Document Type

Thesis

Abstract

Quantitative descriptions of juvenile salmon spatial ecology during the marine phase of their lifecycle are needed to help direct conservation efforts and uncover latent relationships between juvenile salmon and their environment. The NOAA Fisheries Essential Fish Habitat (EFH) program contributes to the conservation of species in the marine waters off of the coast of Alaska by identifying areas of particular importance to reproduction and growth. Salmon EFH definitions have not yet been updated using the species distribution modeling approach taken most recently in defining Alaska groundfish and crab EFH. Current EFH definitions for Alaskan salmon are geographically broad, do not quantitatively describe relationships between marine conditions and salmon, and lack estimates of uncertainty. This research estimated static (time-averaged) and dynamic (time-varying) juvenile salmon distributions in the eastern Bering Sea (EBS) by fitting species distribution models (SDMs) to fishery-independent survey data and environmental indices spanning the years 2002-2019. Following model selection, species distribution maps were generated using predictions based upon the best-performing static and dynamic models to inform EFH definitions for the juvenile life stage. In Chapter 1, roughly parallel nested model structures were fit within two SDM frameworks: generalized additive models (GAMs) and Vector Autoregressive Spatio-temporal (VAST) models. Results from Chapter 1 indicate that while there was evidence for spatial variation in juvenile salmon distributions through time, this variation was encompassed by the EFH boundaries predicted by static models. In terms of performance, GAM and VAST frameworks were largely comparable, although VAST models appeared to be more robust to issues associated with spatial imbalance in survey data. In Chapter 2, nested GAM model structures tested the influence of environmental covariates on variability in the abundance and/or distribution of five species of juvenile Pacific salmon in the EBS. Model selection results and mapping indicated that in-situ environmental covariates significantly influenced abundance of juvenile salmon, while annual covariates significantly influenced the distributions of juvenile salmon. Center of gravity estimates found some evidence for species distribution shifts inshore/offshore and to the north/south in response to specific hindcast environmental conditions. This work might be leveraged to update current salmon EFH definitions and inform future model-based conservation efforts.

Publication Date

12-17-2023

Handle

http://hdl.handle.net/11122/14957

Download

Share

COinS