terraServe API

Land managers, investors, agencies and researchers who model landscapes need consistent, long-term information on how land cover and ecosystems change over large areas. terraServe is terraPulse's application programming interface (API) to the terraPulse data cube, a record of the cover, structure and function of terrestrial ecosystems that terraPulse mines from satellite measurements with machine learning. The cube holds gridded map layers such as tree and forest cover, surface water, impervious surface, aboveground biomass, vegetation indices and phenology, wildfire probability and species habitat, at spatial resolutions from 1 to 500 metres. Measurements extend from 1984 to the present and are updated daily or annually depending on the variable. terraPulse names Landsat and Sentinel imagery among the sources behind its layers, and trains its biomass estimates on spaceborne lidar and other reference measurements.

Through the API, the cube can be subset spatially, temporally or thematically to fit a study area, subject and time span. Numerical, science-grade values are served in tabular and gridded formats including CSV, JSON and GeoTIFF, and any layer can be exposed as XYZ map tiles (256 x 256 PNG, EPSG:3857) for display in ArcGIS, QGIS, Google Maps, Mapbox, Leaflet, OpenLayers and other mapping applications. terraPulse also hosts third-party layers, such as the USDA NASS Cropland Data Layer, for download through terraServe. The API runs on Amazon Web Services, using Amazon API Gateway with AWS Lambda.

terraServe is sold as a subscription to developers and data scientists through individual or institutional accounts. terraPulse states that its platform, of which the API is one of three components, is used by organisations including the World Bank, the US Forest Service, the Bureau of Land Management and NASA.