Next-level Species Distribution Modeling

Species distribution models (SDMs) are statistical and mechanistic models used to define the geospatial distribution of species based on the combination of biological and ecological variables (such as the biotic and abiotic environment, dispersal ability, etc.) that provide favorable conditions and opportunity for their occurrence.

By projecting SDMs onto future environments, scientists can determine where and when those conditions will align to provide a prediction of future species distributions. These predictions are often forecast months, years, or decades into the future, and are static with respect to both the algorithm and the predicted occurrences.

However, the factors that affect species and their movement are not static. Imagine if you could apply those models to the changing world in real time! That is precisely what we are helping scientists to do using Geomatys’ EXAMIND on-the-fly geospatial processing and datascience technology.  

As environmental conditions change, or are affected by perturbations such as a hurricane or development projects which disrupt current habitats, fine-scale SDMs can be applied to predict how animals will disperse. Together with our partners in research and industry, we are working to apply this developing technology to, for instance, manage animal populations. This capacity will become essential in nearly every corner of life across the globe as climate change destabilizes the patterns on which we currently rely for making decisions.

One project where Geomatys’ technology is facilitating such work is through the French association for the management and conservation of the endangered Przewalski’s Horse (TAKH https://www.takh.org/). The association debuted their EXAMIND-powered web portal for visualizing and analyzing Przewalski’s Horse populations, called Shamane, at this year’s IUCN World Conservation Congress on 8 Sept 2021 in Marseille. 

Explore the Shamane platform (https://takh.geomatys.com/)

While the goal is to have machine-learning algorithms trained that can help to predict how the horses will behave in response to environmental factors, perhaps the most time-consuming and valuable work we did to facilitate this project was to build the database, pulling in vast and disparate data sources, ensuring interoperability, and making them accessible to the user all in one environment. 

Thanks to specific functionalities added to its EXAMIND backbone in response to the needs of TAKH researchers, users can now follow individual animals through time, toggle their history and pedigree, explore their habitats in 4D, query related datasets, and launch analyses all from within the spatial data infrastructure environment of the Shamane. The tool thus allows not only for data analysis but also provides intelligence for making real-time population surveillance and management decisions.

The EXAMIND-powered 4D interactive data visualization environment of Shamane.

The screen capture above illustrates how the user can follow the movement of genetically-distinct individual horses (represented by different colors, often grouping into herds) across a 3D view of the landscape. Using the slider at the bottom of the page, they can follow changes in animal positions and also changes in the habitat through time. This allows researchers to determine , for instance, what type of habitat barriers might influence movement. They can also overlay other data, such as meteorological data, onto this view, and run analyses in the left sidebar using a language-agnostic data science notebook. While the tool is available through a web portal, access is restricted to authorized users, secured by the same technology entrusted to Geomatys by the defense industry. This is important for dealing with sensitive data, such as the precise location of endangered species. This tool thus provides a performant and secure platform to manage the conservation of these fragile populations.

Leave a comment

Your email address will not be published. Required fields are marked *