Fig 1.
Topographic map of the study area southern Alberta (CAN) and Montana (USA) where Culicoides sonorensis presence was investigated from 2002 to 2011.
Table 1.
Entomological sampling effort for the detection of Culicoides sonorensis from 2002 to 2011 in Alberta (AB, Canada) and Montana (MT, US).
In table are reported the name of the sampled area (Location), the source of information (data from T. Lysyk or G.D. Johnson), the year of collection (Year), the number of sites in each area (Nr. of sites), duration (weekly, sampling for a week; nightly, sampling for one night only), trapping effort (number of trap nights: number of traps * number of nights), and the number of sites in which C. sonorensis was recorded (positive).
Table 2.
Variables used to characterize habitat for Culicoides sonorensis presence in Alberta (Canada) and Montana (US).
Table 3.
Correlation matrix showing Pearson correlation coefficients (r) for the 10 predictors that obtained more than 1% contribution and permutation importance in the first preliminary run of MaxEnt.
Asterisks indicate highly correlated variables (r > 0.5).
Table 4.
Best models developed with MaxEnt predicting Culicoides sonorensis presence in Alberta (Canada) and Montana (US).
Table includes variables included into each best mode (A–E). Variables, percent contribution, permutation importance, corrected AIC, delta (Δ) and Akaike weights (ω) are provided for each best fit model.
Fig 2.
Response curves plotting the probability of Culicoides sonorensis against the values of top ecological predictors.
(A) vapour pressure deficit of July (; (B) standard deviation of elevation; (C) Land Cover; (D) precipitation of May. The x-axis represents the variable value and the y-axis represents the probability of Culicoides sonorensis presence as estimated by best MaxEnt model.
Fig 3.
Probability distribution for Culicoides sonorensis in southern Alberta and Montana (2010) as estimated by the best performing model (B).
Low probability classes are represented in green (0–40%) and high probability classes in orange (61–80%) and red (81–100%). The intermediate probability class (41–60%) is represented in yellow.
Table 5.
Values for best predictor variables identified by MaxEnt for areas currently above 0.6 probability threshold (60%).
Table 6.
Land cover percent area of landscape of total current (2010) area above 0.6 (60%) threshold category as identified by MaxEnt.
Fig 4.
Probability distribution maps of Culicoides sonorensis under different projected climate change scenarios (Representative Concentration Pathways, RCP).
A, left panel (2010–2030 scenarios): probability distribution map predicted by MaxEnt for 2010–2030 for each scenario. B, right panel (2030–2050 scenarios): probability distribution map predicted by MaxEnt for 2030–2050 for each RCP scenario. Each RCP is displayed (1) RCP 2.6; (2) RCP 4.5; (3) RCP 8.5.
Fig 5.
Maps of pixel change into and out of 0.6 threshold category (expressed as percentage) for Culicoides sonorensis occurrence probability as predicted with the Maximum Entropy algorithm under different climate change scenarios (Representative Concentration Pathways, RCP) projected for 2010–2030.
Each RCP is displayed (A) RCP 2.6; (B) RCP 4.5; (C) RCP 8.5. Green represents those areas moving into lesser than 0.6 probability threshold category (expressed as percentage; 60%), red represents those areas moving into the greater than 60% probability threshold category for Culicoides sonorensis occurrence.
Fig 6.
Change analysis into and out of 60% (0.6 probability) threshold category for Culicoides sonorensis occurrence probability as predicted with the Maximum Entropy algorithm under different climate change scenarios (Representative Concentration Pathways, RCP: RCP 2.6, RCP 4.5, RCP 8.5) projected for 2010–2030.
Table represents both percent change and percent area of entire study area for loss, gain, and net change in pixels for each probability class category expressed as percentage (Low-low = 0–20%; low-mod = 21–40%; Moderate = 41–60%; Mod-high = 61–80%; and high-high = 81–100% probability of occurrence). Adjacent table visually indicates gains and losses (of percent change) for each class category. Percent change is defined by the number of pixels changed for a class divided by area of a class in later image, multiplied by one hundred. Percent area change is defined by the number of pixels changed for a class, divided by the total area of the land cover map, multiplied by one hundred.
Fig 7.
Maps of pixel change analysis into and out of 0.6 threshold category (expressed as percentage) for Culicoides sonorensis occurrence probability as predicted with the Maximum Entropy algorithm under different climate change scenarios (Representative Concentration Pathways, RCP) projected for 2030–2050.
Each RCP is displayed (C) RCP 2.6; (B) RCP 4.5; (C) RCP 8.5. Green represents those areas moving into lesser than 60% probability threshold category, red represents those areas moving into the greater than 60% probability threshold category for Culicoides sonorensis occurrence.
Fig 8.
Change analysis into and out of 60% (0.6 probability) threshold category for Culicoides sonorensis occurrence probability as predicted with the Maximum Entropy algorithm under different climate change scenarios (Representative Concentration Pathways, RCP: RCP 2.6, RCP 4.5, RCP 8.5) projected for 2030–2050.
Table represents both percent change and percent area of entire study area for loss, gain, and net change in pixels for each probability class category expressed as percentage (Low-low = 0–20%; low-mod = 21–40%; Moderate = 41–60%; Mod-high = 61–80%; and high-high = 81–100% probability of occurrence). Adjacent table visually indicates gains and losses (of percent change) for each class category. Percent change is defined by the number of pixels changed for a class divided by area of a class in later image, multiplied by one hundred. Percent area change is defined by the number of pixels changed for a class, divided by the total area of the land cover map, multiplied by one hundred.