Darwin Core: An Evolving Community-Developed Biodiversity Data Standard — TEI-tagged view

John WieczorekpersName, David BloompersName, Robert GuralnickpersName, Stan BlumpersName, Markus DöringpersName, Renato GiovannipersName, Tim RobertsonpersName, David VieglaispersName (2012date). “Darwin Core: An Evolving Community-Developed Biodiversity Data Standard.” PLoS ONEtitle·j.

Object of study — Darwin Coreterm as a community-developed standard for publishing, sharing, and integrating biodiversity dataterm

Research question — how Darwin Coreterm defines a common semanticterm·var language, is implemented and governed, and evolves across biodiversity research communities

Contribution — an account of Darwin Coreterm's semanticterm·var structure, history, implementation patterns, governance, extensions, and knowledge-producing infrastructure

Full text — tagged

The reference list is not part of the text we hold for this paper. Our copy was extracted from the publisher's page and the bibliography did not survive that extraction, so the article below ends before its references. Nothing has been removed by us.

Darwin Coreterm: An Evolving Community-Developed Biodiversity Dataterm Standard Abstract Biodiversity dataterm derive from myriad sources stored in various formats on many distinct hardware and software platforms. An essential step towards understanding global patterns of biodiversity is to provide a standardized viewterm of these heterogeneous data sources to improve interoperabilityterm. Fundamental to this advance are definitions of common termsterm. This paper describes the evolution and development of Darwin Coreterm, a data standardterm for publishing and integrating biodiversity informationterm. We focus on the categoriesterm·var of terms that define the standard, differences between simple and relational Darwin Coreterm, how the standard has been implemented, and the community processes that are essential for maintenance and growth of the standard. We present case-study extensions of the Darwin Coreterm into new research communities, including metagenomics and genetic resources. We close by showing how Darwin Coreterm records are integrated to create new knowledge products documenting species distributions and changes due to environmental perturbations.

Introduction Concern over global loss of biodiversity [1] – [5] has resulted in widespread demand for quick, reliable access to high-quality data on the spatio-temporal occurrence of species and their relation to the environment. Initial studies have documented the response by species to diverse agents of environmental change [6] – [7] , including alarming vertebrate extinctions [8] . The signature of climate and other environmental changes and their effects on biodiversity is now well documented, and the evidence is overwhelming [9] – [14] .

The documentation of global patterns of biodiversity change is an important first step toward a wise, sustainable policy of conservation and management in the face of a changing environment [15] . Such documentation has emerged as a global priority [16] – [18] . A major impediment to creating this documentation is the lack of easily accessible data at the scope and at the scales needed. Most studies documenting biodiversity changes are limited to a few well-studied species or to small temporal and spatial scales [19] . To make effective use of existing data for broad-scale analyses, community coordination is required. Information must be in digital form, accessible, discoverable, and integrated. Each of these criteria presents distinct challenges, some of which are unique to biodiversity-related data.

A general challenge of biodiversity dataterm that is shared with many other information domains is the lack of a coordinated publishing and integration systemterm; data repositoriesterm tend to be isolated from each other in the absence of standards for data and communication protocols. Heterogeneity in meaning and content of termsterm creates obstacles in every aspect of data integration and useterm, including discovery, comparison, and quality assessment. Information communities, especially Library Information Sciences, have a long history of meeting these challenges through the creation and maintenance of standards.

Darwin Coreterm [20] is a standard for sharing data about biodiversityterm – the occurrence of life on earth and its associations with the environment. Darwin Coreterm first emerged around 1999 as a loosely defined set of termsterm, and progressed through several iterations by different groups resulting in many different variants [21] . A formal set of termsterm and processes to manage changes were necessary to ensure utility for data integration. These aspects were developed within the Darwin Coreterm Task Group [22] of the Taxonomic Databases Working Group (TDWG; www.tdwg.org ) and ratified as a standard in October 2009. The philosophy for Darwin Coreterm development has been to keep the standard as simple and open as possible and to develop terms only when there is shared demand. Darwin Coreterm has a relatively long history of community development and is deployed widely [23] – [24] . For example, the Global Biodiversity Informationterm Facility currently indexes approximately 300 million Darwin Core-formatted records published by more than 340 organizations in 43 countries. Increasingly, Darwin Coreterm is being incorporated in communities beyond that of natural history collections ( Figure 1 ), in which the standard has its roots.

Scope of Darwin Coreterm: The Standard, deriving from previous standards work (e.g., Dublin Coreterm), describes core sets (e.g., organismal, taxonomic) of characteristics of biodiversity, which are applicable in many biological domains (e.g., Paleontology, Botany). The standard can be extended to cover details of specific sub-disciplines (e.g., Genetic Resources, Herbaria, Taxonomic Checklists). Collaborations with other standards organizations (Genomics Standards Consortium (GSC) extend Darwin Coreterm for new disciplines (Genomics, Metagenomics, Gene Marker Sequences.

In this paper, we describe the Darwin Coreterm standard, its history, status, relationship to other standards, and prospects for further evolution. We also discuss how data sets are brought into Darwin Coreterm compliance and how Darwin Coreterm records are currently shared in a distributed publishing platform. We will demonstrate how Darwin Coreterm continues to extend beyond its original formulation for natural history collections and will present, via use cases, how the standard continues to be reshaped and extended through new collaborations. We close by discussing innovative computational and informatics approaches to improving the scale, scope, and usability of networks that rely on Darwin Coreterm compliant records.

Methods Defining Darwin Coreterm The primary purpose of Darwin Coreterm is to create a common languageterm for sharing biodiversity dataterm that is complementary to and reuses metadata standards from other domains wherever possible. Creating this common languageterm has been particularly challenging, since natural history data curation practices have been developed locally and organically over hundreds of years, have varied between disciplines as well as institutions, and have had limited culture of data sharing.

Fundamentally, Darwin Coreterm is a set of termsterm having clearly defined semanticsterm that can be understood by peopleterm or interpreted by machinesterm, making it possible to determine appropriate uses of the data encoded therein. The terms are organized into nine categoriesterm (often referred to as “classes”, Figure 2 ), six of which cover broad aspects (event, location, geological context, occurrence, taxon, and identification) of the biodiversity domain. The remaining categoriesterm·var cover relationships to other resources, measurements, and generic information about records. Especially for the record level, Darwin Coreterm recommends the use of a number of terms from Dublin Coreterm (type, modified, language, rights, rightsHolder, accessRights, bibliographicCitation, references). The full set of current Darwin Coreterm terms with their descriptions is available in the Quick Reference Guide ( http://rs.tdwg.org/dwc/terms/ ).

Darwin Coreterm Categoriesterm·var: Simple Darwin Coreterm is comprised of seven categories of termsterm (green). This subset of Darwin Coreterm terms represents descriptive data about organisms that can be represented in one file with one row per record and one column per term. Two additional categoriesterm·var (orange) expand Darwin Coreterm with concepts that require a more complex data structure, such as multiple measurements from a single specimen, and cannot be represented easily in Simple Darwin Coreterm.

The authoritative formterm of the Darwin Coreterm is a downloadable archive [25] , which contains various documents that are also available via web pages ( http://rs.tdwg.org/dwc/ ). Key among the documents is http://rs.tdwg.org/dwc/rdf/dwctermshistory.rdf . This document includes descriptions of all terms, their histories, the relationships among them, and the relations between them and terms from other standards, all in a single file. The document is written in the Resource Description Frameworkterm (RDF) format, which is a standard for encoding the semantic structure of web-based resourcesterm. The same information is rendered in human-readable form on web pages (Quick Reference Guide, Type Vocabulary, Complete History, Mapping to ABCDterm, and Mapping to Old Versions). The standard also includes recommendations for implementation ( Introduction , Simple Darwin Coreterm, XML Guide, and Text Guide), a document describing how the standard is to be maintained (Namespace Policy), and a document describing the rationale behind any changes that are made to the standard (Decision History). In addition to the authoritative standard, Darwin Coreterm consists of non-normative documents, commentaries, and tools tracked separately ( http://code.google.com/p/darwincore/ ).

Though the scope and governance of Darwin Coreterm are distinct from Dublin Coreterm [26] , readers will recognize the ancestry of Darwin Coreterm in their shared traits: mission, principles of operation, and inherited terms. Indeed, if words were nucleotides, the Darwin Coreterm mission is the same as that of Dublin Coreterm except for one base-pair addition (in bold), namely, “to provide simple standards to facilitate the finding, sharing and management of biodiversity informationterm.” This relationship is not accidental. The Dublin Coreterm Metadata Initiative (DCMI) has done a very good job of maintaining the metadata specifications at the core of web-based resources. The success of DCMI encouraged TDWG to adopt similar policies for the specific domain of biodiversity through the Darwin Coreterm. This process, in turn, acts as the basis for the development of data standardsterm for even more specific sub-disciplines (e.g., plant genetic resources) as well as a basis for collaboration with complementary related domains (e.g., metagenomics).

Implementation Guidelines TDWG, as with other information standards bodies, has a history of developing data sharing standards that are bound to specific technologies, such as eXtensible Markup Language (XML). While Darwin Coreterm is currently maintained in RDF, the enduring value of the standard lies in simple definitions of termsterm and their relationships to other terms, independent of technical implementationterm. As with Dublin Coreterm, the idea is to promote use of the common termsterm in every appropriate context, and to leave the implementation details to specific applications. Data can be shared using Darwin Coreterm in a variety of encoding schemesterm (Comma Separated Values, XML, JavaScript Object Notation, RDF, etc.). To aid in using Darwin Coreterm in different contexts, there are documents with recommendations on how to share information as Darwin Coreterm using various technologies.

The Simple Darwin Coreterm ( http://rs.tdwg.org/dwc/terms/simple/index.htm ) document explains the capacity to share records with properties that do not repeat, as simple text files or XML. There is also an XML Guide with reference XML schemas for highly structured data and a Text Guide explaining the construction of Darwin Core Archivesterm (a combination of CSV files and a simple XML document describing the semanticsterm·var of the data file columns and their relationships to each other). Darwin Core Archivesterm can support structured data that conform to a star schematerm (a single core set of records in one or more files, with one-to-one or one-to-many relationships to records in other files). The Darwin Core Archiveterm is becoming ever more popular as an easy means for data sharing, largely due to the release of supporting tools, in particular the next-generation data publishing system – the Integrated Publishing Toolkitterm ( http://www.gbif.org/informatics/infrastructure/publishing ) – an open source application developed by the Global Biodiversity Informationterm Facility discussed in further detail below. Darwin Core Archivesterm have been adopted by Pensoft® ( http://www.pensoft.net/journals/ ) as the preferred form for data appendices to taxonomic publications in MycoKeys, PhytoKeys, and ZooKeys.

Results Development and Ratification as a Standard Darwin Coreterm's development spans more than a decade, but rests on the earlier work of academic societies that addressed how to document or catalog specimens in natural history collections. Those societies put forward prescriptive guidelines about how to catalog specimens properly, first on paper and later in computer databases [27] – [29] . The purpose was to promote completeness in capture and consistency in representation or encoding. Actual data exchange and integration was anticipated as a future capability, but these guidelines preceded the Internet and World Wide Web.

In the early 1990's, entity-relationship modelingterm had become accepted as a prerequisite for designing any database serving multiple users. A normalized database structureterm was the primary technique for enforcing data integrity. The Association of Systematic Collections (ASC) model [30] was the result of the first group effort to create a conceptual model for biological collections that accommodated the distinct taxonomic disciplines commonly found in natural history museums. The ASC modelterm was an ontological descriptionterm of the domain, but was determined at the time to be beyond the technical capabilities to implement [31] . Eventually, this model was then used as the point of departure for several collection management applications, including Specify ( http://specifysoftware.org/ ), Biota ( http://viceroy.eeb.uconn.edu/biota ), and Arctos ( http://arctos.database.museum/ ). The emphasis at that time was to develop best practice for data capture and management, not data exchange or integration.

The first successful tool to enable data integration in the biodiversity domain came to be known as the Species Analystterm. A prototype based on the Z39.50 protocolterm [32] , which was popular in the Library Science community at that time, was deployed in a pilot project, the Z39.50 Biology Implementers Group (ZBIG, U.S. National Science Foundation NSF-DBI-9811443). An outcome of the first meeting of that project was a list of terms with definitions to be included in the profile. The name “Darwin Coreterm” was coined by Allen Allison and adopted at that meeting. The Species Analystterm, a project developed under the NSF-funded ZBIG project at the University of Kansas [33] , led to an international deployment of data servers [34] for ornithological (NSF-DBI-9808739) and ichthyological (NSF-DEB-9985737) [35] collections.

The initial purpose of the Darwin Coreterm was to facilitate the exchange of information about the geographic and temporal occurrence of organisms in the natural world and the physical existence of specimens in biological collections. In contrast to earlier guidelinesterm, Darwin Coreterm was not a prescription for how to manage collection information. Instead, it was designed to circumscribe a conceptual model for a research community aiming to create a loose federation of databasesterm and advance its research capabilities through data integration. The barriers to publishing data in Darwin Coreterm were purposefully kept as low as possible.

The informal Darwin Coreterm terms developed for Species Analystterm were further refined under the NSF-funded Mammal Networked Information System project (MaNIS, NSF-DBI-0108161, http://manisnet.org ) [36] and the Ornithological Information System project (ORNIS NSF-DBI-0345448, http://ornisnet.org ) [37] . Major goals of the MaNIS project were to employ HTTP as the transport protocol instead of the less widely used Z39.50, to create a message protocol to support web-based distributed queries (Distributed Generic Information Retrieval – DiGIR), and to enhance the set of termsterm to include information such as georeferences based on guidelines that address the capture of data qualityterm metrics [38] . The goal under ORNIS was to make sure that the Darwin Coreterm underwent the process of becoming a formal standard.

Darwin Coreterm was developed in parallel with efforts in Europe to establish a comprehensive model for biological collections information — Access to Biological Collections Data (ABCDterm) [39] . The philosophies of the two approaches were distinct, even if they both had the goal of promoting sharing of biodiversity informationterm. Darwin Coreterm was designed to be minimalterm (only terms shared in common throughout natural history collections) and flatterm (no relational structure), while ABCDterm was highly structured and sought to capture the wide variety of biodiversity dataterm and their relationships. Both groups defined their data models in XML schemas. ABCDterm was ratified as a standard by TDWG in 2005. Darwin Coreterm departed from defining the standard in XML schema and followed the DCMI model, incorporating the experience of the previous decade to establish a set of termsterm in wide use in practice for data sharing.

Prior to ratification, Darwin Coreterm had a history of versions [21] , each attempting to adapt to expanding community-defined needsterm·var. The earliest versions were limited to defining terms for biological specimens in collections. Continued development in sectors of the community produced versions with added capabilities targeting observations [40] and paleontology [41] . Standard Darwin Coreterm reconciles the omissions from, incompatibilities between, and inconsistencies within previous versions and provides mappings to the out-dated terms as well as mappings to concepts in the ABCDterm Schema. In addition, terms were added to enable the transmission of taxonomic name and classification data sets. The result is a relatively simple and stable specification, which serves the needs of data publishers, data consumers, and application developers to work with primary biodiversity dataterm.

Darwin Coreterm underwent a year of document development by the Darwin Coreterm Task Group with support from the National Science Foundation and the Global Biodiversity Informationterm Facility. The draft standard underwent review with public commentary before being ratified by the TDWG Executive Committee in October 2009.

Maintenance and Evolution In combination with openness and consensusterm building (key aspects of the TDWG constitution: http://www.tdwg.org/about-tdwg/constitution/ ), two of the guiding principles behind Darwin Coreterm are flexibilityterm and adaptabilityterm. The standard has to be flexible to accommodate a variety of ever evolving technical contexts in which it might be used (spreadsheet columns, relational database fields, XML schemas and documents, non-SQL data stores, and semantic webterm resources). The standard has to be adaptable to accommodate growth (new terms, additional meanings) and interoperabilityterm (connections to related information). To thrive, the standard must have a means to adapt to changing needs.

Darwin Coreterm development is community driventerm and follows the process published in the Darwin Coreterm Namespace Policy ( http://rs.tdwg.org/dwc/2009-09-23/terms/namespace/index.htm ). Open commentaries on issues related to Darwin Coreterm are expressed through a public forum (at tdwg-content@lists.tdwg.org ) to which any subscriber may contribute. Discussions that result in consensusterm for change (new terms, new definitions, new relationships, changes in documentation) are submitted as change requests using templates requesting the appropriate information on the issue tracker of the Darwin Coreterm Project site ( http://code.google.com/p/darwincore/wiki/SubmittingIssues ). Sometimes this process is reversed, with a submitted issue being copied to the discussion forum for elaboration and consensusterm before refining the issue as submitted. All changes undergo a 30-day public review, announced on the listserv. Changes are reviewed by TDWG's Technical Architecture Group (TAG: http://www.tdwg.org/activities/tag/ ), a set of volunteer members with broad interest and expertise in web-based interoperabilityterm. The TAG determines if the proposed changes are compatible with Darwin Coreterm and that they do not reinvent existing standards. Changes that pass this review process are made as version changes to the affected documents, the latest of which always constitute the current standard.

Data Qualityterm and Constraining Darwin Coreterm Data qualityterm and fitness for useterm are primary concerns in the biodiversity community [42] , where information comes from heterogeneous sources spanning the globe over hundreds of years. Both data discovery and data qualityterm assessment are hampered by this heterogeneity. Part of the solution depends on the use of common vocabulariesterm, part on dictionaries of synonymous termsterm, and a great deal on applications that can facilitate discovery and assessment in place of or to aid in otherwise labor-intensive “cleanups”. For example, having a dictionary that equates “voucher” and “sample” with an accepted term “PreservedSpecimen” would allow applications to make this substitution automatically and allow users to discover relevant data using any of the synonyms regardless of how they were stored in the original data.

Darwin Coreterm makes recommendations about constraints on data content, whether about data typesterm·var, valid valuesterm·var, or controlled vocabulariesterm·var. For example, decimalLatitude is recommended to be a floating-point number falling between 90 and −90, inclusive; countryCode is recommended to be a value from ISO 3166-1-alpha-2 (two-letter country codes). Though these recommendations are made in the standard, the philosophy is to relegate enforcement to applications where they make sense. For example, in a collaboration where the purpose is to automate data qualityterm improvement on shared data, it is essential to be able to share the warts-and-all data that need to be investigated and improved.

Discussion Frontiers Though the Darwin Coreterm is defined in an RDF document, integration of biodiversity dataterm in the semantic webterm is in its early stages. One of the major challenges for Darwin Coreterm in the semantic webterm context is the lack of a well-defined ontologyterm - a formal definition of relationships between terms in a defined domain. An ontologyterm would define the relationships between concepts such as biological entities, the events that document where and when they occurred, and the processes through which they are identified as being representative of a taxonomic concept. Without rigorous relationships between concepts and the properties that define them, connections between biodiversity dataterm and related semantically rich information, such as literature and genomes, are difficult to traverse. This creates obstacles to cross-disciplinary semanticterm·var inquiry, such as in the Linked Data distributed data community ( http://linkeddata.org/ ). Current research is trying to address this gap (see TaxonConcept Knowledge Base, http://www.taxonconcept.org/ and BiSciCol, http://biscicol.blogspot.com/ ), active discussions on various aspects of the challenge take place regularly on the Darwin Coreterm public forum ( http://lists.tdwg.org/mailman/listinfo/tdwg-content ), and an RDF task group has been established within TDWG “to adapt TDWG vocabularies for use as RDF classes and properties and to integrate those resources with other well-known vocabularies and ontologies outside TDWG for use in describing biodiversity resources.” [43] Tools and Infrastructure for Publishing Darwin Coreterm Data A variety of toolsterm has been developed to simplify the mobilization and sharing of biodiversity dataterm using the Darwin Coreterm. Many of these tools focus on the creation or validation of Darwin Core Archivesterm ( http://code.google.com/p/darwincore/wiki/ToolsAndApplications#Tools ). The Global Biodiversity Informationterm Facility (GBIF), an international organization promoting the free and open access to biodiversity dataterm online, has invested heavily in the open and collaborative development of tools ( http://tools.gbif.org/ ) to make the creation and publication of Darwin Core Archives relatively simpleterm.

The Integrated Publishing Toolkitterm (IPT; http://www.gbif.org/informatics/infrastructure/publishing/ ) is an open-source software development projectterm, led by GBIF, designed to address the challenges of sharing biodiversity dataterm. By producing and providing easy access to Darwin Core Archivesterm and resource metadata in Ecological Markup Language [44] , the IPT works in tandem with a Global Biodiversity Resources Discovery System (GBRDS), a data registry, to facilitate discovery and retrieval of shared data throughout the world. The IPT complements other tools that have been created to assist data users to explore, organize, and mobilize data sets. A range of spreadsheet and archive creators and validators has been developed by GBIF and others. These tools can be accessed via the GBIF Publishing Software web page ( http://www.gbif.org/informatics/standards-and-tools/publishing-data/publishing-software/ ).

Whereas tools such as the IPT facilitate data mobilization using Darwin Coreterm data, other tools build on these sources to integrate and index biodiversity dataterm. These aggregators are thematic networks with either geographic (e.g., Atlas of Living Australia, http://www.ala.org.au/ ) or taxonomic focus (e.g., VertNet, http://vertnet.org ). Developing these systems to work sustainably and at large scales is an area of active research. For example, VertNet ( http://vertnet.org/ ), an aggregation of vertebrate specimen and observation data into a cloud-based data store, makes use of the Simple Darwin Coreterm in comma separated value (.csv) files with header rows containing Darwin Coreterm term names. These files are published to the data store where the records are indexed for high-performance querying.

Extensions Because Darwin Coreterm aims to cover the common ground in biodiversity, it inevitably lacks terms that are of interest to more specialized groups. A Darwin Core Extensionterm consists of additional terms describing a complementary, related domain, or guidance on the use of Darwin Coreterm within a specific sub-domain of biodiversity. Over the last decade, Darwin Coreterm has evolved by means of extensions, some of which have been incorporated in the standard. For example, the terms of a paleontology extension related to an earlier, pre-standard version of Darwin Coreterm became the terms of the current GeologicalContext class. Similarly, terms originating from an early version of the Darwin Coreterm created for the Ocean Biogeographic Information System (OBIS; http://www.iobis.org/ ) have also been added to Darwin Coreterm. Going forward, specific projects or disciplines may find it convenient to extend the scope of Darwin Coreterm outside the existing namespace and governance process for proof-of-concept research. Once the understanding of the new terms has been tested and proven of utility in a broader context, incorporation into the standard can be accomplished through community consensusterm.

Integration of Darwin Coreterm with New Research Communities Darwin Coreterm continues to be integrated into new research communities. For example, the Nordic Genetic Resource Center (NordGen; http://www.nordgen.org ) and Bioversity International (formerly IPGRI) ( http://www.bioversityinternational.org ) have sought to take advantage of Darwin Coreterm on behalf of the European plant genetic resources and genebank community and, at the same time, share rich discipline-specific genetic resource information. A careful review was made of Darwin Coreterm and a set of additional terms that did not overlap or conflict with existing Darwin Coreterm terms was proposed to meet the community's needs. The new terms, modeled directly from FAO/IPGRI Multi-Crop Passport Descriptors [45] (2001) and the proposed EPGRIS3 trait data standardterm for characterization and evaluation data [46] , were expressed and published openly in an XML schema for germplasm data (DwC-germplasm, http://code.google.com/p/darwincore-germplasm/ ) [47] , resulting in the first successfully deployed extension to the Darwin Coreterm Standard.

The success of DwC-germplasm has encouraged other communities to supplement or work in parallel with Darwin Coreterm to meet specific needs. For example, the Apiary Project ( http://www.apiaryproject.org/ ) extended the Darwin Coreterm to capture basic specimen information along with detailed annotation data (e.g., annotatedBy, dateAnnotated) found on herbarium sheets. Other communities are focusing on providing guidelines on how to use Darwin Coreterm within a discipline without the need for extensions. The Apple Core Initiative ( http://code.google.com/p/applecore/wiki/Introduction ), organized by Canadensys, ( http://www.canadensys.net/ ), is aimed at providing guidance on best practices for the content of Darwin Coreterm terms for vascular plant specimens.

A set of taxonomic extensions collectively known as the “GNA Profileterm” (named for the Global Names Architecture) [48] – [49] have been designed by GBIF with input from Catalogue of Life ( http://www.catalogueoflife.org/ ), Encyclopedia of Life ( http://eol.org/ ), nomenclators, and other taxonomic initiatives. These extensions ( http://rs.gbif.org/extension/gbif/1.0/ ), together with the taxonomic terms of Darwin Coreterm, allow sharing of rich taxonomic data often found in species checklists. The extensions are designed to be used in a one-to-many relationship with core taxon records, allowing the sharing of structured data such as vernacular names, species range distributions, textual descriptions, type species and specimens, image data, and bibliographies.

An effort is underway within TDWG to create a standard for biodiversity multimedia resources and collections, called the Audubon Coreterm [50] . The proposal introduces vocabularies covering the management and content of biodiversity-related media and their taxonomic, geographic, and temporal scope. Audubon Coreterm adopts by inclusion a number of terms from Darwin Coreterm, which can reduce the burden on existing media publishers who may have already used Darwin Coreterm terms for describing the media content.

In collaboration with the Genomic Standards Consortium (GSC; http://gensc.org/ ) [51] , Darwin Coreterm is now being extended to cover DNA-level observations (genomes, metagenomes, and gene marker sequences). The GSC, a standards body similar in scope to Biodiversity Informationterm Standards (TDWG), is working to “standardize the description of genomic data and promote the exchange and integration of genomic data. [52] .” Advances in technology have made sequence-based diversity assessments increasingly routine in biodiversity research. The GSC has developed the MIxSterm (Minimum Information about any (x) Sequence) standard containing checklists for describing genomes, metagenomes, and marker gene studies. Conceptually this approach is similar to Darwin Coreterm in that it defines terms for the description of data, but differs in that it requires the reporting a certain minimum set of descriptors such as environmental and geographic origin and details about sample and sequencing processing in addition to any kind of genomic sequence.

Currently, Darwin Coreterm and GSC developers are collaborating (supported by NSF-RCN4GSC) to harmonize Darwin Coreterm and MIxSterm to aid the growing field of sequence-based diversity research. Darwin Coreterm and GSC developers have concluded that the standards are conceptually similar and that the distinct sets of terms are complementary. This allows the development of a bi-directional (see Figure 2 ) solution, where in the future the GSC will be able to incorporate terms from Darwin Coreterm and add GSC-specific reporting requirements, while Darwin Coreterm will be able to offer a genomics extension using the results of this activity.

Scope and Scale of Darwin Coreterm for Biodiversity Science The creation of a standard format for biodiversity dataterm, the development of aggregation toolsterm, and the increasing use of Darwin Core Archivesterm have helped to overcome the challenge of limited availability of dataterm to answer questions about biodiversity. As more data become digitized and discoverable, biocollections can become a window into broader analyses of ecological, climate, niche, environmental, and biological research questions and critical issues.

Though predated by examples of the effective use of Darwin Coreterm for distribution modeling, influence of natural area preservation, and estimates of climate change impact [53] , the utility of data shared using Darwin Coreterm is illustrated by two recent examples. First, the University of California, Berkeley, and the Wildlife Conservation Society have collaborated to create Réseau de la Biodiversité de Madagascar (REBIOMAterm, http://www.rebiomaterm.net/ ).The effort seeks to make biodiversity dataterm about Madagascar available via a single web-based resource. Primary data on biodiversity in Madagascar come from researchers and institutions all over the world. REBIOMAterm works with each collection to standardize the data using the Darwin Coreterm, to mobilize and aggregate the data, and to vet the data for quality, completeness, and fitness for useterm. Data passing automated quality checks and review by a board of taxonomic experts are analyzed in combination with environmental data to produce maps showing models of where species might occur. The success of this project has made it possible for researchers, policy makers, government officials, and conservationists to capture a wealth of data to address issues of Malagasy biodiversity. The simplicity and flexibilityterm of Darwin Coreterm has made it possible for REBIOMAterm to provide immediate access to high-quality data and tools for monitoring and assessing conservation efforts.

A second example of innovative use of Darwin Coreterm data is the Map of Lifeterm Project ( http://www.mappinglife.org/ ), and sister projects such as LifeMapper ( http://www.lifemapper.org/ ) [54] , which aggregate and mobilize sources of species distribution information and provide tools to visualize and analyses those data. In Map of Lifeterm, species occurrence data points formatted as Darwin Coreterm are queried and brought into the application with related data sources such as range maps, assemblage checklists, and habitat preference data. Integration of these data typesterm·var and modeling will lead to summary products describing species distributions and provide the basis for distribution change assessments in the face of human-induced global changes [55] . The Map of Lifeterm will be a knowledge base of hundreds of thousands of high-resolution species distribution maps covering a wide range of taxa and geographic locations.

Conclusions The demand for Darwin Coreterm data is increasing. The corpus of information spanning three centuries of biodiversity exploration is being digitized at an increasing rate. Major efforts such as the new US National Science Foundation program Advancing Digitization of Biological Collections ( http://nsf.gov/funding/pgm_summ.jsp?pims_id=503559&org=DBI&from=home ) and the Atlas of Living Australia ( http://www.ala.org.au ) are two recent examples. A flood of qualitatively new global biodiversity dataterm is also being generated through the application of high-throughput gene sequencing of environmental samples.

The President's Council of Advisors on Science and Technology (PCAST) report on sustaining environmental capital noted recently that the challenge is not only one of the volume of data, but also its heterogeneity, as traditional biotic surveying practices blend with new approaches, such as remote sensing [56] or metagenomics [57] . A key passage in that PCAST report captures the challenge perfectly: “temporal, spatial, and methodological heterogeneity leads to a depth and richness in biodiversity and ecosystems data not found in other fields. It also makes dataset interoperabilityterm a problem that is at the same time particularly challenging and highly important to solve [58] .” Darwin Coreterm fills an essential role in describing some of these data in a standard format based on community input and demonstrated utility.

Darwin Coreterm is a living standardterm. Although its roots were planted in the vertebrate natural history collections community, Darwin Coreterm continues to grow to serve the needs of biodiversity research. This growth, over the relatively short period of a decade, speaks both to the need for such a standard and to the efforts of its champions to increase its utility through a community development process in order to meet increasing demand. Darwin Coreterm greatly increases the value and re-useterm of freely available and accessible biodiversity dataterm so that they can be effectively mobilized, integrated, and incorporated into other “grand challenge” scientific endeavors.