A new geospatial platform seeks to record which natural areas remain on the planet and how well areas of particular importance to biodiversity and ecosystems are protected. Global Safety Net v3, known as GSN3, combines satellite images, artificial intelligence and mapping techniques. Karl Burkart, director of the initiative at the US research organization Nature Data Lab, tells Mongabay that the team wants to track both the extent of nature and the quality of its conservation.
The tool is linked to an international commitment that needs measurable monitoring. In 2022, at the conference of the UN Convention on Biological Diversity, 196 countries adopted the Kunming–Montreal Global Biodiversity Framework, with a central goal of protecting 30% of land and oceans by 2030. For Burkart, the crucial question is which areas will be included in that percentage. As he says, priority should go to areas essential to ecosystems, biodiversity, communities and their livelihoods.
The current platform emerged from successive efforts to turn scientific priorities into maps. According to Burkart, the study “Global Deal for Nature” helped shape proposals for the 30% target by linking the ecological importance of areas to their level of protection. The first version of the Global Safety Net combined data from 13 scientific publications, removing overlaps between map layers. At that time, 50.3% of the world's land surface was recorded as important for some ecological function, but the one-kilometer resolution limited the map's detail.
The second version highlighted a problem: areas shown as critical for rare or threatened species had already been converted to farmland or other human uses. The team tried to remove these areas using data at a resolution of 300 meters, but the relevant European Space Agency product stopped being provided that year. For GSN3, mapping started afresh, at a resolution of 90 meters and in collaboration with the scientific team at The Nature Conservancy. The new method combines the ecological picture with the history of human interventions, such as roads, mining, urban development and agricultural expansion.
Burkart explains that the view from space is not always enough to distinguish natural vegetation from areas that have undergone substantial changes. The history of interventions helps the team exclude areas it estimates have been converted to human uses. The result remains an estimate of areas likely to be important for biodiversity and ecosystem services, using a common methodology worldwide. According to him, the ability to reassess annually allows changes in natural areas and their protection to be compared across countries.
The platform offers three views: ecological importance, land categories and conservation status. In addition to protected areas and other effective conservation measures, it records documented conservation areas that are not included in the official UN database, such as certain community or private lands. Areas belonging to Indigenous peoples and traditional communities have a separate place. Burkart says the team counts them as equivalent to protected areas in its own system, citing research on the effectiveness of traditional management; this does not mean they acquire equivalent official status.
One finding that surprised the team was the extent of land within protected areas that did not meet the tool's criteria for high ecological importance. In previous versions, researchers had assumed that protected land met these criteria. In the third version, they assessed ecological importance and protection status separately, revealing this discrepancy. Burkart clarifies that the finding does not mean these areas have no value, but shows why formal protection needs to be examined alongside ecological data.
More encouraging was the presence of small pockets of biodiversity even in areas that, on a broad map, appear almost devoid of important natural habitats. In Europe, for example, zooming in reveals many such small refuges where development did not proceed for various reasons. The team wants to examine whether these pockets can help connect larger refuges. Burkart suggests using artificial intelligence to investigate possible connections, presenting this application as a future field of research.
According to Burkart, the data were available for the global conservation assessment that began at the UN in September and could be used there. The platform also supports action planning and the selection of priority areas for funding by philanthropic organizations. He describes the initiative as an independent observer that can check countries' progress, much like external assessments of climate commitments. Another application being explored is corporate accountability, through the identification of ecoregions with little protection and extensive degradation.
Comments