I’m presenting at Coastal GeoTools 2025:
“Determining Shoreline Migration Rates with a Change-Point Detection Model” David Forrest. William & Mary’s Batten School of Coastal & Marine Science and the Virginia Institute of Marine Science.
Large-spatial-scale shoreline change data typically has limited observations over long time scales. The Shoreline Studies Program at VIMS has a rich collection of coastlines spanning from 1937 to 2021 along the Virginia portion of Chesapeake Bay. End point and least squares regression rates have been used in the past to describe change, but these do not always capture significant modifications along the shore nor the recent acceleration in the rates due to sea-level rise. Endpoint estimates are controlled wholly by their end dates, ignoring intervening rate changes. Least squares models are more inclusive of intervening data but do not represent structural changes or slope changes over time. The tool under development uses interrupted time-series analyses to model shore change through time by determining change points between early and later rates of change; it then estimates the timing and extent of the change points. This piecewise regression technique is a machine learning method for fitting separate models to partial, covering sets of observations and to determine the relationship of those models. In contrast with endpoint and least squares estimates, piecewise linear models can represent some level of curvature or change and provide estimates of the magnitudes of the changes. Additionally, extrapolations from piecewise linear models are weighted more towards their time-local observations, and their effects can be less dramatic than extrapolations made by higher-order polynomial models. We developed this analysis in Python and applied it to a series of digitized shoreline data for York River and Mobjack Bay in Chesapeake Bay. A detailed understanding of how and why these rates are changing in varying environments (marsh, upland, residential, agricultural, protected, etc.) is critical to protecting habits and increasing overall coastal resiliency.
This work uses VIMS Shoreline Studies Program’s collection of Chesapeake Bay Shorelines (viewer) to determine changes in migration rates of the Gloucester County, VA shorelines.
Leave a Reply