Professor Scott Steinschneider - firstname.lastname@example.org
Flood risk assessments
In this project, students will conduct work to support the development of a quantitative flood risk assessment tool for shoreline communities along Lake Ontario that is part of an ongoing collaboration between Cornell University and New York Sea Grant. This project will consist of two primary objectives: 1) compare FEMA flood hazard maps to those developed by the Cornell tool to help determine how these approaches differ in their quantification of flood risk; and 2) develop a statistical model that can help predict areas of the shoreline where flood risk estimated by these tools most likely diverges.
|Course number||Course title|
|BEE 4110/6110||Hydrologic engineering in a changing climate|
|BEE 4310/6310||Multivariate statistics for environmental applications|
|BEE 4730||Watershed engineering|
|CEE 4110||Applied remote sensing and GIS for environmental resource inventory and analysis|
|CEE 4350||Coastal engineering|
|CEE 5970||Risk analysis and management|
|CEE 6200||Water resources systems engineering|
|CRP 5080||Introduction to GIS for planners|
Over the last 3 years, shoreline communities on Lake Ontario have experienced two major flood events. In June 2017, water levels reached 75.88 m, the highest in the 100-year record. This record was again broken in May 2019, when water levels reached 75.92 m. These floods occurred soon after the establishment of a new lake level management plan (Plan 2014) that influences water levels on the lake through releases at the Moses Saunders dam on the St. Lawrence River. These floods also followed record setting precipitation in the Great Lakes basin.
Shoreline communities blame the flooding on new operations under Plan 2014, while the international board that manages the lake argues that the flooding was caused by unprecedented precipitation and would have occurred under the previous management plan. Regardless of the cause, the recent flooding has introduced a heightened urgency among municipal, county, and state officials to better prepare shoreline communities for an evolving and uncertain flood regime.
In parallel to the above developments, FEMA has been updating the flood hazard maps used to establish floodplains along the lakeshore and set flood insurance requirements for new construction. FEMA’s Great Lakes Coastal Flood Study uses state-of-the-art hydrodynamic modeling to integrate water level, storm surge, and wave runup processes into aggregate measures of flood risk. However, it is unclear whether and how the updated FEMA maps have accounted for shifts in flood risk under new lake level management or recent changes in basin-scale climate.
In this project, M.Eng students will conduct work to support the development of a quantitative flood risk assessment tool for shoreline communities that is part of an ongoing collaboration between Cornell University and New York Sea Grant. This project will consist of two primary objectives: 1) compare FEMA flood hazard maps to those developed by the Cornell tool to help determine how these approaches differ in their quantification of flood risk; and 2) develop a statistical model that can help predict areas of the shoreline where flood risk estimated by these tools most likely diverges.
For Objective 1, students will collect a variety of existing data, including flood elevations for individual shoreline parcels estimated using legacy and updated FEMA mapping products, as well as the Cornell flood risk tool that accounts for water level variability under the updated lake level management plan. Students will also collect key attributes of these parcels, including nearshore and backshore slope, shoreline type, and the presence of shoreline protection structures. Students will use ArcGIS to determine the differences in key flood hazard elevations and will diagnose major discrepancies between the products.
This comparison will form the basis for Objective 2, in which students will use data analytics in the R statistical software environment to develop a model that can predict major discrepancies between the different flood risk products. Using standard statistical regression models and machine learning algorithms (e.g., regression trees), students will determine which shoreline attributes are most associated with differences in parcel-level estimates of flood hazard. This model will then be used to predict discrepancies between the FEMA and Cornell approaches at locations along the shoreline, including those without recent FEMA data.
Students will document data collection, model development, and prediction accuracy in a final project report. The results of this analysis will help NYSG understand the degree to which the Cornell tool can be used for screening level assessments of flood risk, particularly under different scenarios of water level management and climate change that are not accounted for in the FEMA assessments.