Automatic super-resolution shoreline change monitoring using Landsat archival data: a case study at Narrabeen–Collaroy Beach, Australia

Abstract

Long-term monitoring of shoreline changes is of significant importance for coastal erosion prediction and coastal planning. This paper presents the use of the Landsat archival dataset over 29 years to monitor shoreline changes at Narrabeen–Collaroy Beach, Australia. For each scene of Landsat data, a vector-based shoreline at subpixel level was automatically extracted by employing the super-resolution border segmentation method. Monthly cross-shore beach profile surveys at five locations along the beach are used as ground truth data to assess the results. The experimental results show that after the correction of tidal effects the RMSEs of the time-series shorelines at the five locations are all less than 10 m; RMSEs of annual mean shorelines are within 5.7 m. These results were achieved within the context that the Narrabeen–Collaroy Beach has been observed to vary in width by up to 80 m during this same period. In addition, the observed variabilities in shoreline change at different locations along the 3.6-km long embayment are also clearly observed from Landsat-derived results. The results indicate the feasibility of Landsat data for long-term shoreline change monitoring.