Out-of-date datasets hamper conservation of species close to extinction

International databases and data aggregators on species conservation status are powerful tools supporting the efforts of conservation biologists and practitioners in reducing the loss of biodiversity. However, out-of-date information and poor interoperability of databases can hamper conservation of highly threatened species or in extreme cases can result in their removal from conservation frameworks. Lack of common standards for database updates, slow update timing and incongruencies among datasets in terms of taxonomy, threats, conservation status and holding prevent proper conservation prioritization and actions. A simple survey to solve incongruencies between the BGCI PlantSearch Database and the IUCN Red List resulted in a change of status of 16 plant species, including the 'rehabilitation' of 12 species thought to be extinct.

Botanic Garden Conservation International (BGCI)'s PlantSearch database provides information on the plant collections (living collections and seed banks) of botanic gardens around the world. It presently holds data from over 1,000 institutions, consisting of more than 1.5 million occurrence records, representing over 600,000 taxa. BGCI also maintains the ThreatSearch database which lists global, regional and national red list assessments for plants obtained from a range of sources, including the IUCN Red List. ThreatSearch currently includes over 300,000 conservation assessments, representing over 180,000 taxa.
Comparing data from these two databases showed that a number of taxa reported as extinct were also reported as being in cultivation ex situ in botanic garden collections. This demonstrates the clear incongruency of out-of-date information.
With the aim of assessing the impact of out-to-date information on species conservation, we therefore checked the status of 24 non-European species recorded as Extinct (according to the IUCN Red List and 5, 7) but were also recorded as present in BGCI's PlantSearch database in at least one ex situ collection. Synonyms for each species were identi ed using the World Checklist of Vascular Plants and the threat status for each species (using both accepted names and synonyms) was checked using BGCI's ThreatSearch database. In each case, where more than one assessment was available, the most recent assessment was used. Similarly, records for each species were checked against PlantSearch and the gardens recorded as cultivating any of the species were identi ed. Finally, updates for these species were obtained by contacting all the institutions identi ed to have them in collection. Therefore, species reported as EW in table 1 have at least a veri ed (this article) ex situ living collection in a botanic garden belonging to the BGCI network. Additional data have been collated by checking local oras or by contacting local experts for those species still extant in their native range (see notes to Table 1). For four species we could not get any updates as the institutions involved where uncontactable (Delissea subcordata Gaudich., D. undulata Gaudich, Grevillea divaricata R.Br., Proboscidea spicata Correll).
In summary, Table 1 shows that 14 species deemed to be extinct are extinct in the wild or extant (with overlaps among sources): three species listed in the IUCN Red List as EX are EW and one (Pimpinella schweinfurthii) is still present in its native range in the Arabian Peninsula; four out of fourteen species declared extinct by (5) are actually extinct (plus one possibly extinct), four species are EW and ve are still extant; two species reported as EX by (6) and (7) should be reclassi ed as EW.
These fourteen species are therefore back into a conservation framework.
Some inaccuracies are surprisingly coarse, like for instance, the case of Micrantheum micranthemoides, declared extinct but very common in the market of commercial plants. In that case, despite its EW status, the conservation of M. micranthemoides may deserve a lower conservation priority than other EW species cultivated in a single botanic garden and with very small populations. A more subtle example of the confusion generated by out-of-date datasets is the case of Cynometra beddomei Prain, that was declared EX in 1998 (9), a status still reported in the PlantSearch database despite the indication that an ex situ collection exists. The extinct status was con rmed by (5), but a more recent IUCN assessment indicates that the species is not extinct at all, and it occurs in the wild and it is classi ed as Endangered (10).
Interestingly, three specimens of this species grow at Nehru Tropical Botanic Garden & Research Institute, India (A. Dyhani pers. comm.), which is not the same BGCI institution where the unique ex situ collection reported in PlantSearch was thought to be. In such a case, the perception of conservation needs for this species differs greatly depending on which dataset is considered. A further example of confusion due to out-of-date dataset is the case of Trochetiopsis melanoxylon s. str., a species already extinct at the time of its description by (11). The latter author split T. melanoxylon s.l. into two species T. melanoxylon s. str. and T. ebenus Cronk. Since collections of the putative T. melanoxylon were made before 1995, this species is still reported as cultivated ex situ, while current living specimens belong to T. ebenus.
Looking at the data turnover of PlantSearch, it emerges that many institutions do not consider updating datasets as an important task, and do not refresh their data regularly; for example, there are 257 plant lists in PlantSearch, from European gardens, mostly provided since 2010 but a few are from earlier. Most of these lists have not been updated since they were rst submitted. Seventy-seven gardens have uploaded plant lists since January 2019 and of these, only 14 have subsequently been updated, despite regular requests for updates at European meetings.
In the speci c case, one of the most serious consequences of out-to-date information is the underestimation of the collection value. For instance, despite Trochetiopsis melanoxylon was originally cultivated in six botanic gardens, our survey suggests that only one BGCI garden currently hosts veri ed living specimens of this species (Table 1). If follows that the only institution hosting the last individuals of T. melanoxylon may be unaware of the role and responsibility it has in the conservation of this species. More generally, a clear perception of a collection value achieved through fresh and accurate information is key to proper ex situ population management, like for example material duplication and exchange (see the case of T. melanoxylon; 12). This work shows how periodical updates of species databases and data aggregators can contribute to species conservation by excluding species that do not need conservation anymore (i.e., extinct species) or that deserve high conservation attention like the thirteen species here rehabilitated. Updates should involve several aspects such as the taxonomy, for which we need a de nitive taxonomic backbone to link datasets such as the World Flora Online (13), the critical review of historical information, the scienti c literature on species phylogenesis (14) and distribution (e.g., 15,16), and for databases like PlantSearch a simple periodical physical check of collections. An increased consideration of validated citizen science datasets (e.g., iNaturalist; www.iNaturalist.com) would also bene t the retrieval of new data, especially on species distribution.
Guidelines on what, when, how to update datasets should be developed to better integrate data management into conservation practice. In this perspective, new approaches deriving from other elds such as medicine and data management should be applied to obtain a more e cient system of data gathering with links among institutions, and datasets interoperability (17)(18)(19). Table 1 Update of status for 20 extinct species with at least an ex situ collection in BGCI PlantSearch database.