Presenter: Elizabeth Pendleton (US Geological Survey)

Description:
Climate change is amplifying environmental hazards. Extreme weather events, nuisance flooding, and sea-level rise are persistent coastal hazards requiring synthesis of geologic, geographic, ecologic, oceanographic, and sociologic information to meet the needs of managers tasked with mitigating coastal change in the coming decades. To support the need for coastal hazard assessments, the U.S. Geological Survey is improving its coastal vulnerability index (CVI) to sea-level rise assessment conducted 20 years ago. The CVI overhaul preserves the strengths of the original study, which include a color-coded risk map, a synthesis of factors that influence coastal change, and national-scale coverage, while adding data sources, increasing resolution and domain coverage, reducing the timescale of expected change (from 50 to 10 years), and capitalizing on spatial data analysis techniques such as machine learning classification. A decision tree framework is used to manage the physical characteristics or landscape classification (e.g., landcover type, elevation, marsh stability) of the coastal zone, defined for a pilot area in the Northeast U.S. as +10 to -10 meters relative to mean high water. A database of coastal hazards is then coupled with the landscape classification in a supervised machine learning regression, which predicts coastal change likelihood as a function of landscape endurance and location-specific coastal hazards. Initial output indicates that areas with the highest propensity to change exist where landcover can be refined  with metrics (e.g. shoreline change) and multiple hazards occur. In addition to a coastal change likelihood planning tool for managers that distills over a dozen disparate datasets, this effort serves as a gap analysis to identify where data sources are insufficient to support confidence in change likelihood predictions. This information can be used to prioritize future research and data needs.