The important challenge of quantifying tropical diversity
© Magurran 2017
Published: 7 March 2017
The tropics are the repository of much of the world’s biodiversity, yet are undersampled relative to temperate regions. To help fill this knowledge gap, a paper in BMC Biology explores diversity patterns in tropical African plants, as revealed by the RAINBIO database. The paper documents spatial variation in diversity and data coverage, but also highlights the challenges faced in quantifying diversity patterns using data collated from a range of sources including herbaria.
See research article: http://bmcbiol.biomedcentral.com/articles/10.1186/s12915-017-0356-8.
The gold standard for biodiversity assessment is representative sampling using consistent and appropriate methodologies . Robust field sampling demands expert taxonomic knowledge and the ability to sample localities selected using stratified random sampling . Unfortunately, in many parts of the world, the resources needed for such surveys are limited or absent, while the terrain within which these poorly documented assemblages occur can be unsafe or inaccessible. Citizen science is often hailed as a means of plugging knowledge gaps but it too is concentrated in regions that are already relatively well documented. Researchers therefore need to turn to other sources of evidence, including information gleaned from specimens deposited in museums and herbaria, ad hoc inventories, and personal records, as exemplified in the RAINBIO resource , a curated database of vascular plant species in tropical Africa.
The analysis of RAINBIO data by Sosef and colleagues  seeks to understand how plant diversity is distributed across tropical Africa. In this it demonstrates the challenges, and also the opportunities, afforded by databases assembled from different sources. One of Sosef et al.’s goals is to estimate species richness at two levels of spatial resolution: 0.5° grid cells and country level. To do this the investigators employ a non-parametric species richness estimator developed by Anne Chao (see ). In essence this measure, known as Chao 1, weights the number of observed species by the fraction recorded as either singletons (one specimen) or doubletons (two specimens). As the relative frequency of singletons increases, so too does the estimate of richness, the logic being that the observation of many rare taxa is indicative of a large number of unseen species. Chao’s estimator provides a minimum estimate of richness and assumes homogeneity amongst samples . This latter expectation is likely to be violated in databases constructed from a range of sources. Indeed, investigations in other tropical ecosystems (for example, ) highlight the extent to which variation in sampling effort biases species richness estimation. Sosef et al.  find, as did Engemann and colleagues , that estimated richness correlates with sampling effort. Since herbaria and museums value breadth and rarity in their collections, databases populated from these sources are prone to over-emphasizing singleton and doubleton species. However, it is worth noting that rigorous surveys of tropical ecosystems similarly report high numbers of singleton taxa .
Other metrics employed in the paper, such as the Neilsen diversity metric, are less sensitive to sampling effort than richness estimators. Nonetheless, it is important to remember that the value the Neilsen statistic calculates—dubbed ‘effective species’—refers to an attribute of the data set and does not necessarily provide an ecologically meaningful insight into the structure of the assemblages from which the data were sourced. Even turnover metrics, which evaluate compositional change over space or time and might seem to be less obviously affected by sampling effort, can be influenced by inconsistencies in sampling .
These caveats could suggest that analyses based on heterogeneous databases are so fraught with problems as to be uninformative. Yet, provided that the results are interpreted cautiously, such analyses can provide a crucial initial step towards a better understanding of diversity patterns and help steer future investigations towards productive quests. The first contribution of these analyses is that they make plain the extent of heterogeneity in data collection over space and time. Although collecting dates in RAINBIO range from 1782 to 2015, collecting intensity is not constant. For instance, recording in the Democratic Republic of Congo occurred mostly from the 1930s to the 1960s . Clear recommendations for improving understanding of tropical diversity, such as improved data exchange between datasets, are identified.
A second contribution of analyses of databases such as RAINBIO is that firmer estimates of richness, endemicity, range extent and patterns of commonness and rarity begin to emerge. While is it wise to be careful about estimates of richness and reported values of diversity statistics, especially when making comparisons with other regions and systems, these results help shape hypotheses and guide objectives for future studies. For example, Sosef and colleagues’  analysis of diversity patterns and data coverage is well placed to inform new investigations on a range of topics, such as the consequences for biodiversity of climate change, and to highlight the groups and regions in which taxonomic research is most urgently needed.
New insights into functional diversity represent a third contribution of an analysis such as this . Sosef et al. report geographic variation in plant growth form dominance across tropical Africa, a field of research in which data have previously been scarce. This advance is possible because the observed diversity patterns are underpinned by georeferenced data on named species.
Finally, a macroecological analysis such as Sosef et al.’s, can be viewed alongside other investigations of data with a similar geographical scope, for example Marshall et al.’s  recent assessment of rarity amongst the tropical African flora. Of course different compilations of data may share many of the same shortcomings, not least because some overlap in source material is likely, but as long as they are not interpreted too simplistically, complementary analyses such as Sosef et al.’s  and Marshall et al.’s  jointly advance understanding and are well placed to support the conservation of biodiversity in our rapidly changing world. In particular they underline the pressing need for better understanding of the distribution and nature of tropical diversity. Innovative and insightful analyses of data in RAINBIO and other databases will be key to achieving this aim.
I thank Faye Moyes for constructing Fig. 1.
AEM is supported from ERC grants BioTIME (250189) and BioCHANGE (727440).
AEM wrote this article.
The author declares that she has no competing interests.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Dornelas M, Gotelli NJ, McGill BJ, Shimadzu H, Moyes F, Sievers C, et al. Assemblage time series reveal biodiversity change but not systematic loss. Science. 2014;344:296–9.View ArticlePubMedGoogle Scholar
- Gaston KJ, May RM. The taxonomy of taxonomists. Nature. 1992;356:281–2.View ArticleGoogle Scholar
- Engemann K, Enquist BJ, Sandel B, Boyle B, Jørgensen PM, Morueta-Holme N, et al. Limited sampling hampers “big data” estimation of species richness in a tropical biodiversity hotspot. Ecol Evol. 2015;5(3):807–20.View ArticlePubMedPubMed CentralGoogle Scholar
- Sosef MSM, et al. Exploring the floristic diversity of tropical Africa. BMC Biol. 2017. http://bmcbiol.biomedcentral.com/articles/10.1186/s12915-017-0356-8.
- Gilles D, Zaiss R, Blach-Overgaard A. Catarino Ls, Damen T, Deblauwe V, et al. RAINBIO: a mega-database of tropical African vascular plants distributions. PhytoKeys. 2016;74:1–18.View ArticleGoogle Scholar
- Buckland ST, Baillie SR, Dick JM, Elston DA, Magurran AE, Scott EM, et al. How should regional biodiversity be monitored? Environmental Ecological Stat. 2012;19(4):601–26.View ArticleGoogle Scholar
- Colwell RK, Coddington JA. Estimating terrestrial biodiversity through extrapolation. Phil Trans R Soc Lond B. 1994;345:101–18.View ArticleGoogle Scholar
- Magurran AE. Measuring biological diversity. Oxford: Blackwell Science; 2004.Google Scholar
- Coddington JA, Agnarsson I, Miller JA, Kuntner M, Hormiga G. Undersampling bias: the null hypothesis for singleton species in tropical arthropod surveys. J Animal Ecol. 2009;78:573–84.View ArticleGoogle Scholar
- Marshall CAM, Wieringa JJ, Hawthorne WD. Bioquality hotspots in the tropical African flora. Curr Biol. 2016;26(23):3214–9.View ArticlePubMedGoogle Scholar