[目的] 预测未来土地利用/覆盖变化(land use and land cover change，LULCC)及其对生态系统碳储量的影响,为区域土地利用决策和碳管理提供科学依据。[方法] 基于30 m分辨率的海南岛1990，2000，2010，2020年土地利用遥感解译数据，运用ArcGIS与InVEST模型，探究土地利用时空演变及碳储量响应状况，并引入GeoSOS-FLUS模型预测研究区2030年土地利用多情景变化特征及其对未来不同情景下生态系统碳储量的影响机制。[结果] ①1990—2020年研究区耕地、林地、草地和未利用地面积减少，水域和建设用地面积增加。未利用地和耕地面积持续减少，建设用地面积持续增加。②30 a间LULCC导致区域碳储量持续减少，达到1.50×106 t且年变化率为5.00×104 t/a。建设用地的大肆扩张及林地退化是导致碳储量下降的重要原因，“未利用地→草地”为碳储量增加中最明显的图谱变化，“草地→林地(人工林地)”是碳储量减少中最显著的图谱变化。③2020—2030年的3种预测情景中，林地仅在生态优先情景下得到了有效保护，且面积增加了11.91 km2。建设用地在3种预测情景中均呈现不同程度扩张态势，且发展优先情景涨幅最大。[结论] 海南岛大面积高碳密度的天然草地转换为低碳密度的人工林地，高碳区转变为低碳区，区域固碳能力削弱。应采取提高林地、草地等地类比重等一系列的土地利用调控政策，加大区域碳源向碳汇转换的优化发展。
[Objective] The future land use and land cover change (LULCC) and its impact on ecosystem carbon storage were predicted in order to provide a scientific basis for regional land use decision-making and carbon management. [Methods] Based on the quantitative analysis of 30 m resolution remote sensing images of Hainan Island in 1990, 2000, 2010, and 2020, the spatio-temporal evolution of land use and its impact on carbon storage were analyzed. Combined with geographic information system (GIS) spatial superposition analysis, the spatial patterns and mechanisms of land use changes in coastal zones and the impacts of complex land use changes on tropical and subtropical coastal zones were assessed. The InVEST model was used to estimate carbon storage in Hainan Island’s coastal zone. The estimation was further linked to the GeoSOS-FLUS model to predict the distribution characteristics of multi-scenario land use changes in the study area in 2030, and to analyze the influence mechanism on ecosystem carbon storage under different future scenarios. [Results] ① From 1990 to 2020, the areas of cropland, forest land, grassland, and unused land in the study area decreased, while the areas of water and built-up land increased. Among them, the areas of unused land and cropland consistently decreased every year, while the area of built-up land increased year by year. ② Land use/cover change led to a continuous decrease in carbon storage for the study area, reaching 1.50×106 t with a decline rate of 5.00×104 t/a over the past 30 years. The increase in built-up land and the decline in forestland were the key causes for the decrease in carbon storage. The increase in carbon storage due to the transformation from unused land to grassland was the most significant change, with an increase of 1.25×105 t. On the other hand, the change from grassland to forest land (artificial forest land) was the most prominent cause for the decrease in carbon storage, accounting for a decrease of 5.68×105 t. ③ Predicted carbon storage in the Hainan Island coastal zone based on the FLUS model was evaluated. The accuracy met the research requirements verified by historical data. Among the three prediction scenarios, forestland will be effectively protected by 2030 only under the ecological priority scenario with an increase in area of 11.91 km2. Built-up land will increase to different degrees under the three scenarios, and the development priority scenario had the largest increase. [Conclusion] A large area of natural grassland with high carbon density on Hainan Island has transformed into artificial forest land with low carbon density, and the high carbon area has transformed into a low carbon area, which weakened the regional carbon sequestration ability. A series of land use regulation policies should be adopted to increase the proportions of woodland, grassland, and other land types, and to optimize the transformation of regional carbon sources into carbon sinks.