Nonlinear characteristics and driving factors of vegetation-soil moisture feedback at fine scale in Northeast China

Understanding the feedback mechanisms between vegetation and soil moisture (SM) is crucial for elucidating the coupled carbon-water processes of terrestrial ecosystems under climate change. However, existing studies are limited by coarse spatial resolution and an over-reliance on linear methods, thereby hindering the exploration of their nonlinear interactions for highly heterogeneous areas. In this study, we employed high-resolution (1 km) satellite-based datasets and a Random Forest-based nonlinear Granger causality analysis to investigate the feedback relationship between gross primary productivity (GPP) and SM at different depths (0-100 cm) in Northeast China from 2000 to 2022, along with the driving factors influencing their interactions. The results show that both GPP and SM showed a significant increasing trend during the study period, with a dominant pattern of synergistic growth (GPP + SM +), accounting for more than 94%. This proportion of synergistic increase slightly decreased with increasing soil depth. The bidirectional causal relationships were observed in 41.24% to 71.87% of the study area, with the proportion declining as soil depth increased. Among these, the influence of SM on GPP was generally stronger than the feedback from GPP to SM. The mutual feedback exhibited a lagged response of 2-3 months, showing a nonlinear pattern that first decreased and then increased with soil depth. Further analysis revealed that this feedback relationship was jointly regulated by solar radiation, precipitation, and temperature, and varied significantly across different vegetation and soil types.This study reveals the complex interaction mechanisms between vegetation and SM, providing a basis for regional agricultural optimization and ecological restoration.