Land cover and change analytics

The problem

Planning, agriculture and disaster-risk teams need to know how land is changing: where cropland is left fallow, where settlements expand, how land cover shifted after a policy or event. Global land cover products rarely match local classes or administrative boundaries, and raw change maps are hard for decision-makers to read without summary statistics and a way to explore them.

What you get

• Land cover or change layers for the agreed area and time periods (GeoTIFF / Cloud-Optimised GeoTIFF)
• Change statistics summarised by administrative unit or other client boundaries (CSV and PostGIS tables)
• Derived thematic products where relevant, such as fallow-land maps
• An interactive web view or GeoServer layers for exploring the results
• Methods note covering data sources, classes, accuracy checks and limitations

How it works

1. 1

   Define

   Agree the classification scheme, periods of interest, reporting units and the decision the analysis supports.

2. 2

   Assemble data

   Select and preprocess satellite imagery or existing land cover products (for example Landsat Collection 2 or Sentinel-2 L2A composites) in Google Earth Engine.

3. 3

   Analyse

   Produce classifications or change maps, derive thematic layers such as fallow land, and compute zonal statistics.

4. 4

   Publish

   Serve the layers and statistics through GeoServer and PostGIS with a lightweight web interface for browsing and download.

Tools

• Google Earth Engine
• Landsat Collection 2
• Sentinel-2 L2A
• QGIS
• GDAL
• PostgreSQL / PostGIS
• GeoServer
• React / Next.js