文摘
Decisions affecting the management of natural resources in agricultural landscapes are influenced by both social and ecological factors. Models that integrate these factors are likely to better predict the outcomes of natural resource management decisions compared to those that do not take these factors into account. We demonstrate how Bayesian Networks can be used to integrate ecological and social data and expert opinion to model the cost-effectiveness of revegetation activities for restoring biodiversity in agricultural landscapes. We demonstrate our approach with a case-study in grassy woodlands of south-eastern Australia. In our case-study, cost-effectiveness is defined as the improvement in native reptile and beetle species richness achieved per dollar spent on a restoration action. Socio-ecological models predict that weed control, the planting of trees and shrubs, the addition of litter and timber, and the addition of rocks are likely to be the most cost-effective actions for improving reptile and beetle species richness. The cost-effectiveness of restoration actions is lower in remnant and revegetated areas than in cleared areas because of the higher marginal benefits arising from acting in degraded habitats. This result is contingent on having favourable landowner attitudes. Under the best-case landowner demographic scenarios the greatest biodiversity benefits are seen when cleared areas are restored. We find that current restoration investment practices may not be increasing faunal species richness in agricultural landscapes in the most cost-effective way, and that new restoration actions may be necessary. Integrated socio-ecological models support transparent and cost-effective conservation investment decisions. Application of these models highlights the importance of collecting both social and ecological data when attempting to understand and manage socio-ecological systems.