Combining financial costs and statistical power to optimize monitoring to detect recoveries of species after megafire

Abstract

AbstractAimMegafire plays a crucial role in driving the distribution of biodiversity around the world. Long‐term monitoring is vital for understanding how species are impacted immediately by megafire and subsequently respond over time. However, monitoring should be designed with sufficient statistical power to detect impact and recovery. In this study, we developed a simulation framework for optimizing the design of biodiversity monitoring programmes to detect population recoveries after megafire.LocationVictoria, Australia.Time period2019–2020.Major taxa studiedVertebrates.MethodsWe collated species distribution models for 45 priority vertebrates most likely to respond to management after the 2019–2020 megafires in Victoria, Australia. We combined these models with fire severity maps to optimize the location of monitoring sites in and around the fire footprint. We simulated the impact of the megafires on species distributions and modelled plausible recoveries over the next 10 years. Using estimates of detectability for a suite of preferred sampling methods, we simulated monitoring at pairs of burnt and unburnt sites to evaluate the statistical power to detect the modelled recoveries. We tested the sensitivity of power to alternative monitoring designs, rates of recovery and monitoring budgets.ResultsPriority regions to establish monitoring sites varied by taxonomic group. Power to detect population recoveries increased as the monitoring budget increased, as the recovery rate increased and when the proportion of sites in burnt compared with unburnt habitat increased. According to the optimal monitoring design, an AUD $9M budget could detect 90% of recoveries to pre‐fire levels in 40% of species with >80% power. Power was highest for mammals, followed by birds, reptiles and amphibians.Main conclusionsOur simulation approach allowed us to test the relative performance of alternative post‐fire monitoring designs ahead of time. Although we focused on megafire, our approach could easily be applied to detect population recoveries after any large‐scale catastrophic disturbance.