Exploring the spatiotemporal variation of carbon storage on Hainan Island and its driving factors: Insights from InVEST, FLUS models, and machine learning

Abstract

Land use/cover changes (LUCCs) significantly influence the carbon cycle. However, as an important conservation area in China, the impact of the LUCCs on carbon storage (CS) in Hainan Island has not been systematically studied. This study integrates the InVEST and FLUS models to analyze the historical and future changes in CS, and uses machine learning algorithms to explore the driving forces behind the spatial heterogeneity of CS. The main results are as follows: (1) The CS distribution on Hainan Island shows a spatial pattern of higher levels in the central mountainous areas and lower levels along the coast. From 1990 to 2020, CS decreased by 2.28 × 10⁶ t, primarily in coastal regions. (2) Natural factors, such as elevation and normalized difference vegetation index, play a decisive role in the spatial heterogeneity of CS, while anthropogenic factors, such as population density and gross domestic product, also have a significant impact on CS. (3) According to predictions for three scenarios (natural development, rapid development, and ecological protection), CS is expected to decrease by 3.11 × 10⁶ t and 4.06 × 10⁶ t in the natural development and rapid development scenarios, respectively, by 2050. However, under the ecological protection scenario, the decline in CS is effectively controlled, with a decrease of only 0.27 × 10⁶ t. This study combines the InVEST, FLUS, and CatBoost models for an in-depth analysis of the spatiotemporal variations in CS and the underlying driving mechanisms. “It offers a novel framework for carbon management in tropical islands and other similar regions. Based on these findings, we recommend strengthening ecological protection on Hainan Island, limiting the unchecked expansion of urban areas, and striking a balance between economic development and ecological conservation to support the achievement of carbon neutrality goals.