Data and code from: Climate driven changes in wetland hydroperiods predict losses in habitat suitability for amphibian breeding
Data files
Jun 24, 2026 version files 232.22 MB
-
Dare.et.al.2026.zip
232.22 MB
-
README.md
4.43 KB
Abstract
This data and code are for the manuscript titled "Climate driven changes in wetland hydroperiods predict losses in habitat suitability for amphibian breeding" by Dare et al. 2026. Here we combine hydrometric model outputs and occupancy modelling to produce estimates of future breeding habitat loss for three subalpine species due to changes in hydrology. R scripts 1 to 2 predict the mean hydroperiod for our 488 sites. R scripts 3-4 run our occupancy models for each species and produce estimates of historic and future breeding occupancy. R scripts 5-8 summarize the results in a series of figures and tables seen in the manuscript.
Dataset DOI: 10.5061/dryad.9cnp5hr0r
Description of the data and file structure
Files and variables
File: Dare.et.al.2026.zip
Description: Zipped folder containing all R scripts and required data to rerun the analysis for Dare et al. 2026 in Ecosphere.
This code and dataset combine occupancy and hydrology models to estimate historic and future amphibian breeding habitat for three species in the Pacific Northwest of the USA. The first two scripts use daily soil moisture estimates (both historical and under a 2080s climate) over a 66-year period and empirical measurements of wetland depth to predict hydroperiod for each of our 488 sites for both time periods. These soil moisture estimates are outputs of a variable infiltration capacity model (VIC), which was previously implemented over the Mountain West at 1/16th degree resolution. The remaining scripts run occupancy models for each species, predict occupancy for each site for both time periods, and estimate the loss of breeding habitat based on changes in hydroperiod between the two time periods.
R Script Summary
1. Thresh@0DepthModel.R
Takes the soil moisture at 0 depth for each site in Lee et al.’s 2015 dataset and runs a linear model with GIS-derived and field variables. Computes a model average that is used to predict the soil moisture at 0 depth at the remaining sites.
2. HistFutHydroperiodEstimate.R
Uses the daily estimates of soil moisture (historical and perturbed to future conditions) found in the fluxes folder from the variable infiltration capacity model, paired with the estimated soil moisture at 0 depth, to estimate a mean hydroperiod length for each site. For each day (row) in the flux file, if the soil moisture is below the soil moisture threshold at 0 depth, the waterbody is deemed dry that day, and vice versa if above the threshold.
3. OccuModels.R
Combines presence-absence data from multiple sources, as well as hydroperiod, to run the occupancy models for each species. The script for the models themselves is ‘3.1.jags.model.R’.
4. OccupancyPredictions.R
Plugs in the hydroperiod estimations for each time period into the model equations to get estimates of future and historic occupancy for each species.
5. Figure4.R
Summarises the site occupancy and hydroperiod data into figures showing breeding site loss and how our conclusions change based on different occupancy cutoffs and species development times.
6. Figure 5.R
Breaks down the results spatially to show the differences in losses across the study area and between species.
7. Figure 3.R
Produces coefficient plots for the occupancy models.
8. Figure S1.R
Produces a figure of species response to the quadratic effect of hydroperiod.
Data Summary
2012.2013.Combined.Breeding.Binary.csv- observations and visit covariates from the 2012-2013 data collection effort.
2019.occu.syle.csv- observations and visit covariates from the 2019 data collection effort.
Armi.breeding.csv- observations and visit covariates from the ARMI dataset in Olympic National Park.
armi.mora.breeding.csv- observations and visit covariates from the ARMI dataset in Mount Rainier National Park.
all.vic.regression.1950.csv- Lee et al.’s (2015) 76 sites with site level data, hydroperiod estimations, and soil moisture at 0 depth.
Site csv’s- All files that end in 'site'. Site-level data that accompanies each of the observation datasets.
fluxes.csv- A table of every site and the appropriate VIC file (flux) for that site, based which 1/16th degree VIC cell the site falls within.
regression_results_Lee_et_al._2015.csv- A summary of Lee et al.’s (2015) hydroperiod models for each of their 76 sites.
slope.aspect.csv- A table of slope and aspect values for each of our sites.
Flux folder- contains the variable infiltration capacity (VIC) model outputs across 47 different gridded cells (1/16th degree size) over the study area, separated into (fut) future (2080s) and historic (hist) outputs by subfolder. Each file represents a different cell, and within each file, rows are dates and columns are different variables from the VIC model, including soil moisture.
Code/software
R, JAGS, and WinBUGS are the only software required.