An Unreconciled Double-Observer Method for Estimating Detection Probability and Abundance

• Published in: The Auk 2010-10, Vol.127 (4), p.841-849
• Main Authors: Riddle, Jason D, Pollock, Kenneth H, Simons, Theodore R
• Format: Article
• Language: English
• Subjects: abundance; Aviculture; Bird songs; Birds; Breeding; detection probability; double-observer; Estimates; Estimation bias; Estimation methods; Methods; Methods and Statistics; Parametric models; Population estimates; Population size; PRESENCE; Probability; repeated counts; Warblers
• ISSN: 0004-8038; 1938-4254; 2732-4613
• DOI: 10.1525/auk.2010.09118

Multiple-observer point-count methods allow estimation of detection probability and have some advantages over other point-count methods. We introduce the unreconciled double-observer method, an independent double-observer method that does not require observers to match or otherwise reconcile individual observations. The modeling of the counts uses the repeated-counts model (Royle 2004). We compared estimates of detection probability and abundance from the unreconciled double-observer method with estimates from the traditional independent double-observer method (which requires matching of individual animals) using field-simulated bird-point-count data. The unreconciled double-observer method provided point estimates of detection probability and abundance that were essentially identical to the results of the independent double-observer method, even though much less effort was required to collect data with the unreconciled method. Estimates of detection probability from the unreconciled double-observer method were usually less precise—and estimates of abundance always less precise—than those from the independent double-observer method, because there is less information available in the unreconciled double-observer approach. We also evaluated the unreconciled double-observer method on 12 Breeding Bird Survey (BBS) routes. The unreconciled double-observer method provided estimates of route-level abundance for most routes and for most of the 20 species we considered. We believe that this method has potential for established bird-monitoring programs such as the BBS because no changes in historical data-collection methods are required other than collecting data simultaneously using two or more observers.