Data for: Climate warming and projected loss of thermal habitat volume in lake populations of brook trout
Data files
Mar 06, 2024 version files 141.31 KB
-
Lake_Climate_Data_RCP45_30yr.csv
76.48 KB
-
Lake_Climate_Data_RCP85_30yr.csv
33.79 KB
-
Lake_Data_Brook_Trout_THV.csv
7.14 KB
-
Lakes_THV_change_RCP45.csv
6.18 KB
-
Lakes_THV_change_RCP85.csv
6.16 KB
-
README.md
8.21 KB
-
Secchi_data_Brook_Trout_THV.csv
3.34 KB
Abstract
We applied an ensemble of climate warming models and the seasonal temperature profile model for lakes (STM) to assess changes in brook trout thermal habitat volume (THV) among lakes (N=100) within a large, protected area under two climate warming scenarios, RCP 4.5 and RCP 8.5. Brook trout thermal habitat was defined as 9-17°C. Climate warming projections for the balance of this century, regardless of RCP category, will result in the loss of brook trout habitat in lakes that range widely in size. THV loss will be most extensive in lakes that are relatively shallow given their surface area. By 2071-2100 under RCP 4.5, the 90th percentile of THV loss = 31% vs. 63% under RCP 8.5. By the century's end under RCP 8.5, the protected area landscape will be a matrix of lakes with some serving as climate refugia (with reduced THV) and others having severe reductions in THV (>90th percentile THV loss).
README: Data for: Climate warming and projected loss of thermal habitat volume in lake populations of brook trout
The tables include a list of input variables into the STM model used to evaluate brook trout thermal habitat volume. See the published article for model details, the source for downloading, and a description of the lake thermal structure illustrated for an example lake = Scott Lake in Algonquin Park.
Description of the data and file structure.
Tables (*.csv) included here contain the following information:
Lake_Data_Brook_Trout_THV: The list of lakes in Algonquin Park included in evaluating the loss of brook trout thermal habitat volume (THV) using the seasonal temperature profile model (STM). Physical parameters such as perimeter, surface area, mean depth and max depth are lake morphometry variables used in STM. Lakes are mapped in Figure 1. Lake data were used in determining relative shallow vs deep lakes given their surface area in Figure 2.
Secchi_data_Brook_Trout_THV: Historic (1970s) and recent (2010s) Secchi depth data for Algonquin Park lakes. 62 lakes in this dataset are used in the evaluation of THV loss. The purpose is to estimate secchi for lakes that did not have a reading in the 2010s. Secchi reflects darkness or in recent times brownification of lakes that in turn affects the depth of thermocline, potentially. Notice how the data 'center of gravity' is below the 1:1 line in Figure 3.
Lake_Climate_Data_RCP45_30yr: Output from the ensemble climate model set for RCP4.5 for use in the STM, partitioned into the 30-year periods. For each lake and period, monthly temperature and August precipitation are averaged over each period. Lake-level data are derived from the nearest node used in spatially mapping climate change. Map of node locations are in Supplement 1.
Lake_Climate_Data_RCP85_30yr: Output from the ensemble climate model set for RCP8.5 for use in the STM, partitioned into the 30-year periods. For each lake and period, monthly temperature and August precipitation are averaged over each period. Lake-level data are derived from the nearest node used in spatially mapping climate change. Map of node locations are in Supplement 1.
Lakes_THV_change_RCP45: A summary of proportional THV loss per lake under the RCP4.5 scenario of climate warming. Results are summarized per 30-year period. This data is summarized in Figure 5. Parameters for the general extreme value fit are in Supplement 1.
Lakes_THV_change_RCP85: A summary of proportional THV loss per lake under the RCP8.5 scenario of climate warming. Resuls are summarized per 30 year period. This data is summarized in Figure 5. Parameters for the general extreme value fit are in Supplement 1.
Column parameter labels:
Lake_Data_Brook_Trout_THV
Lake Name;
Latitude;
Longitude;
Elevation = above sea level, meters;
Area = surface area, hectares;
Shoreline = total shore length, kilometers;
Depth_Max = maximum lake depth, meters;
Depth_Mean = mean lake depth, meters;
DOC = dissolved organic carbon concentration, milligrams per liter;
dif_max_min = difference between maximum depth and mean depth, meters.Secchi_data_Brook_Trout_THV
Lake Name;
Secchi depth (m) 1970s = observed maximum depth of visible Secchi disc from lake surveys in 1970s, meters;
Predicted_secchi_1970s = predicted depth of secchi disc in 1970s based on regression, meters. NA = not available because of direct observation;
Secchi (m) 2010s = observed maximum depth of visible Secchi disc from lake surveys in 2010s, meters;
Predicted_secchi_2010s = predicted depth of Secchi disc in the 2010s based on regression, meters. NA = not available because of direct observation.Lake_Climate_Data_RCP45_30yr = climate variables under RCP 4.5 scenario
Lake Name;
Period = one of three periods used to summarize warming effects, 1) 1981-2010, 2) 2011-2040, 3) 2041-2070, 4) 2071-2100;
Tjan = average January air temperature per period from ensemble model set (same for each month), Celsius;
Tfeb = average February air temperature per period from ensemble model set, Celsius;
Tmar = average March air temperature per period from ensemble model set, Celsius;
Tapr = average April air temperature per period from ensemble model set, Celsius;
Tmay = average May air temperature per period from ensemble model set, Celsius;
Tjun = average June air temperature per period from ensemble model set, Celsius;
Tjul = average July air temperature per period from ensemble model set, Celsius;
Taug = average August air temperature per period from ensemble model set, Celsius;
Tsep = average September air temperature per period from ensemble model set, Celsius;
Toct = average October air temperature per period from ensemble model set, Celsius;
Tnov = average November air temperature per period from ensemble model set, Celsius;
Tdec = average December air temperature per period from ensemble model set, Celsius;
Tann = average annual air temperature per period from ensemble model set, Celsius;
Paug = average August precipitation as rain, mm.Lake_Climate_Data_RCP85_30yr = climate variables under RCP 8.5 scenario
Lake Name;
Period = one of three periods used to summarize warming effects, 1) 1981-2010, 2) 2011-2040, 3) 2041-2070, 4) 2071-2100;
Tjan = average January air temperature per period from ensemble model set (same for each month), Celsius;
Tfeb = average February air temperature per period from ensemble model set, Celsius;
Tmar = average March air temperature per period from ensemble model set, Celsius;
Tapr = average April air temperature per period from ensemble model set, Celsius;
Tmay = average May air temperature per period from ensemble model set, Celsius;
Tjun = average June air temperature per period from ensemble model set, Celsius;
Tjul = average July air temperature per period from ensemble model set, Celsius;
Taug = average August air temperature per period from ensemble model set, Celsius;
Tsep = average September air temperature per period from ensemble model set, Celsius;
Toct = average October air temperature per period from ensemble model set, Celsius;
Tnov = average November air temperature per period from ensemble model set, Celsius;
Tdec = average December air temperature per period from ensemble model set, Celsius;
Tann = average annual air temperature per period from ensemble model set, Celsius;
Paug = average August precipitation as rain, mm.Lake_THV_change_RCP45 = proportional loss of brook trout thermal habitat volume per lake under RCP 4.5 scenario. Relative to the baseline period, 1981-2010
Lake Name;
Area_ha = lake surface area, hectares;
Vol_2SD_loss_2011_2040 = proportion loss of brook trout thermal habitat defined as the upper and lower 2SD boundaries of thermal habitat. Lower 2SD = 9-12.99 C and upper 2SD = 13-16.99 C;
Vol_2SD_loss_2041_2070 = proportion loss of brook trout thermal habitat defined as the upper and lower 2SD boundaries of thermal habitat. Lower 2SD = 9-12.99 C and upper 2SD = 13-16.99 C;
Vol_2SD_loss_2071_2100 = proportion loss of brook trout thermal habitat defined as the upper and lower 2SD boundaries of thermal habitat. Lower 2SD = 9-12.99 C and upper 2SD = 13-16.99 C;Lake_THV_change_RCP85 = proportional loss of brook trout thermal habitat volume per lake under RCP 8.5 scenario. Relative to the baseline period, 1981-2010
Lake Name;
Area_ha = lake surface area, hectares;
Vol_2SD_loss_2011_2040 = proportion loss of brook trout thermal habitat defined as the upper and lower 2SD boundaries of thermal habitat. Lower 2SD = 9-12.99 C and upper 2SD = 13-16.99 C;
Vol_2SD_loss_2041_2070 = proportion loss of brook trout thermal habitat defined as the upper and lower 2SD boundaries of thermal habitat. Lower 2SD = 9-12.99 C and upper 2SD = 13-16.99 C;
Vol_2SD_loss_2071_2100 = proportion loss of brook trout thermal habitat defined as the upper and lower 2SD boundaries of thermal habitat. Lower 2SD = 9-12.99 C and upper 2SD = 13-16.99 C;
Methods
Lake morphology data for Algonquin Park (area, shoreline length, maximum and mean lake depth) were used in the seasonal temperature profile model for lakes (STM) to evaluate how climate warming under the RCP 4.5 and RCP 8.5 scenarios reduces brook trout thermal habitat volume. An ensemble of climate models was used as input for the STM and summaries of effects on lake thermal structure are provided in 30-year periods: 1981-2010, 2011- 2040, 2041-2070, and 2071-2100. The input data are provided here for each of these periods. The Github source for the STM is provided in the article. Additional supplementary data and maps are provided as a supplement to the published article.
Usage notes
The seasonal temperature profile model for lakes (STM) was used to run this analysis. Model download sites and publications evaluating the model performance are described in the article. The distribution of changes in brook trout thermal habitat volume, relative to baseline, was modeled using general extreme value models to capture the extending right tail of each distribution in each time period. The R package fisdistPLUS was used for fitting the general extreme value distribution.