How many herbarium specimens are needed to detect threatened species?
Highlights
► We assess how many herbarium specimens are needed to detect threatened species. ► Fifteen specimens produce consistent range estimates for conservation assessments. ► None of the threatened species were misclassified. ► <3% species to receive conservation support were falsely identified as threatened. ► This approach could reduce digitisation effort by up to half.
Introduction
It is clear we are entering a time of immense environmental change. For us to gauge more accurately the impact of humans on biodiversity and associated ecosystem services, knowledge of species’ conservation status is critical (Rockström et al., 2009). In the past, biodiversity conservation has been primarily focused on megafauna, and especially vertebrates such as mammals and birds. A current goal is to address this imbalance and give more emphasis to non-vertebrate species such as plants, fungi and invertebrates, which represent the majority of earth’s biodiversity (Stuart et al., 2010).
The revised Global Strategy for Plant Conservation (GSPC) (2011–2020) calls for an assessment of the conservation status of all known plant species, as far as possible, to guide conservation action (target 2, UNEP, 2010b). A full conservation assessment of all known plant species to a consistent international standard, such as the International Union for Conservation of Nature (IUCN) Red List (www.redlist.org; IUCN, 2010), is the longer-term aim to facilitate conservation action; however, it is not realistic to assess all species by this method in the near future (UNEP, 2010a). Major constraints on achieving this target include: lack of funding for field work and data compilation, and lack of expertise for assessment activities leading to incomplete and scattered outputs (Secretariat of the Convention on Biological Diversity, 2009). To date, less than 4% (12,914) of the world’s plants species have been fully assessed and included on the IUCN Red List (IUCN, 2010).
The IUCN Red List is widely recognised as the international standard for assessing conservation status, comparable across taxonomic groups, and used extensively in setting conservation priorities by policy makers and other stakeholders. Full conservation assessments require extensive knowledge of the ecology and population dynamics of the taxa concerned, but for the majority of plant species, especially those from the tropics, this information is not available. The IUCN Red List states that “the absence of high quality data should not deter attempts at applying the criteria […]. The Red List criteria should be applied to a taxon based on the available evidence concerning its numbers, trend and distribution” (IUCN, 2001). Although the criteria are highly quantitative and defined, projections, assumptions and inferences can be used in order to place a taxon in the appropriate category (IUCN Standards and Petitions Working Group, 2010).
Georeferenced herbarium specimens can be used to model features such as species range (Willis et al., 2003) and population structure (Rivers et al., 2010), and such information can be evaluated against thresholds set for different IUCN threat categories, and used together with the appropriate subcriteria to assign an assessment of conservation status based on range (criterion B). Preliminary conservation assessments based on range estimates are objective, repeatable and independent of any projections, assumptions or inferences on the part of the assessor, provided that the set of herbarium specimens gives an accurate picture of the true range of the species. Over time, as more specimens are collected and more information becomes available, these assessments can be repeated, independently verified or corrected, and potentially upgraded.
Range estimates are essential for a species to be listed under criterion B, although other subcriteria also need to be fulfilled. For reference, on the IUCN Red List today, 80% of plants (assessed in the last 10 years) are listed on the basis of their range estimate (criteria B and D2) (IUCN, 2010). Many plant conservation assessments are carried out using herbarium specimen data; it is therefore important to assess how many specimens are needed to have confidence in the range estimates that underpin the majority of plant assessments.
Section snippets
Study area and species information
Madagascar, the focal region for this study, is one of the world’s biodiversity hotspots, based on the number of endemic species and loss of natural vegetation (Brummitt and Nic Lughadha, 2003, Mittermeier et al., 2005, Myers et al., 2000). Species endemism in Madagascar reaches over 80% among many animal groups (Goodman and Benstead, 2005) and similar levels are found among plants, such as palms (>95%) (Dransfield and Beentje, 1995), legumes (80%) (Du Puy et al., 2002) and orchids (>85%) (
Threatened or not threatened
The number of specimens needed for a reliable range estimate for use in the conservation assessment of species not under threat is seen in Fig. 2. Using only 10 specimens per species, the range-based assessments of not threatened legumes were estimated with a consistency of 87%, relative to the complete dataset; while orchids were estimated with a consistency of 98%. Using 15 specimens, then the range measures consistently estimated the threat status of legumes 96% of the time, and all of the
Discussion
Overall, 15 herbarium specimens are needed to have high confidence (95%) that the range estimates used in conservation assessments are consistent with those of the full set of specimens. First, we assessed the not threatened species, in order to ensure the species had sufficient sampling and had reached their final preliminary rating based on range. However, as we can be confident that the range of a not threatened species is identified after 15 specimens, we also have confidence in the range
Conclusion
Many digitisation projects are underway in herbaria around the world, increasing the availability of georeferenced collection data for plants. The critical next step is to integrate and analyse these data to make better informed conservation decisions. Despite the fact that the majority of the world’s plants have poorly known distributions represented by few specimens, we can still make robust preliminary conservation assessments. This study will help to focus priorities and make digitisation
Acknowledgements
We would like to thank the Legume section, the Orchid section and the GIS unit at Kew for technical help and advice. The late J.-N. Labat at the Muséum National d’Histoire Naturelle in Paris, and Missouri Botanical Garden kindly contributed legume specimen data. We also acknowledge P. Cribb, C. Hermans, J. Hermans, D. Houghton, J. Moat, L. Rajaovelona, and M. Rakotoarinivo for compiling and digitising the orchid dataset. M.R. was supported by the Natural Environment Research Council
References (30)
- et al.
Orchids
Current Biology
(2008) - et al.
Applications of the IUCN Red List: towards a global barometer for plant diversity
Endangered Species Research
(2008) - et al.
Biodiversity: where’s hot and where’s not
Conservation Biology
(2003) - et al.
The Palms of Madagascar
(1995) - et al.
The Leguminosae of Madagascar
(2002) - et al.
The sizes of species’ geographic ranges
Journal of Applied Ecology
(2009) - et al.
Updated estimates of biotic diversity and endemism for Madagascar
Oryx
(2005) - et al.
Orchids of Madagascar
(2007) - IUCN, 2001. IUCN Red List Categories: Version 3.1. IUCN Species Survival Commission, IUCN, Gland, Switzerland and...
- IUCN, 2010. 2010 IUCN Red List of Threatened Species. <http://www.iucnredlist.org> (downloaded October...
Legumes of the World
Improving the targeting of seed collecting programmes for ex situ conservation
Quantification of extinction risk: IUCN’s system for classifying threatened species
Conservation Biology
Cited by (99)
Habitat suitability assessment for Saunders's Gull (Saundersilarus saundersi) in the Yellow River Delta, China
2024, Ecological InformaticsConservation assessments using herbarium data reveal a genus at risk: The case of Cordia L. (Cordiaceae) in South America
2024, Journal for Nature ConservationFilling the gap to avoid extinction: Conservation status of Brazilian species of Epidendrum L. (Orchidaceae)
2023, Journal for Nature ConservationIssues with species occurrence data and their impact on extinction risk assessments
2022, Biological ConservationCitation Excerpt :It is worth noting that changes in the risk category result from changes in the number of locality data available to each record due to spatial, taxonomic, or temporal issues not corrected or due to an incomplete data collation. We also assessed the impact of taxonomic, spatial, and temporal issues on preliminary assessments using three additional reference datasets representing subsets of the CNCFlora data, including 1) species with up-to-date assessments carried out post-2018, 2) species with older assessments presented in the Red Book of the Brazilian Flora (Martinelli and Moraes, 2013), and 3) species with >15 occurrence records, a number suggested as the minimum for reliable automated assessments (Fig. S1; Rivers et al., 2011). We used R (v. 4.02; R Core Team, 2020) to perform all analyses, the package bdc for data-cleaning (Ribeiro et al., 2022; https://cran.r-project.org/web/packages/bdc/index.html), and ggplot2 (Wickham, 2016) to create all figures, except the Sankey diagram, which was generated using networkD3 package (Allaire et al., 2017).