Spatio-temporal dynamics of mangrove extent in coastal Bekasi Regency, Indonesia

Abstract

One of West Java’s largest mangroves is found alongside the shore of Bekasi Regency. Nevertheless, the primary issues are the high degree of waste contamination, extraction of mangrove products, and conversion of mangrove forests into aquaculture ponds. To determine the present state of mangroves on the coast of Bekasi Regency, research on the dynamics of the mangrove extent spatio-temporally must be conducted. Mangrove mapping has made extensive use of machine learning and satellite images. This study aims to calculate the mangrove area in the villages of Pantaibahagia and Pantaibakti along the coast of Bekasi Regency using Random Forest (RF) classification based on the Google Earth Engine (GEE) Platform. The RF classification results showed a significant loss in mangrove extent over a short period of time (seven years). In 2017, the tidal zone was primarily covered by mangroves. From the linear trend line, it is known that from to 2017-2023, the mangrove area tended to decrease, while in 2023, there was a decrease in the mangrove area, especially in the north coast area. In 2017, the total mangrove area was 305.03 ha. Until 2019, there has been a slight increment, reaching 366.41 ha of mangrove are. The most significant loss was found in 2020, in which the total loss reached 85.58 ha in one year. After 2020, the mangrove area has slightly improved, but it is not significant. We tested the produced map of the RF classification using a confusion matrix and kappa coefficient, which resulted in an Overall Accuracy (OA) of 90.50% and a Kappa coefficient of 0.8105.