Identifying Scenario Spaces, their Outliers and Redundancies, with Application to Bowhead AWMP Evaluation Trials

• Main Authors: Geof H. Givens, Marsha S. Huang
• Other Authors: The Pennsylvania State University CiteSeerX Archives
• Format: Text
• Language: English
• Published: 2000
• Subjects: bowhead whale
• Online Access: http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.31.9146; http://www.stat.colostate.edu/~geof/documents/newoutliers.ps

This paper describes how to characterize groups of simulation trial scenarios designed for testing Aboriginal Whaling Management Procedures (AWMPs). Specific features of the trials established for testing bowhead whale AWMPs are also examined. Trial scenarios are determined by a very high-dimensional parameter vector. The dimensionality and potential for important interactions between parameters mean that it is very difficult to identify outlying or redundant scenarios, or to see which scenarios share common characteristics that are appropriate for certain scenario classes. The four classes of scenarios identified by the Scientific Committee are reviewed, along with the appropriate relationships between these classes. We propose characterizing scenarios by a five-statistic summary based on population trajectory features rather than on the input parameter vector. This five-number summary can be used for identifying outlying or redundant scenarios, as we demonstrate in an example. We als.