|
Activity Number:
|
232
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 5, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistics and the Environment
|
| Abstract - #301423 |
|
Title:
|
Using a Bayesian Hierarchical Measurement Error Model To Estimate Individual Observer Detection Rates in Aerial Transect Surveys
|
|
Author(s):
|
Mark C. Otto*+
|
|
Companies:
|
U.S. Fish and Wildlife Service
|
|
Address:
|
Patuxent Wildlife Research Center, Laurel, MD, 20708-4002,
|
|
Keywords:
|
Poisson-Log Normal
|
|
Abstract:
|
We compare visibility corrections in the North American Breeding Waterfowl Survey using design-based and Bayesian hierarchical model estimators. The survey currently uses a combined ratio estimator to assess the proportion of ducks missed from the fixed wing counts compared to either ground or helicopter counts of the same segments. Counts are pooled over enough years and observer areas to provide stable estimates of detection. A Bayesian hierarchical model can estimate detection rates using a measurement error model that includes random individual observer effects. Since pilot and observer count adjacent sides of the same segments, all the data can be used to estimate observer difference rather than just the segments with ground counts. The hierarchical model yield more stable detection rates and population totals.
|
