Predicting the future of biodiversity using Essential Biodiversity Variables

Study proposes workflow for informing indicators of biodiversity change for alien invasive species

GBIF-mediated data resources used : 732,113 species occurrences
Bubulcus ibis
Cattle egret (Bubulcus ibis) by David Morales Ramirez (CC BY-NC 4.0)

This feature is also published in the GBIF Science Review 2019, which highlights important and noteworthy examples of the use and reuse of GBIF-mediated data in research and policy.

What’s the weather going to be like tomorrow, next week—or even in ten years? This is a question researchers have worked hard on for many years. Thanks to high demand from commercial and political interests, such research has received a great deal of interest and funding, and today we have excellent weather and climate models based on standardized variables like temperature, wind speed, etc.

But what if the question was: what is biodiversity going to be like tomorrow or in ten years? It’s difficult to measure biodiversity in the same way you can measure temperature. What standardized variables can be used to derive models to accurately predict changes in biodiversity?

Proposed in 2012, Essential Biodiversity Variables (EBVs) are a concept with the idea of defining standardized measurements that can be repeated and reproduced accurately in order to inform indicators of change in biodiversity. Involved in biodiversity informatics since 2003, Alex Hardisty, Director of Informatics Projects at the School of Computer Science & Informatics at Cardiff University in the UK, has led the work of preparing what is thought to be the first EBV data product, described in a publication from early 2019:

“I was particularly interested in the challenges involved in interoperating between research infrastructures, and in previous projects we became aware of EBVs and their use as a potential unifying use-case for driving interoperability between different infrastructures.”

Working in close collaboration with two research infrastructures—GBIF and the Atlas of Living Australia (ALA)—Hardisty and his team developed a workflow in three stages aimed at producing an EBV for measuring changes in the distributions of alien invasive species.

In the first stage, they gathered, filtered and harmonized occurrence data from the infrastructures, followed by the second stage, in which they merged data from the two sources and performed a range of quality controls. In the final stage, they generated tables of time-series, calculated areas of occupancy—or AOO—by decade and produced maps and graphs of the results.

The team tested the EBV workflow on three candidate species with established alien ranges: Sydney golden wattle (Acacia longifolia), European wasp (Vespula germanica) and cattle egret (Bubulcus ibis). For all three species, the metrics derived from the EBV showed an exponential increase in AOO over the last few decades.

Obtaining the data and preparing it for the EBV, however, wasn’t trivial and the team faced a number of challenges along the way, as described by Hardisty:

“Much to our surprise, a lot of manual manipulation was required to get data ready, and it took much longer than anticipated. There were many differences between GBIF and ALA—in data retrieval, tools available for data handling and in terms used to describe the data. As a result, the level of automation we wanted wasn’t possible and we had to do a lot of manual work.”

As the point of EBVs is to create reliable variables that can be measured repeatedly, steps should be taken to harmonize data formats, quality checks and assertions across research infrastructures, and to automate—a point to which Hardisty suggests working more intensely to overcome some of the consistencies encountered by their team:

“Some human expertise will always be necessary, but at the same time, humans are incredibly bad at keeping an accurate record of what they do, which means it can be impossible to repeat the precise steps. For EBVs, automated workflows are critical.”

As to the demand for EBVs and who should be responsible for producing data products, Hardisty returns to the weather and climate analogy:

“Why do we have good weather forecasting abilities today? Because of demands from the aviation and space industry, East/West political tensions in 1950-70s, even nuclear bomb development—all driving the need for predictable models. We’ve yet to see a real demand for routine EBV data products, however, our study has laid the groundwork for future developments. If the world wants EBVs with which it can create indicators and measure changes and trends in biodiversity, then the world has to provide the infrastructure to do that. This is driven by the political agenda—and only when this becomes hot enough will funding become available.”

Hardisty AR, Belbin L, Hobern D, McGeoch MA, Pirzl R, Williams KJ and Kissling WD (2019) Research infrastructure challenges in preparing essential biodiversity variables data products for alien invasive species. Environmental Research Letters. IOP Publishing 14(2): 25005.
Available at: