Development of a predictive spatial distribution model for Erioderma pedicellatum (boreal felt lichen) for the island of Newfoundland

The worldwide population of Erioderma pedicellatum is currently listed as critical by the IUCN, with over 95% of the current population residing on the island of Newfoundland. Surveys of E. pedicellatum habitats and populations have primarily been opportunistic, rather than systematic in nature. Bor...

Full description

Bibliographic Details
Main Author: Skinner, Randolph
Format: Thesis
Language:English
Published: Memorial University of Newfoundland 2011
Subjects:
Gam
Online Access:https://research.library.mun.ca/9529/
https://research.library.mun.ca/9529/1/Skinner_Randolph.pdf
Description
Summary:The worldwide population of Erioderma pedicellatum is currently listed as critical by the IUCN, with over 95% of the current population residing on the island of Newfoundland. Surveys of E. pedicellatum habitats and populations have primarily been opportunistic, rather than systematic in nature. Boreal felt lichen is listed as a species of special concern and vulnerable under COSEWIC and the Newfoundland and Labrador provincial listings respectively. By using a Geographic Information System (GIS) and compiled occurrence data and pseudo-absence data, I developed the first systematic predictive spatial distribution model for E. pedicellatum on the island of Newfoundland. A suite of 19 models using 4 different parameters were developed; the model with distance from coastline and aspect was the best candidate. Testing with reserve data and using a confusion matrix showed that the model displayed low model sensitivity (i.e., a low ability to predict false presence), but high model specificity (a strong ability to predict true absence). The final predictive model can assist future COSEWIC status assessments and provincial conservation management decisions that require information on probable species distribution. -- Keywords: Erioderma pedicellatum; boreal felt lichen; predictive habitat model; presence absence model; GAM; Newfoundland