How sampling design of GPS collar deployment influences consistency of mapped migration corridors over time

Abstract

Federal and state agencies within the United States have recently issued directives prioritizing the conservation of ungulate migration corridors and winter ranges. The ability to identify and delineate the spatial distribution of seasonal ranges underpins these policies. While such delineations are often derived from global positioning system (GPS) collar data collected for a few years on a focal population, they are being used in long-term conservation planning. Our objectives were to quantify consistency in migration corridors from year to year and cumulatively across multiple years and identify which aspects of the sampling design of GPS collar deployment will delineate a consistent and relatively complete migration corridor. We used data from 6 sub-herds of mule deer (Odocoileus hemionus), a species known to have high migratory fidelity, located in Wyoming and northern New Mexico, USA, monitored for 5–7 years (510 unique individuals). We calculated 2 types of migration corridors over time: cumulative corridors where each new year of data was added to all previous years and yearly corridors where each year was based only on data collected in that year. We then calculated the year-to-year consistency in the 2 types of migration corridors by calculating the percent overlap between corridors calculated in sequential years. We found that collaring a higher proportion of a sub-herd increased the consistency in migration corridors, whereas collaring new individuals via redeployments in a subsequent year of monitoring caused corridors to shift. To obtain a corridor with ≥90% consistency (i.e., approaching the complete area used by a population in our data), our results suggest that biologists should strive to collar ≥6% of a sub-herd for a minimum of 2 years. However, if ≥6% of a sub-herd cannot be collared, monitoring for longer (3–4 years) will provide roughly 90% consistency in a migration corridor estimate for mule deer. Furthermore, adding 16–25% new individuals each year will help capture variation among individuals while maintaining corridor consistency of ≥90%, leading to a more accurate delineation of the corridor. Our results provide managers with a logistical framework for collaring projects aimed at delineating migration corridors that are durable into the future.