a2v006/. Internal output: bd_gis/outputs/climate/lst_timeseries.csv.
• Sea level rise rate 4.58 mm/yr: Springer Nature, "Assessing the correlation between sea level rise, temperature, and erosion-accretion along the coastline of Bangladesh," Discover Geoscience, 2025. https://link.springer.com/article/10.1007/s44288-025-00129-2
• IPCC AR6 sea-level projection (Bay of Bengal ~0.77 m by 2100, SSP5-8.5): IPCC Sixth Assessment Report, WG1, as summarized for the Bay of Bengal; effective coastal rise reaches 0.85-1.40 m once local subsidence is added. https://www.ipcc.ch/report/ar6/wg1/
• Rice yield vs night temperature: Peng, S. et al., "Rice yields decline with higher night temperature from global warming," PNAS, 2004. https://www.pnas.org/doi/10.1073/pnas.0403720101
• Bangladesh Delta Plan 2100 (spending 0.8% to planned 2.5% of GDP; coordinated by General Economics Division): General Economics Division, Planning Commission, BDP 2100. https://bdp2100kp.gov.bd/
• National area reference: 147,570 km2, BBS/FAO.

Data Sources

• LECZ elevation: SRTM 30m DEM, processed via Google Earth Engine. Output: bd_gis/outputs/coastal/lecz_areas.csv. Thresholds: 1m, 2m, 3m, 5m, 10m.
• Population in LECZ: WorldPop 2020 (100m resolution), clipped to 5m elevation zone.
• Rainfall: CHIRPS 5.5km precipitation, 1985-2023 (biennial composites). Output: bd_gis/outputs/climate/rainfall_timeseries.csv, monsoon_rainfall.csv.
• Land surface temperature: MODIS Terra LST (MOD11A2), 1km, 2000-2024 (biennial composites). QA-masked. Output: bd_gis/outputs/climate/lst_timeseries.csv.
• National area reference: 147,570 km2 (BBS/FAO).