Use of occupancy surveys to monitor Clark’s Nutcrackers

Background/Question/Methods

Resource managers recognize the immediate need for determining an effective survey method for Clark’s Nutcrackers (Nucifraga columbiana) because some populations are apparently declining, likely due to the high mortality of whitebark pines (Pinus albicaulis). The population status and habitat use of nutcrackers are poorly understood, in part due to the lack of a reliable method of surveying populations. My objectives were to carry out occupancy surveys that incorporate detectability to determine if (1) habitat is a significant cause of heterogeneity of detectability of nutcrackers, (2) occupancy probabilities corrected for detectability improve occupancy estimates, and (3) occupancy surveys are a reliable means to survey nutcrackers. I carried out 208 hours of nutcracker surveys, as well as cone counts, cone bearing tree counts and a modified point quarter method at 88 points over two years.

Results/Conclusions

My results demonstrate that it is necessary to correct for imperfect detectability when surveying nutcrackers to increase reliability of occupancy and abundance estimates. The likelihood of detecting nutcrackers varies seasonally, and is conditional on time of day, local habitat, and point count radius. For example, the tree density has a strong effect on detectability in early August, but not during the harvest season in late August. These results show that because there is not similar detectability across seasons and habitats, surveys in different seasons and habitats cannot be compared without taking detectability into account. After accounting for detectability, model results suggested that nutcracker occupancy varied seasonally, and was variably impacted among seasons by habitat characteristics. For example, the number of whitebark pine cones per hectare had a strong effect on nutcracker occupancy in early and late August, but not in July or September. Occupancy probabilities corrected for detectability do improve occupancy estimates: I determined that the number of occupied sites would be underestimated if detectability is not taken into account. Therefore, results from surveys that do not control for detectability would likely be biased and inaccurate conclusions about population status or trends could result. Traditional monitoring methods typically have a long history of use, and are often assumed to be accurate due to their prevalence. However, many of these methods have not been evaluated for the impact of variable detectability. My data suggest that changes to naïve occupancy and abundance estimates resulting from detectability corrections must be evaluated when establishing monitoring programs for not only nutcrackers, but many other species.