Abstract
Extreme environmental disturbances such as fires are predicted to increase in severity and extent with climate change. To better understand ecological impacts of unprecedented scales of disturbance, empirical monitoring data are urgently needed. Broadscale passive networks of camera traps or acoustic recorders are an increasingly popular approach for monitoring. Australia's 2019–2020 Black Summer megafires burned over 8 million hectares, and a network of camera traps across eastern New South Wales provided a Before-After-Control-Impact opportunity to examine impacts of this event on biodiversity. We evaluated impacts of severe fire on mammals using dynamic occupancy models to characterise the colonisation of unoccupied sites, and extinction of occupied sites between 2018 and 2021. We used Generalized Linear Mixed Models to identify species traits correlated with detections, predicting widely distributed species with broadest habitat breadths would be most frequently detected, and threatened species would be least frequently detected. Our analyses provided limited insight into impacts of the fires. Forty-two mammal species were detected as 'definite' but approximately 91 % of records were attributed to just eight species. Extinction or colonisation trends were detected for four species, one was an extinction trend. Endangered species were less likely to be detected than Vulnerable or Least Concern species. Decreasing costs for sensors and artificial intelligence will increasingly encourage passive monitoring networks. However, our results caution that large volumes of sensor data will not necessarily overcome many shortcomings of passive monitoring. Bespoke designs, progressive analyses, and method refinement will remain important to ensure the greatest value can be derived from sensor network data.
Original language | English |
---|---|
Article number | 110378 |
Number of pages | 12 |
Journal | Biological Conservation |
Volume | 289 |
DOIs | |
Publication status | Published - Jan 2024 |