Authors:

Maggie McLeod

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

Between 2013 and 2020, Great Lakes shorelines experienced considerable change from a record-setting water level increase. Previous work quantifies shoreline changes and environmental conditions during this period using multispectral satellite imagery combined with water gauge data, topography, and other sources. This study harnesses satellite imagery with machine learning modeling to identify where shoreline has recovered since the 2013-2020 changes and quantify the extent of shoreline rebuilding. Shoreline recovery locations are mapped in Saginaw Bay, Lake Huron, using 3-meter resolution, daily satellite imagery from Planet Labs. Satellite data is downloaded for 2023-2025 as input for a shoreline detection algorithm. Open-source Landsat imagery served as a testbed to ensure the algorithm's compatibility with publicly available data before applying the algorithm to modeling efforts. Horizontal shoreline positions are delineated with this pre-existing algorithm based on a Direct Difference Water Index (DDWI). Next, geoprocessing tools find rates of shoreline change across Saginaw Bay, and use these rates to classify locations as eroding, rebuilding, or stable. This information serves as a basis for a machine learning model that predicts future shoreline response. Predicted shorelines are tested on dates that have already passed and then validated with the forecasted time's actual imagery. The relationship between shoreline response and environmental factors allows for a robust analysis of Saginaw Bay's dynamic change from extreme water level fluctuations. This study provides results on the effectiveness of this model, and guidance on an algorithm that can be harnessed for finer temporal scale analysis, to inform coastal communities of beach response. Keywords: [no keywords provided]

Source:

Purdue University / 2025

Topics:

No topics listed

Co-authors:

Maggie McLeod