Constructing a Bayesian belief network to provide insights into the dynamic drivers of ecosystem service relationships

Abstract

Vegetation restoration has induced complex spatiotemporal changes in ecosystem service (ES) relationships. However, dynamic analysis on the driving mechanisms underlying ES relationships remains insufficient, constraining the comprehension on sustainable vegetation restoration. In this study, a comprehensive index (ESI) representing the coordination among four ESs (carbon sequestration, water yield, baseflow regulation, and soil conservation) was quantified in Yan’an City, a typical vegetation restoration area. Using the Bayesian belief network, a social-ecological framework was constructed to explore changes in driving factors of ESI during 1990 to 2020. The results revealed that: (1) the coordination of ES relationships showed a significant enhancement over the past 30 years, which primarily attributed to areas with the highest values of water yield. Areas with low values of baseflow regulation posed a risk to the ESI. (2) The combined contribution of FVC and precipitation was consistently close to 70 %, emerging as the dominant driving factors for the ESI. But in the last decade, contributions from human activities have increased significantly. (3) Scenario reasoning highlighted that expanding vegetation cover alone would exhibit threshold effects on the ESI improvement. Additionally, the results revealed a promising prospect of enhancing baseflow and maintaining the ESI through human actions. This study provides an example of using Bayesian belief network to analyze social-ecological interactions, enhancing the understanding of changes in ES relationships under vegetation restoration.

Keywords

• Ecosystem services
• Coupling relationships
• Driving forces
• Bayesian belief network
• Vegetation restoration