Mapping 40 Years of Flood Exposure in Bangladesh

A Data-Driven Look Using JRC Global Surface Water, Landsat, and HydroRIVERS (1984–2024)

Bangladesh is one of the most flood-affected countries in the world. Its landscape—woven by the Padma, Jamuna, Meghna, and hundreds of tributaries—is shaped each year by monsoon rainfall, upstream flows, and seasonal inundation. But how has flood exposure changed over time? And which regions experience the most persistent flooding?

To answer these questions, I created a nationwide Flood Exposure Map covering nearly four decades (1984–2024) using Google Earth Engine, combining data from:

• JRC Global Surface Water (Landsat-based, 1984–2021)

• Landsat-8/9 Surface Reflectance (2013–2024)

• HydroSHEDS FreeFlowingRivers (major river network)

This dataset helps visualize long-term flood patterns, recent monsoon flooding, and how they align with the river systems of Bangladesh.

Why This Map Matters

Flooding in Bangladesh is not only a natural process; it’s a central part of the country’s geography, agriculture, and risk landscape. Communities, croplands, embankments, and roads are shaped by:

• monsoon rainfall

• Himalayan snowmelt

• upstream dam releases

• regional sea-level rise

• riverbank migration and erosion

A multi-decade flood exposure map helps:

• identify chronically inundated areas

• separate seasonal water from permanent rivers/lakes

• support agricultural planning

• guide disaster risk reduction

• reveal shifts in river dynamics

Data Sources Used

1. JRC Global Surface Water (1984–2021)

Derived from the entire Landsat archive, this dataset provides:

• Water Occurrence — percentage of time a pixel is water

• Seasonality — number of months per year water is present

We use these to detect long-term flood-prone areas, excluding permanent rivers.

2. Landsat-8 & Landsat-9 (2013–2024)

Using NDWI (Normalized Difference Water Index), we calculate:

• Recent monsoon flood frequency

• Only considering June–October

• Masks out clouds/shadows

This shows how flooding has behaved in the last decade, useful for current planning.

3. HydroSHEDS FreeFlowingRivers

Includes major rivers such as:

• Padma

• Jamuna

• Meghna

• Teesta

• Surma

• Brahmaputra

Used for geographic context and to explain spatial flood patterns.

What the Map Shows

Long-Term Flood Exposure (1984–2021)

The JRC-based map highlights:

• High-exposure floodplains along the Brahmaputra, Jamuna, and Padma

• Seasonal wetlands (haors) in northeastern Bangladesh

• Riverine sandbars (chars) that appear and disappear over decades

• Areas where rivers have migrated or widened

The colors reflect the proportion of Landsat images where water was detected:

• Dark blue → frequently underwater

• Light blue → occasional flooding

• White → mostly dry / stable land

Recent Monsoon Flood Frequency (2013–2024)

The Landsat NDWI map reveals:

• Increasing monsoon inundation in the north-central floodplain

• Stable but recurring flooding in Sylhet haor basins

• Coastal flooding signatures from tidal surges

This gives a real-time view of how flood patterns are shifting.

Key Insights

• The Jamuna–Brahmaputra floodplain remains Bangladesh’s most consistently inundated region.

• Seasonal wetlands (haors) in the northeast show very high historical water presence.

• Coastal flood exposure is lower in the long-term record but increasing in recent years.

• Flooding around the Padma has shifted over time due to river migration and sediment loads.

• Landsat-based monsoon detection reveals more frequent short-term flooding in central Bangladesh.