Validating the use of landsat Imagery to Automatically Map Shorelines and an Applied Case Study for Sergipe, Brazil
Josh Kelly
SDSU Ph.D. Candidate
Host: Allen Gontz
Wednesday, April 19, 2017
CSL 422 – 1 pm
Abstract
Satellite remote sensing has been used extensively in shoreline studies and validated using aerial photography. This ground truth method only represents an instantaneous depiction of the shoreline at the time of acquisition and doesn’t account for the spatiotemporal variability of the dynamic shoreline boundary. Landsat’s capability to accurately delineate a shoreline is assessed by comparing all Landsat water index-derived shorelines with a GPS-surveyed neap intertidal zone that coincides with the satellite flyover date. The shoreline boundary is best defined as the area above low water and below high (intertidal zone) and is at its narrowest during the monthly neap tide, thus providing the most rigorous assessment of Landsat’s ability to accurately map the shoreline. Calculating the proportion of the shoreline that was located within the intertidal zone assessed the performance of each water index. The newly-developed WI2015 index delineated the most accurate shoreline as 94.8% of its interpreted shoreline fell within the neap intertidal zone. As a result of this analysis, WI2015 was used to map the modern and historical shorelines of Sergipe, Brazil using Landsat data dating back to 1984. Shoreline change analysis has revealed significant erosion of beaches and landforms likely due to human interventions such as pier and dam construction. This investigation has proven the ability of Landsat to automatically and accurately delineate the shoreline boundary with the proper water index and the case study shown here is an example of how valuable the Landsat data repository can be for multi-spatiotemporal shoreline assessments.