FloodWatch Ghana

Data: SRTM DEM - GPM IMERG Rainfall - ESA WorldCover - OSM | GeoBuilders Africa

Layers

Opacity 80%

Flood Risk Score

Low (0) High (1)
Validated — May 18, 2025 Flood Event
Greater Accra Floods · 132mm · 4 deaths · 3,000+ displaced
All 7 flooded districts correctly flagged as high risk. Recall 1.00 · F1 0.44.
Rainfall powered by NASA GPM IMERG — actual observed data, not climatological averages. Static structural model — event-driven dynamic layer planned for v1.1.
View methodology →
FloodWatch Ghana
GeoBuilders Africa - v0.1 - Greater Accra Region

What is this map?

FloodWatch Ghana is an open source flood risk intelligence system for Greater Accra. In its v0.1 "raw" state, it provides a high-resolution, static baseline of structural flood vulnerability. This methodology is optimized for long-term urban planning, infrastructure assessment, and city-wide resilience strategy.

What does the risk score mean?

0.0 - 0.33Low Risk
0.33 - 0.67Moderate Risk
0.67 - 1.0High Risk

The score reflects chronic structural vulnerability - areas that are low-lying, poorly drained, and near water bodies. This static baseline acts as a foundation for long-term planning, identifying where the city's infrastructure is most at risk regardless of individual weather events.

How is the score calculated?

Data layer Source Weight
Elevation (DEM) NASA SRTM 30%
Precipitation NASA GPM IMERG Final Run 25%
Terrain slope Derived from SRTM 20%
Land cover ESA WorldCover 2021 15%
Water proximity OpenStreetMap 10%

Validation & Accuracy

FloodWatch v0.1 was validated against the May 18, 2025 Greater Accra flood event (132mm rainfall, 4 deaths, 3,000+ displaced). Quantitative results across all 29 districts:

  • Recall: 1.00 — all 7 flooded districts correctly flagged as high risk
  • F1 Score: 0.44 — up from 0.31 in the original model
  • Flooded district mean risk (0.811) now correctly exceeds non-flooded (0.774)

Rainfall is powered by NASA GPM IMERG Final Run — actual satellite-observed monthly precipitation bias-corrected against ground gauges — replacing the climatological averages used in earlier versions. This makes risk scores responsive to real rainfall conditions rather than historical means.

As a static structural model, it identifies chronic vulnerability — areas that are low-lying, poorly drained, and exposed. Event-specific flash flood dynamics (drainage overload, extreme single-day rainfall) are planned for v1.1 as a dynamic risk layer.

How to use

Click any district to see detailed risk statistics. Use the layer controls (top left) to toggle layers and adjust opacity. The map updates monthly as new rainfall data becomes available.

Built by

FloodWatch Ghana is an open source initiative by GeoBuilders Africa, founded by Rachel Atia, a Geospatial Data Engineer based in Accra, Ghana.