Predictive Habitat Modeling
Deep-sea coral ecosystems provide essential habitat for many commercial fish species as well as a variety of other organisms, but very few of these ocean areas have been explored or mapped due to the significant time and expense of studying the deep seafloor.
Predictive habitat modeling is a tool that helps scientists better map the areas where these mysterious and spectacular deep-sea ecosystems will likely be found. These models allow scientists to use the limited data that exist, often remotely sensed data collected over large areas of the seafloor using a combination of satellites and ships, in order to predict where to find ecologically or economically important species.
The maps produced are used to inform management and conservation decisions to protect these fragile, unique ecosystems. For example, we can use habitat modeling to identify candidate sites for new marine protected areas and work with management agencies to stop damaging fishing practices in particularly sensitive areas. These maps can also help find and prioritize biodiversity hotspots during field expeditions that can only survey a small area of the seafloor.
Marine Conservation Institute uses predictive habitat models to advocate for the protection of deep-sea coral and sponge habitats across our oceans. These models are used to brief fishery managers, fishing industry representatives, non-governmental organizations and the US Congress about policy relevant results, and conservation and management measures. The models we produce have been used to advocate for the improved management of deep-sea ecosystems in New Zealand waters, the U.S. West Coast, and the North Pacific Ocean. We interact closely with NOAA’s Deep Sea Coral Research and Technology Program to provide the scientific information necessary to effectively manage U.S. deep-sea coral ecosystems. Predictive habitat models also provide critical data and analyses that support Marine Conservation Institute’s High Seas Protection Portal and Blue Parks initiative.