The development of the Flood and Erosion Prediction System is essentially the culmination of many of the coastal process investigations being carried out in the LMPDS. The objective is to develop a system for the prediction of shoreline recession (or accretion) on a lakewide basis at a 1 kilometer resolution.

Baird have completed development of the core of a GIS linked Flood and Erosion Prediction System. A summary of progress by Baird & Associates and is presented here.

A Flood and Erosion Prediction System has been developed to provide a graphical, geographically referenced database of information required to predict future shoreline position. To date, the system the following key components: ArcView GIS, Nearshore wave transformation, coastal process, flood and erosion predictions (the latter with the COSMOS model). Linkages, Graphical User Interfaces and data analysis requirements have been addressed through the development of tools using Avenue scripting to enhance the ArcView GIS. The system allows users to create "projects" for studying any specified section of shoreline using the built-in data and coastal process modeling tools. Projects for eight detailed study sites around Lake Michigan have been created and the coastal process tools have been used to simulate the historical trends at these sites.

Base data that has been integrated to form part of the system includes:

• shoreline classification data (geomorphic, nearshore subaqueous, and structural classifications);
• "single value" shoreline recession rates;
• more detailed estimates of shoreline recession rates at study sites;
• lakewide digital bathymetry from NOAA;
• base map information and topography for study sites (lakewide pending);
• SHOALS bathymetry for 7 of 8 study sites (Fischer Creek, WI is not available);
• hourly WIS hindcast data for the period 1957-1987 (1987 to 1997 is pending);
• hourly water levels at 8 gauge stations for varying periods;
• weekly ice cover data for the period 1972-1995;
• ground penetrating radar at study sites (pending);
• grain size information at study sites;
• boreholes and stratigraphy at study sites; and
• available dredging records;

Added data analysis functionality has been incorporated into the GIS with the development of automated procedures to:

• extract cross-shore profiles from bathymetric and topographic data;
• conduct surface and volume comparisons of bathymetric data;
• create digital scatter plots of wave and water level data for any reach (both nearshore and offshore information), with the option of omitting periods when shore ice is present; and
• generate historical hourly water level files for any reach based on the water levels measured at the 8 available gauges.

The COSMOS coastal process model has been linked to LMPDS GIS system. The model link provides the ability to:

• automatically create the COSMOS model input file using reach-specific data and profiles;
• automatically create a cohesive profile at each reach based on the shoreline classification (sand cover);
• estimate the longshore transport at each reach which is used in the sediment budget calculations;
• estimate profile change in response to storm conditions;
• estimates long-term bluff recession rates (and confidence intervals on rates); and
• manually edit model parameters to provide flexibile input for modelling "what" if scenarios;

The system as described above has been applied to the study sites to model historical shoreline trends. Analysis completed to date includes:

• surface comparisons of offshore bathymetry for the New Buffalo, Warren Dunes, and Shoreham study sites for the period from 1945 to 1997;
• generation of historical wave/water level/ice data for each reach at each study site;
• estimation of longshore transport rates and shoreline change at each study site using the COSMOS model and developed tools;
• calibration of the newly-revised bluff recession capability of COSMOS using erodibility factors derived from physical model tests (performed during FY98) and recession rates measured from historical aerial photos; and
• development of a sediment budget for each study site.

1999-2000 Activities

A key task in FY99 is to test the ability of FEPS to predict the historic recession rates for the prototype counties. This task is critical to building confidence in the ability of the system to predict future erosion. Baird & Associates has explained that there are two primary sources of error in prediction of recession rates: 1) input data for the shoreline classification scheme being the weakest component; and 2) how processes are predicted in the model.

A Classification and Regression Tree (CART) analysis is being completed to assess the quality and caliber of input data as it relates to recession rates. It is expected that an extension of this CART investigation from the five prototype counties to a lake-wide assessment will not be required as the selected prototype counties comprise a wide range of backshore and nearshore conditions.

FEPS is being applied to predict historic recession rates for each reach in the five prototype counties. Discrepancies between predicted and measured recession rates are being investigated to resolve any problems with the input data and the prediction models. FEPS is also being refined to be able to shift a bluff crest line landward to represent the shore recession for a given prediction period and to then add an uncertainty band about this recession line.

Thus far, Baird has completed several site specific tests of the FEPS system using older bathymetry data sets and have completed lake bed surveys in Wisconsin. They are now awaiting new SHOALS data in order to develop new grids to complete the erosion and longshore transport estimates for all reaches in the prototype counties.

Click to return to the What's New Home Page.