Corridors and Elk Migration: A Comparative Analysis of Landscape Connectivity Models and GPS Data in the Greater Yellowstone Ecosystem

Landscape connectivity models aim to map the links or corridors that wildlife would or do use between patches of habitat. Migratory species such as elk traverse between such patches which serve as seasonal ranges. The goal of this study was to compare and contrast the suitability of several landscape connectivity models for describing and predicting migration in a long-distance migrant. We measured the suitability of connectivity models for covering and predicting the migratory movements of elk in the Greater Yellowstone Ecosystem. GPS point data was converted to sequential networks for multiple populations of elk. GPS data was also used to delineate the summer and winter ranges of each population. The kernel density of routes in the networks was measured for comparison to connectivity models. The ranges served as the patches to be connected by such models. A resistance surface was produced using reclassified landcover data for mapping habitat suitability and linear road data for human presence or obstruction to movement. Landscape connectivity was measured for eleven migratory elk populations using three distinct models. The first measured connectivity using circuit theory; the second, agent based modeling; the third, least cost corridors. The model results were compared to the migratory network density by measuring correlation. This was followed by a new method of measuring the influence of autocorrelation between the models and networks. Some of the models were then altered to test for suspected influences. This study shows that least cost corridors and circuit theory can are limited in their ability to predict the migratory movements between summer and winter ranges but only so much. They lack the ability to predict exploratory movements that do not link conspicuous ranges to each other. They also lack the ability to account for all avoidance behaviors in the landscape. Our results suggest that connectivity models need improvement by accounting for exploration outside of prime habitat. It also suggests connectivity models are not adequate predictors of migratory movements and not suited to conservation planning of migratory networks. This supports Sawyer's (et al. 2009) ungulate conservation planning of considering connectivity but basing priority on migratory landscape usage. It is assumed that fragmentation or loss in connectivity impedes seasonal migration, cutting off wildlife from resources (Rudnick et al. 2012). This study shows that migratory elk are actually using less than prime and supposedly fragmented habitat in migration and that there is more than connectivity at play