Detection of Coastline Changes using Multi-Temporal Satellite Images: A case study of Gianyar and Klungkung Regency, Bali

Abstract

Coastal areas are one of the essential regions for the ecosystem. Consequently, monitoring and detecting the changes of coastal areas is thus necessary in order to provide valuable information for future mitigation and adaptation. Recently, the utilization of satellite images for coastal areas has been used widely and effectively as a viable alternative to in-situ observation and monitoring. In this study, multi-temporal satellite imagery data from Sentinel-2 have been used to find out the changes in the coastal area of Gianyar and Klungkung Regency - Bali, considering the capability which has good resolution with a wide swath and high revisit time. Sentinel-2 imagery data taken from 2016 and 2019 at an interval of about 4 years were collected to determine coastline changes. There are two main steps to detect the coastline. First is image classification by supervised classification where each pixel has been labeled with one or three classes. And the second is to make border segmentation between sand and water by applying the Modified Normalized Difference Water Index (MNDWI). The areas that are experiencing erosion and accretion were determined by using the overlay method between two satellite images. Results show that about 9,248 m2 of coastline were eroded in Gianyar Regency and 121,000 m2 in Klungkung Regency for the period of 2016 to 2019. Meanwhile, the accretion area reached 158,615 m2 in Gianyar and 32,159 m2 in Klungkung Regency.