Using ideal distributions of the time since habitat was disturbed to build metrics for evaluating landscape condition

Developing a standardized approach to measuring the state of biodiversity in landscapes undergoing disturbance is crucial for evaluating and comparing change across different systems, assessing ecosystem vulnerability and the impacts of destructive activities, and helping direct species recovery actions. Existing ecosystem metrics of condition fail to acknowledge that a particular community could be in multiple states, and the distribution of states could worsen or improve when impacted by a disturbance process, depending on how far the current landscape distribution of states diverges from pre-anthropogenic impact baseline conditions. We propose a way of rapidly assessing regional-scale condition in ecosystems where the distribution of age classes representing increasing time since last disturbance is suspected to have diverged from an ideal benchmark reference distribution. We develop two metrics that (1) compare the observed mean time since last disturbance with an expected mean and (2) quantify the summed shortfall of vegetation age-class frequencies relative to a reference age-class distribution of time since last disturbance. We demonstrate the condition metrics using two case studies: (1) fire in threatened southwestern Australian proteaceaous mallee-heath and (2) impacts of disturbance (fire and logging) in the critically endangered southeastern Australian mountain ash Eucalyptus regnans forest on the yellow-bellied glider Petaurus australis. We explore the effects of uncertainty in benchmark time since last disturbance, and evaluate metric sensitivity using simulated age-class distributions representing alternative ecosystems. By accounting for and penalizing too-frequent and too-rare disturbances, the summed shortfall metric is more sensitive to change than mean time since last disturbance. We find that mountain ash forest is in much poorer condition (summed shortfall 38.5 out of 100 for a 120-yr benchmark disturbance interval) than indicated merely by loss of extent (84% of vegetation remaining). Proteaceaous mallee-heath is in worse condition than indicated by loss of extent for an upper benchmark interval of 80 yr, but condition almost doubles for the minimum tolerable time since last disturbance interval of 20 yr. To fully describe ecosystem degradation, we recommend that our summed shortfall metric, focused on habitat quality and informed by biologically meaningful baselines, be added to existing condition measures focused on vegetation extent. This will improve evaluation of change in ecosystem states and enhance management of ecosystems in poor condition.