Translating a species distribution model into management recommendations for an at-risk lichen. Issue 2 (3rd April 2022)

• Record Type: Journal Article
• Authors: Setchell, Alexandra; Haughian, Sean R.
• Is Part Of: Ecoscience. Volume 29:Issue 2(2022)
• Page Start: 159
• Page End: 165
• Publication Date: 2022-04-03
• Subjects: Decision-support tools -- Fuscopannaria leucosticta -- species at risk -- New Brunswick -- forest management; Outils d'aide à la décision -- Fuscopannaria leucosticta -- espèce en péril -- Nouveau-Brunswick -- aménagement forestier; Ecology -- Periodicals; 577
• DOI: 10.1080/11956860.2021.1949821
• Languages: English
• ISSNs: 1195-6860

Abstract

ABSTRACT: Species distribution models are powerful tools on their own for the management of rare and at-risk species, but their utility as such can be limited if their accuracy is low, or if management policies are applied without consideration of the structure of the data. We tested the accuracy of a MaxEnt species distribution model for white-rimmed shingle lichen ( Fuscopannaria leucosticta ) in New Brunswick, Canada, and proposed an additional decision support-tool that matches the species probability of occurrence (PO) to possible management actions. The model was tested using two years of new field observations. Classification trees were used to define four groups across the range of PO values. Chi-squared tests were used to test the accuracy of predictions in each class. The model was found to accurately predict occurrences in very low and moderate classes, but underpredicted occurrences in low and high classes. We recommend using similar, data-driven support tools in the conservation decision-making process to minimize losses of the species and maximize efficacy of management.