Design Attributes for Sampling Rare Ecological Events in Forest Ecosystems: Lichens in the Pacific Northwest
Thomas C. Edwards, (US Geological Survey), email@example.com, and
Richard Cutler, (Utah State University), firstname.lastname@example.org
Despite the fact that sampling for rare events would seem to be a common problem in ecology, there is surprisingly little literature on the subject. This paper describes, and discusses the advantages and disadvantages of, a number of the most commonly used probability and non-random sampling methods for sampling rare ecological events in forest ecosystems. Focus is on a suite of species in the US Pacific Northwest referred to as Survey and Manage species, and includes fungi, lichens, bryophytes, mollusks, a few vascular plants, arthropods and two vertebrates. All represent species for which little information is known, and this lack of knowledge make their management quite difficult. Sampling methods considered include stratified random sampling with disproportionate allocation, network and snowball sampling, adaptive methods such as adaptive cluster sampling, and model-based ppz-sampling. Information needs of the US Forest Service and logistical concerns that constrained the eventual choice of sample designs are discussed. The probability sampling framework that was finally implemented for the Survey and Manage program, and how it addresses information needs for forest management, are described in detail. Some important elements of the framework are the use of an existing vegetation survey grid for sampling locations, and stratification (with disproportionate allocation) by stand age class and late successional reserve status. The paper concludes with some recommendations for future probability sampling efforts for the Survey and Manage program.