Incorporating between-expert uncertainty into state-and-transition simulation models for forest restoration

Abstract

Anthropogenic utilization of forests alters vegetation structure and can negatively affect flora and fauna. Ecological simulation models assist in restoring such degraded forests by predicting the effects of alternative management and restoration actions on vegetation structure. However, the lack of empirical ecological data to parameterize models often necessitates the use of expert opinion. Differences in opinion between experts may be large and lead to uncertainty in model predictions, but this is rarely acknowledged in forest simulation models. In this paper, we constructed state-and-transition simulation models (STSMs) based on expert opinion to predict whether management will expedite the development of stands with large, old trees – a habitat resource considered critical to forest-dependent biodiversity – in the Box-Ironbark forests and woodlands of Victoria, Australia. Three candidate management actions are modified timber harvesting regulations, establishing no-cutting National Parks, and a new ecological thinning prescription. We also investigated the importance of uncertainty caused by differences between experts when constructing STSMs. Model results predict that mature woodlands with large trees will likely not develop sooner with either ecological thinning or modified harvesting than natural development alone, and may lead to a loss of desired vegetation. In addition, differences in model predictions caused by between-expert variation are substantial and can be used to identify biological processes requiring better understanding in this ecosystem.