Data and code from: Long-term trends of striped bass habitat suitability under a temperature-oxygen squeeze in Lake Texoma, Oklahoma and Texas
Data files
Jun 25, 2026 version files 204.34 MB
-
README.md
7.70 KB
-
Whiteis_et_al_2026_Data.zip
204.33 MB
Abstract
Warmer epilimnetic water temperatures and reduced dissolved oxygen levels in deeper water are common stressors for fishes in freshwater reservoirs throughout the summer. We investigated the effects of a temperature–oxygen squeeze - a phenomenon that squeezes fish into a narrow, livable zone - on striped bass from 1995 to 2020 in Lake Texoma. This reservoir is located on the border of Texas and Oklahoma, possesses one of the largest self-sustaining populations of inland freshwater striped bass in the US, and is of high economic importance to the region. We used a trait-based approach to quantify 1) how the total volume of suitable habitat for striped bass has changed seasonally and across years, 2) the relative influence of temperature and oxygen on habitat suitability, and 3) whether a temperature-oxygen squeeze has affected the striped bass population and body condition. We found that the volume of suitable habitat for striped bass decreased over the course of each summer and was greatly influenced by day-to-day persistence and water level, but did not exhibit a significant linear decline over the years from 1995 to 2020. Additionally, the relative influence of temperature and oxygen on habitat suitability varied depending on the thresholds used to define the temperature-oxygen squeeze. We found no significant linear relationships between habitat volume, striped bass population, and body condition. One hypothesis that would explain this null result is that striped bass in Lake Texoma may be adapted to the seasonally stratified conditions, especially since habitat volume did not show a monotonic decline over time. We also detail other biophysical and statistical hypotheses that might explain why we did not detect an effect of a temperature-oxygen squeeze on striped bass. This study provides insights into future management of Lake Texoma and other reservoir fisheries as global change alters patterns in temperature, oxygen, and reservoir habitat.
Dataset DOI: 10.5061/dryad.3r2280gwb
Description of the data and file structure
This research aims to develop a habitat suitability model for managing reservoirs and shaping water conservation strategies that protect reservoir fisheries in the face of climate change. This challenge is explored through a study of striped bass in Lake Texoma. Specifically, a trait-based approach is utilized to model the amount of water volume associated with suitable habitat conditions for striped bass from 1995 to 2020 through the concept of a temperature-oxygen squeeze. Three main research questions (RQs) will be addressed. First, how is striped bass habitat suitability in Lake Texoma changing seasonally and interannually with a temperature-oxygen squeeze from 1995 to 2020 (RQ1)? Second, what are the combined and individual influences of temperature and DO on suitable habitat volume (RQ2)? Third, what is the impact of a temperature-oxygen squeeze in Lake Texoma on striped bass populations and body condition (RQ3)? Ultimately, this study of striped bass in Lake Texoma intends to provide a framework to explore reservoir fish species dynamics and determine best management practices of reservoir waters under a changing climate.
Files and variables
File: Whiteis_et_al_2026_Data.zip
Description: A zipped file containing R code, input files for running the code, outputs of code, and a READ_ME file describing the files and documents within the zipped file.
The Inputs file includes a bathymetry raster of Lake Texoma, a DEM of Lake Texoma and the surrounding area, and the final raster used for analysis that was created by subtracting the bathymetry raster from the DEM. Map services and data available from U.S. Geological Survey, National Geospatial Program. The map projection of this raster was UTM Zone 14N. Additionally, our analysis made use of the outputs of a deep learning model created by Suaza-Sierra (2024) that predicted vertical temperature and DO profiles by depth using observed data for training and testing. We also utilized fish catch per net night (C/f) and relative weights (Wr) data for striped bass (Morone saxatilis) sampled annually by the Oklahoma Department of Wildlife Conservation (ODWC; Sager et al., 2024). This includes data for the total number of striped bass and striped bass of < 12 in, 12-20 in, and > 20 in size groups. Input data spans 1995 to 2020, with some missing information for the striped bass relative weights and deep learning model outputs.
The Outputs file contains results at each step. This includes outputs showing:
- the cell count and surface area at each water level (texoma_sim_subtract_data.csv),
- suitable habitat per depth for each temperature-oxygen squeeze range (texoma_striped_bass_habitat.csv),
- daily suitable habitat volume (striped_bass_volume_summary.csv),
- mean habitat volume by day of year (striped_bass_mean_habitat_summary.csv),
- mean habitat volume with standard deviation (striped_bass_sd_mean_habitat_summary.csv),
- full dataframe with date, daily habitat volumes for each squeeze range, daily water level, and one day lagged volumes for each squeeze range used for GAMs (daily_volume_and_lags_for_GAMs.csv),
- the yearly minimum habitat volume (minimum_habitat_per_year.csv),
- yearly minimum 2-week average habitat volume (minimum_2week_average_habitat_per_year.csv),
- yearly minimum 30-day average habitat volume (minimum_1month_average_habitat_per_year.csv),
- and the suitable volume associated with each habitat scenario for each squeeze range per depth and daily summary (any file starting with "habitat_scanarios").
The above data files were used in different steps of analysis. This included (1) plotted mean seasonal habitat volumes (m3) under each temperature-oxygen squeeze range, (2) plotted mean seasonal habitat volumes (m3) of individual drivers (no stressors, temperature, DO, and both) under each range, (3) GAMs showing the overall model fit and individual influences of a one day lag (m3), seasonality, date, and water-level (m) on daily suitable striped bass habitat volume (m3), and (4) 54 linear regression models using combinations of the 3 temperature-oxygen squeeze ranges, 3 indicators of yearly suitable habitat (minimum habitat volume (m3), minimum average habitat volume of a 14-day period (m3), and minimum average habitat volume of a 30-day period (m3)), and 6 striped bass descriptors (total fish per net night (C/f), fish per net night of < 12 in. striped bass (C/f), fish per net night of 12-20 in. striped bass (C/f), fish per net night of > 20 in. striped bass (C/f), relative weights of < 12 in. striped bass (Wr), and relative weights of > 20 in. striped bass (Wr)).
Finally, this file includes the R code project, titled “Lake_Texoma_Final.R.” This code shows the step-by-step process of inputting data, creating the habitat suitability model using a trait-based approach for striped bass, and data analysis to determine trends in suitable habitat volume and the impact of a temperature-oxygen squeeze on striped bass in Lake Texoma.
Code/software
File: Lake_Texoma_Final.R
Description: This R file details the workflow used to determine trends in suitable habitat volume and the impact of a temperature-oxygen squeeze on striped bass in Lake Texoma.
Packages: raster, dplyr, ggplot2, cowplot, lubridate, terra, sf, readxl, devtools, tidyverse, patchwork, zoo, scales, and mgcv
Workflow:
- Load packages.
- Bring in bathymetry and DEM rasters to create a combined raster of full Lake Texoma bathymetry.
- Bring in simulated vertical temperature and DO profiles by depth in Lake Texoma created by Suaza-Sierra (2024).
- Run a function to calculate cell count and surface area at each unique water level.
- Define striped bass temperature-oxygen squeeze ranges and run functions to calculate suitable water volume at each depth and daily summaries of suitable water volume across depths.
- Quantify and plot seasonal variability in the volume of suitable habitat.
- Run generalized additive models (GAMs) to capture nonlinear seasonal dynamics, long-term trends, and temporal dependence by using variables of day of year, date, water level, and a one day habitat lag to account for temporal autocorrelation. Plot overall model fit and partial effects of each variable.
- Pull out representative values of the annual temperature-oxygen squeeze based on pulse and press disturbances and run linear regression models to investigate relationships between habitat volume and striped bass population and body condition data.
- Quantify suitable habitat volume under different scenarios of full reservoir volume, volume based on temperature, volume based on DO, and volume based on temperature-oxygen squeeze.
- Create additional descriptive plots for striped bass data and water level.
Access information
Data was derived from the following sources:
- DEMs: Map services and data available from U.S. Geological Survey, National Geospatial Program.
- Suaza Sierra, Isabela, "Predictive Understanding of Lake Water Temperature and Dissolved Oxygen Profiles Across the Red River Basin Through Interpretable Machine Learning" (2024). Open Access Theses & Dissertations. 4151.
https://scholarworks.utep.edu/open_etd/4151 - Sager, C., Hull, N., & Mauck., M. (2024). Lake Texoma Fisheries Management Plan.
Our analysis involved three main steps. First, we used modeled profiles of temperature and DO to estimate the daily volume of suitable habitat for striped bass in Lake Texoma for the period of 1995 to 2020. Second, we quantified seasonal and interannual variability in the volume of suitable habitat using a generalized additive model. Third, we fit linear regression models to quantify the relationship between three representations of suitable habitat volume under a summer temperature-oxygen squeeze and striped bass population size and condition during the following winter. These steps were conducted under three potential striped bass thermal limits (25 °C, 28 °C, and 30 °C) and DO conditions of at least 3 mg/L.