Predictive Habitat Modeling
Predictive habitat modeling is the process by which relationships are mathematically defined between habitats (e.g., sand with seagrass) and associated environmental or geographic variables (e.g., depth, slope, distance to the shoreline) of those habitats. NCCOS uses predictive modeling approaches to semi-automate habitat characterization and classification processes. This approach provides a framework to more efficiently process increasing volumes of data. Recent advances in artificial intelligence now make it easier to implement predictive modeling approaches and generate continuous predictive maps for benthic habitats. These approaches enhance our ability to characterize habitats in support of NCCOS science and the coastal and marine management community.
Over the last decade, natural and human-caused threats have changed the composition of shallow-water (<30 m) benthic habitats. Looking to better understand these changes, local resource managers asked NCCOS to develop new habitat maps. Using satellite imagery, underwater photographs, and predictive modeling, we can produce original and updated benthic habitat maps to visualize current and future benthic habitat composition.
Mapping Habitat Change in Saipan Lagoon, CNMI NCCOS provides a critical spatial framework to better understand benthic habitat changes over the last decade in Saipan Lagoon. Future projects or management actions in Saipan Lagoon will benefit greatly from the updated imagery, habitat map, and underwater photos and videos compiled during this project.
Mesophotic coral reef ecosystems, found at depths of 30 to 150 meters, provide many ecological and economic benefits to nearby islands, including supporting fisheries and enhancing shoreline protection. Like their shallow-water counterparts, these ecosystems are increasingly under threat from a variety of stressors and threats, like overharvesting, pollution, coastal development, invasive species, ocean acidification, and climate change. In recent years, there has been a greater effort to document mesophotic coral reef ecosystems, and collect baseline information about their spatial distributions, compositions, and health. This baseline spatial information, including habitat maps, are needed to monitor changes in these ecosystems and to develop strategies to mitigate these threats moving forward.
Developing A Benthic Habitat Map for the Insular Shelf South of St. Thomas and St. John NCCOS uses existing acoustic data sets and underwater videos to produce maps of mesophotic coral reef ecosystems (30 to 100 m) on the insular shelf. These data sets can be used to make more informed decisions about sampling, permitting activities, and the management of marine protected areas in the U.S. Caribbean.
Deep-sea coral and sponge communities, found at depths greater than 150 meters, are also vulnerable to natural and human-caused stressors and impacts. However, collecting data on these communities is logistically challenging and cost prohibitive, especially in open ocean areas such as the Gulf of Mexico, or remote locations like those in the Pacific Ocean. To fill this information gap for these remote areas, NCCOS produces maps of deep-sea coral and sponge communities (100–3,500 m), providing a cost-effective tool for identifying potential distributions of these communities and other deep-sea organisms in areas where data are sparse yet information is needed to support conservation and management decisions.
Predicting and Characterizing Deep-Sea Coral Habitats within the Papahānaumokuākea Marine
National Monument, Hawai‘i
NCCOS provides critical information on the extent of vulnerable habitat within the national monument in the form of easily digestible map products for habitat characterization and future reviews of the monument expansion and protections.
Multiple branches across NCCOS are working together to make predictive modeling a core capability in NCCOS's habitat characterization and classification process. This effort enables NCCOS to:
Make habitat maps more objective, quantitative, and probabilistic
Operationalize HMT’s habitat mapping and characterization processes
Transition mapping to a monitoring tool, which could be used to detect and quantify changes in seafloor habitats over time