TL;DR
This paper critically assesses the global landscape of species occurrence data, revealing significant spatial and taxonomic biases that hinder effective biodiversity monitoring and conservation planning. It highlights the urgent need for standardized data infrastructure, improved data curation, and targeted efforts to fill persistent data gaps, particularly in under-sampled tropical regions and marine environments.
What Problem Does This Paper Address?
The paper addresses the critical challenge of understanding and mitigating global biodiversity loss, which is hampered by the uneven and often inadequate availability of species distribution data. While massive amounts of data have been digitized, significant gaps and biases exist in their spatial coverage, taxonomic representation, and temporal depth. This makes it difficult to accurately assess biodiversity patterns, track trends, and inform conservation strategies, especially for global targets like those set by the Convention on Biological Diversity. The authors aim to identify these gaps and biases across major biodiversity databases to guide future data collection and improve the usability of existing information.
Methods
The study employs a multi-faceted approach to analyze the adequacy and representativeness of species distribution data. This includes:
- Database Analysis: Examining the content and coverage of major global species occurrence databases such as GBIF (Global Biodiversity Information Facility), OBIS (Ocean Biodiversity Information System), BIEN (Botanical Information and Ecology Network), and the IUCN Red List.
- Spatial Analysis: Using GIS tools (ArcMap) to map data density, assess coverage across different geographic designations (e.g., countries, ecoregions, EEZs, Longhurst regions), and analyze spatial biases in relation to factors like distance to infrastructure and human modification.
- Taxonomic Analysis: Evaluating the representation of different taxonomic groups within datasets and identifying areas where certain taxa are under-represented.
- Temporal Analysis: Examining data growth over time to identify regions and taxa that have seen recent improvements in data coverage.
- Comparative Analysis: Comparing data from different sources (e.g., BIEN vs. WCVP for plants) to identify discrepancies and potential sampling biases.
Datasets
The paper utilizes data from several key global biodiversity repositories:
- GBIF (Global Biodiversity Information Facility): The largest repository of species occurrence data, providing a vast collection of records across taxa.
- OBIS (Ocean Biodiversity Information System): Focuses on marine biodiversity data.
- BIEN (Botanical Information and Ecology Network): A comprehensive database of plant occurrence records.
- IUCN Red List: Contains species distribution data used for threat assessments.
- World Checklist of Vascular Plants (WCVP): Used as a reference for plant species richness.
- Other monitoring datasets: Including BioTIME, Living Planet Index (LPI), PREDICTS, and RivFishTIME, which focus on species abundance changes and population trends.
Key Insights and Ideas
- The "Data Rich, Data Poor" Paradox: While overall data volume is increasing, the distribution is highly uneven. High-income countries and well-trafficked areas are over-represented, while biodiverse tropical regions and remote marine environments remain severely undersampled.
- Marine Data Challenges: Marine data is particularly sparse, with significant gaps in the high seas, deep ocean, and tropical coastlines. Much of the perceived improvement in marine data coverage comes from tracking data of a few species, which doesn't necessarily reflect broader biodiversity.
- Infrastructure vs. Usability: The proliferation of open-access databases is positive, but fragmentation, lack of standardization, and data curation issues hinder true interoperability and usability for conservation and policy.
- Human Impact as a Sampling Proxy: Areas with high human modification tend to have the most data, suggesting that sampling efforts are often biased towards accessible and disturbed environments, rather than biodiversity hotspots.
- Taxonomic Skew: Bird data, often contributed by citizen science initiatives, dominates many datasets, masking the under-representation of other crucial taxa like invertebrates and many plant groups.
Key Results
- GBIF Coverage: Global spatial coverage in GBIF has increased from 6.19% in 2015 to 15.52% in 2025 (at 5km resolution), but 78 countries still have under 50% of their area represented. Terrestrial areas are better covered than marine areas.
- Marine Data Gaps: Even with combined GBIF and OBIS data, only 19% of the world's surface and 13% of the ocean have at least one species occurrence record. High-sea regions have only 2-4% coverage.
- BIEN vs. WCVP Discrepancies: Figure 3 illustrates significant regional differences in species counts between BIEN and WCVP, indicating potential under- or over-sampling by one or both databases in various parts of the world.
- IUCN Red List Data Sources: For animals, 83% of IUCN Red List assessments rely on polygon data only, with only 10% using point locality data, highlighting a reliance on expert knowledge rather than direct observations for many species.
Limitations and Future Work
- Data Quality and Standardization: The paper repeatedly emphasizes the lack of universal quality control standards for species records and the challenges of data curation, leading to potential misidentifications and inaccuracies.
- Accessibility and Fragmentation: Data is scattered across numerous repositories with varying accessibility, making it difficult to assemble a comprehensive picture.
- Need for Targeted Efforts: The authors call for concerted efforts to improve data coverage in under-sampled regions (tropics, high seas, Central Asia, North Africa) and for under-represented taxa.
- Policy and Funding: Sustained government investment and support are crucial for long-term monitoring initiatives and for developing standardized data infrastructure.
- Technological Advancement: The paper suggests that emerging technologies like AI and machine learning could enhance biodiversity surveillance and data processing.
- Journal Mandates: Journals are encouraged to mandate the public sharing of data in standardized repositories to prevent further fragmentation and improve reproducibility.
Summary updated