Human activities and climate change are threatening mycorrhizal fungi at a rate never before seen, but we lack the tools to locate where these threats are happening. To address this, we are using geo-spatial data to identify and quantify threats to underground biodiversity.

Biodiversity mapping uses data to make predictions about the biodiversity across all of Earth’s ecoregions. With this kind of machine learning work there is always a degree of uncertainty. SPUN sees areas with high uncertainty as opportunities to “ground truth” our predictions — collecting real-life samples from areas where our model has made predictions allows us to check those predictions against the biodiversity we find and strengthens our models.

How does SPUN choose locations for sampling expeditions? To produce the most valuable data, we make strategic decisions when we choose where to sample for mycorrhizal fungi. We prioritize regions that have unique characteristics and low sampling intensity — this way we make sure to collect samples in areas that will have the highest impact on our effort to map global mycorrhizal biodiversity.