COASTAL EROSION HAZARD MAPPING: THE STATE OF THE SCIENCE

In response to a Congressional mandate in 1994, the Federal Emergency Management Agency (FEMA), in partnership with the H. John Heinz Center for Science, Economics and the Environment, undertook a landmark national study to evaluate the economic toll of long-term coastal erosion. The two-phase study examined the methods by which coastal States determine past and project future shoreline change and investigated the economic impacts of including erosion information in the National Flood Insurance Program (NFIP). Results for many of the States participating in Phase One, the pilot mapping component of the study, were published in a special issue of the Journal of Coastal Research in 1999. In addition to documenting the high number of current structures likely to be threatened by erosion over the next 60 years, the study results revealed significant heterogeneity in the data available to States for erosion assessments, as well as in the methods used to evaluate these data.

The most notable differences among individual Phase One studies include the temporal record length for measuring shoreline change, the mathematical method used to calculate rates of change, and the issue of whether or not to include immediate post-storm shorelines in long-term assessments. These differences are not restricted to the Phase One States; many of the same issues that emerged from the Heinz Center study are relevant to the other coastal and Great Lakes States. Little work to date has been aimed at determining the prediction error and uncertainty associated with each of the various methodologies. With FEMA authorized to develop a plan to map erosion hazards under the NFIP, the need for critical review of mapping methods by the scientific and management communities is greater than ever before. A novel approach for determining the error associated with several common methods was applied to data from Delaware and New York. The best erosion forecasts were those based on rates derived from linear regression of 19th and 20th century shoreline positions, excluding shorelines surveyed immediately after major storms.