USGS Home
Contact USGS
Search USGS

Fig.1. Global Positioning Systems radiocollar tested on wolves and white-tailed deer. Molded plastic box is 8.2 x 11.9 x 5.4 cm; the entire collar weighs 920 g.

Accuracy of the GPS boards was tested under open sky prior to collar deployment. On 3 tests with 58, 53, and 56 locations, uncorrected error was 76.5 m, 72.5 m, and 62.9 m respectively ( = 70.6 m ). This measure of error is the "2DRMS" value, or 2 * square root ( (SD of X errors )² + (SD of Y errors)²).

We attached the collars to a desktop DOS computer and programmed them to make location attempts at 1 per 15 minute, 1 per 30 minute, 1 per 45 minute, 1 per hour, or 1 per any number of hours specified. The GPS device turned on at the programmed intervals and turn off after each attempt. Each location attempt lasted until a fix was obtained or 96 sec. If no location was recorded, the GPS collar tried again in 15 min and 30 min. If all 3 attempts failed, further attempts were not made until the next programmed interval. If an attempt was successful, the VHF signal produced was 3 single beeps and a double beep; otherwise single beeps only were transmitted. ATS estimated that the collars could receive and record 1,200 +/- 400 locations each, depending on energy used during each location attempt.

A transmitting device (triggering transmitter) allowed communication with the collars to determine status of the GPS batteries and to release the collar at any time (Mech and Gese 1992). If no such signal was sent, the collars were programmed to drop off automatically about 65 days after GPS battery expiration. None of our collars dropped off in this fashion. Once a collar was released, the VHF signal was transmitted at 30 beeps per minute for about 25 days. When battery levels were too low to make a GPS location attempt, the VHF signal pulse rate changed from 60 beeps per minute to 45 beeps per minute.

After we retrieved the collars, location data stored in the collar was downloaded to a personal computer (in ASCII). Downloaded data included time, date, location (latitude and longitude or UTM), identity of satellites used, and 2 indices of the estimated accuracy of the locations (whether the locations were 2-dimensional or 3-dimensional and PDOP, the Positional Dilution of Position; Logsdon 1992).

We placed GPS collars on 4 wolves (23-39 kg) in February 1997 at Camp Ripley, on 3 white-tailed deer (58-76 kg) in the Superior National Forest during April 1997, on 2 wolves (32 and 35 kg) in Denali in May 1997, and on 2 wolves in the Superior National Forest in July 1997 (Table 1).

^a Ripley = Camp Ripley National Guard Training Site, Little Falls, MN; SNF = Superior National Forest, MN; Denali = Denali National Park, AK.
^b Minimum is based on number of extra attempts the GPS made to obtain a location. Maximum is based on total number of locations the GPS was programmed to obtain. Collars with attempt rates at 4/hour have only one number because "retry" intervals are also programmed intervals.
^c Based on 1,200 + 400 attempts per collar as predicted by ATS, Inc.
^d Short life caused by condensation inside battery box.
^e The drop-off mechanism on this collar failed; the collar was retrieved when the wolf was trapped.
^f Based on 1,200 + 400 attempts per collar, minus 10% for VHF current drain as predicted by ATS, Inc.
^g Based on 1,200 + 400 attempts per collar, minus 20% for VHF current drain as predicted by ATS, Inc.

We estimated the expected longevity of each collar (Table 1) for all sampling rates from the manufacturer's estimate that each collar was capable of making 1,200 + 400 location attempts. However, the relationship between location-attempt rate and GPS life is not linear. It is influenced by variable VHF and microprocessor current drain between location-attempt intervals of different lengths and by increased current drain per GPS attempt at longer intervals. These are therefore approximations of GPS life. Actual life will probably be 10-20% lower for collars programmed to attempt locations at intervals >3 hours, and closer to estimated values for collars programmed to attempt locations at <3 hours (based on calculations by the manufacturer).

The manufacturer estimates that with no GPS location attempts, the VHF transmitter and microprocessor current would deplete the battery in 512 days. However, with 1 GPS location attempt per day and retries of once per week, for example, they estimate the GPS battery should last 354 days; thereafter, only the VHF transmitter would operate.