DEVELOPMENT AND EVALUATION OF SIGHTABILITY MODELS FOR SUMMER ELK SURVEYS

We developed 2 sightability models from summer helicopter surveys of r adiocollared elk (Cervus elaphus) in Grand Teton National Park, Wyomin g. Significant variables (P < 0.05) included elk group size, activity, and percent vegetation cover for Model A, and elk group size and perc ent vegetation cover for Model B. We compared these 2 summer models an d a winter elk sightability model developed in Idaho that incorporates group size, percent vegetation cover, and percent snow cover. We base d model comparisons on predicted detection rates and model performance when applied to well-documented elk populations at Starkey Experiment al Forest and Range, Oregon (SEF), and Wind Cave National Park, South Dakota (WCNP). Pre dieted sightability was similar from summer Models A and B for active elk in <60% vegetation cover, but was lower from Mo del A for bedded elk. Model estimates of elk abundance (WCNP, SEF) and composition (SEF) usually were more accurate and consistently more pr ecise from Model B, suggesting elk activity had little influence on es timates of sum.mer elk population characteristics. Comparisons between Model B and the Idaho model indicated predicted sightability of small groups (less than or equal to 10 elk) was similar; the Idaho model pr ovided better accuracy and precision for validation tests of populatio ns consisting of predominantly small elk groups (WCNP: = 4.7 elk/group ; SEF: = 6.3 elk/group). The Idaho model, however, overestimated detec tion of large elk groups (30-45 elk/group) in moderate-dense vegetatio n (>30% vegetation cover), but this overestimation was accounted for b y Model B. Thus, we recommend application of the Idaho model during su mmer surveys where elk are less gregarious (<20 elk) and recommend app lication of summer Model B to high-density elk populations where elk o ccur in larger groups.