Research
| Title： | Predicting the ecosystem-wide impacts of eradication with limited information using a qualitative modelling approach |
|---|---|
| First author： | Han, Yi; Kristensen, Nadiah P.; Buckley, Yvonne M.; Maple, Dion J.; West, Judith; McDonald-Madden, Eve |
| Journal： | ECOLOGICAL MODELLING |
| Years： | 2020 |
| DOI： | 10.1016/j.ecolmodel.2020.109122 |
| Abstract： | Conservation-motivated eradications may cause unexpected perverse effects, and these undesirable consequences can be difficult to predict due to the paucity of information on species interactions. A probabilistic qualitative approach, which does not require extensive model parameterization, is becoming increasingly accepted and applied to conservation scenarios when information is limited. However, recent work has criticized this approach on philosophical grounds and proposed an alternative non-probabilistic Boolean analysis method, which circumvents the philosophical difficulties. There is a need for exploring the ability of this novel approach for informing conservation decisions. To do so, we applied the first real-world test of the non-probabilistic Boolean approach using a case study of management of Fells catus (feral cat) and Rattus rattus (black rat) on Christmas Island. We also applied the probabilistic approach as a contrast. Our modeling results showed that the probabilistic approach generated ambiguous outcomes, making it impractical to draw management recommendations. In contrast, the non-probabilistic Boolean approach revealed interpretable rules governing species responses, suggesting that while cat management alone is a risky strategy, the risk of negative effects of cat management on native species can be reduced by the addition of rat management. Thus, given limited resources, in combination with cat management it is prudent to prioritize rat management efforts in the habitats of potentially impacted native species of high concern and value. We conclude that the Boolean approach can be very useful when little information is available to model an ecological system and that it provides a way of identifying the potential risks and benefits of management strategies, enabling better informed conservation decision-making in the face of limited knowledge. |
