Estimation of grassland aboveground biomass in northern China based on topography-climate-remote sensing data

Accurately assessing grassland aboveground biomass (AGB) over large areas is essential for the conservation of grassland resource and the improvement of ecological environment. In this study, the models for estimating grassland AGB were structured using Google Earth Engine (GEE), based on the large-scale and long-term measurement data of grassland AGB in northern China, combined with remote sensing, topography and climate data. The distribution pattern of grassland AGB was produced, and the spatial -temporal evolution pattern and future change trend of grassland AGB were analyzed. The results show that: (1) the random forest (RF) model, incorporating topography, climate and remote sensing data, exhibited good simulation accuracy (training dataset: R 2 = 0.73, RMSE = 44.85 g/m 2 , MAE = 30.92 g/m 2 ; testing dataset: R 2 = 0.64, RMSE = 57.54 g/m 2 , MAE = 44.74 g/m 2 ). (2) the grassland AGB in northern China increased spatially from west to east, and the annual average was between 16.02 -272.84 g/m 2 . The grassland AGB showed a fluctuating increasing trend at a rate of 0.59 g/m 2 / a from 2000 to 2020, and the coefficient of variation of the grassland AGB was small, ranging from 0.00 to 0.69, which was in a more stable state. (3) the grassland AGB in northern China generally exhibited an upward trend before 2020 (67.99 %). However, the upward trend may weaken after 2020 (38.95 %). The research results provide a quantitative perspective for studies on grassland livestock carrying capacity and climate change in northern China.