Executive Summary Background Features of Unit Tested Setting for the Test Results Data Summary Conclusions Acknowledgements & Further Information

DATA SUMMARY PreparationBeyond the one broken leg, no influence to the well-being of the sheep wearing the collars was ever observed. The herders never expressed concern, and in fact, seemed genuinely interested in the technology. For the purpose of the test, the information was not intended to be applied to 2001 allotment permit administration, but only in the use and utility of the technology. Over the 128 days between early June and mid-October 2001, the sheep fitted with the collars traveled 232 miles (373 kilometers). Their routes depict a slow, deliberate circulation through the allotment—up and down drainages, and ascending and descending several ridges. Nearly half of the total grazing season was spent between 7,000 and 7,500 feet (2,134 and 2,286 meters) on the valley floor, and less than 5 percent above 9,000 feet (2,743 meters).

The two collars attempted 2,288 satellite position fixes over the duration of the test. Of these, 94 percent were successful. When a failure occurs, the collar records a descriptor of the most likely cause. Of the causes listed, the lack of at least three satellites was the most common, followed by the threshold for dilution of precision (DOP-basically the uncertainty in position resulting from the particular geometry of the satellite constellation) being exceeded. Over the test period, failures were most common in July. On a daily basis, failures were most common during the late morning to afternoon period. It can be speculated that both of these tendencies are related to overhead obstructions (i.e., trees). When trees were more dominant in the landscape, position failures were more common. Day to day, sheep seek shade during the peak heat periods, and generally bed in the open during the night. The most successful fix periods of the day occurred at night between 10:00 p.m. and 4:00 a.m. (See table 5) Some subtle differences between the individual sheep wearing the collars may be evident in the data. The sheep wearing collar 378 appears to have routinely ascended higher when grazing ridge-tops, and was quicker to enter mid-day cover than the sheep wearing collar 379.  The grazing habitats of neither sheep appear to have had any effect on the integrity of the collars. Post-inspection of the collars revealed nothing beyond expected scuffs and scratches on the external case. (See figure 1.) Table 5: Summary of position fix success Collar 378 379 fix frequency 4 hours 2 hours total attempts 763 1,525 total successful 709 1,431 percent successful 93 94 percent of memory capacity used 8 15 Figure 1: Collar 378 full season GPS position results The only difference tested between the two collars was the sampling frequency. The results are a matter of fractal resolution. The path of the sheep wearing collar 379, which sampled at twice the temporal resolution of collar 378, would appear to wander to a greater extent than did 378. In fact, for the full season, the plotted route of collar 378 is 176 miles compared to 232 miles for collar 379, or only 76 percent. This difference, however, is more likely a simple expression of the greater temporal resolution of collar 379. Figure 2 displays comparative routes recorded by the collars during September 16 and 17, 2001. These dates are a random example, with the exception that all position fix attempts were successful for both collars over the 48-hour period. Figure 2—Comparison between position sampling frequencies of Collars 378 (4 hour) and 379 (2 hour). The example in Figure 2 demonstrates both the natural variation between animals, as well as the more detailed route of collar 379 with the greater temporal resolution.