EO Capability Benefits
Cloud masking, i.e. filtering out pixels where clouds block the spectral information from the actual object of analysis, such as land surfaces, is a critical pre-processing step for many Earth Observation applications. It allows users to study only images where the percentage of cloud cover falls below a user-defined threshold. Moreover, cloud monitoring is useful in itself for weather forecasting and tracking cloud systems, providing critical information for disaster management and early warning. Long-term cloud trends inform climate studies, while cloud cover data also improves solar energy forecasting and site selection for solar farms.
EO Capability Description
Cloud detection algorithms exploit cloud properties observed through multispectral sensors. These properties include high reflectivity in the visible spectrum and low temperature in thermal bands. Clouds also exhibit higher reflectance in the near-infrared compared to snow and ice, which also appear bright and could otherwise easily be confused for clouds.
When downloading pre-processed multispectral imagery such as from Sentinel-2, datasets usually include a pixel quality layer. In the case of Sentinel-2, this is the Scene Classification Layer (SCL) which includes, among other classes, information on pixels covered by clouds, whereby high and medium probability clouds are differentiated; areas affected by cloud shadows; thin, high-altitude cirrus clouds.
Indicative Cost Range Details
As cloud detection and masking is an integral part of processing chains for multi- and hyperspectral imagery, it is usually not sold as a separate data product. EUMETSAT’s Satellite Application Facility on Climate Monitoring (CM SAF) offers various cloud products as daily or monthly composites.
Indicative Cost Range
Free