Risk assessment of forest landscape degradation using Bayesian network modeling in the Miyun Reservoir catchment (China) with emphasis on the Beijing–Tianjin sandstorm source control program

Abstract As the local strategic surface water source for Beijing City, the Miyun Reservoir catchment is suffering from the challenges of forest landscape degradation (FLD) under various natural and anthropogenic disturbances. This study developed a cause–effect Bayesian network model, consisting of three tiers of nodes, site indicators, external disturbance, and the outcome of FLD, to perform a regional ecological risk assessment to better understand the FLD risks (1998–2013) in the catchment. The effects of the Beijing–Tianjin sandstorm source control (BTSSC) Phase I program, a typical forest landscape restoration program, on the FLD were further examined. Overall, the higher level of either current occurrence or risk probability for FLD suggests the urgency of the following restoration efforts. The dominance of soil erosion for FLD indicates that soil erosion control should be a priority in the following Phase II program. Meanwhile, the uneven spatial distribution of high‐risk areas suggests that the following efforts should be focused within the upstream Hebei Province. In addition, based on the comparison of site indicators between different risk areas, the study concluded that community livelihood activities and land degradation due to climate change have been the other two dominant driving forces contributing to FLD. Furthermore, the study found that the forest landscape has not been restored into the best state by the BTSSC program due to the multiple and complex challenges. Thus, the technical concept of nature‐based solutions, including close‐to‐nature forest management technology and a comprehensive community‐involvement governance model, was proposed for the following Phase II program.