Alaska Landbird Monitoring Survey - Population Trend Analysis

This code release implements three hierarchical Bayesian population trend models for Alaskan breeding landbirds, following the methods of Handel and Sauer (2017) and Amundson et al. (2014). The models use point-count data from the Alaska Landbird Monitoring Survey (ALMS; off-road points in Bird Conservation Regions 4 and 5) and the North American Breeding Bird Survey (BBS; roadside routes), and estimate annual population indices and period trends for individual species.

The three models are:
1) ALMS-only model estimates off-road population trends from ALMS data using a hierarchical N-mixture model with time-removal (availability) and half-normal distance sampling (perceptibility) detection corrections.
2) BBS-only model estimates roadside population trends from BBS data using a hierarchical Poisson model with observer × route and year random effects.
3) Joint ALMS + BBS model combines both datasets in a single model that shares annual stratum × year random effects (gamma) between the two surveys. The joint model explicitly estimates the roadside effect (rho): the difference in trend slope between BBS roadside routes and ALMS off-road points within the same stratum. A credible interval for rho that excludes zero indicates a significant and potentially habitat-mediated discordance between roadside and off-road trends.

This code release implements the model framework used in Handel and Sauer (2017) as a set of R nimble models. It also provides helper scripts for pre-processing ALMS and BBS data.