Catalog Overview#

This page gives a complete overview of all collections currently registered in the EOForestSTAC catalog, organised by theme.

Themes#

The catalog is divided into four thematic groups. Use DiscoveryProvider.list_themes() to retrieve this list programmatically.

Biomass & Carbon (`biomass-carbon`)#

Collection ID	Title	Description
`CCI_BIOMASS`	ESA CCI Biomass – Global annual aboveground biomass	Global aboveground biomass (Mg/ha) at 100 m, multiple epochs.
`SAATCHI_BIOMASS`	Saatchi & Yu – Global aboveground biomass	Global aboveground biomass at 100 m (Saatchi & Yu 2020).
`LIU_BIOMASS`	Liu et al. – Europe aboveground biomass, canopy cover, and canopy height	European aboveground biomass, canopy cover, and canopy height at 30 m.
`ROBINSON_CR`	Robinson et al. – Chapman-Richards growth-curve parameters	Secondary-forest aboveground carbon dynamics at 1 km.

Disturbance & Change (`disturbance-change`)#

Collection ID	Title	Description
`EFDA`	European Forest Disturbance Atlas (EFDA)	Annual disturbance maps for European forests (fire, wind, bark beetle, …) at 30 m.
`JRC_TMF`	JRC Tropical Moist Forests (TMF)	Forest change (degradation, deforestation, regrowth) at 30 m.
`HANSEN_GFC`	Hansen Global Forest Change (GFC)	Tree cover and annual loss/gain at 30 m.
`JRC_GFC2020`	JRC Global Forest Cover 2020 (GFC2020)	Forest presence/absence at 10 m.
`RADD_EUROPE`	RADD Europe	Monthly forest disturbance occurrence from Sentinel-1 at 10 m (2020–2025).

Structure & Demography (`structure-demography`)#

Collection ID	Title	Description
`GAMI`	Global Age Mapping Integration (GAMI)	Global forest age ensemble at 100 m.
`GAMI_AGECLASS`	GAMI – Global Forest Age-Class Fractions	Forest age-class fractions (12 classes, 20 ensemble members) at 5 resolutions: 1°–0.0833°. Reference years 2010 and 2020.
`POTAPOV_HEIGHT`	Global Canopy Height (Potapov et al.)	Global forest canopy height at 30 m.
`FORESTPATHS_GENUS`	ForestPaths – European Tree Genus Map 2020	European tree genus classification at 10 m.

Land Use & Land Cover (`land-use-land-cover`)#

Collection ID	Title	Description
`RESTOR_LANDUSE`	Annual Land Use and Land Cover maps for the Brazilian Amazon	Annual LULC classification for the Brazilian Amazon (2000–2022) at 30 m.
`Potapov_LCLUC`	Potapov Land Cover and Land Use Change 2000–2020	Global cropland and land use change maps at 30 m.

Direct access without installing eoforeststac#

All datasets are publicly accessible via HTTPS (i.e, no Python package or credentials required). Browse the STAC Browser to find the Zarr asset URL for any collection/version, then stream it directly with xarray (Python) or Rarr / rstac (R).

Python

Install dependencies:

pip install xarray zarr

Open and subset a Zarr store over HTTPS:

import xarray as xr

s3_path = (
    "https://s3.gfz-potsdam.de/dog.atlaseo-glm.eo-gridded-data"
    "/collections/CCI_BIOMASS/CCI_BIOMASS_v6.0.zarr"
)

# Open lazily — no data is downloaded yet
ds = xr.open_zarr(s3_path)
print(ds)

# Spatial + temporal slice
agb = ds.sel(
    time="2010-01-01",
    latitude=slice(-3, -6),
    longitude=slice(-68, -62),
).aboveground_biomass

agb.compute()   # triggers the actual S3 read

# Optional: export to GeoTIFF
import rioxarray
agb.rio.set_spatial_dims(x_dim="longitude", y_dim="latitude")
agb.rio.write_crs("EPSG:4326", inplace=True)
agb.compute().rio.to_raster("agb_2010_subset.tif")

R

Install dependencies:

install.packages("rstac")
BiocManager::install("Rarr")   # Bioconductor
install.packages("terra")

Browse the catalog with rstac:

library(rstac)

s    <- stac("https://simonbesnard1.github.io/eoforeststac")
cols <- s |> collections() |> get_request()
print(cols)

# Extract the Zarr href for a specific item
items     <- s |> collections("CCI_BIOMASS") |> items() |> get_request()
zarr_href <- items$features[[1]]$assets$zarr$href

Read and export with Rarr over HTTPS (no credentials needed):

library(Rarr)
library(terra)

zarr_path <- zarr_href   # from rstac above, or paste directly

# Read coordinate arrays first (small/fast)
lat  <- read_zarr_array(file.path(zarr_path, "latitude"))
lon  <- read_zarr_array(file.path(zarr_path, "longitude"))
time <- read_zarr_array(file.path(zarr_path, "time"))

time_dates <- as.Date(time, origin = "2007-01-01")
time_idx   <- which(time_dates == as.Date("2010-01-01"))
lat_idx    <- which(lat >= -6  & lat <= -3)
lon_idx    <- which(lon >= -68 & lon <= -62)

# Read only the slice you need
agb_subset <- read_zarr_array(
  file.path(zarr_path, "aboveground_biomass"),
  index = list(lat_idx, lon_idx, time_idx)
)

# Convert to SpatRaster and export
agb_2d <- drop(agb_subset)
r      <- rast(agb_2d)
ext(r) <- c(min(lon[lon_idx]), max(lon[lon_idx]),
            min(lat[lat_idx]), max(lat[lat_idx]))
crs(r) <- "EPSG:4326"
writeRaster(r, "agb_2010_subset.tif", overwrite = TRUE)

Note

Rarr supports HTTPS endpoints directly — pass the full URL as zarr_path. No Python bridge or S3 credentials are needed for public datasets.

Adding new collections#

To register a new product in the catalog, add entries to:

eoforeststac/products/<product>.py — product metadata and S3 path config.
eoforeststac/catalog/<product>.py — STAC collection and item factory.
eoforeststac/catalog/root.py — register in DEFAULT_VERSIONS and _product_specs().

See Contributing to EOForestSTAC for a full contributor guide.

Catalog Overview#

Themes#

Biomass & Carbon (biomass-carbon)#

Disturbance & Change (disturbance-change)#

Structure & Demography (structure-demography)#

Land Use & Land Cover (land-use-land-cover)#