Spatiotemporal variations and driving forces of regional-scale NPP based on a multi-method integration: a case study in the Beibu Gulf Economic Zone

作者：Lv Zhou, Qiulin Dong, Yuanjin Pan, Fei Yang, Meilin He, Xiang Huang, Jiao Xu · 发表于：Ecological Indicators · 年份：2025 · DOI：10.1016/j.ecolind.2025.113453 · 被引用次数：19 · 研究领域：Climate change impacts on agriculture、Environmental Impact and Sustainability、Water Resources and Sustainability

• A 250-meter spatial resolution NPP dataset was estimated using the CASA model and remote sensing data. • The robust CASA model was applied to regional NPP estimation, and the results remained highly reliable. • GeoDetector and STL decomposition revealed spatiotemporal NPP patterns. Net Primary Productivity (NPP) serves as a critical ecological indicator for evaluating carbon sequestration and overall ecosystem functionality. While widely studied at large scales, its estimation accuracy, spatiotemporal dynamics, and driving factors at regional scales require further exploration. This study employed high-resolution (250 m) remote sensing data and an improved Carnegie-Ames-Stanford Approach (CASA) model to estimate monthly NPP from 2017 to 2021, which was cross-validated against the MOD17A3HGF dataset (a MODIS-derived product widely used for estimating NPP). Subsequently, we applied STL decomposition to detect seasonal trends in NPP, Moran’s I was used to assess spatial autocorrelation, and the Mantel test combined with GeoDetector was employed to identify the primary topographic, meteorological, and anthropogenic drivers. The main findings are as follows: (1) Model Performance: The improved CASA model effectively captured ecological variations at finer scales, demonstrating high consistency and accuracy relative to the MOD17A3HGF dataset. (2) Spatial Distribution: Moran’s I analysis indicated a significant spatial autocorrelation in NPP distribution, displaying an increasing ...