Projecting Species Distributions Using Size-Structured and Size-Aggregated Models

The is part of an NSF funded project that will create species distribution maps using all life stages of West Coast fishes. Portions of this work are found in Chapter 4 of my PhD dissertation.

Data

Multiple data sources have potential applications to this project. Currently we are using:

• The NOAA Rockfish Recruitment and Ecosystem Assessment Survey, data for which are available upon request. This survey largely target pre-recruits but it initially started with a focus on rockfish. The survey operates in the late spring/early summer off the California Coast.
• The NOAA Pre-recruit Survey, data for which are available upon request. This survey follows the same protocols as the Rockfish Recruitment and Ecosystem Assessment Survey, but operates off the Oregon and Washington Coasts.

Methods

Variable Coefficient Generalized Additive Models (VGAMs)

VGAMs allow a coefficient to vary with change in another variable. In this application, the coefficient is location of a species which varies with a given environmental variable. Our initial model use climate indices as a stand-in for this variable while future models will likely use sea surface temperature.

• VGAMs applied to the RREAS data

Predictive-Process Generalized Linear Mixed Models (GLMMs)

This is a work in progress and will include future applications of the sdmTMB package.

• Code for these models is found here