Aboveground biomass (AGB) is a key indicator of the physiological status and productivity of grasslands, and its accurate estimation is essential for understanding regional carbon cycles. In this study, we developed a suitable AGB model for grasslands in Xinjiang based on the random forest algorithm, using AGB observation data, remote sensing vegetation indices, and meteorological data. We estimated the grassland AGB from 2000 to 2022, analyzed its spatiotemporal changes, and explored its response to climatic factors. The results showed that (1) the model was reliable (R2 = 0.55, RMSE = 64.33 g·m-2) and accurately estimated the AGB of grassland in Xinjiang; (2) the spatial distribution of grassland AGB in Xinjiang showed high levels in the northwest and low values in the southeast. AGB showed a growing trend in most areas, with a share of 61.19%. Among these areas, lowland meadows showed the fastest growth, with an average annual increment of 0.65 g·m-2·a-1; and (3) Xinjiang's climate exhibited characteristics of warm humidification, and grassland AGB showed a higher correlation with precipitation than temperature. Developing remote sensing models based on random forest algorithms proves an effective approach for estimating AGB, providing fundamental data for maintaining the balance between grass and livestock and for the sustainable use and conservation of grassland resources in Xinjiang, China.
Keywords: aboveground biomass; climatic factors; grassland type; random forest.