Can habitat surrogates predict the response of target species to landscape change?

Surrogates are commonly used for monitoring biodiversity under a wide range of scenarios. However, surrogates are not often evaluated under diverse ecological conditions, and this hinders the identification of spatial and temporal boundaries of a surrogate's relationship with its biodiversity metric, including whether a surrogate can predict biodiversity responses to environmental change. We adapted a causal framework from the medical literature and applied this framework to investigate the consistency of a well-established habitat surrogate of arboreal marsupials: hollow-bearing trees. We tested the consistency of the relationship between hollow-bearing trees and arboreal marsupials across four long-term studies (>10 years) covering different habitat types and environmental disturbance. We also tested the ability of the change in hollow-bearing trees over time to predict the change in arboreal marsupials over time. We found a somewhat consistent relationship between hollow-bearing trees and relative abundance of arboreal marsupials, but the mechanistic details of this relationship varied both by location and by species of arboreal marsupial. Similarly, the surrogate approach was not able to predict trends over time, a result likely due to differences in natural temporal variation between the surrogate and target species of interest. Our investigation demonstrates that habitat surrogates can be very useful for certain aspects of monitoring programs, but that serious limitations prevail when trying to monitor changes over time, or if information on species-specific responses is required. Our new framework can be readily applied to any biodiversity surrogate with an established mechanistic link between the surrogate and biodiversity metric of interest.