Policy ForumBig Data and Biodiversity

Filling in biodiversity threat gaps

See allHide authors and affiliations

Science  22 Apr 2016:
Vol. 352, Issue 6284, pp. 416-418
DOI: 10.1126/science.aaf3565

The diversity of life on Earth—which provides vital services to humanity (1)—stems from the difference between rates of evolutionary diversification and extinction. Human activities have shifted the balance (2): Species extinction rates are an estimated 1000 times the “background” rate (3) and could increase to 10,000 times the background rate should species threatened with extinction succumb to pressures they face (4). Reversing these trends is a focus of the Convention on Biological Diversity's 2020 Strategic Plan for Biodiversity and its 20 Aichi Targets and is explicitly incorporated into the United Nations' 2030 Agenda for Sustainable Development and its 17 Sustainable Development Goals (SDGs). We identify major gaps in data available for assessing global biodiversity threats and suggest mechanisms for closing them.

Reducing rates of biodiversity loss and achieving environmental goals requires understanding what is threatening biodiversity, where risks occur, how fast threats are changing in type and intensity, and what are the most appropriate actions to avert them (5). A UN report proposed specific policy recommendations for mobilizing the “big data” revolution for sustainable development and environmental protection (7). The combination of crowd-sourced data, large-scale ground-based monitoring schemes, and satellite earth-observation missions is seemingly capable of unprecedented insight into global threats to biodiversity and how human interventions are altering those threats [e.g., (7)].

DELUGE OR DROUGHT? We used a threat classification scheme (8) (see the graph) that, although not without shortcomings (9, 10), has been widely deployed for tens of thousands of conservation assessments for species, sites, and projects. By “threat,” we mean “The proximate human activities or processes that have caused, are causing, or may cause the destruction, degradation, and/or impairment of biodiversity targets” (8). Determining the impact of a threat on a species or ecosystem is a separate process often included in a conservation assessment. We followed a structured data collection procedure and associated each data set with one or more classes of threat [see supplementary materials (SM) for details]. We omit three threat classes from our analysis: two (Geological Events; Other Options) are not exclusively anthropogenic; one (Climate Change and Severe Weather) received comprehensive treatment by the Fifth Assessment Report for the Intergovernmental Panel on Climate Change. We restricted our search to spatial data sets with a global extent. We assume that the data sets identified by this initial search will grow as additional data sets and metadata become known or are created. Over time, we recommend inclusion of the numerous available regional data sets (even if they do not meet data set attributes identified here) to create more globally representative information.

We identified 290 unique data sets (table S1) across nine threat classes from data sources ranging from remote sensing via satellites to citizen-science initiatives (fig. S1). Six data providers account for more than a fifth of the entire catalog of data sets. This apparent data deluge is misleading: Our analysis reveals how little is actually available, at the global level, about the spatial and temporal distribution of anthropogenic threats to biodiversity.

In order to assess whether data on different threats were available in proportion to their importance for biodiversity, we used threat information (for threatened taxa that have been comprehensively assessed) from the International Union for Conservation of Nature's Red List of Threatened Species (IUCN Red List), the repository of information on the global extinction risk of species. We find that the frequency of threats to marine or terrestrial and inland water species on the Red List is disproportionate to the availability of data sets on those threats (see the graph and table S2). Biological Resource Use (including direct and indirect impacts of hunting, fishing, and logging) is one of the most common threats to species, yet accounts for just 5% of threat data sets.

To assess how much threat information is available and actionable, we examined the data sets with respect to five desirable data attributes (see the table above and table S1). We note that determining accurate attribute values was often dif cult because of a lack of formal metadata, which creates uncertainty in the absolute number of data sets that might satisfy all criteria. Regardless, only 14 data sets (5%) satisfy all five attributes and not all threat classes are represented (see fig. S2, SM, and table S1 for details). Data sets that do comply are often applicable to only a few taxa or habitats.

Qualifying attributes of biodiversity data sets

Five data-set attributes considered key for use in biodiversity threat assessments.


BUSINESS MODELS. The conservation community should aspire to at least one “gold-standard” data set—that meets at a minimum all five attributes in the table and is applicable to as many taxa as possible—for each class and subclass of threat. This will require working with data providers to develop business models that leverage new, longer-term funding mechanisms and partnerships with government and the private sector.

Partnerships with data owners and creators. In certain instances, data required for effective conservation policy already exist but are not accessible [e.g., owing to access cost, commercial considerations, or intellectual property (IP) arrangements] to organizations or agencies mandated to conserve biodiversity. Sometimes these data result from taxpayer-funded initiatives that can result in major success stories (6). In 2008, NASA announced the free, public release of the Landsat image archive, dating back to 1978. This empowered the scientific community to begin studies of land cover change at an actionable resolution. Since then the European Space Agency opened the Sentinel Scientific Data hub, a free and open-access data portal for imagery from the Copernicus Sentinel missions, and the French Space Agency declared 5-year-old or older SPOT satellite data free of charge to noncommercial users.

Private-sector data also have potential to fill major gaps. Gaining access will require partnerships that respect the IP of companies and the right of conservation organizations to use data for conservation actions. One such agreement between the UN Environment Programme (UNEP) World Conservation Monitoring Center and the IHS Company enables detailed and comprehensive data on oil and gas activity worldwide to be used for biodiversity assessments. More broadly, the conservation community should emulate the UN's Data for Climate Action initiative, which is laying the groundwork for working with the private sector to access big data—with options ranging from companies making data freely available to arrangements for scientists to access data within the company's protected network.

Data sets and types of threats

The percentage of all threat data sets (dark blue) that relate to each threat class and the percentage of threatened terrestrial and inland water (medium blue) and marine (light blue) species on the IUCN Red List affected by each threat class. Number of data sets or species in each class is indicated beside each bar. Threat classes not covered by a single data set are denoted by an * in the figure labels. See table S2 for details on species included.


Funding mechanisms. In July 2015, the UN's Third International Conference for Financing for Development produced a comprehensive framework—the Addis Ababa Action Agenda (AAAA). The AAAA specifies >100 measures for how to finance the sustainable development agenda and explicitly recognizes the need to fund “science, technology, innovation and capacity building,” as well as “data, monitoring and follow-up” (11). The AAAA “encourage[s] the mobilization of financial resources from all sources and at all levels to conserve and sustainably use biodiversity and ecosystems.” This is an important recognition of the need to finance the achievement of SDG 15 (the most relevant to halting the loss of biodiversity), although critically missing is any specific mention of the need to fund the data required to achieve that goal.

THE DATA PIPELINE. For many threat classes the creation of a gold-standard data set need not start from scratch. Existing data sets and data pipelines, if provided with appropriate resources or mandates, can be scaled up. We highlight this potential with two issues where data scarcity on threats is a major obstacle.

Invasive and problematic species. Invasive alien species homogenize global biodiversity and are a significant threat to native species, particularly those endemic to islands and specific ecosystems. National and regional policy mechanisms are in place to prevent, control, and minimize the impact of alien species. Effective policy must be empowered with comprehensive data on which species are where and pathways by which they move (as the European Union's legal framework explicitly requires). These data allow implementation agencies to monitor transmission routes, prevent invasive species' entry or departure, and respond rapidly to early detections. The Threatened Island Biodiversity Database and the IUCN's Global Invasive Species Database are backed by international institutions and networks of experts and, if appropriately resourced, are capable of scaling up to meet the five key data attributes in the table.

Land use and cover change. Habitat loss is a leading cause of biodiversity decline, and most countries have local, regional, and national legislation protecting natural landscapes. Yet globally, we do not have a standard land use and cover change assessment tool for biodiversity conservation end users. New and standardized land cover change detection approaches for the 2000–2010 interval are emerging, at both high (30-m) (12) and moderate (300-m) resolution (13). Although these products have promise, it is impossible to obtain a global and standardized overview of how natural landscapes are changing on a time scale that allows appropriate conservation action. Changing this requires breaking the practice of repeatedly modifying remote-sensing algorithms—interesting for the field itself but exasperating for end users—and, instead, agreeing to a series of global maps comparable through time and space.

To be useful, threat data sets must be integrated with conservation assessment processes. The IUCN Red List compiles input from >10,000 species experts into easily and freely available conservation assessments for nearly 80,000 species that influence international and national policy mechanisms. Connecting such efforts to gold-standard data sets for each major class of threat will help bring actionable insights into what conservation actions are needed, and where, for the most imperiled species and populations. In so doing, we can better leverage the technology of the Information Age to counter biodiversity loss, a defining feature of the Anthropocene.

References and Notes

  1. Independent Expert Advisory Group, A World That Counts: Mobilising the Data Revolution for Sustainable Development (2015); http://bit.ly/Data4SustDev.
  2. Addis Ababa Action Agenda, Third International Conference on Financing for Development, 13 to 16 July 2015, Addis Ababa (2015); http://bit.ly/AAAAFundDev.
  3. Acknowledgments: See supplementary materials for complete listing of acknowledgments.
View Abstract

Stay Connected to Science

Navigate This Article