Spatiotemporal evolution and driving forces of long-term aboveground biomass in grasslands of Xinjiang

Abstract

[Objective] Grassland is an essential component of terrestrial ecosystems, and the aboveground biomass (AGB) of grassland can directly reflect the current status of grassland resources. Accurately assessing the aboveground biomass of grassland and revealing its long-term change trends are fundamental to maintaining and enhancing grassland productivity, determining reasonable livestock carrying capacities, and ensuring sustainable grassland utilization. [Methods] Based on the grassland zoning in China, this study used 11703 records of accumulated field survey quadrat data to establish random forest regression models and multiple stepwise linear regression models for aboveground biomass in seven grassland zones of the Xinjiang Uygur Autonomous Region. The optimal model was determined through accuracy assessment. Subsequently, Landsat imageries were employed to invert the results of 30 m resolution grassland aboveground biomass in Xinjiang from 1990 to 2020. After revealing the spatiotemporal change of grassland aboveground biomass over 31 years, 14 potential driving factors were selected from four aspects: meteorology, terrain, soil, and human activities. Geographical detector was then used to analyze the primary driving factors, aiming to provide a scientific evidence for the future management, protection, and sustainable utilization of grassland resources. [Results] The main findings are as follows: (1) The average R² of the random forest regression models for the seven grassland zones was 0.74, with an average RMSE of 786.89 kg/hm², outperforming the multiple stepwise linear regression models. (2) From 1990 to 2020, Xinjiang’s grassland AGB showed an overall increasing trend, with an average AGB of 2137.31 kg/hm², and an annual average change of 15.05 kg/hm²/a. (3) The spatial distribution pattern of AGB in Xinjiang indicated higher values in mountainous areas compared to plains and higher values in northern Xinjiang compared to southern Xinjiang. The Ili River Valley and Altay region had higher AGB, whereas the Junggar Basin and the southeastern Tarim Basin had lower AGB. (4) Geodetector analysis results for three time periods showed that precipitation and soil organic carbon content significantly influenced Xinjiang’s grassland AGB. Notably, the influence of the human footprint factor has intensified during the period from 2010 to 2020. [Conclusion] Xinjiang’s grassland AGB has shown a continuous growth trend from 1990 to 2020, driven by both meteorological and soil factors. In the future, greater attention should be given to the impact of human activities on AGB to ensure the sustainable development of Xinjiang’s grassland ecosystems.

Keywords

• grassland|aboveground biomass|google earth engine|random forest|landsat 8|geodetector|xinjiang uygur autonomous region