Integrated Shoreline Dynamics Assessment Using QSCAT and Random Forest Along the Phetchaburi Coastline, Thailand

Abstract

The Phetchaburi coastline faces significant erosion risks, exacerbated by human activities and climate change. This research evaluates shoreline shifts and mangrove cover modifications along the 88.5 km coastline, utilizing the QGIS Shoreline Change Assessment Tool (QSCAT) plugin and Random Forest (RF) classification on Landsat 8-9 OLI/TIRS images from 2017 to 2023. The integration of these methods provides a robust framework for monitoring coastal changes in a data-scarce, microtidal environment. Results indicate a highly dynamic coastline with a maximum erosion distance of -108.71 m and a maximum accretion of +122.95 m. Statistical analysis of 1,766 transects reveals a critical erosion trend, with a maximum erosion rate reaching -38.66 m/year. Overall, 50.74% of the transects experienced erosion, while 48.98% showed accretion. The RF model demonstrated high effectiveness in land use classification, achieving an overall accuracy of 96.50% and a Kappa coefficient of 0.92. Spatiotemporal analysis reveals a strong link between mangrove degradation and increased erosion zones, emphasizing the protective function of vegetation. These findings provide essential quantitative benchmarks for developing targeted coastal management and restoration strategies in the Gulf of Thailand.