Abstract: Vegetation net primary productivity (NPP) is an important indicator for evaluating the responses of ecosystems to climate change and anthropogenic activities. The monthly NPP datasets (250 m resolution) of Hunan Province from 2000 to 2019 were generated through combining the Carnegie-Ames-Stanford Approach (CASA) model and a series of spatial datasets such as the MODIS monthly vegetation index product, TerraClimate datasets, and land-cover map (i.e., ChinaCover2010). The NPP trend and its relationships with climatic factors at the watershed scale were investigated using the Mann-Kendall test and multiple regression model. The results showed that annual mean NPP of Hunan Province was 566.92 g C·m^-2·a^-1 between 2000 and 2019. The headwater of Xiangjiang River basin had the highest value of annual mean NPP of 625.28 g C·m^-2·a^-1, while the Dongting Lake basin had the lowest NPP value of 492.11 g C·m^-2·a^-1. The annual NPP fluctuations of the five watersheds were generally similar, with an increasing trend before 2008 and then a shortly but dramatically decreasing trend in 2009 with slowly NPP recovery since 2010. About 38% area exhibited a monotonic increasing trend and about 10% area a monotonic decreasing trend, which was mainly distributed in the Dongting Lake basin and the “Changsha-Zhuzhou-Xiangtan” urban agglomeration, respectively. The relative importance of the meteorological factors varied among the five watersheds, with an overall ranking of vapor pressure (0.331), temperature (0.318), solar radiation (0.299), and precipitation (0.062). The correlation between precipitation and NPP was not statistically significant. Our results demonstrated that extreme climate events (e.g., snow disasters, floods) and anthropogenic activities (e.g., land development, returning cropland to the lake) decreased vegetation NPP of Hunan Province. It is necessary to distinguish the differences in the ecological effects of such NPP changes when managing ecosystems in practice.

Key words: CASA model, net primary productivity, Mann-Kendall test, threshold effect, Hunan Province.