Date of this Version
Johnson, Douglas H., Drewien, Roderick C., and Benning, Douglas S. Counting cranes: how much effort is enough? In: Ellis, David H., ed., Proceedings of the Eighth North American Crane Workshop, 11–14 January 2000, Albuquerque, New Mexico (Seattle, Wash: North American Crane Working Group, 2001), pp. 203-210.
Accurate population estimates of sandhill cranes (Grus canadensis) are important because management decisions, such as establishing hunting regulations, are based on those estimates. Counts often are made during aerial surveys when the cranes are congregated. A complete census may be feasible if the area to be surveyed is small and adequate resources are available. For large areas, resources may be inadequate for a census so partial counts (sample surveys) are made. Because cranes are gregarious, the counts in a sample of units may contain either a disproportionately large, or a disproportionately small, fraction of the total, leading to high variation among units and a very imprecise estimate of population size. Here we address the issue of survey accuracy by considering the Rocky Mountain population of greater sandhill cranes (G. c. tabida), which were surveyed each March in the San Luis Valley, Colorado. The entire population in the valley was surveyed for 12 years. We determined the accuracy of various sampling plans: these involved sampling fractions ranging from 20% to 50% to illustrate the potential for sampling a population that is spatially aggregated and to evaluate the feasibility of reducing the survey effort. We also compare systematic and simple random sampling and evaluate the efficacy of stratification. Our recommendations generally are pertinent to other populations of cranes, as well as other spatially aggregated species.