High-resolution land-surface temperature

The images show an agricultural area in southern Germany in late May 2017. The Copernicus Sentinel-2 image (left) was acquired at 20 m spatial resolution. This allows agricultural parcels and other landscape features such as roads to be distinguished. The Copernicus Sentinel-3 image (centre) captures the land-surface temperature, which is essential for estimating evapotranspiration, but here with a pixel size of around 1 km. By using advanced machine-learning algorithms, data from the two sensors can be fused, thus obtaining a 20 m representation of land-surface temperature (right) which can then be used to produce 20 m evapotranspiration maps.

Read more: Satellites and machine learning for water management