![projects/nature-trace/assets/species_distribution_models/australia_mammals/v0](https://developers.google.com/earth-engine/datasets/images/nature-trace/projects_nature-trace_assets_species_distribution_models_australia_mammals_v0_sample.png) info


This dataset is part of a Publisher Catalog, and not managed by Google Earth Engine.

Contact nature-trace@google.com

for bugs or [view more datasets](https://developers.google.com/earth-engine/datasets/publisher/nature-trace.md.txt)
from the Nature Trace Catalog. [Learn more about Publisher datasets](https://developers.google.com/earth-engine/datasets/publisher.md.txt).
[![Catalog Owner Logo](https://developers.google.com/static/earth-engine/datasets/logos/nature-trace_logo.png)](https://deepmind.google)

Catalog Owner
:   Nature Trace

Dataset Availability
:   2017-01-01T00:00:00Z--2025-01-01T00:00:00Z

Dataset Producer
:


    [Google](https://deepmind.google/)

Tags
:
    [alphaearth-derived](https://developers.google.com/earth-engine/datasets/tags/alphaearth-derived.md.txt) [biodiversity](https://developers.google.com/earth-engine/datasets/tags/biodiversity.md.txt) [conservation](https://developers.google.com/earth-engine/datasets/tags/conservation.md.txt) [ecosystems](https://developers.google.com/earth-engine/datasets/tags/ecosystems.md.txt) [nature-trace](https://developers.google.com/earth-engine/datasets/tags/nature-trace.md.txt) [publisher-dataset](https://developers.google.com/earth-engine/datasets/tags/publisher-dataset.md.txt) [species](https://developers.google.com/earth-engine/datasets/tags/species.md.txt)

#### Description

These species distribution model outputs, developed by Google in
collaboration with [QCIF](https://www.qcif.edu.au/) and
[EcoCommons](https://www.ecocommons.org.au/), represent estimates of
species' relative occurrence likelihoods (i.e., higher values indicate a
higher likelihood that the species would be detected at a location, for a
given survey methodology and given survey effort, compared to lower values).

They were generated using a multi-species
heterogenous graph neural network model
trained jointly on species observation data from the
[Global Biodiversity Information Facility](https://www.gbif.org/) for 174
Australian mammals from 2017 - 2024 (inclusive), combined with species
[trait](https://figshare.com/articles/dataset/COMBINE_a_Coalesced_Mammal_Database_of_Intrinsic_and_extrinsic_traits/13028255/4) and taxonomy data for the same species from COMBINE,
and satellite and environmental as embedded by
[AlphaEarth Foundations](https://deepmind.google/discover/blog/alphaearth-foundations-helps-map-our-planet-in-unprecedented-detail/).
[A previous phase of this work and the model architecture are described here](https://ar5iv.labs.arxiv.org/html/2503.11900).
Estimates of species' geographical distributions are made at c. 1km resolution (S2 cells level 13).

These distributions are intended to support conservation initiatives like
the [Global Biodiversity Framework](https://www.cbd.int/gbf/targets), which
includes [30x30](https://en.wikipedia.org/wiki/30_by_30). It is likely that
distributions could be significantly improved via fine tuning with
additional species observation data, where this is available.

Limitations:

- Experimental Status: These results are intended as demonstrations of early research and have not yet been peer-reviewed.
- Aggregate vs. Specific Validation: Model outputs for an earlier run described in the paper have been evaluated collectively against expert-produced Area of Habitation maps (mean [ROC AUC](https://bookdown.org/pjhanly/fw840hanly/week-4-species-distribution-models.html#area-under-the-curve-auc) = 0.89 over all species, AUC ROC \> 0.9 for each of the released species), but have not been vetted by species experts on a species-by-species basis.
- Presence-Only Data Constraints: These models are trained primarily on presence-only data. Outputs represent a relative likelihood of detection given current survey efforts, rather than true probability of occupancy.
- Input Data Bias: The accuracy of these maps is intrinsically linked to the input datasets. Biodiversity observation data is known to contain spatial, temporal, and taxonomic biases (often reflecting accessibility near roads or population centers rather than true biological distribution). The availability of species observation data has also been found to [reflect
  social and political inequities](https://www.science.org/doi/abs/10.1126/science.adh8874). These biases may be propagated into the final model outputs.

This experimental preview was released November 5, 2025.

### Bands

**Bands**

Pixel size: 1000 meters (all bands)

| Name | Min | Max | Pixel Size | Description |
|---|---|---|---|---|
| `AgileAntechinus` | 0 | 254.48 | 1000 meters | Likelihood of Agile Antechinus occurrence (scaled to \[0-250\]) |
| `AustralianSwampRat` | 0 | 253.85 | 1000 meters | Likelihood of Australian Swamp Rat occurrence (scaled to \[0-250\]) |
| `BlackFlyingFox` | 0 | 254.94 | 1000 meters | Likelihood of Black flying fox occurrence (scaled to \[0-250\]) |
| `BrownAntechinus` | 0 | 254.79 | 1000 meters | Likelihood of Brown Antechinus occurrence (scaled to \[0-250\]) |
| `Brush-TailedPhascogale` | 0 | 242.99 | 1000 meters | Likelihood of Brush-tailed Phascogale occurrence (scaled to \[0-250\]) |
| `Brush-TailedRock-Wallaby` | 0 | 252.78 | 1000 meters | Likelihood of Brush-tailed Rock-wallaby occurrence (scaled to \[0-250\]) |
| `BushRat` | 0 | 254.55 | 1000 meters | Likelihood of Bush Rat occurrence (scaled to \[0-250\]) |
| `CommonRingtailPossum` | 0 | 254.98 | 1000 meters | Likelihood of Common Ringtail Possum occurrence (scaled to \[0-250\]) |
| `CommonWombat` | 0 | 253.88 | 1000 meters | Likelihood of Common Wombat occurrence (scaled to \[0-250\]) |
| `EasternGreyKangaroo` | 0 | 253.93 | 1000 meters | Likelihood of Eastern Grey Kangaroo occurrence (scaled to \[0-250\]) |
| `GreaterGlider` | 0 | 254.84 | 1000 meters | Likelihood of Greater Glider occurrence (scaled to \[0-250\]) |
| `HastingsRiverMouse` | 0 | 243.4 | 1000 meters | Likelihood of Hastings River Mouse occurrence (scaled to \[0-250\]) |
| `LeadbeatersPossum` | 0 | 254.99 | 1000 meters | Likelihood of Leadbeater's Possum occurrence (scaled to \[0-250\]) |
| `Long-NosedBandicoot` | 0 | 254.13 | 1000 meters | Likelihood of Long-nosed Bandicoot occurrence (scaled to \[0-250\]) |
| `Long-NosedPotoroo` | 0 | 252.91 | 1000 meters | Likelihood of Long-nosed Potoroo occurrence (scaled to \[0-250\]) |
| `LumholtzsTree-Kangaroo` | 0 | 253.14 | 1000 meters | Likelihood of Lumholtz's Tree-kangaroo occurrence (scaled to \[0-250\]) |
| `MountainBrushtailPossum` | 0 | 254.95 | 1000 meters | Likelihood of Mountain Brushtail Possum occurrence (scaled to \[0-250\]) |
| `NorthernBrownBandicoot` | 0 | 247.76 | 1000 meters | Likelihood of Northern Brown Bandicoot occurrence (scaled to \[0-250\]) |
| `Platypus` | 0 | 254.59 | 1000 meters | Likelihood of Platypus occurrence (scaled to \[0-250\]) |
| `Short-EaredPossum` | 0 | 254.36 | 1000 meters | Likelihood of Short-eared Possum occurrence (scaled to \[0-250\]) |
| `SouthernBrownBandicoot` | 0 | 254.41 | 1000 meters | Likelihood of Southern Brown Bandicoot occurrence (scaled to \[0-250\]) |
| `SquirrelGlider` | 0 | 234.98 | 1000 meters | Likelihood of Squirrel Glider occurrence (scaled to \[0-250\]) |
| `SwampWallaby` | 0 | 253.3 | 1000 meters | Likelihood of Swamp Wallaby occurrence (scaled to \[0-250\]) |
| `WesternGreyKangaroo` | 0 | 254.99 | 1000 meters | Likelihood of Western Grey Kangaroo occurrence (scaled to \[0-250\]) |
| `Yellow-FootedAntechinus` | 0 | 254.58 | 1000 meters | Likelihood of Yellow-footed Antechinus occurrence (scaled to \[0-250\]) |

### Terms of Use

**Terms of Use**

All materials are licensed under the [Creative Commons Attribution 4.0 International
License (CC-BY 4.0)](https://creativecommons.org/licenses/by/4.0/).

Attribution: "Dataset developed by Google in collaboration with QCIF and
EcoCommons."

Unless required by applicable law or agreed to in writing, all materials distributed here
under the CC-BY license are distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR
CONDITIONS OF ANY KIND, either express or implied.
See the [license](https://creativecommons.org/licenses/by/4.0/legalcode)
for the specific language governing permissions and limitations under those licenses.

This is not an official Google product.

### Citations

Citations:

- Google LLC, "Australian mammal species distributions" (2025). Available online at https://developers.google.com/earth-engine/datasets/catalog/projects_nature-trace_assets_species_distribution_models_australia_mammals_v0.

### Explore with Earth Engine

> [!NOTE]
> **Important:** Earth Engine is a platform for petabyte-scale scientific analysis and visualization of geospatial datasets, both for public benefit and for business and government users. Earth Engine is free to use for research, education, and nonprofit use. To get started, please [register for Earth Engine access.](https://console.cloud.google.com/earth-engine)

### Code Editor (JavaScript)

```javascript
var dataset = ee.Image(
    'projects/nature-trace/assets/species_distribution_models/australia_mammals/v0');
var species = dataset.select('GreaterGlider');
var occurrenceVis = {
  min: 0,
  max: 250,
  palette: [
    '0000FF', '004FFF', '009DFF', '00ECFF', '00FFC4', '00FF75', '00FF27',
    '28FF00', '76FF00', 'C5FF00', 'FFEB00', 'FF9C00', 'FF4E00', 'FF0000'
  ]
};
Map.setCenter(133.4, -26.6, 5);
Map.addLayer(species, occurrenceVis, 'GreaterGlider');
```
[Open in Code Editor](https://code.earthengine.google.com/?scriptPath=Examples:Datasets/nature-trace/projects_nature-trace_assets_species_distribution_models_australia_mammals_v0) [Species Distribution, Australia Mammals](https://developers.google.com/earth-engine/datasets/catalog/projects_nature-trace_assets_species_distribution_models_australia_mammals_v0.md.txt) These species distribution model outputs, developed by Google in collaboration with QCIF and EcoCommons, represent estimates of species' relative occurrence likelihoods (i.e., higher values indicate a higher likelihood that the species would be detected at a location, for a given survey methodology and given survey effort, compared to lower values). ... projects/nature-trace/assets/species_distribution_models/australia_mammals/v0, alphaearth-derived,biodiversity,conservation,ecosystems,nature-trace,publisher-dataset,species 2017-01-01T00:00:00Z/2025-01-01T00:00:00Z -43.7 112.9 -10.4 153.7 Google Earth Engine https://developers.google.com/earth-engine/datasets

- [](https://doi.org/https://deepmind.google/)
- [](https://doi.org/https://developers.google.com/earth-engine/datasets/catalog/projects_nature-trace_assets_species_distribution_models_australia_mammals_v0)