USGS Home
Contact USGS
Search USGS

Kathryn Bennett, James Biggs, and P. R. Fresquez, Los Alamos Laboratory, ESH-20, M887, Los Alamos, NM 87545 USA

In 1996, we initiated a study to assess seasonal habitat use and movement patterns of Rocky Mountain elk (Cervus elaphus nelsoni) using global positioning system (GPS) radio collars. As part of this study, we attempted to assess the accuracies of GPS (non-differentially corrected) positions under various vegetation canopies and terrain conditions with the use of a GPS "test" collar. The test collar was activated every 20 min to obtain a position location and continuously up-linked to Argos satellites to transfer position data files. We used a Telonics, Inc. up-link receiver to intercept the transmission and view the results of the collar in real time. We placed the collar on a stand equivalent to the neck height of an adult elk and then placed the stand within three different treatment categories: (1) topographical influence (canyon and mesa tops), (2) canopy influence (open and closed canopy), and (3) vegetation type influence (ponderosa pine and pinon pine-juniper). The collar was kept at each location for 1 hr (usually obtaining 3 fixes). In addition, we used a hand-held GPS to obtain a position of the test collar at the same time and location. The hand-held unit was differentially corrected. Previous tests of the hand-held unit indicated that the accuracy was within two meters of an actual position. To determine Locational error of the test collar within the different treatments, we made comparisons between the test collar and the hand-held GPS following correction. The overall mean Locational error was 106 ± 16 m (354 ± 53 ft). There were no statistical differences (α = 0.05) in Locational errors between ponderosa pine and pinon pine-juniper vegetation types (P = 0.8199), open and closed canopies (P = 0.8672), or canyons and mesa tops (P = 0.9874). Observation rate was also calculated for each treatment category (e.g., the number of positions obtained in a defined period of time divided by the maximum number of positions possible). There were no statistical differences (α = 0.05) between observation rates of the three treatments (topographical influences: P = 0.1120; canopy influences: P = 0.3897; and vegetation type influences: P = 0.6282).