Next Steps Towards Better Living Atlas Deployments and Maintenance

The Atlas of Living Australia (ALA) has undertaken several user surveys since 2008. These ranged from user needs analyses, functional reviews, user polls and user satisfaction surveys. Following each of these surveys, successive improvements to the ALA’s web presence and functionality have been made. There are still, however, several pain points that our stakeholders are identifying, which we can improve by refining our user experience and user interface (UX/UI). ALA recently consolidated previous findings and undertook additional stakeholder engagement regarding the current look-and-feel of the ALA’s web presence. Peer review of similar sites including the Global Biodiversity Information Facility (GBIF), and their approaches to design were included in this review, which culminated in a report summarising the perceptions and pain points of our users and provided recommendations to address these. Users perceptions of the ALA website are generally positive: users trust our website and there is no other resource in Australia, which consolidates such a large amount of biodiversity data. Despite this, it was clear that users still find navigating the different ALA services frustrating, and more complex than is needed. Users sometimes find it hard to locate the information they need and to understand what other questions ALA can be used to answer. There is also a disparity between the look and feel of ALA applications and the main website, with some applications requiring updating to use current ALA style. A set of user personas for ALA was developed (Suppl. material 1) as part of this review process, which informed user journeys, requirements and user stories for the species pages, home page and basic search designs. Work commenced in August 2023, focussing on designs for ALA species pages, home page, and basic search. With this UX/UI project, ALA staff collaborated with a design company to build and enhance in-house skills, reducing reliance on outsourcing for design needs and incorporating user considerations more effectively into our product designs. These designs have now been finalised and delivered, and work is in progress to implement them on the ALA's website. Comprehensive user testing was part of this process, which included validating low-fidelity and high-fidelity designs with users, incorporating changes and addressing identified pain points. This was a new approach to design for the ALA. The design project was a great success, and will result in a more modern site, which is much simpler and clearer for users to navigate and use. Focussing on external user feedback and input throughout the design project, rather than relying solely on feedback from within ALA, was key to this success. Consecutive to the UX/UI project, a program to update and improve compilation and management of ALA’s taxonomy was undertaken to address many issues with the ALA's taxonomic backbone. Three main issues were identified: The merged, composite nature of the taxonomy used in the ALA created duplications, misplacements and simple errors, exacerbated by flaws in name-matching data The static nature of the taxonomy (taxonomic changes in the system at the time required the entire ALA to be reindexed, making editing very difficult and time intensive) A subset of intractable geographically based taxonomic changes that cannot be handled by the current system including identification of populations and geographic splits The merged, composite nature of the taxonomy used in the ALA created duplications, misplacements and simple errors, exacerbated by flaws in name-matching data The static nature of the taxonomy (taxonomic changes in the system at the time required the entire ALA to be reindexed, making editing very difficult and time intensive) A subset of intractable geographically based taxonomic changes that cannot be handled by the current system including identification of populations and geographic splits This exercise to update the taxonomic backbone has improved the species pages by removing duplicate pages, increasing the accuracy of linking taxa to correct conservation status, and other content. ALA's taxonomy is fundamental master data that underpins the majority of applications and services that the ALA provides. Improving this greatly enhances the user experience and provides greater transparency regarding how the taxonomy was compiled. For the species pages, we were able to identify changes to the current page design and content to improve the quality and completeness of the content on them e.g., removing duplicate/irrelevant literature links, implementing the ability to be able to select focal images, removing unnecessary/superfluous content. Sources of authoritative descriptive information and images were sourced and approval to re-publish content negotiated with data providers. This provides benefits not only to our users but also to the content providers, as it will increase traffic to and discoverability of their website content as well. Once we have implemented new designs, we will monitor our support tickets, test the designs with users, and analyse our website metrics to see if there has been a significant change in visitation to or use of our website.

GBIF Togo, hosted at the University of Lomé, has published more than 62,200 occurrence records from 37 datasets and checklists. As a node participant of Global Biodiversity Information Facility (GBIF) since 2011, it has participated actively in several projects including the Biodiversity Information for Development (BID) programme. GBIF facilitates collaboration between nodes at different levels through its Capacity Enhancement Support Programme (CESP). One of the actions included in the CESP guidelines is called ‘Mentoring activities’. Its main goal is the transfer of knowledge between partners, such as information, technologies, experience, and best practices. Sharing architecture and development is the key solution to solving some the technical challenges and impediments (e.g. hosting, staff turnover, etc.) that GBIF nodes occasionally face. The Atlas of Living Australia (ALA) team have developed a feature called ‘data hub’, which allows the creation of a standalone website with a dedicated occurrence search engine that supports data discovery (e.g. specific genus, geographic area) published by particular GBIF nodes. In 2017, a CESP project between the GBIF Benin and the GBIF France led to the creation of a new portal: Atlas of Living Beninises. This portal shared the same back-end database as the Atlas of Living France portal, while at the same time, each portal displayed and managed information relevant only to its region. In 2018, another CESP project between GBIF France and GBIF Togo shared the same goal as the previous one: implement a new Atlas of Living Australia portal for Togo. This goal will be fulfilled using a similar implementation as the previous project: a shared back-end and different front-end. Togo will be the second African GBIF node to implement this kind of infrastructure. This poster will highlight the architecture specific to the Atlas of Living Togo, and present the management procedure that distinguishes data coming from the three different countries.

Understanding the biology underpinning the natural regeneration of plant species in order to make plans for effective reforestation is a complex task. This can be aided by providing access to databases that contain long-term and wide-scale geographical information on species distribution, habitat, and reproduction. Although there exists widely-used biodiversity databases that contain structured information on species and their occurrences, such as the Global Biodiversity Information Facility (GBIF) and the Atlas of Living Australia (ALA), the bulk of knowledge about biodiversity still remains embedded in textual documents. Unstructured information can be made more accessible and useful for large-scale studies if there are tools and services that automatically extract meaningful information from text and store it in structured formats, e.g., open biodiversity databases, ready to be consumed for analysis (Thessen et al. 2022). We aim to enrich biodiversity occurrence databases with information on species reproductive condition and habitat, derived from text. In previous work, we developed unsupervised approaches to extract related habitats and their locations, and related reproductive condition and temporal expressions (Gabud and Batista-Navarro 2018). We built a new unsupervised hybrid approach for relation extraction (RE), which is a combination of classical rule-based pattern-matching methods and transformer-based language models that framed our RE task as a natural language inference (NLI) task. Using our hybrid approach for RE, we were able to extract related biodiversity entities from text even without a large training dataset. In this work, we implement an information extraction (IE) pipeline comprised of a named entity recognition (NER) tool and our hybrid relation extraction (RE) tool. The NER tool is a transformer-based language model that was pretrained on scientific text and then fine-tuned using COPIOUS (Conserving Philippine Biodiversity by Understanding big data; Nguyen et al. 2019), a gold standard corpus containing named entities relevant to species occurrence. We applied the NER tool to automatically annotate geographical location, temporal expression and habitat information contained within sentences. A dictionary-based approach is then used to identify mentions of reproductive conditions in text (e.g., phrases such as "fruited heavily" and "mass flowering"). We then use our hybrid RE tool to extract reproductive condition - temporal expression and habitat - geographical location entity pairs. We test our IE pipeline on the forestry compendium available in the CABI Digital Library (Centre for Agricultural and Biosciences International), and show that our work enables the enrichment of descriptive information on reproductive and habitat conditions of species. This work is a step towards enhancing a biodiversity database with the inclusion of habitat and reproductive condition information extracted from text.

Building on centuries of research based on herbarium specimens gathered through time and around the globe, a new era of discovery, synthesis, and prediction using digitized collections data has begun. This paper provides an overview of how aggregated, open access botanical and associated biological, environmental, and ecological data sets, from genes to the ecosystem, can be used to document the impacts of global change on communities, organisms, and society; predict future impacts; and help to drive the remediation of change. Advocacy for botanical collections and their expansion is needed, including ongoing digitization and online publishing. The addition of non‐traditional digitized data fields, user annotation capability, and born‐digital field data collection enables the rapid access of rich, digitally available data sets for research, education, informed decision‐making, and other scholarly and creative activities. Researchers are receiving enormous benefits from data aggregators including the Global Biodiversity Information Facility (GBIF), Integrated Digitized Biocollections (iDigBio), the Atlas of Living Australia (ALA), and the Biodiversity Heritage Library (BHL), but effective collaboration around data infrastructures is needed when working with large and disparate data sets. Tools for data discovery, visualization, analysis, and skills training are increasingly important for inspiring novel research that improves the intrinsic value of physical and digital botanical collections.

Currently Russia doesn't have a national biodiversity information system, and is still not a GBIF (Global Biodiversity Information Facility) member. Nevertheless, GBIF is the largest source of biodiversity data for Russia. As of August 2020, >5M species occurrences were available through the GBIF portal, of which 54% were published by Russian organisations. There are 107 institutions from Russia that have become GBIF publishers and 357 datasets have been published. The important trend of data mobilization in Russia is driven by the considerable contribution of citizen science. The most popular platform is iNaturalist. This year, the related GBIF dataset (Ueda 2020) became the largest one for Russia (793,049 species occurrences as of 2020-08-11). The first observation for Russia was posted in 2011, but iNaturalist started becoming popular in 2017. That year, 88 observers added >4500 observations that represented 1390 new species for Russia, 7- and 2-fold more respectively, than for the previous 6 years. Now we have nearly 12,000 observers, about 15,000 observed species and >1M research-grade observations. The ratio of observations for Tracheophyta, Chordata, and Arthropoda in Russia is different compared to the global scale. There are almost an equal amount of observations in the global iNaturalist GBIF dataset for these groups. At the same time in Russia, vascular plants make up 2/3rds of the observations. That is due to the "Flora of Russia" project, which attracted many professional botanists both as observers and experts. Thanks to their activity, Russia has a high proportion of research-grade observations in iNaturalist, 78% versus 60% globally. Another consequence of wide participation by professional researchers is the high rate of species accumulation. For some taxonomic groups conspicuous species were already revealed. There are about 850 bird species in Russia of which 398 species were observed in 2018, and only 83 new species in 2019. Currently, the number of new species recorded over time is decreasing despite the increase in observers and overall user activity. Russian iNaturalist observers have shared a lot of archive photos (taken during past years). In 2018, it was nearly 1/4 of the total number of observations and about 3/4 of new species for the year, with similar trends observed during 2019. Usually archive photos are posted from December until April, but the 2020 pandemic lockdown spurred a new wave of archive photo mobilisation in April and May. There are many iNaturalist projects for protected areas in Russia: 27 for strict nature reserves and national parks, and about 300 for others. About 100,000 observations (7.5% of all Russian observations) from the umbrella project "Protected areas of Russia" represent >34% of the species diversity observed in Russia. For some regions, e.g., Novosibirsk, Nizhniy Novgorod and Vladimir Oblasts, almost all protected areas are covered by iNaturalist projects, and are often their only source of available biodiversity data. There are also other popular citizen science platforms developed by Russian researchers. The first one is the Russian birdwatching network RU-BIRDS.RU. The related GBIF dataset (Ukolov et al. 2019) is the third largest dataset for Russia (>370,000 species occurrences). Another Russian citizen science system is wildlifemonitoring.ru, which includes thematic resources for different taxonomic groups of vertebrates. This is the crowd-sourced web-GIS maintained by the Siberian Environmental Center NGO in Novosibirsk. It is noteworthy that iNaturalist activities in Russia are developed more as a social network than as a way to attract volunteers to participate in scientific research. Of 746 citations in the iNaturalist dataset, only 18 articles include co-authors from Russia. iNaturalist data are used for the management of regional red lists (in the Republic of Bashkortostan, Novosibirsk Oblast and others), and as an additional information source for regional inventories. RU-BIRDS data were used in the European Russia Breeding Bird Atlas and the new edition of the European Breeding Bird Atlas. In Russia, citizen science activities significantly contribute to filling gaps in the global biodiversity map. However, Russian iNaturalist observations available through GBIF originate from the USA. It is not ideal, because the iNaturalist GBIF dataset is growing rapidly, and in the future it will represent more than all other datasets for Russia combined. In our opinion, iNaturalist data should be repatriated during the process of publishing through GBIF, as it is implemented for the eBird dataset (Levatich and Ligocki 2020).

Benin became member of the Global Biodiversity Information Facility (GBIF) in 2004 and acceded to the status of voting member in 2011. GBIF Benin through the constant efforts of its node is now very active in GBIF community with respect to capacity building, data mobilization and data uses. GBIF Benin published more than 400 000 occurrence data from about 125 datasets on GBIF portal . As for capacity building, GBIF Benin yearly organizes at least 2 (two) workshops to enhance the capacities of national and regional partners in data mobilization and data uses. At regional level, GBIF Benin is leading a consortium of many countries (Senegal, Côte-d’Ivoire, Niger, Democratic Republic of Congo, Guinea, and Madagascar etc.) to help overcome the challenges of data mobilization and data uses at regional level. From the academic year 2017-2018, GBIF Benin, through its node manager, successfully cooperated with the University of Kansas to create a master program in biodiversity informatics. Biodiversity informatics is a field of investigation relatively new in science and is concerned with massive occurrence data collection on biodiversity as well as on environment; data treatments, analysis, and representations so as to derive sound research products to inform decisions on biodiversity conservation and sustainable uses in the context of climate and global changes. In Benin, the master program in biodiversity informatics is a permanent two-year program structured in teaching units with the following contents: 1) Basics concepts of biodiversity; 2) Biodiversity data capture; 3) Biodiversity inventories; 4) Biodiversity data analysis; 5) Climate change and biodiversity; 6) Ecological niche modeling and strategies for biodiversity conservation; 7) Data-science-policy interface; 8) Public Health and Applications of biodiversity data etc. At completion of their studies, students graduated in the program will be capacitated so as to achieve the following innovative objectives: 1) Use Geographic Information System to map spatial distribution of species; 2) Model the current and the future ecological niche of species in the context of climate and global changes; 3) Characterize biodiversity on scales ranging from local to global; 4) Assess geographic patterns among suites of species (i.e., communities); 5) Refine the knowledge on particular taxonomic groups; 6) Define priority zones of biodiversity conservation; 7) Develop strategies of species conservation; 8) Implement biodiversity conservation strategies; 9) Predict the risks of propagation of infectious diseases (Lassa fever, Ebola fever etc.) which vectors are living organisms, so as to support preventive actions, etc. With such capacities, the graduated students of the master program are obviously the new generation of biodiversity information scientists who are able to address the needs of information so as to contribute to biodiversity conservation and its sustainable uses. Furthermore, in their respective countries and the rest of Africa, they will successfully contribute to the achievements of the Sustainable Development Goals as defined by the United Nations in 2015. With respect to data uses, more and more research products are piling up in Benin and are being integrated into decision makers’ arena. In 2018, the results of our data uses were integrated in the elaboration of the second communication on climate change of Benin.

Building on the development of Biodiversity Informatics, the Global Biodiversity Information Facility (GBIF) undertook the task of enabling access to the world’s wealth of biodiversity data via the Internet. To date, GBIF has become, in many respects, the most extensive biodiversity information exchange infrastructure in the world, opening up a full range of possibilities for science. Science has benefited from such access to biodiversity data in research areas ranging from the effects of environmental change on biodiversity to the spread of invasive species, among many others. As of this writing, more than 7,000 published items (scientific papers, reviews, conference proceedings) have been indexed in the GBIF Secretariat’s literature tracking programme. On the basis on this database, we will represent trends in GBIF in the users’ behaviour over time regarding openness, social structure, and other features associated to such scientific production: what is the measurable impact of research using GBIF data? How is the GBIF community of users growing? Is the science made with, and enabled by, open data, actually open? Mapping GBIF users’ choices will show how biodiversity research is evolving through time, synthesising past and current priorities of this community in an attempt to forecast whether summer—or winter—is coming.

Global patterns of fern species diversity: An evaluation of fern data in GBIF

ABSTRACTAs the world's climate changes, species are undergoing range shifts. Range shifts are generally documented using databases such as the Global Biodiversity Information Facility (GBIF), which largely contain data from monitoring schemes and wildlife surveys. Such databases have two major limitations: (i) data may be spatially biased because traditionally surveyed areas are in rural habitats and (ii) there is a time lag between formal monitoring and survey data collection and assimilation into GBIF, which means rapid range shifts cannot be tracked. Alternative data sources, such as social media, could provide information on species distributions and range shifts that compensate for spatial biases in GBIF records because social media data may be collected outside traditionally surveyed areas. Such data are also usually shared online immediately after a wildlife sighting. The complementarity of GBIF and social media data, however, has rarely been assessed, particularly when tracking range shifts. Despite their potential utility, social media data may be particularly prone to temporary trends or geographic variation in behaviour that are not understood. We lack tools with which to counter these biases. To address these knowledge gaps, we compare the habitat usage revealed by biological records of the Jersey tiger moth from GBIF and from social media data sources (Instagram and Flickr). We develop a novel method to investigate recorder bias in social media data and compare between data sources. We find that biological records from Instagram reveal greater than expected occurrence in urban environments. Recorder effort differs notably between data sources and Instagram complements GBIF by recording species in areas unaccounted for by GBIF. By incorporating recorder effort metrics, data from social media sources could be used to improve monitoring of range‐shifting species in urban spaces.

Global Registry of Scientific Collections, or GRSciColl, is a comprehensive registry of natural history collections maintained by the Global Biodiversity Information Facility (GBIF), building upon previous efforts developed by the Consortium of the Barcode of Life (CBOL) and other organizations. Its primary purpose is to provide a centralized platform for accessing information on institutions, their collections and associated staff members. While both GBIF and GRSciColl aim to encompass data from all around the globe, Asian and African collections are often underrepresented. To address this disparity, initiatives like the Biodiversity Information for Development (BID) and the Biodiversity Information Fund for Asia (BIFA) have been established. These initiatives specifically focus on assisting institutions from underrepresented regions in sharing their data through GBIF. In addition to providing support in the framework of these initiatives, the GBIF Regional support teams are actively engaged in efforts to improve access to relevant data about Asian and African collections. This involves updating information on Asian and African institutions and their collections in GRSciColl. Furthermore, the teams assist towards publishing biodiversity data held by these institutions for open access. To support this objective, a Collection Mobilization repository has been established on GitHub, with more than 1,900 issues created at an institutional level and a systematic workflow is developed and followed. The overall aim of this workflow adheres to the goal of GRSciColl in providing an updated information on scientific collections, their host institutions, and relevant staff, also making the specimen data from those institutions open access when possible. Here are a few key points of this ongoing work: The process begins by compiling relevant information on institutions and their collections by browsing the institution’s web pages to gather accurate and up-to-date data. Based on the findings, necessary changes or additions to the existing data are made in the institution and collection page of the GRSciColl registry, and simultaneously noted in related GitHub issues in the collection mobilization repository, ensuring transparency and record keeping. When required, contact is established with the staff of the institution, to seek clarifications and ensure accuracy of the information in the registry. Appropriate GitHub labels are then assigned to the institutions based on their data mobilization or digitization status, to help prioritize the ongoing work. Institutions that are familiar with GBIF data infrastructure or those that have digitized records of the specimens are given priority for contact to gauge their interest in mobilizing their data for open access. Institutions interested in publishing their data are offered support in the process, and their datasets are added to the Data Mobilization Repository in GitHub as an issue, using the Suggest a Dataset tool on GBIF.org. Finally, progress on data preparation and then publication is tracked and relevant comments and labels are added in the GitHub issues. The process begins by compiling relevant information on institutions and their collections by browsing the institution’s web pages to gather accurate and up-to-date data. Based on the findings, necessary changes or additions to the existing data are made in the institution and collection page of the GRSciColl registry, and simultaneously noted in related GitHub issues in the collection mobilization repository, ensuring transparency and record keeping. When required, contact is established with the staff of the institution, to seek clarifications and ensure accuracy of the information in the registry. Appropriate GitHub labels are then assigned to the institutions based on their data mobilization or digitization status, to help prioritize the ongoing work. Institutions that are familiar with GBIF data infrastructure or those that have digitized records of the specimens are given priority for contact to gauge their interest in mobilizing their data for open access. Institutions interested in publishing their data are offered support in the process, and their datasets are added to the Data Mobilization Repository in GitHub as an issue, using the Suggest a Dataset tool on GBIF.org. Finally, progress on data preparation and then publication is tracked and relevant comments and labels are added in the GitHub issues. The ongoing work ensures continuous improvement and expansion of GRSciColl, while promoting open access to valuable scientific data. This presentation will highlight the work done so far, as well as the challenges and lessons learnt from this ongoing project.

AimThe International Union for Conservation of Nature (IUCN) expert range maps and the Global Biodiversity Information Facility (GBIF) species occurrence data are commonly used to estimate species’ geographic range. Macroecological studies often cross‐reference geographic range data with a climate dataset, to extract the mean environmental conditions encountered by a species within its geographic range. We aimed to assess the consistency of the environmental data estimates derived from IUCN versus GBIF geographic range data, and to test whether such differences may affect estimates of ecogeographical relationships, such as Bergmann's rule.LocationWorldwide.Time periodAround 2000.TaxaRodents (Rodentia).MethodsWe first assessed the correlation between environmental data estimates (19 bioclimatic variables and elevation) derived from IUCN versus GBIF geographic range data of 1,315 rodent species, comparing how range size, conservation status, habitat, zoogeographic realm or elevation changed these correlations. Then, we compared the association between body mass and climate (mean temperature and precipitation) when the latter are derived from IUCN or GBIF data.ResultsThere was high correlation between environmental data estimates derived from IUCN versus GBIF data, especially when excluding GBIF occurrences outside of IUCN polygons. Species’ characteristics, or using the mean or median, did not change the consistency between estimates. Overall, GBIF occurrence data and IUCN range maps produced similar patterns of body mass—climate correlations.Main conclusionsAt the large spatial and taxonomic scale employed in this study, there does not seem to be any considerable differences in the average environmental data estimates derived from IUCN versus GBIF geographic range data. This result indicates that both sources of geographic range data could be used independently or in concert for macroecological inferences that involve summarizing species’ niches by a single estimate of the average of their used environments.

A discussion session held at a National Science Foundation-sponsored Herbarium Networks Workshop at Michigan State University in September of 2004 resulted in a rallying objective: make all botanical specimen information in United States collections available online by 2020. Rabeler and Macklin 2006 outlined a toolkit for realizing this ambitious goal, which included: a review of relevant and state-of-the-art web resources, data exchange standards and, mechanisms to maximize efficiencies while minimizing costs. a review of relevant and state-of-the-art web resources, data exchange standards and, mechanisms to maximize efficiencies while minimizing costs. Given that we are now in the year 2020, it seems appropriate to examine the progress towards the objective of making all US botanical specimen collections data available online. Our presentation will attempt to answer several questions: How close have we come to meeting the original objective? What fraction of “digitized” specimens are minimally represented by a catalog number, a determination, and/or a photograph? What fraction has been thoroughly transcribed? How close have we come to attaining a seamlessly integrated, comprehensive, and national view of botanical specimen data that guides a stakeholder to appropriate resources regardless of their entry point? What “holes” in this effort still exist and what might be required to fill them? How close have we come to meeting the original objective? What fraction of “digitized” specimens are minimally represented by a catalog number, a determination, and/or a photograph? What fraction has been thoroughly transcribed? How close have we come to attaining a seamlessly integrated, comprehensive, and national view of botanical specimen data that guides a stakeholder to appropriate resources regardless of their entry point? What “holes” in this effort still exist and what might be required to fill them? Given our interest in the success of both the Global Biodiversity Information Facility (GBIF) and the Integrated Digitized Biocollections (iDigBio), as well as the overwhelming likelihood that either one of these initiatives is the usual entry point for someone seeking US-based botanical data, we approached the answers to the above questions by first crafting a repeatable data download and processing workflow in early July 2020. This resulted in 25.6M records of plant, fungi, and Chromista from 216 datasets available through GBIF and 32.8M comparable records available through iDigBio from 525 recordsets. We attempted to align these seemingly discordant sets of records and also chose Darwin Core terms that were best suited to match the four hierarchical levels of digitization defined in the Minimal Information for Digital Specimens (MIDS) (van Egmond et al. 2019). During the analysis/comparison of the datasets, we found several examples where the number of data records from an institution seemed much lower than expected. From a combination of analyzing record content in GBIF/iDigBio and consulting regional/taxonomic portals, it became evident that, besides datasets only being included in either GBIF or iDigBio, there was a significant number of records in regional/taxonomic portals that were not yet made available through either GBIF or iDigBio. Progress on digitization has benefited greatly from the US National Science Foundation's creation of the Advancing Digitization of Biodiversity Collections (ADBC) program, and funding of the 15 Thematic Collection Networks (TCN). The launching of new projects and the ensuing digitization of herbarium collections have led to a multitude of new specimen portals and the enhancement of existing software like Symbiota (Gries et al. 2014). But, it has also led to insufficient data sharing among projects and inadequately aligned data synchronization practices between aggregators. Consistency in terms of data availability and quality between GBIF and iDigBio is low, and the chronic lack of record-level identifiers consistently restricts the flow of enhancements made to records. We conclude that there remains substantial work to be done on the national infrastructure and on international best practices to help facilitate collaboration and to realize the original objective of making all US botanical specimen collections data available online.

Digital point‐occurrence records from the Global Biodiversity Information Facility (GBIF) and other data providers enable a wide range of research in macroecology and biogeography. However, data errors may hamper immediate use. Manual data cleaning is time‐consuming and often unfeasible, given that the databases may contain thousands or millions of records. Automated data cleaning pipelines are therefore of high importance. Taking North American Ephedra as a model, we examined how different data cleaning pipelines (using, e.g., the GBIF web application, and four different R packages) affect downstream species distribution models (SDMs). We also assessed how data differed from expert data. From 13,889 North American Ephedra observations in GBIF, the pipelines removed 31.7% to 62.7% false positives, invalid coordinates, and duplicates, leading to datasets between 9484 (GBIF application) and 5196 records (manual‐guided filtering). The expert data consisted of 704 records, comparable to data from field studies. Although differences in the absolute numbers of records were relatively large, species richness models based on stacked SDMs (S‐SDM) from pipeline and expert data were strongly correlated (mean Pearson's r across the pipelines: .9986, vs. the expert data: .9173). Our results suggest that all R package‐based pipelines reliably identified invalid coordinates. In contrast, the GBIF‐filtered data still contained both spatial and taxonomic errors. Major drawbacks emerge from the fact that no pipeline fully discovered misidentified specimens without the assistance of taxonomic expert knowledge. We conclude that application‐filtered GBIF data will still need additional review to achieve higher spatial data quality. Achieving high‐quality taxonomic data will require extra effort, probably by thoroughly analyzing the data for misidentified taxa, supported by experts.

Bionomia, https://bionomia.net previously called Bloodhound Tracker, was launched in August 2018 with the aim of illustrating the breadth and depth of expertise required to collect and identify natural history specimens represented in the Global Biodiversity Information Facility (GBIF). This required that specimens and people be uniquely identified and that a granular expression of actions (e.g. "collected", "identified") be adopted. The Darwin Core standard presently combines agents and their actions into the conflated terms recordedBy and identifiedBy whose values are typically unresolved and unlinked text strings. Bionomia consists of tools, web services, and a responsive website, which are all used to efficiently guide users to resolve and unequivocally link people to specimens via first-class actions collected or identified. It also shields users from the complexity of casting together and seamlessly integrating the services of four giant initiatives: ORCID, Wikidata, GBIF, and Zenodo. All of these initiatives are financially sustainable and well-used by many stakeholders, well-outside this narrow user-case. As a result, the links between person and specimen made by users of Bionomia are given every opportunity to persist, to represent credit for effort, and to flow into collection management systems as meaningful new entries. To date, 13M links between people and specimens have been made including 2M negative associations on 12.5M specimen records. These links were either made by the collectors themselves or by 84 people who have attributed specimen records to their peers, mentors and others they revere. Integration With ORCID and Wikidata People are identified in Bionomia through synchronization with ORCID and Wikidata by reusing their unique identifiers and drawing in their metadata. ORCID identifiers are used by living researchers to link their identites to their research outputs. ORCID services include OAuth2 pass-through authentication for use by developers and web services for programmatic access to its store of public profiles. These contain elements of metadata such as full name, aliases, keywords, countries, education, employment history, affiliations, and links to publications. Bionomia seeds its search directory of people by periodically querying ORCID for specific user-assigned keywords as well as directly though account creation via OAuth2 authentication. Deceased people are uniquely identified in Bionomia through integration with Wikidata by caching unique 'Q' numbers (identifiers), full names and aliases, countries, occupations, as well as birth and death dates. Profiles are seeded from Wikidata through daily queries for properties that are likely to be assigned to collectors of natural history specimens such as "Entomologists of the World ID" (= P5370) or "Harvard Index of Botanists ID" (= P6264). Because Wikidata items may be merged, Bionomia captures these merge events, re-associates previously made links to specimen records, and mirrors Wikidata's redirect behaviour. A Wikidata property called "Bionomia ID" (= P6944), whose values are either ORCID identifiers or Wikidata 'Q' numbers, helps facilitate additional integration and reuse. Integration with GBIF Specimen data are downloaded wholesale as Darwin Core Archives from GBIF every two weeks. The purpose of this schedule is to maintain a reasonable synchrony with source data that balances computation time with the expections of users who desire the most up-to-date view of their specimen records. Collectors with ORCID accounts who have elected to receive notice, are informed via email message when the authors of newly published papers have made use of their specimen records downloaded from GBIF. Integration with Zenodo Finally, users of Bionomia may integrate their ORCID OAuth2 authentication with Zenodo, an industry-recognized archive for research data, which enjoys support from the Conseil Européen pour la Recherche Nucléaire (CERN). At the user's request, their specimen data represented as CSV (comma-separated values) and JSON-LD (JavaScript Object Notation for Linked Data) documents are pushed into Zenodo, a DataCite DOI is assigned, and a formatted citation appears on their Bionomia profile. New versions of these files are pushed to Zenodo on the user's behalf when new specimen records are linked to them. If users have configured their ORCID account to listen for new entries in DataCite, a new work entry will also be made in their ORCID profile, thus sealing a perpetual, semi-automated loop betwen GBIF and ORCID that tidily showcases their efforts at collecting and identifying natural history specimens. Technologies Used Bionomia uses Apache Spark via scripts written in Scala, a human name parser written in Ruby called dwc_agent, queues of jobs executed through Sidekiq, scores of pairwise similarities in the structure of human names stored in Neo4j, data persistence in MySQL, and a search layer in Elasticsearch. Here, I expand on lessons learned in the construction and maintenance of Bionomia, emphasize the criticality of recognizing the early efforts made by a fledgling community of enthusiasts, and describe useful tools and services that may be integrated into collection management systems to help churn strings of unresolved, unlinked collector and determiner names into actionable identifiers that are gateways to rich sources of information.

There has been major progress over the last two decades in digitising historical knowledge of biodiversity and in making biodiversity data freely and openly accessible. Interlocking efforts bring together international partnerships and networks, national, regional and institutional projects and investments and countless individual contributors, spanning diverse biological and environmental research domains, government agencies and non-governmental organisations, citizen science and commercial enterprise. However, current efforts remain inefficient and inadequate to address the global need for accurate data on the world's species and on changing patterns and trends in biodiversity. Significant challenges include imbalances in regional engagement in biodiversity informatics activity, uneven progress in data mobilisation and sharing, the lack of stable persistent identifiers for data records, redundant and incompatible processes for cleaning and interpreting data and the absence of functional mechanisms for knowledgeable experts to curate and improve data. The first Global Biodiversity Informatics Conference (GBIC) in 2012 delivered the Global Biodiversity Informatics Outlook (GBIO, Hobern et al. 2012), an architectural vision for the major components of a distributed global infrastructure for biodiiversity information, but realigning the work of existing organisations and projects to achieve this vision remains challenging. Recognising the need for greater alignment between efforts at all scales, the Global Biodiversity Information Facility (GBIF) convened the second Global Biodiversity Informatics Conference (GBIC2) in July 2018 to propose a coordination mechanism for developing shared roadmaps for biodiversity informatics. GBIC2 attendees reached consensus on the need for a global alliance for biodiversity knowledge, learning from examples such as the Global Alliance for Genomics and Health (GA4GH) and the open software communities under the Apache Software Foundation. These initiatives provide models for multiple stakeholders with decentralised funding and independent governance to combine resources and develop sustainable solutions that address common needs. GBIF was asked to coordinate next steps following GBIC2, including publication of a paper, Connecting data and expertise: a new alliance for biodiversity knowledge (Hobern et al. 2019). The supplementary materials for the paper include PDF brochures explaining the concept in eleven languages. During 2019, GBIF is coordinating further consultations to establish an optimal model for the governance and operations of the alliance and to advance collaboration around some of the major building blocks of the GBIO. Collaboration at this scale, and across all aspects of biodiversity information, is essential for effective delivery of important information products such as the Essential Biodiversity Variables and the planned pan-European natural history collections infrastructure, DiSSCo. This presentation explains the goals for this alliance and updates on progress during 2019 in operationalising the concept.