BVR GLM-AED

Files for the GLM-AED (General Lake Model - Aquatic EcoDynamics) model configured for Beaverdam Reservoir, Vinton, Virginia, USA from 2015-2022 that accompany the analysis by Wander et al. "Warming air temperatures alter the timing and magnitude of reservoir zooplankton biomass." This repository includes the configuration files, driver datasets, and calibrated parameter files of the GLM-AED water quality model as well as the code used to generate the figures in the manuscript.

Repository structure

This repository contains 9 main folders: 1) /analysis contains the summarized water quality, zooplankton, and phytoplankton output from the GLM-AED model, as well as the R scripts needed to recreate all manuscript analyses and figures; 2) /calibration contains .csv files, calibrated parameters, and associated parameter ranges in each AED module that are used to perform the sensitivity analysis presented in the manuscript; 3) /field_data contains both the raw and cleaned .csv files for different water quality variables that are simulated for the reservoir; 4) /glm.app contains the model files and executables necessary to run the GLM-AED v.3.3.1 model on a local computer (only the Mac version of the model is supported); 5) /inputs contains: a) R scripts for interpolating reservoir bathymetry data to create inflow driver files, b) an .Rmd file for creating a daily water level .csv file used in the InflowPreparation.R, and c) associated .csv files necessary to run these scripts; 6) /modeling contains all of the scripts that were used for running GLM-AED, performing sensitivity analyses, calibration, validation, and visualizing preliminary model output; 7) /results contain .csv files and figures summarizing calibration results for each AED water quality module; 8) /sensitivity contains .csv files for the sensitive parameters and their ranges that were identified following the sensitivity analysis for each module; and 9) /sims contains the configuration files for the spin-up period vs. non-spin-up period for each of the climate warming scenarios presented in the manuscript (baseline, plus1, plus5, and plus10).

Instructions to reproduce figures and analyses

Most scripts that are needed to reproduce figures and analyses are located in the /analysis folder, except for step 8 below, which requires running part of a script in the /field_data folder.

1. Run 01_bvr_glm_ms_figs.R to visualize modeled vs. observed water quality variables at the surface (Fig. 2 and Fig. S5) and bottom (Fig. S6) of Beaverdam Reservoir, boxplots of water quality variables for baseline and warming temperature scenarios in the surface (Fig. 4), and boxplots of bottom water quality variables (Fig. S7).
2. Run 02_zoop_scenarios.R to visualize modeled zooplankton density for cladocerans, copepods, and rotifers from 2015-2022 (Fig. 3), zooplankton taxon biomass for baseline conditions vs. the plus5 scenario (Fig. 6), relative zooplankton biomass for all scenarios and functional groups (Fig. S12), stacked annual zooplankton biomass bar plot (Fig. S8), smoothed monthly zooplankton biomass for each scenario (Fig. 7) and across each year (Fig. S17), boxplots of mean zooplankton biomass for each scenario and taxon (Fig. 5), lineplots of mean annual zooplankton biomass for each scenario (Fig. S11), lineplots of mean zooplankton diagnostic rates and area plots of phytoplankton relative biomass for each taxon in baseline conditions vs. the plus5 scenario (Fig. 9), lineplots of mean annual phytoplankton biomass for each scenario and taxon (Fig. S16), phytoplankton raw biomass for each scenario from 2015-2022 (Fig. S13), mean day of year analyses to calculate the maximum zooplankton biomass for each taxon and scenario (Fig. 8), and lineplots summarizing the changes in zooplankton and phytoplankton monthly biomass across scenarios (Fig. S18). Additionally, this script runs the Kruskal-Wallis tests across scenarios for each taxon for magnitude and biomass, as well as the associated post-hoc analyses for these tests (Tables S7-8).
3. Run 03_GOF_tables.R to generate goodness-of-fit tables for water quality variables aggregated across the full water column (Table S3), only for the surface (Table 1), and for zooplankton taxa aggregated across the full water column (Table S6).
4. Run 04_zoop_par_sensitivity.R to run the 10% parameter sensitivity analysis for several zooplankton parameters that differ among taxa and generate supplemental Figures S4, S14, and S15. Additionally, this code can be used to run the cyanobacteria sensitivity analysis (Fig. S3).
5. Run 05_temperature_sensitivity.R to plot the grazing temperature sensitivity function and exponential respiration and mortality function for all zooplankton taxa (Fig. S19).
6. Run 06_maca_rcp_temp.R to generate a table that summarizes the mean difference in air temperature between historical mean temperatures vs. RCP 4.5 and 8.5 projected temperatures (Table S4).
7. Run code from lines 34-78 in run_glm_scenarios.R to plot surface water temperature over the simulation period from 2015-2022 (Fig. S9) and lines 121-234 to plot stratification duration for each scenario (Fig. S10).
8. Run code from lines 517-684 in field_data/FieldDataPrep.R to generate size and biomass distributions for all zooplankton taxa (Fig. S2).
9. Run ROA_annual_temp.R to generate mean annual air temperature plot from the Roanoke Airport meteorological station (Fig. S1).

Note that there were several other scripts used to create files used in this analysis (e.g., plankton_aggregations.R, zoop_taxa_diagnostics.R, plankton_diagnostics.R). Additional code is provided to visualize limiting variables for each of the three phytoplankton taxa across scenarios (phyto_limiting_vars.R) and create glm nml files and meteorological files necessary to run a 15-year spin-up period in the model and visualize output across scenarios (spin-up.R). These scripts do not need to be run to recreate manuscript figures or analyses, but are provided for reference.