NCCOS Assessment: Modeling at-sea density of marine birds to support renewable energy planning on the Pacific Outer Continental Shelf of the contiguous United States (NCEI Accession 0242882)

Submission Package ID: 4LFC6T

Methods:
This analysis relied mainly on two types of data: counts of marine birds at sea from sighting surveys and information about the environment in the Pacific OCS region. Sighting datasets were provided by various state and federal agencies, academic institutions, and private organizations. Available spatial information describing the environment in the Pacific OCS region was compiled and synthesized by NCCOS. Environmental data came from a range of sources including remote sensing datasets and an ocean model dataset. Spatial environmental variables were characterized as spatial rasters, with dynamic variables represented by seasonal long-term climatologies. Spatial predictive modeling was applied to the sighting data to account for spatial and temporal heterogeneity in survey effort, platform, and protocol. An ensemble machine-learning technique, component-wise boosting of hierarchical zero-inflated count models, was used to relate the counts of each species or taxonomic group to the environmental predictor variables while accounting for survey heterogeneity and the aggregated nature of sightings. The modeling technique allowed for complex non-linear relationships between response and predictor variables and interacting effects among predictors. Bootstrapping was used to derive estimates of the uncertainty in model predictions. For a complete description of the methods see Leirness et al. (2021).