TRACKING WILDLIFE BY SATELLITE 



13 



Table 7. Proportion of locations in each of the 3 location quality index (LQ) categories for platform transmitter 

 terminals (PTT's) on various species, April 1987 -September 1988. LQ3 locations are the best quality, followed by 

 LQ2 andLQl . Sample sizes are the total number of locations receivedfrom all PTT's during the period. LQO locations 

 are excluded. 



In addition, we addressed the following hypotheses 

 concerning location bias: 



Directional biases in locations resulted from using map- 

 ping systems other than that used by Argos. 



Slope and aspect of an animal's true location were cor- 

 related with the magnitude or azimuth of error. 



Elevations of PTT's were correlated with the magnitude 

 or azimuth of location error. 



We also produced histograms of overall error, combin- 

 ing factors that produced bias with those that produced 

 variability. 



Methods 



Data were received from three fixed PTT's at known 

 locations in Inuvik, Northwest Territories, and at Nome 

 and Fairbanks, Alaska, in early 1987. Elevations of the 

 first two PTT's were at sea level; the Fairbanks PTT was at 

 approximately 152 m. All three PTT's had external anten- 

 nas and power supplies, but otherwise they used the same 

 electronic components as PTT's used on animals. Only 

 LQ2 and LQ3 locations were processed because these 

 PTT's were used primarily as reference platforms; there- 

 fore, quantity of location estimates was less important 

 than quality. The PTT at Inuvik was located indoors in a 

 protective housing and kept at near-constant ambient tem- 

 perature. The PTT at Nome was housed indoors, but ambi- 

 ent temperature was not kept constant. The PTT at Fair- 

 banks was subject to daily fluctuations in ambient 

 temperature. 



Data were also received from Telonics collars placed at 

 locations for which the true coordinates were known to 

 within 50 m. In one experiment, collars were placed simul- 



taneously at three sites of varying elevation (two at U.S. 

 Geological Survey benchmarks). In another experiment, 

 three randomly chosen collars were placed on caribou in a 

 small (100 X 100 m) pen at the University of Alaska in 

 Fairbanks. At the same time, three collars that transmitted 

 during the same hours were placed adjacent to the pen on a 

 wooden post. Later, one of the collars used on a caribou 

 was placed on the same wooden post, while an additional 

 PTT that had become available was placed on a caribou. 



Location estimates from experiments to calibrate ac- 

 tivity sensors were also available from these caribou, as 

 well as muskoxen, moose, mule deer, and elk. We also 

 used data from sites in the field whenever a study animal 

 died, taking care in each case to map the site carefully. 

 Finally, locations for free-ranging animals were gathered 

 from the ground (at times coincidental with satellite over- 

 passes) on Dall sheep in the Brooks Range, Alaska, by M. 

 Hansen, and elk in Yellowstone National Park by D. Vales. 



Because most data were available from PTT's that were 

 tested before deployment in the field, we were not able to 

 adhere to a rigorous, factorial design to test the things that 

 might contribute to telemetry error. Instead, we attempted 

 to isolate biases by employing data sets free of confounding 

 variables associated with each factor. All data were proc- 

 essed by Argos after their April 1987 processing changes. 



Results and Discussion 



Precision. Locations from the constant-temperature, 

 fixed PTT at Inuvik formed a nearly symmetrical distribu- 

 tion around the mean value, with standard deviations in 

 the longitudinal and latitudinal directions slightly greater 

 than those specified by Argos (Tables 2 and 8). Standard 



