Data from: Soilscapes of mortality risk suggest a Goldilocks effect for overwintering ectotherms
Data files
Oct 04, 2024 version files 2.37 GB
-
cleanSCAN.csv
1.54 GB
-
cleanSCAN.RDS
73.06 MB
-
cleanSCANcc.RDS
737.70 MB
-
groupedPROBs.cutoff.12.RDS
11.40 KB
-
groupedPROBs.cutoff.8.RDS
11.57 KB
-
groupedPROBs.RDS
88.42 KB
-
groupedPROBs.varthresh.RDS
12.43 KB
-
lipids.fall.spring.csv
3.08 KB
-
MRtempclean.csv
1.14 KB
-
MRtempclean.RDS
769 B
-
README.md
11.39 KB
-
SCAN2047.csv
1.70 MB
-
SCAN2117.csv
1.69 MB
-
scp.csv
5.55 KB
-
SCP.RDS
1.55 KB
-
topo2.tif
17.74 MB
-
varSCAN.RDS
47.10 KB
Abstract
Changing climates are driving population declines in diverse animals worldwide. Winter conditions may play an important role in these declines but are often overlooked. Animals must not only survive winter but must also preserve body condition which is a key determinant of growing season success. We hypothesized that ectotherms overwintering in soil face a tradeoff between the risks of cold damage (including freezing) near the surface and of elevated energy use at deeper depths. To test this hypothesis, we developed landscapes of mortality risk across depth for overwintering bumble bee queens. These critical pollinators are in decline in part due to climate change, but little is known about how climate effects their overwintering mortality. We developed a mechanistic modeling approach combining measurements of freezing points and the temperature dependence of metabolic rates with hourly soil temperatures from across the US to estimate mortality risk across depth under historic conditions and under several climate change scenarios. We found that, under current climate conditions, overwintering queens face a Goldilocks effect: temperatures can be too cold at shallow depths due to substantial freezing risk, but too hot at deep depths where they risk prematurely exhausting lipid stores. Models suggest that increases in mean temperatures and in seasonal and daily temperature variation will increase risk of overwinter mortality. Our work suggests that ectotherms overwintering in the soil must select depths that balance competing risks of temperatures that are too cold or too hot. Better predictions of effects of changing climate on dormant ectotherms will require more measurements of physiological responses to temperature during dormancy across diverse taxa.
README for Waybright & Dillon Soilscapes code and data
Sarah A. Waybright* & Michael E. Dillon
Department of Zoology and Physiology and Program in Ecology
University of Wyoming, Laramie, WY 82071, USA
*Corresponding author: *swaybrig@uwyo.edu, (+12604387821)
In this study we developed a mechanistic modeling approach combining measurements of freezing points and the temperature dependence of metabolic rates with soil temperatures from across the US to estimate mortality risk across depth under historic conditions and under several climate change scenarios.
SAW and MED collaborated to write this code, while soil temperature data was sourced from the Soil Climate Analysis Network.
All below code was run with R version 4.1.2, with the following additional packages loaded: maps_3.3.0, spData_0.3.10 , raster_3.5-2 , sp_1.4-5, glmmTMB_1.1.3, mgcv_1.8-38, nlme_3.1-153, sf_1.0-2, lubridate_1.7.10, foreach_1.5.1, forcats_0.5.1, stringr_1.4.0, dplyr_1.0.7, purrr_0.3.4, readr_2.0.1, tidyr_1.1.3, tibble_3.1.3, ggplot2_3.3.5, and tidyverse_1.3.1.
Soilscapes.R: main script loading data, doing calculations, running all statistical models. Everything is run and depends on this script. Follow this script and the embedded comments for the entire workflow.
Soilscapes-Figures.R: script with code to make figures; depends on code in Soilscapes.R
cleanSCAN.csv: raw temperature data from SCAN sites after cleaning described in Supplementary materials. Includes the following columns:
depth: soil depth in cm
temp: temperature (°C)
datetime: date and time (year/month/day hours:minutes:seconds)
site: unique SCAN site names
state: state the site is located
lat: latitude (°N)
lon: longitude (°W)
alt.m: elevation of the site (meters)
obs.pre.clean: number of temperature records for each site and depth before data “cleaning”
depth_ok: column with Y (yes) or N (no) in which we determined if the depth contained enough values to be included in the analysis
notes: any additional notes about cleaning process
cleanSCANcc.RDS*: raw temperature data with additional columns of temperature data manipulated for four climate change scenarios; Note that this is a large file (~750 MB) with 5 years of temperature data for each of over 100 stations at 4 depths under 4 climate scenarios in addition to the original data (~86 million rows). Includes the following columns:
depth: soil depth in cm
datetime: date and time (year/month/day hours:minutes:seconds)
site: unique SCAN site names
state: state the site is located
lat: latitude (°N)
lon: longitude (°W)
alt.m: elevation of the site (meters)
obs.pre.clean: number of temperature records for each site and depth before data “cleaning”
obs.post.clean: number of temperature records for each site and depth after data “cleaning”
depth_ok: column with Y (yes) or N (no) in which we determined if the depth contained enough values to be included in the analysis
notes: any additional notes about cleaning process
cc.scenario: soil temperature data manipulated to reflect one of 4 climate change scenarios: mean temperatures increase by 3°C (M3A0D0), mean temperatures and annual temperature variation increase by 3°C (M3A3D0), mean temperatures and daily temperature variation increase by 3°C (M3A0D3), or mean temperatures, annual temperature variation, and daily temperature variation increase by 3°C (M3A3D3).
temp: temperature (°C)
SCP.csv: data on supercooling point measurements of Bombus from Keaveny et al., 2022, https://www.sciencedirect.com/science/article/pii/S0306456522000109. Includes the following columns:
bee.id: unique bee ID
species: species of bumble bee
caste: male, worker, or queen
date: date of experiment (day/month/year)
time.of.capture: time of capture (military time)
capture.location: location of capture
scp.type: type of supercooling point experiment (wet vs. dry)
season: season bee was captured
colony: colony name (if commercial species B. impatiens)
mass.g: mass of bee (g)
scp.c: supercooling point (°C)
ramp.rate: ramp rate (°C/min)
MRTtempclean.csv: data on metabolic rates as a function of temperature for Bombus impatiens queens. Includes the following columns:
expid: unique bee ID
settemp: set temperature (°C)
temp: temperature (°C)
VCO2m: volume of CO2 produced per min
groupedPROBs.RDS*: estimates of freezing and lipid depletion probabilities for all sites, depths, and climate change scenarios, where the dormancy transition threshold is 10°C. The code that generates these data is provided but these calculations are the most time-intensive part of the script, so the result is provided for those wanting to avoid many hours of computing time on a fast desktop computer. Includes the following columns:
site: unique SCAN site names
depth: soil depth in cm
cc.scenario: soil temperature data manipulated to reflect one of 4 climate change scenarios: mean temperatures increase by 3°C (M3A0D0), mean temperatures and annual temperature variation increase by 3°C (M3A3D0), mean temperatures and daily temperature variation increase by 3°C (M3A0D3), or mean temperatures, annual temperature variation, and daily temperature variation increase by 3°C (M3A3D3).
lat: latitude (°N)
lon: longitude (°W)
alt.m: elevation of the site (meters)
prob.freeze: predicted probability of a queen bumble bee freezing
wdays: dormancy duration (days)
gfat.600mean: predicted mean of total lipid depleted by queen bumble bee
gfat.600sd:predicted standard deviation of total lipid depleted by queen bumble bee
prob.gt18: probability of a queen bumble bee depleting excess energy reserves when she started with 20% of her body mass as lipid
prob.gt08: probability of a queen bumble bee depleting excess energy reserves when she started with 10% of her body mass as lipid
temp.mean: extracted estimate of mean temperature for temperature series
temp.tr: extracted trend of temperature for temperature series
annamp: amplitude of annual temperature variation
dayamp: amplitude of daily temperature variation
varSCAN.RDS*:
site: unique SCAN site names
depth: soil depth in cm
cc.scenario: soil temperature data manipulated to reflect one of 4 climate change scenarios: mean temperatures increase by 3°C (M3A0D0), mean temperatures and annual temperature variation increase by 3°C (M3A3D0), mean temperatures and daily temperature variation increase by 3°C (M3A0D3), or mean temperatures, annual temperature variation, and daily temperature variation increase by 3°C (M3A3D3).
temp.mean: extracted estimate of mean temperature for temperature series
temp.tr: extracted trend of temperature for temperature series
annamp: amplitude of annual temperature variation
dayamp: amplitude of daily temperature variation
SCAN2047.csv, SCAN2117.csv: example of raw temperature data before cleaning, used for Supplemental figure 2. Includes the following columns:
depth: soil depth in cm
temp: temp: temperature (°C)
date: date and time (year/month/day hours:minutes:seconds)
topo2.tiff: Topo map downloaded from US Geological Survey website, used for figures which include maps.
lipids.fall.spring.csv: total lipid measurements and wet masses of queen bumble bees collected in the Fall of 2019 and Spring of 2020 in Laramie, WY. Includes the following columns:
bee_num: unique bee ID
trt: season bee was captured
capture_date: date of capture
capture_location: location of capture
capture_mass.g: mass at capture (g)
capture_notes: additional notes about bee at capture
wet_weight_g: wet mass of bee (g)
wet_head_g: wet mass of head (g)
dry_head_g: dry mass of head (g)
wet_thorax_g: wet mass of thorax (g)
dry_thorax_g: dry mass of thorax (g)
wet_abd_g: wet mass of abdomen (g)
dry_abd_g: dry mass of abdomen (g)
total_ug_H: total mass of head (μg)
total_ug_T: total mass of thorax (μg)
total_ug_A: total mass of abdomen (μg)
Std_Curve_R2: R2 of standard curve
total.fat.g: total lipid (g)
total.dry: total wet mass (g)
total.wet : total dry mass (g)
prop.fat.dry: proportion of dry mass as fat
prop.fat.wet: proportion of wet mass as fat
groupedPROBs.cutoff.8.RDS*: estimates of freezing and lipid depletion probabilities for all sites, depths, and climate change scenarios, where the dormancy transition threshold is 8°C. The code that generates these data is provided but these calculations are the most time-intensive part of the script, so the result is provided for those wanting to avoid many hours of computing time on a fast desktop computer. Includes the following columns:
site: unique SCAN site names
depth: soil depth in cm
lat: latitude (°N)
lon: longitude (°W)
alt.m: elevation of the site (meters)
prob.freeze: predicted probability of a queen bumble bee freezing
wdays: dormancy duration (days)
gfat.600mean: predicted mean of total lipid depleted by queen bumble bee
gfat.600sd: predicted standard deviation of total lipid depleted by queen bumble bee
prob.gt08: probability of a queen bumble bee depleting excess energy reserves when she started with 10% of her body mass as lipid
groupedPROBs.cutoff.12.RDS*: estimates of freezing and lipid depletion probabilities for all sites, depths, and climate change scenarios, where the dormancy transition threshold is 12°C. The code that generates these data is provided but these calculations are the most time-intensive part of the script, so the result is provided for those wanting to avoid many hours of computing time on a fast desktop computer. Includes the following columns:
site: unique SCAN site names
depth: soil depth in cm
lat: latitude (°N)
lon: longitude (°W)
alt.m: elevation of the site (meters)
prob.freeze: predicted probability of a queen bumble bee freezing
wdays: dormancy duration (days)
gfat.600mean: predicted mean of total lipid depleted by queen bumble bee
gfat.600sd: predicted standard deviation of total lipid depleted by queen bumble bee
prob.gt08: probability of a queen bumble bee depleting excess energy reserves when she started with 10% of her body mass as lipid
var.thresh.grouped.RDS*: estimates of freezing and lipid depletion probabilities for all sites, depths, and climate change scenarios, where the dormancy transition threshold creates a counter-gradient with latitude. The code that generates these data is provided but these calculations are the most time-intensive part of the script, so the result is provided for those wanting to avoid many hours of computing time on a fast desktop computer. Includes the following columns:
site: unique SCAN site names
depth: soil depth in cm
lat: latitude (°N)
lon: longitude (°W)
alt.m: elevation of the site (meters)
prob.freeze: predicted probability of a queen bumble bee freezing
wdays: dormancy duration (days)
gfat.600mean: predicted mean of total lipid depleted by queen bumble bee
gfat.600sd:predicted standard deviation of total lipid depleted by queen bumble bee
prob.gt08: probability of a queen bumble bee depleting excess energy reserves when she started with 10% of her body mass as lipid
Data file not necessary as the calculations that generate the data are in the script but provided to save processing time.
We provide cleaned soil temperature datasets based on the Soil Climate Analysis Network (SCAN), and physiology data combined with scripts to replicate analyses presented in this manuscript. The associated README file explains folder and file contents.