MULTISCALE HABITAT SELECTION BY BLACK PINE SNAKES (PITUOPHIS MELANOLEUCUS LODINGI) IN SOUTHERN MISSISSIPPI

With the historical and continued fragmentation of Longleaf Pine (Pinus palustris) forests, an increasing need emerges to identify remaining Black Pine Snake, Pituophis melanoleucus lodingi, populations and understand species environment correlations so that conservation efforts for this taxon are properly directed and appropriate. This study had three main objectives: (1) to use radiotelemetry data in conjunction with GIS modeling to assess multiple scale resource selection for P. in. lodingi; (2) to use occurrence data to validate our habitat suitability model; and (3) to quantify microhabitat characteristics of P. in. lodingi locality records and evaluate differences in land cover between historical and recent records. At the landscape scale, P. m. lodingi selected evergreen forests in upland areas lacking cultivated crops, pasture and hay fields, developed areas, and roads. Within home ranges, P. in. lodingi were found closer to scrub/shrub habitat and open areas than expected. Our Boolean model reduced the area deemed suitable for P. in. lodingi by 84.3%, while accurately categorizing 75% of recent P. m. lodingi occurrences, demonstrating the potential for robust model development despite small sample sizes. We found significant habitat differences between recent and historical P. in. lodingi locality records, which corroborated the findings of our Boolean model. Recent sites of occurrence were, on average, 597 in from areas with a model score of "optimal," whereas sites with no recent records were, on average, 1582 in from areas with an optimal score. Recent sites were characterized by significantly less canopy cover, less basal area, less midstory cover, greater percentages of grass, bare soil, and forbs in the ground cover, less shrubs and litter in the ground cover, and a more recent buns history than sites with no recent records. Areas containing historical P. In. lodingi populations are simultaneously becoming unsuitable at both the landscape and microhabitat scales. Our Boolean modeling approach represents a cost-effective way to plan for informed management actions for snake taxa of conservation concern.