"#### All sourced input datasets might not contain the every 10 x 10 degree tile to cover the entire globe. For example, the Primary Humid Tropical Forest layer is provided only for the tropics. As such, for missing tiles, psuedo layers with 0 values are created to cover the entire surface of the earth. These pseudo layers ensure that the forest classification can run without returning errors for any 10 x 10 degree tile in the world.\n",
"#### All sourced input datasets might not contain the every 10 x 10 degree tile to cover the entire globe. For example, the Primary Humid Tropical Forest layer is provided only for the tropics. As such, for missing tiles, psuedo layers with 0 values are created to cover the entire land surface of the earth. These pseudo layers ensure that the forest classification can run without returning errors for any 10 x 10 degree tile in the world.\n",
"\n",
"#### Below is an example of such layers created for where the JRC Tropcal maps have no detected degradation"
The following are notes that provide the steps followed for the download and pre-processing of input datasets (Table 1). The end product are the datasets resampled and aligned to a 30 m x 30 m grid in 10 x 10 degree tiles. These steps are reproducible for batch processing on the AWS DPS cloud-computing system that supports the NASA MAAP. For ease of use, a single 10 x 10 degree tile (00N_000E) is exemplified so the steps are comprehensible and easily implementable on local machines. Cells are executable in either Python or R depending on the opertation to be performed, and it is indicated which to use where.
#### Available at https://glad.umd.edu/dataset/primary-forest-humid-tropics
#### Note, data is available for download for each continent covering the tropics. 10 x 10 degree tiles are hence produced separately for Asia, Africa and South America.
#### GEZ 2010 layer was accessed at: https://www.fao.org/forest-resources-assessment/remote-sensing/global-ecological-zones-gez-mapping/en/
#### Desert ecoregions are removed, the vectors are merged with cotinent boundaries and the output "EcoCont_clean.shp" prodocued in the study is accessible at https://drive.google.com/drive/folders/1l3mrSmR4fFMRcUuKLsOTBDbuFdzYIUfk?usp=drive_link
#### The ESA WorldCover 2021 dataset was accessed through the NASA MAAP STAC end-point. Steps for this are provided below, and executed in Python. Users outside the NASA MAAP are directed to donwload the original data at https://worldcover2021.esa.int/download, and then follow the pre-processing steps described in the next cell.
#### All sourced input datasets might not contain the every 10 x 10 degree tile to cover the entire globe. For example, the Primary Humid Tropical Forest layer is provided only for the tropics. As such, for missing tiles, psuedo layers with 0 values are created to cover the entire surface of the earth. These pseudo layers ensure that the forest classification can run without returning errors for any 10 x 10 degree tile in the world.
#### All sourced input datasets might not contain the every 10 x 10 degree tile to cover the entire globe. For example, the Primary Humid Tropical Forest layer is provided only for the tropics. As such, for missing tiles, psuedo layers with 0 values are created to cover the entire land surface of the earth. These pseudo layers ensure that the forest classification can run without returning errors for any 10 x 10 degree tile in the world.
#### Below is an example of such layers created for where the JRC Tropcal maps have no detected degradation