Biodiversity e-infrastructures not only increase access to and usability of data for science, but also support education and effective public policies. This point is particularly true regarding sustainable development, in which a paucity of relevant data about species’ numbers, distributions, and status hinders and biases key biodiversity metrics [1]. For megadiverse countries like Brazil, free and open access to research-grade data in usable formats is crucial to addressing such challenges and making informed decisions and policies [2]. However, for users to be able to rely on information systems, it is crucial for them to operate with uninterrupted, long-term funding. The greater the size and complexity of a system, the higher the costs required to sustain it, but also the greater its potential value. This point holds for all infrastructures, be it a railway or an information system. E-infrastructures require long-term maintenance and constant development, continuous and dynamic evaluation and planning, and efficient governance models to assure continuity of the network and its services; otherwise, they perish just as would their physical counterparts [3].

In addressing conservation and sustainable development, policies and actions must be able to scale from local to global. Too many important decisions directly linked to biodiversity conservation, such as choice of sites for protected areas and assessment of species’ conservation status, or that affect conservation directly, such as the construction of roads, hydroelectric dams, and mining activities, are based on partial information or even on criteria unrelated to the objectives of the planned decision [2,4,5]. The massive global infrastructure expansion that has been occurring in the last decades and is expected to take place until 2050 can damage the environment beyond repair [5]. Many of these infrastructures will be built in developing countries that hold most of the planet’s biodiversity, emphasizing the need for relevant data and decision-making tools to be at hand [5]. It is estimated that 25 million kilometers of new roads will be built up until 2050, and the initial construction of roads in forested areas often sets off a chain of serious problems, including habitat fragmentation, illegal mining, and land speculation [5].

Every country should commit to increase its knowledge base to strengthen local, national, and global e-infrastructures by making data and information openly available (a commitment of the Biodiversity Convention, Article 17, and many other international agreements). Effective thematic and geographically delimited e-infrastructures provide a framework for networking and collaboration among data providers and users. Within this environment, shared data and knowledge are more likely to be used, not only for scientific development, but also in policies and decision support systems [6]. These e-infrastructures often reflect and respond directly to local challenges, besides feeding data and tools into global systems. The effective integration of science communities can be appreciated, for instance, in the formation of the VertNet initiative, which links North American institutions with vertebrate holdings [7]. Frequently, such communities collaborate in overcoming important challenges, such as georeferencing of large datasets. The outcome is a storehouse of primary, research-grade data, associated tools and knowledge that are useful from local to global scales, and a well-knit collaborative network of scientists.

Such e-infrastructures are a reality in Latin America and include Comissión Nacional para el Conocimiento y uso de la Biodiversidad (CONABIO), a permanent interdepartmental commission, created in 1992 in Mexico; Instituto Nacional de Biodiversidad (INBio), a nongovernmental organization founded in 1989 in Costa Rica; and Humboldt Institute (Instituto de Investigación en Recursos Biológicos Alexandre Von Humboldt), a nonprofit civil organization of public nature, under the Ministry of Environment and Sustainable Development, created in 1993 in Colombia—each with different strategies and experiences but all openly sharing biodiversity data locally and with Global Biodiversity Information Facility (GBIF). All three institutions have more than 20 years of history and services, but whereas CONABIO as a governmental institute and Humboldt, linked to the Ministry of Environment, have full support, INBio is struggling to survive [8].

Centro de Referência em Informação Ambiental (CRIA) was established in 2000 as a nonprofit association with the aim of disseminating scientific information and in this way contributing to the conservation and sustainable use of Brazil's biological resources. To fulfill its mission, CRIA is responsible for the development of speciesLink, a major biodiversity e-infrastructure conceived in 2001 within the Biota-FAPESP Program-The Virtual Institute of Biodiversity, focused on the biodiversity of São Paulo state, and later expanded to the entire country thanks to funding from Brazil’s Ministério da Ciência, Tecnologia e Inovação and to the involvement of scientific societies and biological collections. Currently, speciesLink provides free and open access to ~7.4 million primary, research-grade, biodiversity data records from 382 datasets from 135 national institutions, covering all Brazilian states (Fig 1B), and from 13 institutions from abroad.

Besides the open dissemination of data and tools, as most biodiversity e-infrastructures rely on a great number of data providers, an important output of these systems is the establishment of social networks, where each node has an important role to play. In the case of speciesLink, each data provider is responsible for digitizing, updating, and improving its data quality and has full control over its data. RNP is responsible for Brazil’s National Research and Education Network, which links more than 400 institutions through 41 metropolitan networks and hosts all speciesLink’s equipment and servers, guaranteeing data security, access, and uninterrupted operation. CRIA is responsible for maintaining and continuously developing the information system, attending the demands of both data providers and users.

All components of the network—data providers, users, RNP, and CRIA—are equally important. The strength of this collaborative network lies in integrating biological collections associated with research institutions and graduate programs, researchers, decision makers, software developers, and many other actors, all responsible for the intensive development, sharing, and usage of data and tools (S1 Text). In 2014, speciesLink had an average of 1.4 million records retrieved per day (S1 Table). About 95% of the access to the network is from Brazil, and its distribution is strongly correlated with the metropolitan networks and location of data providers (Fig 1A and 1C).

One of Brazil’s Institutos Nacionais de Ciência e Tecnologia, the Brazilian Virtual Herbarium of Flora and Fungi, is one of speciesLink’s thematic nodes. Currently 54% of Brazil’s active herbaria participate in the network, thanks to the coordinated effort of the Virtual Herbarium and support from the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq). The marked increase of online data in Brazil’s Virtual Herbarium results from the Brazilian government’s investment in training human resources in taxonomy and curation since 2005 (Protax–CNPq) and in its support to the Virtual Herbarium between 2009 and 2014 (Fig 2A). This investment also accelerated the description of new species of plants and fungi between 2002 and 2013, when 2,372 new species of flowering plants were described [9] (updated to 2013 using The International Plant Names Index [IPNI] as the data source). More important is the apparent correlation between the increased rate of species description with the availability of online data and the fact that in the last decade 74% of new species were described by taxonomists based in Brazil (Fig 2B). These numbers are strong indications that Brazil is beginning to reap the benefits of investments in taxonomy and that access to information, tools, and knowledge and the networking environment translates into academic power, which in turn enables the use of data for informed decision making. For example, in 2013, speciesLink data was used to produce a new version of Brazil’s Red List of Plants by the Centro Nacional de Conservação da Flora under the coordination of the Ministério do Meio Ambiente [10].

The large number of online records associated with the continuous update of the List of Species of the Brazilian Flora (http://floradobrasil.jbrj.gov.br) has also enabled the development of new analytical tools such as Lacunas and BioGeo. Lacunas helps identify gaps of data and knowledge in the Virtual Herbarium for different taxonomic groups (S1 Fig, S2 Table, and S3 Table) [11], while BioGeo facilitates the generation of distribution models of Brazilian plant and fungus species through a predefined workflow (S2 Fig) [12]. Both systems are available through public interfaces and were specifically developed to facilitate the design of research and collection strategies and to support planning and monitoring of conservation policies.

Part of the problem that e-infrastructures have been facing is that agencies prioritize short-term innovative projects and rarely fund ongoing and long-term initiatives [13]. Options envisioned to ensure stable funding include creation of dedicated grants by funding agencies [13], the integration of e-infrastructures into larger collaborative information systems [3], and the creation of new independent public agencies to guide the work of producing, evaluating, and updating a formal legal framework for e-infrastructures’ governance [14]. In 2009, the European Research Infrastructure Consortium was created to provide a new legal support system for European research infrastructures, highlighting the importance of new governance frameworks that provide lasting arrangements for secured funding [15]. A recent study analyzing 16 e-infrastructures across different disciplines such as physics, biology, and computing sciences showed that only six had reached sustainable funding, while nearly half lacked long-term funding [15]. Even global projects such as Encyclopedia of Life and Barcode of Life [16] were imperiled in the past, because they simply ran short of funding or the time to meet their aims was miscalculated and the money was not enough to accomplish all goals.

From the outset, speciesLink has been sustained through project-based funding, normally public grants for 2–4-year periods. Each new project promotes new developments and attracts new data providers and users but also increases the demand for support services in the long-term. In between projects, there are short-term crises in trying to keep the system operational. A strategy to consolidate a biodiversity e-infrastructure in Brazil was developed by three scientific societies in 2005 under the coordination of the Ministério da Ciência, Tecnologia e Inovação [17]. An essential element of the plan was the expansion and further development of speciesLink to include new partners, more collections, and developers, establishing a truly comprehensive national network. This plan has guided most developments of speciesLink and was the basis for a component of the Global Environment Facility–funded project to establish the Sistema de Informação sobre a Biodiversidade Brasileira, launched in November 2014. Yet, a strategy for long-term funding has not materialized. Quite simply, no public mechanisms exist to fund the maintenance and continuous development of such systems of public interest so, in spite of speciesLink’s success, scientists and developers are struggling to keep the initiative alive.

The time has come for countries to seriously support biodiversity information e-infrastructures: some must support initial steps in implementation, whereas other countries (e.g., Brazil, Mexico, Colombia, Costa Rica, and many others) have the luxury of having facilities already in place. If effective steps to secure the permanence of e-infrastructures are not taken soon, we will risk having biological data, which are currently organized and made available globally, once again inaccessible. In the case of Brazil, speciesLink is in immediate peril of disappearing. Brazil is one of the most diverse countries in the planet [18], holding ~19% of all existing plant species [19]; thus, speciesLink is not only of interest to Brazilian people and government anymore but has acquired importance in the global scenario as well. Not only will the hundreds of thousands of users of this system miss this crucial research and policy infrastructure, but the social scientific network linked to the e-infrastructure may lose strength.