An inventory of biodiversity data sources for conservation monitoring

protected areas, fisheries and certain aspects of climate change (greenhouse gas emissions, sea temperature) are better represented in the data sources than other forms of biodiversity infor- mation. There are more data for terrestrial biomes than for marine or freshwater. Even for well-studied taxa like birds, time-series data are mostly from Europe and North America. These findings reflect recognised global trends, with major gaps in biodiversity data in high- biodiversity tropical countries [13, 29, 30]. There are generally more data for Europe and North America than other regions, and more for terrestrial birds, mammals and trees than for other taxa, but there is a lot of local variation [4, 29, 31–34]. There was an uneven spread of databases across CMP pressure categories, with a dispropor- tionately large number for climate change and biological resource use and none that were rele- vant or easily accessible for energy production & mining and human intrusions & disturbance. For responses, there were numerous data sources relating to livelihood, economic & moral incentives and conservation designation & planning, but none for awareness raising, law enforcement & prosecution, and education & training. There are limited data at global level that allow direct measurement of the benefits and livelihood gains accrued by people due to conservation. This may partly be an issue of lack of data collection [35], but may also reflect the fact local data from site-level projects are not easy to aggregate to regional or global scales.

Only data from satellite-based remote sensing is updated on a daily basis (and therefore provides almost real-time monitoring). WRI’s Global Forest Watch updates information every day on forest loss and active fires. NOAA’s Coral Reef Watch updates data daily on sea temper- ature, flagging coral reefs where bleaching is most likely. Some data sources provide access to existing maps (e.g. SEDAC). However, there appears to be a growing trend for organisations to produce portals where people can click on different data sets to make them appear as overlays on a world map (examples include Global Forest Watch, Map X, OBIS-SEAMAP). In most instances, however, mapping has limited use for monitoring, primarily due to the lack of scope to track trends over time in quantifiable ways. It was also often difficult to ascertain how easy it is to access the source data behind the maps. There are about 15 data sources, with regularly updated time-series data, that were consid- ered to have the most potential for helping conservation agencies with biodiversity monitoring (Table 1). These data sources are easily (if not always instantly or freely) accessible and provide data on key issues including coral bleaching, fisheries, forest cover, greenhouse gas emissions, Key Biodiversity Areas, protected areas, species distribution and abundance, wetlands extent and wildlife trade. However, for the full data landscape to become more manageable and acces- sible to people who need it most, and to provide all the types of information required, several

actions need to be taken by conservationists, scientists and other end users. What needs to happen next Define terms and relationships. An ontology of databases, data sets, data platforms, data portals and data sources relating to biodiversity would help clarify types of data available and their interlinkages. The system should clarify which elements of biodiversity monitoring (state, pressure, response) the data are, can or could be used for. The CMP/IUCN threat and response categories can help to some extent, although they may need adapting for the monitor- ing context by selecting those most relevant to different scales (e.g. highlighting that data on direct benefits to people may be best collected at local levels rather than global levels). Existing thinking around data classifications should be built on. For example, Bingham et al. [11] identified three types of elements in the bioinformatics landscape: elements with a single specific focus (such as taxonomic backbones); higher-level elements that rely on one or more other elements (e.g. AquaMaps, which harvests species occurrence data from the Global PLOS ONE Biodiversity data sources PLOS ONE | https://doi.org/10.1371/journal.pone.0242923 December 2, 2020 6 / 14