Predictive models that can forecast biodiversity decline under anthropogenic climate change used to be too simplistic, as these ignored crucial biological mechanisms such as demography, dispersal, evolution, and species interactions (for instance species competition and ecosystem dependence).

Fortunately these climate-biodiversity models are improving. But in order to also improve the quality of their biodiversity predictions these improved models now require more – and better data.

When it comes to data collection, simply documenting a change in biodiversity is not good enough. Modern predictive models require data about the specific mechanisms of change too, in order to better project the response to climate change.

This a group of 22 ecologists conclude in a joint paper that was published in Science in 2016, titled ‘Improving the forecast for biodiversity under climate change’ – that was based on a review of available predictions for biodiversity decline under anthropogenic climate change that were published in scientific literature.

Such metastudies show many interesting patterns. An earlier publication, also in Science, revealed that biodiversity decline accelerates with climate warming – and that some continents (South America, Australia) seem to be more vulnerable than others.

In that study, from 2015, it was also early concluded that the vast majority of available model-based biodiversity projections exclude complex ecological mechanism with which species (while interacting) respond to climate change – ending with the following outlook:

“We must cautiously interpret the predictions underlying this meta-analysis. The majority of studies extrapolate correlations between current climate and species distributions to novel conditions and omit important biological mechanisms, including species interactions, evolution, landscape dispersal barriers, habitat degradation, and intraspecific trait variation. Depending on the mechanism, its consideration can either increase or decrease predicted risks. For instance, evolution can decrease extinction risks by allowing populations to adapt to changing climates, whereas anthropogenic landscape barriers can increase risks by limiting dispersal into newly suitable habitats. Next-generation models for estimating extinction risks should incorporate these factors in order to increase biological realism and therefore the accuracy of future predictions.”

When assessing outcome quality models that do incorporate biological mechanisms prove to be more inaccurate, while correlative climate-biodiversity models become increasingly inaccurate.

How to improve biodiversity forecasts under climate change?

These next-generation models that do incorporate more complex ecological mechanism do however require more sophisticated data in order to further improve their predictions.

In the newer Science publication the research team – that was led by the same author, Mark Urban of the University of Connecticut – also suggests practical ways to improve the data feed to get to these improved forecasts, differentiating data quality for 6 different components: evolution, environment, physiology, demography, dispersal and species interactions (competition).

A review of available literature showed that thus far the vast majority of model-based climate-biodiversity predictions (77% of 131 reviewed studies) used models that included none of these 6 important biological mechanisms – and of the remaining studies most included just one mechanism. Species interaction was almost never included in model-based forecasts for biodiversity decline.

Even well-studied individual species show data quality gaps. For instance the European beech tree has good overall climate research data, with high data quality for mechanisms of environment, physiology, demography and species interaction, but omissions (medium quality) for specific data about beech tree evolution and dispersal changes under climate change. The fence lizard (a medium-sized lizard species from the US) has poor data quality to feed climate-biodiversity model components of species dispersal, interaction and evolution – under influence of climate change.

The authors recommend a trial & error approach of model testing and revisions in order both to fine-tune new biodiversity models and to reveal what type of data is of particular importance to improve outcomes. They also recommend usage of different independently devised models to produce ensemble forecasts, for better comparison and quicker model improvements. [For further recommendation – there are many – please read the initial publication.]

© Rolf Schuttenhelm | www.bitsofscience.org