// GBIF Data
How EcoViewer uses the Global Biodiversity Information Facility to bring species occurrence data into ecological context — matched to Ecosystem Functional Groups from the IUCN Global Ecosystem Typology.
// GBIF Ecosystem Indicator Species
The Global Biodiversity Information Facility (GBIF) is the world's largest open repository of biodiversity data — hundreds of millions of species occurrence records contributed by institutions, researchers, and citizen scientists worldwide.
In EcoViewer, GBIF data serves a specific purpose: placing species occurrences in their ecosystem context. Rather than treating each record in isolation, EcoViewer links occurrence data to the Ecosystem Functional Group it belongs to — giving users a way to explore not just where a species was observed, but what kind of ecosystem that location likely represents.
Not all GBIF records can be meaningfully linked to a specific ecosystem. EcoViewer focuses on ecosystem indicator species — a carefully filtered subset whose occurrences can be reliably matched to an Ecosystem Functional Group or biome from the IUCN Global Ecosystem Typology.
The IUCN Global Ecosystem Typology (GET) classifies all of Earth's ecosystems into a six-level hierarchy. An Ecosystem Functional Group (EFG) is the third level — a group of related ecosystems within a biome sharing common ecological drivers and biotic traits. For example, intertidal mangrove forests (MFT1.2) are an EFG: distinct from tropical rainforests, yet sharing the coastal realm with saltmarshes and river deltas.
We define ecosystem indicator species as species whose physiological characteristics
allow them to thrive in specific Ecosystem Functional Groups and help shape those
ecosystems. Rhizophora spp. (mangroves), for example, have adaptations
that allow them to survive at the interface of marine, freshwater, and terrestrial realms;
because they actively build and maintain intertidal forests, their presence strongly signals the MFT1.2 EFG.
Important caveat: the presence of an ecosystem indicator species
does not mean that species is only found in that Ecosystem Functional Group,
nor that the ecosystem is definitively present at that location. Rather, it indicates
a higher likelihood of finding that ecosystem type there. Occurrence records should
be interpreted as signals, not definitive maps.
// Species–EFG matching
Each indicator species in EcoViewer is matched to an Ecosystem Functional Group — or to a broader biome where a specific EFG cannot be confidently determined — using peer-reviewed ecological literature. Matching draws on published species distribution studies, habitat association research, and biogeographic classifications relative to the IUCN Global Ecosystem Typology. A selection of key references is listed below.
| Title | Authors | Year | Link |
|---|---|---|---|
| IUCN Global Ecosystem Typology 2.0 : descriptive profiles for biomes and ecosystem functional groups | Keith, D.A. et al. | 2020 | DOI ↗ |
| Recent range expansion in Australian hummock grasses (Triodia) inferred using genotyping-by-sequencing | Anderson, B.M. et al. | 2019 | DOI ↗ |
| The diversity of post-fire regeneration strategies in the cerrado ground layer | Pilon, N.A.L. et al. | 2020 | DOI ↗ |
| Peatlands and the Boreal Forest | Wieder, R.K. et al. | 2006 | DOI ↗ |
| Species Composition, Distribution, and Diversity of Woody Species in a Tropical Dry Forest of India | Chaturvedi, R.K. & Raghubanshi, A.S. | 2014 | DOI ↗ |
| Water points and their influence on grazing resources in central northern Namibia | Klintenberg, P. & Verlinden, A. | 2008 | DOI ↗ |
This is a selection of key references. The full list of literature used in the GBIF ecosystem indicator species methodology is maintained in the GBIF Processing repository.
// Data pipeline
Raw GBIF plant occurrence records for the last ten years number in the tens of millions. EcoViewer processes them through a multi-step pipeline that filters, classifies, and validates each record against ecological, spatial, and environmental constraints — reducing the dataset to a high-confidence subset of ecosystem indicator species occurrences. The full pipeline code is open source and available in the GBIF processing repository.
Starting with ~172 million raw records from the GBIF dataset, occurrences are spatially thinned through deduplication and gridding. This process removes duplicate and overly dense records to mitigate spatial bias from heavily sampled regions, reducing the dataset to ~27 million records.
Each species is matched to its corresponding Ecosystem Functional Group or biome using a curated indicator species database, aligned against the IUCN Global Ecosystem Typology taxonomy.
Records falling within urban, industrial, plantation, or cropland environments are reclassified. These land cover types are explicitly reassigned to anthropogenic ecosystem categories (T7 biome) to avoid confounding human-modified landscapes with natural ecosystem signals.
Physical biome boundaries are enforced using latitudinal constraints. Tropical EFGs are restricted to the tropics, boreal to higher latitudes, and polar to polar regions. Records outside their expected range are reassigned or dropped.
SRTM elevation and a global aridity index sampled via Google Earth Engine at 30 m resolution further validate each record. Alpine species outside their elevation range, dryland species in humid environments, and similar mismatches are corrected or removed.
A final set of taxonomic fixes is applied — for example, reassigning specific target taxa to their appropriate ecosystem classifications, and removing genera where EFG matching is unreliable. This completes the pipeline, resulting in a dataset of 1,270,401 validated occurrence records.
The chart below shows the distribution of the 1,270,401 indicator species occurrence records across Ecosystem Functional Groups and biomes in the current dataset. Each bar represents a biome, subdivided by EFG. Biome-level matches — where a specific EFG could not be confidently determined — are shown as a separate segment within the biome bar.