Model-Based Inference for Stratified Rotating Panel Surveys: A Strategy for Estimating the Number of Bald Eagle Nesting Territories in Florida

• Activity Number: 561
• Type: Topic Contributed
• Date/Time: Wednesday, August 1, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Mary C Christman*+
• Companies: MCC Statistical Consulting LLC
• Keywords: model-based estimation ; stratified sampling ; panel surveys ; count data ; environmental

Abstract

The bald eagle is an important species for conservation and so accurate estimation of its true abundance is important for development of adequate conservation strategies. Previously, the state of Florida performed annual aerial censuses of every county in the state for species management purposes. We address the question of precise estimation of productivity and total abundance of nesting territories using panel surveys to replace the census. The estimation procedure is model-based and relies on imputing the number of nesting territories that would have been observed in a county had it been in the sample rotation that year. The imputation is based on a variant of Poisson regression of previously collected counts assuming linearity and county-specific slopes. We consider several approaches to fitting the models, including maximum likelihood estimation of the parameters, least squares estimation (LSE), a modified LSE where MSE in the standard formula is replaced with the predicted value from the regression, and simple forward imputation of the last observed value. The approaches are compared using data collected in censuses of all counties annually from 1990 to 2000.