Radiometrically Terrain Corrected Products for Biomass Mapping

In a European Space Agency funded project to generate annual estimates of the global spatial distribution of the critical climate indicator “biomass” (CCI BIOMASS project), Earth Big Data's software was used to scale radiometric terrain correction of millions of Sentinel-1 data for Gamma Remote Sensing employing the Copernicus 30 m DEM.

Annual global acquisitions of Sentinel-1 data sets acquired in the interferometric wide-swath mode (IW) were processed from the ground range detected format (GRD) into GeoTIFFs and tiled into 1x1 degree tiles for the time series modeling of biomass. As an example, for the year 2017, before the loss of Sentinel-1B, more than 305,000 RTC products at 720 pixel per degree resolution were produced. During processing, the Cloud infrastructure was scaled up to 2,000 AWS elastic compute cloud instances with 32 GB RAM and 4 virtual CPUs. With this fleet of compute instances, a throughput of 30 TB/hour was achieved for reading data from the Alaska Satellite Facility DAAC.

The entire processing job reduced an input volume of approximately 330 TB to an output of 4 TB in about 13 hours. The AWS cloud monitoring graphs for this processing run indicates a resource use of 2,000 ec2 instances.