People depend on agriculture, local water sources, clean air, and local plants and animals in their daily lives. However, the current species extinction rate is ten to hundreds of times higher than it has been during the last 10 million years, which has negative consequences for ecosystem services and the communities that depend on them. In the SILVIS Lab, we are creating maps of species distributions that help agencies and land managers prioritize conservation actions.
For centuries, humans have recognized that our collective actions modify and shape the world around us. These actions also have direct and lasting impacts on the plants and animals that many communities rely on for their livelihoods ─ often referred to as ecosystem services. For example, in the Rocky Mountain West, local communities rely on provisioning from local wildlife (e.g., hunting), safe water and air from local plants and snowmelt, ecosystem cultural services for tribal communities, natural soil creation for agriculture, pastures for grazing livestock, and income from tourists seeking to fish, hike, hunt, ski, and view local wildlife. Similarly, the Great Lakes wetlands provide fisheries habitat that supports wildlife and people, space for recreational activities and tourism, hydro and wind power, shoreline protection, sediment trapping, and storage for nutrients and carbon. The ecosystem services provided by any one region, such as the Rocky Mountain West or the Great Lakes, are intrinsically tied to the health of the species and people in that community. Without plants and animals, many of these services could suffer. Therefore, one of the most challenging questions for scientists is how to ensure that our science advances conservation and subsequently bolsters ecosystem services that support local communities.
In the SILVIS Lab, Kathleen Carroll identifies and averts biodiversity loss by developing complex models to predict and examine biodiversity patterns across the US. Biodiversity, which measures the variety of life in an area, is essential to conservation. If biodiversity declines, there is a risk to both species and ecosystem services. By modeling and predicting biodiversity over broad areas, we can determine where the biggest threat to species is. While this approach is simple in theory, extensive information about where species are and what they select habitat based on is key. Researchers have been limited by access to data. The SILVIS lab has developed new satellite indices, which provide detailed datasets that I am using in complex models. My preliminary results show higher predictive power from relatively new machine learning models, called randomForest models, compared to other approaches. Access to the new satellite datasets and complex models helps me to 1) further evaluate the value of fine-scale satellite data for biodiversity mapping, 2) develop predictive biodiversity models, and 3) provide maps to land trusts and communities to help them protect species and ecosystem services.
Story by Carroll, Kathleen