欢迎访问《生态学杂志》官方网站,今天是

生态学杂志 ›› 2026, Vol. 45 ›› Issue (11): 3774-3782.doi: 10.13292/j.1000-4890.202611.003

• 研究报告 • 上一篇    下一篇

海南岛增强植被指数时空变化特征及其气象影响因子

罗琪1,2,3*,刘少军1,2,3,韩静1,2,3,蔡大鑫1,2,3,赵婷1,2,3   

  1. (1海南省气象科学研究所, 海口 570203; 2中国气象局三沙海洋气象野外科学试验基地, 海南三沙 573199; 3海南省南海海洋气象野外科学观测研究站, 海南三沙 573199)
  • 出版日期:2026-11-10 发布日期:2026-11-18

Spatiotemporal variation characteristics of enhanced vegetation index and its meteorological influencing factors on Hainan Island.

LUO Qi1,2,3*, LIU Shaojun1,2,3, HAN Jing1,2,3, CAI Daxin1,2,3, ZHAO Ting1,2,3   

  1. (1Hainan Institute of Meteorological Sciences, Haikou 570203, China; 2Sansha Marine Meteorological Field Experiment Station of CMA, Sansha 573199, Hainan, China; 3South China Sea Marine Meteorology Hainan Observation and Research Station, Sansha 573199, Hainan, China).

  • Online:2026-11-10 Published:2026-11-18

摘要: 植被是体现区域生态环境演变的关键指标,开展海南岛植被覆盖时空演变规律研究,并分析气象因子对植被覆盖变化的相对贡献,能够为海南岛热带生态系统保护、修复提供科学支撑。本研究基于2000—2024年MODIS-EVI数据和ERA5-Land再分析数据,利用一元线性回归、Pearson相关系数、R/S分析、XGBoost-SHAP可解释性机器学习模型,分析海南岛增强植被指数(EVI)的时空变化和分布特征,并讨论了多种气象驱动因子对海南岛EVI的影响。结果表明:海南岛EVI总体呈上升趋势且表现出明显季节性波动,在年内呈现单峰型分布。海南岛EVI空间分布特征呈现出“中部高、四周低”的分布格局,高值区集中于海南岛中部和南部山区,EVI均值大于0.48,中值区环绕于高值区外围,均值在0.42~0.48,北部沿海为EVI低值区,均值为0.36~0.40。露点温度是影响海南岛EVI最重要的因素,其次是气温、光合有效辐射,地表温度、太阳总辐射等要素对EVI有稳定但非决定性的贡献,风速的贡献最小。从变化趋势看,海南岛EVI将以较慢速度持续增加,Hurst指数和ARIMA模型也预测海南岛EVI月变化呈现较慢的增加趋势。


关键词: 海南岛, EVI, 时空分布, 影响因素, SHAP分析

Abstract: Vegetation serves as a critical indicator of regional eco-environmental variations. Investigating the spatiotemporal variation of vegetation coverage on Hainan Island and quantifying the relative contributions of meteorological factors to vegetation dynamics can provide scientific references for the conservation and restoration of tropical ecosystems on Hainan Island. Based on MODIS-EVI datasets and ERA5-Land reanalysis products covering 2000 to 2024, we adopted univariate linear regression, Pearson correlation coefficient, R/S analysis, and XGBoost-SHAP interpretable machine learning framework to explore the spatiotemporal variation and distribution patterns of the Enhanced Vegetation Index (EVI) on Hainan Island. Furthermore, the effects of multiple meteorological factors on EVI dynamics were elucidated. The results showed that the EVI of Hainan Island presented an overall increasing trend with distinct seasonal fluctuations, exhibiting a single-peak intra-annual pattern. Spatially, EVI displayed a pattern of “high values in central regions and low values in peripheral areas”. High-EVI zones were concentrated in the mountainous areas of central and southern Hainan Island with average EVI  exceeding 0.48. Moderate-EVI regions encircled the high-value zones, with mean values ranging from 0.42 to 0.48. The northern coastal areas were low-EVI regions, with the average EVI between 0.36 and 0.40. Dew point temperature was the dominant factor controlling EVI on Hainan Island, followed by air temperature and photosynthetically active radiation. Land surface temperature and global solar radiation exerted stable but non-deterministic contributions to EVI, whereas wind speed had the minimal impact. Trend analysis indicated that the EVI on Hainan Island would continue to increase at a relatively slow pace, a prediction further corroborated by both the Hurst index and the ARIMA model, which forecasted a gradual upward trend in the monthly EVI.


Key words: Hainan Island, Enhanced Vegetation Index (EVI), spatiotemporal distribution, influencing factor, Shapley additive explanation