Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Ecological Forecasting Under Climate Variability

Fish Community Structure Responses to Environmental Forcing — California Current Ecosystem

Author: Morgan Youngblood
Collaborator/Mentor: E. Meese
Affiliation: University of Hawaiʻi at Hilo — Marine Science Program
Study region: Point Conception biogeographic break, California Current
Status: Analysis complete


Overview

This repository contains the full data cleaning, integration, and modeling pipeline for a multivariate ecological forecasting study examining how environmental forcing variables (SST, wind velocity, chlorophyll-a, kelp biomass) drive fish community structure shifts across the Point Conception biogeographic break in the California Current ecosystem.

The analysis integrates >20 years of subtidal fish abundance surveys with regional and global climate datasets to quantify density responses of warm- and cool-affinity species, with models structured for direct integration with global climate projections through 2100.


Dataset

Parameter Value
Fish survey record >20 years subtidal abundance data (PISCO/UCSB)
Environmental drivers SST, wind (u/v/wspd), chlorophyll-a, kelp biomass
Data sources ERDDAP (SST, Chl-a), PSL/NCEP NARR (wind NetCDF), Landsat/SBC LTER (kelp), PISCO UCSB Fish Recruitment 1.3
Modeling framework GLM, GAM (mgcv, gamm4, FSSgam, MuMIn)
Climate integration GFDL-ESM2M global model, ROMS-NEMUCSC regional ocean model

Note: Raw and integrated data files are excluded from this repository. Scripts are fully documented and reproducible given access to source data.


Repository Structure

thesis/
├── scripts/
│   ├── Data Cleaning
│   │   ├── clean_chlorophyll_data.Rmd      # ERDDAP chlorophyll-a processing
│   │   ├── clean_sst_data.Rmd              # ERDDAP SST processing
│   │   ├── clean_wind_data.Rmd             # NCEP NARR NetCDF wind extraction
│   │   ├── clean_wind_2.Rmd                # Wind data QA/QC continuation
│   │   ├── clean_wind_3.Rmd                # Wind data integration finalization
│   │   ├── clean_fish_data.Rmd             # PISCO fish recruitment cleaning
│   │   ├── clean_kelp_data.Rmd             # Landsat kelp biomass processing
│   │   ├── clean_stress_data.Rmd           # Environmental stress index cleaning
│   │   └── clean_site_locations.Rmd        # Survey site spatial processing
│   ├── Integration & Analysis
│   │   ├── integrate_data.Rmd              # Multi-source data integration pipeline
│   │   ├── environmental_factors_analysis.Rmd  # Environmental driver analysis
│   │   ├── fish_species_density_analysis.Rmd   # Species density response modeling
│   │   ├── glm_build.Rmd                   # GLM construction and selection
│   │   └── gam_build.Rmd                   # GAM construction and selection
│   ├── Modeling & Figures
│   │   ├── glm_gam_modeling.R              # Full GLM/GAM modeling pipeline (E. Meese)
│   │   ├── data_exploration_glm.R          # Exploratory GLM analysis
│   │   └── glm_presentation_figures.R      # Publication-ready figure generation
│   └── Utilities
│       ├── netcdf_wind_extraction.R        # NetCDF wind data extraction utility
│       ├── netcdf_kelp_biomass.R           # NetCDF kelp biomass extraction utility
│       └── wind_speed_analysis.Rmd         # Wind speed analysis (early exploration)
└── data/                                   # Raw data excluded via .gitignore

Technical Stack

  • Language: R
  • Modeling: mgcv, gamm4, FSSgam, MuMIn, MASS, car
  • Data engineering: tidyverse, dplyr, tidyr, readr
  • Spatial/NetCDF: ncdf4, sf, terra
  • Visualization: ggplot2, ggthemes, GGally
  • Climate data: ERDDAP API, PSL/NCEP NARR NetCDF, Landsat SBC LTER

Key Findings

  • Developed GLMs quantifying density responses of warm- vs. cool-affinity species across the Point Conception biogeographic break
  • Quantified shifts in model predictability pre- and post-2014 marine heatwave (warm-affinity species R² increased from 0.12 to 0.23)
  • Models structured for integration with GFDL-ESM2M and ROMS-NEMUCSC to forecast ecological community shifts through 2100
  • Co-authored pending publication utilizing these models for long-range ecological forecasting

Acknowledgments

Modeling pipeline developed in collaboration with E. Meese, whose mentorship and analytical contributions are reflected throughout the scripts in this repository. Fish recruitment data provided by the Partnership for Interdisciplinary Studies of Coastal Oceans (PISCO) at UC Santa Barbara.


License

Code in this repository is released under the MIT License.
Data are not included and remain subject to the terms of the originating data providers (PISCO, NOAA/ERDDAP, SBC LTER).

About

Multivariate GLM/GAM pipeline modeling fish community structure responses to SST, wind, chlorophyll, and kelp biomass forcing in the California Current ecosystem

Topics

Resources

Stars

0 stars

Watchers

1 watching

Forks

Releases

Contributors

Languages