Driving mechanisms of understory herbaceous biomass and forest-grass restoration strategies in riparian silvopastoral systems of arid regions

Abstract

Riparian forests in arid regions interact with human activities, resulting in unique silvopastoral systems. However, high-intensity grazing and environmental pressures have accelerated their degradation, posing threats to both pastoral livelihoods and ecological security. This study is based on field surveys conducted on 207 silvopastoral plots in arid hinterlands from 2022 to 2024. It employs explainable machine learning and causal inference to analyze the drivers of understory herbaceous biomass in both single-species (Sin-sil) and mixed-species (Mix-sil) riparian silvopastoral systems. The optimal model revealed that environmental conditions exert dominant influences on both systems (Sin-sil system: 42.077%; Mix-sil system: 60.251%), followed by stand structure (35.519%; 29.440%) and understory herbs (22.404%; 10.309%). Causal analysis further elucidates the mechanisms behind herbaceous biomass variation: the Sin-sil system exhibits relatively superior environmental conditions with competitive relationships between trees and herbaceous plants, while the Mix-sil system demonstrates poorer environmental conditions with mutual feedback between understory herbaceous vegetation and the local climate. Subsequently, a multi-objective optimization model was employed to synergistically enhance both herbaceous and tree biomass. The results demonstrate significant differences in restoration potential among various watersheds and types of silvopastoral systems. For the Sin-sil system, collaborative restoration strategies primarily include measures such as increasing herbaceous height, enhancing herbaceous density, and raising the proportion of Bassia scoparia. For the Mix-sil system, efforts should focus on improving the density of herbs. These findings indicate that implementing intensive management practices in silvopastoral systems is an effective strategy for promoting the synergistic restoration of forest and grass in arid regions.