Synthesizing tree biodiversity data to understand global patterns and processes of vegetation

Prepared by Gunnar Keppel, Dylan Craven, Patrick Weigelt, Stephen A. Smith, Masha T. van der Sande, Brody Sandel, Sam C. Levin, Holger Kreft & Tiffany M. Knight

Key aspects of biodiversity and the major global databases with relevant information. (Figure 1 from the published paper.)

Trees are conspicuous elements of nature and key to the functioning of ecosystems and to our livelihoods. Because of their high importance, we know that vast information about trees exists in various places, such as online databases, published literature and herbarium records. We recently asked the questions “How much information about tree diversity is available?”, “How can we best integrate existing information?” and “What can we use the available information for?”. Our answers to these questions were recently published in the Journal of Vegetation Science.

We found that databases with information about the different aspects of biodiversity (see the schema above) held large amounts of data, with more than 84% of all tree species having some information about where they occur and their functional traits. Furthermore, the proportion of tree species for which biodiversity information was available was about 10% greater than that for plant species overall. Due to more complete information being available, trees therefore make a good model group for investigating global and regional biodiversity patterns.

While the amount of available information is exciting, we also found important gaps. For example, while the TRY database contained data for at least one plant functional trait for more than 95% of all tree species, only about half of these species (45%) had data available for at least three traits. This is of concern because two traits, growth form and woodiness, can be inferred for all tree species. Furthermore, limited data are digitally accessible about the demography, ecological interactions, and socio-economic role of tree species.

We have the technology and statistical approaches to integrate existing tree biodiversity data and to impute that data to species lacking information and to under-sampled regions. Furthermore, imputations will become more accurate as more resources are digitized and made available, a process known as data mobilization. However, for some taxa and regions, especially in the tropics, additional data collection is urgently needed. Available data can be used to identify remaining gaps to ensure targeted and efficient collection of such data.

The desert oak, Allocasuarina decaisneana (Casuarinaceae), is restricted to sandy desert environments in central Australia, where adults of the species are often the only trees. Photo credit: Gunnar Keppel.

Integrating existing data can help us in many ways. For example, when comparing the occurrence records for the Australian Desert Oak (photo above) with the mapped distribution of the vegetation type it dominates, the desert oak woodlands, it appears likely that the map does not cover the entire distribution of this vegetation type. There are also more applied uses. Refugia are places that may provide buffering from changes in climate and are hence considered increasingly important for biodiversity conservation. However, we still know relatively little about how refugia function through time and integrating existing data on functional traits and genetics of trees could provide vital insights.

Last, but not least, the existing biodiversity data should be readily and freely accessible to researchers and conservation practitioners alike. An interoperable framework that can be openly accessed through an easy-to-use platform would allow this. Therefore, existing data on tree biodiversity can enhance scientific progress and biodiversity conservation and management.

This is a plain language summary for the paper of Keppel et al. published in the Journal of Vegetation Science (