Global Tree Cover on Agricultural Land - 2000-2020 - Algorithm (GEE Java Script)

Read full text
Authors: Robert J. Zomer
Cited by: 0

Abstract

This Google Earth Engine (GEE) JavaScript script computes benchmark-year estimates of percent tree cover on global cultivated and cropland-mosaic agricultural land for five benchmark years (2000, 2005, 2010, 2015, 2020). Tree cover is derived from the MODIS Vegetation Continuous Fields product (MOD44B, Collection 6.1) at 250 m resolution, aggregated to ~1 km using a three-year mean composite centered on each benchmark year to reduce inter-annual variability. The agricultural extent is defined using the GLC2000 global land cover classification (Classes 16, 17, 18), applied as a temporally invariant mask. Outputs are exported as 8-bit unsigned integer GeoTIFF files in MODIS Sinusoidal projection (~1 km resolution) to Google Drive and/or GEE Assets. The script was developed to support the global agroforestry and tree cover analysis published in Zomer et al. (2026, Nature Sustainability) and is was used to develop the accompanying dataset deposited at: https://doi.org/10.6084/m9.figshare.31669636

Fields of Study & Subjects

Remote Sensing in Agriculture Land Use and Ecosystem Services Conservation Biodiversity and Resource Management

Publication details

Journal / source
Figshare
Publication date
12 March 2026
Language
EN