Estimating gray whale abundance from shore-based counts using a multilevel Bayesian model

Counts of southbound migrating whales off California form the basis abundance estimation for eastern North Pacific stock gray (Eschrichtius robustus). Previous assessments (1967–2007) have estimated detection probability (p) from detection-non pods by two independent observers. However, tracking distinct in field can be difficult single observers; resulting biased estimates pod sizes that needed correcting, and matching observations same both observers involved key assumptions. Due to these limitations, a new observation approach has been adopted wherein paired team work together use computerised mapping application better track enumerate tally number passing during watch periods. This produced consistent counts over four recently monitored migrations (2006/7, 2007/8, 2009/10 2010/11), with an apparent increase p compared previous method. To evaluate estimate years, stations operating simultaneously were using hierarchical Bayesian ‘N-mixture’ model jointly without challenge between stations. The baseline detectability po was as 0.80 (95% Highest Posterior Density Interval [HPDI] = 0.75–0.85), which varied conditions, observer effects changes whale migration. Abundance described selection parametric normally distributed common migration trend semi-parametric time trends independently each year; resultant curve weighted compromise models, allowing departures trend. summed ranged 17,820 HPDI 16,150–19,920) 2007/08 21,210 19,420–23,230) 2009/10, indicative stable population.