To begin addressing statistical sufficiency, USGS and Olympic National Park jointly held a workshop with invited statisticians to focus on issues related to the spatial distribution of monitoring sites in large, mountainous parks. The issues addressed were:

• Difficulty in accessing all parts of large, mountainous parks.
• Achieving a statistically powerful yet affordable sample size.
• Integration of monitoring projects to increase understanding.
• The need for both intensive (e.g., heavy equipment involved, high frequency of sampling needed) and extensive sampling (i.e., large spatial distribution).
• An interpretable (i.e., based on present distribution of species) data set that is also flexible in the event of large changes in resource distribution.
• An adequate representation of both rare and common resources.

Another thorny sampling issue is to determine how to sample monitoring sites through time. "Panelized"sampling designs are touted as most efficient by some statisticians, but they are highly sensitive to missing data, are difficult to analyze, and may be unworkable for park monitoring programs. We will be addressing this challenge with a workshop in the spring of 2003.

Finally, we need to have an idea of how great a sampling effort is needed for monitoring of various taxa to achieve statistical power to detect a desirable amount of change with a given level of confidence. "Power" analyses conducted on pilot data are often used to determine a rough estimate of adequate sample numbers. We are conducting these analyses using pilot data sets for birds, bats, small mammals, and vegetation.

Project Status
We held the first workshop in April 2001 to identify a strategy for developing a spatial sampling frame for the Pacific Northwest National Parks. We held the second in November 2003 to address the need for a temporal sampling scheme. We have also completed two contracts with statisticians to analyze the power of individual species in several pilot studies to detect change and to determine the number of subplots needed within larger plots. We are following these with three other contracts to address the specific sample frames that have evolved from this work, as well as the power of FIA data, which describes forest vegetation and is collected by the USDA Forest Service.

Products or Related Links
Woodward, A. and K. J. Jenkins. 2001. Summary of Statistics Workshop - April 2-3, 2001.

Woodward, A. 2005. Temporal Sampling Frames: Summary of a Workshop, November 12-14, 2003.

Acknowledgements
We appreciate the participation of statisticians and NPS personnel at both workshops, the efforts of workshop presenters, excellent notes taken by designated note-takers, and help editing the summary documents. Participants are listed under each workshop summary. Presenters include Paul Geissler, Reed Glesne, Gail Irvine, Trent McDonald, Tony Olsen, Eric Rexstad, and N. Scott Urquhart. Note-takers include Mike DeBacker, John Boetsch, Jerry Freilich, Cat Hoffman, Giovanna Ranalli, Barbara Samora, and Alice Shelly. Other invited statisticians include E. Oz Garton, Phil Larsen, Lyman McDonald, Jean-Yves Courbois, Alice Shelly, and Don Stevens, We give special thanks to Paul Geissler for extra efforts.