Suboptimal is good enough: Aligning thermal sensitivity to habitat temperature across season
Data files
Nov 10, 2025 version files 2.29 MB
-
averagedtemps.csv
76.89 KB
-
Dev_Size_Survival_ByDay.csv
1.32 KB
-
FalseBay1Oct22.csv
295.90 KB
-
FalseBay2Oct22.csv
295.90 KB
-
FalseBay3Oct22.csv
296.16 KB
-
FalseBay4Oct22.csv
296.07 KB
-
FalseBay5Oct22.csv
295.98 KB
-
Haminoea_ThermalSensitivity_Summary.csv
1.24 KB
-
Mechanistic_Modelling.R
7.81 KB
-
README.md
9.35 KB
-
TimetoHatch_DaystoHatch_Survival_MechanisticModel.csv
703.96 KB
-
ToptShift_MechanisticModel_OutputSummary.csv
3.17 KB
-
TPC_Fitting.R
7.08 KB
Abstract
Predicting organismal performance in changing environments is a global challenge. In ectotherms, organismal performance depends on how well-aligned the thermal sensitivities of fitness components are to body temperatures in natural habitat. It is frequently assumed that Topt (temperatures that maximize performance) for important fitness components have evolved to align with habitat temperatures. However, most organisms at temperate latitudes experience significant seasonal temperature variation, and therefore frequently operate at suboptimal temperatures. Importantly, seasonal performance patterns may differ among fitness components if those components have different thermal sensitivities. The extent to which habitat temperature can deviate from Topt before an organism can no longer persist in its environment, and the role of seasonality in driving this deviation, are infrequently explored. Here, we assess how well-aligned the thermal sensitivity (thermal performance curves, or TPCs) of development, growth, and survival are to local habitat temperatures across two seasons using embryos of the marine gastropod Haminoea vesicula. We develop a mechanistic model by integrating thermal sensitivity for development, growth, and survival with habitat temperatures to predict embryo performance across seasons. Our results demonstrate that habitat temperatures mostly fall far below Topt for development and growth. Accordingly, in the cool spring, developmental periods are 20% longer and hatchling sizes are slightly (about 1%) smaller on average than in the summer. Survival risk intensifies in the summer due to an acute high-temperature event, indicating that seasonal temperature variation affects performance differently between fitness components. We horizontally shift TPC functions to assess how well-aligned the thermal sensitivity of each fitness component is to habitat temperatures. Simulations indicate that Topt for survival is well-aligned to habitat temperature, but Topt for development and growth would need to cold-shift by 11° and 16° C, respectively, to achieve optimal performance. However, cold-shifting TPCs to maximize development and growth results in sharp survival reductions due to summer heat spikes. Overall, Topt greatly exceed the most frequent habitat temperatures, leading to suboptimal performance most of the time. However, high Topt allow embryos to survive seasonally driven heat spikes. Sometimes “suboptimal” is good enough.
Dataset DOI: 10.5061/dryad.wpzgmsc0v
Description of the data and file structure
Data Description
This dataset contains code and data files that should allow replication of the workflow to model embryo thermal performance based on habitat temperature. Field data were collected at False Bay, San Juan Island, Washington, USA, and experimental data were generated at the University of Washington's Friday Harbor Laboratories, San Juan Island, Washington, USA.
Files and variables
File: averagedtemps.csv
Description: Habitat temperatures averaged across five different sites at each hourly time window at False Bay, San Juan Island, Washington, USA. Temperatures were recorded from April - September 2022.
Variables
- Temperature: Habitat temperature (in degrees Celsius).
- Date: Date and time in PST, to be converted to a POSIXct object (datetime object in R) in associated R script.
Files: FalseBay1Oct22.csv, FalseBay2Oct22.csv, FalseBay3Oct22.csv, FalseBay4Oct22.csv, FalseBay5Oct22.csv
Description: Habitat temperatures from five different sensors at False Bay, San Juan Island, Washington, USA. Temperatures were recorded from April - September 2022.
Variables
- Temperature: Habitat temperature (in degrees Celsius).
- Date: Date and time in PST, to be converted to a POSIXct object (datetime object in R) in associated R script.
File: Dev_Size_Survival_ByDay.csv
Description: A .csv containing experimentally-generated data on size and survival across each day of embryo development for embryos reared in laboratory conditions.
Variables
- category: Categorical variable describing the temperature treatment, where "lo" = 10 degrees Celsius, "med" = 20 degrees Celsius, and "hi" = 26 degrees Celsius
- temperature: Treatment temperature (in degrees Celsius).
- day: Days since birth.
- stage: Developmental stage.
- size: Embryo size (in microns).
- survival: Proportion of embryos that are alive at each developmental day.
- se_stage: Standard error of developmental stage across replicate egg ribbons.
- se_size_prelim: Standard error of embryo size across replicate embryos and egg ribbons.
- se_size: Standard error of the mean embryo size across replicate egg ribbons.
File: Haminoea_ThermalSensitivity_Summary.csv
Description: A summary of experimentally generated values for mean embryo development, growth, and survival rates across all of embryonic development during exposure to different temperature treatments.
Variables
- Spp: String variable representing species, where "HV" = Haminoea vesicula, the focal species of the study.
- Temperature: The temperature at which embryos were experimentally reared, in degrees Celsius
- Replicate: A numerical identifier for each whole egg ribbon replicate per temperature treatment (n = 3 whole egg ribbons per temperature treatment).
- Growth Rate: Mean embryo growth rate in microns per day
- Developmental Rate: Mean embryo developmental rate in developmental stages per day.
- Survivorship: Mean survival, the % of embryos that survived to hatch
File: Mechanistic_Modelling.R
Description: R script used to generate predictions of time to hatch, size at hatch, and survival of embryos in natural habitat.
File: TPC_Fitting.R
Description: R script used to fit thermal performance curves for embryo development, growth, and survival to raw experimental data.
File: ToptShift_MechanisticModel_OutputSummary.csv
Description: A tabular summary of predicted embryo performance in habitat with cold and hot-shifted thermal performance curves. Generated by running the mechanistic model (detailed in mechanistic file R script) after shifting the position of all thermal performance curves left or right in one degree Celsius increments.
Variables
- shift: Amount that thermal performance curves have been shifted in degrees Celsius, where negative values represent cold (left) shifts, and positive values represent warm (right) shifts.
- Reprodhours: Total number of hours during the reproductive season during which embryos could be born and complete embryonic development without succumbing to heat death.
- Growthmean: Cumulative arithmetic mean of growth rates (in microns per day) of embryos born at every hour across the entire reproductive season.
- Growthscaledmean: The same as the column "Growthmean" but scaled to a maximum value of 1.
- Survivmean: Cumulative geometric mean of survival rates (as a fraction of 1) of embryos born at every hour across the entire reproductive season.
- Fitnessmean: Product of growth and survival means, not representative of true fitness, but rather a loose, exploratory estimate for embryos.
- Devmean: Cumulative arithmetic mean of developmental rates (in developmental stages per day) of embryos born at every hour across the entire reproductive season.
- daysmean: Mean time to hatch (in days) for embryos born at every hour across the entire reproductive season.
- sizemean: Mean size at hatch (in microns, measured as the longest dimension of the larval shell, from the apical tip to the dorsal/posterior end) of embryos born at every hour across the entire reproductive season.
- nodeathdaysmean: Mean time to hatch (in days) for embryos born at every hour across the entire reproductive season, excluding embryos that did not survive to hatch.
- nodeathsizemean: Mean size at hatch (in micros) for embryos born at every hour across the entire reproductive season, excluding embryos that did not survive to hatch.
File: TimetoHatch_DaystoHatch_Survival_MechanisticModel.csv
Description: A mechanistic model that combines thermal performance curves and habitat temperature in order to predict embryo days to hatch, size at hatch, and survival to hatch in natural habitat.
Variables
- Temperature: Temperature (in degrees Celsius) in egg-laying habitat, averaged across 6 sites at hourly time windows across the duration of reproductive season (April - September).
- Date: Date and time in PST, to be converted to a POSIXct object (datetime object in R) in associated R script.
- growthrate_daily: Growth rate (in microns per day) derived from aligning habitat temperature with thermal performance curve for embryo growth.
- devrate_daily: Development rate (in stages per day) derived from aligning habitat temperature with thermal performance curve for embryo development.
- survivorship: Survival (as a fraction of 1) derived from aligning habitat temperature with thermal performance curve for embryo survival.
- growthrate_hourly: Amount of growth (in microns per hour) derived from aligning habitat temperature with thermal performance curve for embryo growth.
- devrate_hourly: Developmental rate (in stages per hour) derived from aligning habitat temperature with thermal performance curve for embryo development.
- status: "hatch" if an embryo born at the time in that row will survive to hatch, and "died" if it succumbs to heat death before hatching
- days: Time to hatch (in days) for an embryo born on the date and time specified in the row.
- finalsize: Size at hatch (in microns) for an embryo born on the date and time specified in the row
- meansurvivorship: Geometric mean survival (as a fraction of 1) of all survival values at each hourly time window during embryonic development for an embryo born on the date and time specified in the row.
- meandev: Arithmetic mean developmental rate (in developmental stages per day) of development rates at each hourly time window during embryonic development for an embryo born on the date and time specified in the row.
- meangrowth: Arithmetic mean growth rate (in microns per day) of growth rates at each hourly time window during embryonic development for an embryo born on the date and time specified in the row.
- meangrowthscaled: Arithmetic mean growth rate (in microns per day) of growth rates at each hourly time window during embryonic development for an embryo born on the date and time specified in the row, scaled to a maximum value of 1.
- scaledfitness: The product of arithmetic mean growth and geometric mean survival, representing a loose, exploratory estimate of fitness for an embyro.
Code/software
Relationship between data files and scripts
Data scripts are listed in roughly the order necessary for the analysis to be replicated from starting with a data file containing habitat temperature data through generating mechanistic model outputs and summarizing outputs. We describe which files are used as input for each script, as well as the expected output.
- R script to fit TPCs to experimental data: "TPCFitting.R"
- Files used: "Haminoea_ThermalSensitivity_Summary.csv"
- R script to predict embryo time to hatch, size at hatch, and survival in natural habitat: "Mechanistic_Modelling.csv"
- Files used: "averagedtemps.csv"
- Files output: "TimetoHatch_DaystoHatch_Survival_MechanisticModel.csv"
- Summary stats of mechanistic model output: "ToptShift_MechanisticModel_OutputSummary.csv" follow R script to shift TPCs in order to generate these outputs.