Global Yearly NDVI Percentile Composite at 10-Meter Resolution

Sentinel-2 satellite data for 2020 is processed into a global color composite representing vegetation dynamics. The image combines the 10th, 50th, and 90th percentiles of the NDVI time series into blue, green, and red channels, respectively. A consortium led by VITO produced this dataset, with contributions from Brockmann Consult, CS SI, Gamma Remote Sensing, IIASA, and Wageningen University.

Use Cases

• Classify land cover types by analyzing color patterns from the NDVI p10, p50, and p90 composite bands.
• Monitor seasonal vegetation phenology by comparing the 10th and 90th percentile NDVI values within a 1x1 degree tile.
• Identify evergreen vegetation areas where the NDVI p10, p50, and p90 values are high and similar, appearing white in the composite.
• Detect surfaces with partial annual vegetation, indicated by color gradients from yellow to red in the three-band image.
• Map unvegetated surfaces, which appear dark due to low NDVI values across all percentile bands.

Strengths

• Global spatial coverage at a 10-meter resolution per pixel.
• Data is cloud-optimized, with artifacts from clouds and shadows removed for the 2020 time series.
• Composite provides three distinct vegetation metrics (p10, p50, p90) in a single GeoTIFF file.

Limitations

• Temporal coverage is limited to the single reference year of 2020.
• No data tiles are provided for ocean areas and inner ice sheets, creating gaps in global coverage.
• The dataset's utility for time-series analysis is restricted without comparable composites for other years.

Provenance

Source

European Space Agency (ESA) WorldCover project, data hosted on NASA Earthdata.

Collection Method

Generated from Sentinel-2 satellite imagery time series for 2020, with cloud removal, followed by calculation of NDVI 10th, 50th, and 90th percentiles.

Time Range

2020

Freshness

null

Geography

Worldwide, divided into 1x1 degree tiles.