Analysis of the Evolution of Mangrove Landscape Patterns and Their Drivers in Hainan Island from 2000 to 2020

Abstract

The ecological services provided by mangroves are of great value and significance in regard of the achievement of the major strategic goals of carbon neutrality and the carbon peak. Here, we first evaluated the uses of five machine learning methods combined with Sentinal-2A data band features to identify and extract mangrove forests in Dongzhai Harbor, northeast Hainan Province, China. Then, the XGBoost algorithm with the highest accuracy was selected to identify and extract information on mangrove forests on Hainan Island, focusing on five periods from 2000 to 2020. The landscape pattern index, dynamic attitude model, and mathematical statistics were integrated to analyze trends over this 20-year period. The results revealed the following: (1) The total mangrove landscape area of Hainan Island between 2000 and 2020 showed a trend of first decreasing and then increasing. In the past 20 years, the mangrove area has increased by 1315.75 ha, with an annual change rate of 65.79 ha/a. (2) From 2000 to 2020, the mangroves in Hainan Island were characterized by increased fragmentation, increased heterogeneity, decreased connectivity, and increased richness, while proportion of each landscape type tends to be equilibrated. (3) Natural factors such as the annual average temperature in the study area were the main factors driving the large-scale reduction in mangroves and the deepening of landscape fragmentation, followed by human factors, and the impact of macro-policies cannot be ignored. The results of this study can provide a significant reference for future remote sensing data extraction from mangrove forests and their ecological protection and restoration on Hainan Island.