The origins of GBIF lie in a diagnosis made by the international scientific community in the years following the 1992 Rio Earth Summit. The newly signed Convention on Biological Diversity created an urgent demand for reliable information about the distribution and status of the world’s species, but the knowledge needed to meet that demand was scattered across a small number of institutions, concentrated overwhelmingly in wealthy countries, with no shared mechanism to bring it together. The Organisation for Economic Cooperation and Development’s Megascience Forum concluded that an international mechanism was needed to make biodiversity data and information accessible worldwide. That conclusion led directly to the establishment of GBIF in 2001, an infrastructure designed from the outset around the principles that would later be formalised as FAIR: data that are findable, accessible, interoperable and reusable.

What began as an effort to digitise the contents of natural history museums has evolved into something far broader. Over 25 years, GBIF has continually expanded to accommodate new kinds of biodiversity information, from digitised natural history specimens and citizen science observations to DNA-based records derived from molecular work. Most recently, the infrastructure has grown to handle structured survey and monitoring data designed to support large-scale biodiversity tracking, allowing researchers to move beyond simple presence records toward understanding population trends over time. In 2026, GBIF adopted the Catalogue of Life as its taxonomic backbone, the result of a multi-year collaboration to build shared infrastructure for reconciling species names across datasets. This taxonomic backbone is technically critical, because the same species can carry many names and the same name can refer to different species; reconciling these labels is what makes billions of heterogeneous records searchable and comparable.

The architecture underpinning this achievement is deliberately distributed. GBIF’s work is carried out through a network of voting country participants and organisational participants, each represented by a national or thematic node that mobilises data, builds local capacity and connects regional biodiversity communities to the global infrastructure. Rather than collecting data into a single central repository, GBIF indexes data that publishers maintain at source. More than 2,700 institutions, including museums, universities, government agencies, citizen science platforms and a growing number of private-sector organisations, have published data through the network, with new publishers joining at a rate of more than two every three days, bringing the total number of organisations involved to nearly 3,500. This model allows a comparatively small coordination office to steward an exponentially growing body of information contributed by thousands of independent institutions.

The scientific return on this investment is measurable and striking. GBIF-mediated data now underpin an average of eight new peer-reviewed research papers per day, and have contributed to more than 15,000 publications to date. The applications span an extraordinary range of fields: climate change research uses occurrence records to model how species ranges shift as temperatures rise; food security studies draw on crop wild relatives data; public health researchers map disease vectors; and invasive species managers track the spread of organisms beyond their native ranges. An independent 2023 assessment estimated that GBIF generates roughly 12 euros in societal benefit for every euro invested, with researcher time savings alone valued at 35 million euros annually. Few scientific infrastructures can document such a direct and quantifiable return on public investment.

The influence of GBIF data extends well beyond academic journals into the machinery of international environmental policy. GBIF-mediated data support multiple indicators under the Kunming-Montreal Global Biodiversity Framework, the landmark agreement through which the world’s nations have committed to halting and reversing biodiversity loss. The same records inform assessments by the International Union for Conservation of Nature and the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services, two of the most influential bodies in global conservation science. Increasingly, the private sector is also turning to GBIF-mediated data to meet emerging nature-related disclosure requirements, a development that signals how thoroughly open biodiversity data has penetrated from the museum basement into corporate boardrooms and regulatory frameworks.

Dr Joe Miller, GBIF’s Executive Secretary, reflected on the anniversary and the anniversary publication that documents it. He described pride in serving as Executive Secretary at this milestone and in contributing to the publication, which details the history of GBIF both as an infrastructure and a global network. In his view, the article pays tribute to everyone involved in the network across a quarter century, including staff, participants, nodes, volunteers, data publishers and data users, representing the past, present and future of the organisation. He expressed the hope that the paper will help strengthen the network and its capacity to respond to biodiversity data needs both now and in the future.

Yet the anniversary is also an occasion for candour about persistent shortcomings. The global loss of biodiversity continues to be described as a crisis in major international assessments, even as the data infrastructure has improved dramatically. The data themselves reflect long-standing imbalances in the practice of science. Regions with high biodiversity, particularly in the Global South, remain underrepresented, while well-studied, charismatic groups such as birds are heavily overrepresented compared with less visible taxa. Many marine species remain inadequately documented, as do hard-to-identify cryptic habitats and organisms. These are not merely technical problems; they mirror inequalities in scientific funding, institutional capacity and research priorities that have shaped the field for generations.

Closing these gaps was, in fact, part of the original motivation for creating GBIF a quarter-century ago, and the organisation’s own assessment is that the network cannot resolve these asymmetries alone. Doing so will depend on broader shifts in global scientific funding, data culture and capacity, alongside GBIF’s continued efforts to expand its participant network and to diversify the types of data it can mobilise. There is also an ongoing technical challenge of data heterogeneity and quality. Because GBIF indexes data rather than directly vetting every record, it relies on its network of data publishers to maintain quality at source and on users to report issues. This distributed model is what keeps the network scalable, allowing billions of records to flow into a single index, but it is not without friction, and users must remain alert to errors, misidentifications and gaps in even the most comprehensive datasets.

Looking toward its next 25 years, GBIF is guided by its 2023 to 2027 Strategic Framework, which prioritises continued growth of the participant network, expanded support for survey and monitoring data, and deeper integration of DNA-derived biodiversity records, all aimed at closing the longstanding geographic and taxonomic gaps in digital biodiversity data worldwide. The stakes of this work could hardly be higher. As nations race toward the targets of the Kunming-Montreal Global Biodiversity Framework, and as environmental change accelerates faster than traditional monitoring can follow, the ability to see life on Earth in near real time has become indispensable. From a proposal born of a post-Rio data crisis to an index of 3.8 billion records consulted dozens of times each day by working scientists, GBIF’s first quarter-century demonstrates what sustained, open, international cooperation can build. Its next chapter will determine whether the world can finally fill in the blanks on the map of life before they disappear.

Subject of Research: The history, impact and future challenges of the Global Biodiversity Information Facility as an open biodiversity data infrastructure

Article Title: GBIF marks 25 years of FAIR biodiversity data, now spanning 70 countries and 3.8 billion records

Article References: GBIF marks 25 years of FAIR biodiversity data, now spanning 70 countries and 3.8 billion records. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: GBIF, biodiversity data, open science, FAIR principles, natural history collections, Kunming-Montreal Global Biodiversity Framework, data infrastructure, citizen science, taxonomic backbone, Catalogue of Life, biodiversity policy, data gaps

A new perspective article published in npj Ocean Sustainability argues that the next leap in ocean science will not come from another sensor or satellite, but from the infrastructure that connects the data we already have. The authors examine how marine spatial data infrastructures, the networks of standards, repositories, and governance frameworks that underpin ocean data sharing, can be aligned with the principles of open science and with the FAIR guidelines, which hold that research data should be findable, accessible, interoperable, and reusable. The central claim is deceptively simple: sustainable ocean management is only as good as the knowledge systems that feed it, and those systems are currently too disjointed to support the decisions the ocean urgently needs.

The timing of this argument is not accidental. The United Nations Decade of Ocean Science for Sustainable Development has entered its second half, and the ambitions set out in 2021 are colliding with the practical realities of data governance. The Kunming-Montreal Global Biodiversity Framework commits nations to protecting thirty percent of land and sea by 2030, the high seas treaty known as the Biodiversity Beyond National Jurisdiction agreement is moving toward implementation, and marine spatial planning initiatives are expanding across every continent. Each of these commitments depends on the ability to ask hard, quantitative questions: Where are the biodiversity hotspots? How do shipping lanes overlap with whale migration corridors? Which seafloor habitats will be most resilient as waters warm and acidify? Answering them requires integrating physical oceanography, biology, geology, human-use data, and socioeconomic information at resolutions and accuracies that no single institution can produce alone.

Marine spatial data infrastructures, often abbreviated as MSDIs, emerged over the past two decades as a response to exactly this integration problem. Modeled in part on terrestrial spatial data infrastructures developed for land management, MSDIs provide the technical scaffolding for storing, cataloging, and serving geospatial ocean data. They define metadata standards that describe where a measurement was taken, when, by what instrument, and under what quality-control procedures. They establish web services that allow a map of coral reef extent in one national database to be layered seamlessly onto shipping density data held by an international maritime body. In principle, an MSDI lets a marine planner in a coastal ministry assemble a complete picture of their waters without ever downloading a raw file. In practice, the authors contend, the promise remains unevenly realized.

The technical barriers are real but solvable, and the article is refreshingly specific about them. Ocean data arrives in a bewildering array of formats: netCDF files from oceanographic models, Darwin Core archives from biodiversity surveys, Shapefiles and GeoJSON from habitat mapping, and proprietary formats from acoustic and optical sensors. Vocabulary mismatches compound the format problem, as one agency’s benthic habitat classification may bear little resemblance to another’s even when describing the same seafloor. Semantic interoperability, the ability of machines to understand not just that two datasets exist but that they describe comparable phenomena, demands shared ontologies and controlled vocabularies such as those maintained by the community behind the NERC Vocabulary Server and the Marine Regions gazetteer. FAIR alignment pushes these requirements further, insisting that data carry persistent identifiers, that metadata be rich enough for independent reuse, and that licenses clearly state what others may do with the material.

What elevates the article beyond a technical audit is its insistence that infrastructure is inseparable from culture and equity. Open science, the authors emphasize, is not merely the removal of paywalls from publications. It is a systemic commitment to transparency across the entire research lifecycle, from preregistered sampling designs to openly licensed datasets to reproducible analysis code. Applied to the ocean, this commitment carries a justice dimension that is easy to overlook from well-funded laboratories in the Global North. Much of the world’s marine data is collected in the waters of developing nations and small island states, yet the analytical capacity to exploit it is concentrated elsewhere. A genuinely aligned knowledge system would ensure that coastal communities who steward coral reefs, mangroves, and fisheries can access, understand, and contribute to the data describing their own seascapes, rather than serving as passive subjects of extraction by international research programs.

The authors sketch what this alignment looks like in operational terms. National and regional MSDIs would adopt FAIR-compliant repositories with persistent identifiers for every dataset and sample. Metadata would follow internationally recognized standards such as ISO 19115 for geospatial information, supplemented by community schemas for biological occurrences. Application programming interfaces, built on open protocols like OGC standards and the SPARQL query language for linked data, would allow researchers and decision-makers to compose datasets on the fly rather than negotiating bilateral data-sharing agreements that can take years. Crucially, the vision includes provenance tracking, so that a sea surface temperature anomaly used in a fisheries closure decision can be traced back through every processing step to the original instrument deployment. In an era when contested science can stall policy for a decade, that traceability is not bureaucratic decoration; it is the foundation of trust.

The article also confronts the hard economics of open infrastructure. Repositories cost money to run, data curation is skilled labor, and long-term stewardship of ocean observations requires funding models that outlast individual grants. The authors point to the tension between mandates for immediate open data and the legitimate needs of researchers to publish from their own collections, arguing for community norms that balance openness with recognition. They highlight the role of coordinated bodies, including the Intergovernmental Oceanographic Commission’s International Oceanographic Data and Information Exchange program and the emerging Ocean Data and Information System, as the connective tissue that can harmonize national efforts without imposing one-size-fits-all solutions on institutions with wildly different capacities.

Perhaps the most compelling thread in the analysis concerns artificial intelligence and the coming wave of machine-driven discovery. Foundation models trained on ocean observations, digital twins that simulate entire ocean basins, and automated anomaly detection for illegal fishing all presuppose data that machines can read, validate, and combine at scale. FAIR infrastructure designed for human users a decade ago is not automatically fit for this purpose. Machine-actionable metadata, semantically annotated datasets, and interoperable services are the raw substrate on which the next generation of ocean intelligence will be built. The authors make clear that this is not a distant hypothetical; it is happening now, and knowledge infrastructures that fail to adapt risk becoming irrelevant bottlenecks in an otherwise accelerating field.

The stakes could hardly be higher. The ocean regulates the planet’s climate, feeds billions of people, and harbors the largest reservoir of life on Earth, yet it is warming, acidifying, deoxygenating, and losing biodiversity at rates that outpace many models. The article’s conclusion is ultimately an optimistic one: the pieces of a global, equitable, and technically rigorous ocean knowledge system already exist in scattered form. Standards bodies, data centers, and open science movements have each built parts of the puzzle. What remains, the authors argue, is deliberate alignment, the patient diplomatic and technical work of making these components interoperable, FAIR-compliant, and responsive to the communities who depend on ocean knowledge most. If that work succeeds, the invisible ocean becomes visible not just to satellites and supercomputers, but to every planner, scientist, and citizen with a stake in its future.

Subject of Research: Integration of marine spatial data infrastructures with FAIR principles and open science for sustainable ocean knowledge systems

Article Title: Aligning Marine Spatial Data Infrastructures and Open Science for FAIR and sustainable ocean knowledge systems

Article References: Aligning Marine Spatial Data Infrastructures and Open Science for FAIR and sustainable ocean knowledge systems. (n.d.). https://doi.org/10.1038/s44183-026-00246-6

Image Credits: AI Generated

DOI: 10.1038/s44183-026-00246-6

Keywords: marine spatial data infrastructure, FAIR principles, open science, ocean sustainability, ocean data governance, semantic interoperability, marine spatial planning, ocean observations, data infrastructure, biodiversity data, ocean decade, digital ocean