Arctic PRISM/double sampling


This protocol was included in this guide because it is an example of double sampling that could be used in the Southeastern U.S. In double sampling, a large sample of plots is surveyed for species’ abundance using a rapid method of unknown accuracy, and a subset of the plots are surveyed using an intensive method that yields unbiased estimates of abundance.  The ratio of counts obtained using the rapid and intensive methods (on the subset of plots surveyed using both methods) is then used to adjust the results from all plots surveyed.  Double sampling has been used for decades on aerial surveys of waterfowl.  During the past decade the method was refined and used in the Arctic PRISM program throughout the arctic regions of Alaska and Canada.  Recently, it has been widely used for landbirds, especially in the southwestern United States.


Double sampling yields unbiased estimates of density, population size, and trend in size that are subject only to the assumptions that the nominal sampling plan is followed and that estimates from the intensive plots are unbiased.  Thus, differences or changes in observer skill, effects of weather or traffic noise, and change in phenology can all occur without causing any bias in the estimates.  Rapid surveys have usually been made using area searches, point counts, or aerial surveys, but any method can be used.  Thus, rapid surveyors could use distance, double observer, or removal methods, and the intensive surveys would then reveal how accurate the rapid method was.  If the rapid method turned out to yield unbiased estimates, then the intensive surveys could be discontinued. The intensive surveys also yield substantial additional information, such as nesting success, that may be useful in other ways.  For example, in the Arctic PRISM surveys, intensive surveyors also conduct predator scans, record plant phenology, and trap invertebrates.


The intensive surveys take time to conduct and may require 25-50% of the total survey effort, so they should only be included if there is uncertainty about accuracy of the rapid counts (or if the other information they provide justifies their inclusion).  The method also requires that plots be thoroughly searched, which means surveyors cannot count solely from roads.

Example of use

In addition to its use in aerial surveys and Arctic PRISM, double sampling has recently been used by the Bureau of Reclamation on the Lower Colorado River, by the Arizona Department of Game and Fish in a Statewide survey of riparian areas, by the Great Basin Bird Observatory in the Nevada Bird Count, by the Sonoran Joint Venture in Mexico, and in a multi-agency survey of birds in the Sonoran desert.  It has also been selected by several Department of Defense installations for upcoming projects.

For more information

Bart, J. and S.L. Earnst.  2002.  Double sampling to estimate density and population trends in birds.  Auk 119:36-45.

Brown, S., J. Bart, R.B. Lanctot, J.A. Johnson, S. Kendall, D. Payer, and J. Johnson.  2007.  Shorebird abundance and distribution on the coastal plain of the Arctic National Wildlife Refuge.  Condor 109:1-14.

Smith, P.A., J. Bart, R.B. Lanctot, B.J. McCaffrey, and S. Brown.  2009. Probability of Detection of Nests and Implications for Survey Design.  Condor 111:414-423.