Date of Award

6-2026

Degree Name

MS in Statistics

Department/Program

Statistics

College

College of Science and Mathematics

Advisor

Trevor Ruiz

Advisor Department

Statistics

Advisor College

College of Science and Mathematics

Abstract

This thesis develops a structured dynamic factor analysis (sDFA) framework for decomposing multivariate environmental time series into latent biological and physical components. The methodology is applied to five years of high-resolution passive monitoring data collected from two sites in Morro Bay, California from 2020 through 2024. Relative contribution indices are developed based on the structured DFA that measure how much each latent process contributes to each observed variable at any given time. Structured DFA models fit to the application data suggest site-specific patterns in how biological and physical processes affect water quality variables. At the bay mouth location, physical processes contribute substantially more to pH variability than to dissolved oxygen, indicating decoupling between these variables, while at the bay south location, DO is more physically driven than pH. Site-level dynamic linear models are fit to relate external environmental variables (including temperature, salinity, and chlorophyll) to temporal variation in pH, dissolved oxygen, and decoupling to identify which environmental conditions drive patterns observed across sites. This framework provides a generalizable approach for investigating complex environmental systems.

Share

COinS