A comprehensive distribution map and habitat suitability model for Great Basin bristlecone pine (Pinus longaeva D.K. Bailey)

Abstract

Great Basin bristlecone pine (Pinus longaeva D.K. Bailey) is a long-lived subalpine conifer native to the western United States. Bristlecone pine is considered a keystone species of high-elevation Great Basin sky island forests, yet detailed distribution data and information about constraining environmental factors for the species are lacking. We compiled a distribution dataset consisting of 490 individual bristlecone pine forest stands covering nearly 100,000 ha in Utah, Nevada, and California to enable improved monitoring of the species and enhanced understanding of forest succession dynamics. These data were sourced using records from herbaria, an internet-based ecological database, government vegetation surveys and extensive field surveys. We field-verified approximately 60% of mapped bristlecone pine stands in cooperation with a network of collaborators. We used this occurrence data to train a Maximum Entropy (MaxEnt) Habitat Suitability Model (HSM) covering the entire known distribution of bristlecone pine. Sixty environmental variables representing climatic, topographic, and edaphic factors were considered in the model and ranked by relative importance in constraining the MaxEnt model to the current known distribution of bristlecone pine. The model was regenerated using the 10 most influential variables that showed an insignificant correlation (|r| < 0.7) with all other variables to reduce possible effects of collinearity. A second model was produced using only climatic and topographic variables to fill coverage gaps in our soil variables, and the two resulting probability maps were merged. Both models performed well using the AUC metric of model fit--the models achieved values of 0.967 and 0.954, respectively. The variables that contributed the most predictive power to both models were mean dewpoint temperature and mean precipitation values for the months of January and February, respectively (> 80% predictive power), although other climatic, edaphic, and topographic factors were also shown to be useful for refining the distribution model. Our model’s occurrence probability map will serve as a valuable resource for conservation of bristlecone pine populations and assessing risk in the face of multiple threats, such as fire disturbance, MPB, and WPBR. Additionally, we intend to project the model to characterize the potential response of the species distribution to anticipated climate change in the region.

Citation

Burchfield, David R.; De Groff, Otto W.; Kitchen, Stanley G.; Charlet, David A.; Page, Douglas H.; Millar, Constance I.; Merkler, Douglas J.; Taylor, Gregory W.; Ortiz, Héctor G.; Petersen, Steven L. 2021. A comprehensive distribution map and habitat suitability model for Great Basin bristlecone pine (Pinus longaeva D.K. Bailey). In: Burchfield, David. Assessment of Great Basin bristlecone pine (Pinus longaeva D.K. Bailey) forest communities using geospatial technologies. Provo, UT: Brigham Young University: 57-99. Chapter 3. Ph.D. dissertation.