The magnitude of selection on growth varies among years and increases under warming conditions in a subarctic seabird

Abstract Because of ongoing rapid climate change, many ecosystems are becoming both warmer and more variable, and these changes are likely to alter the magnitude and variability of natural selection acting on wild populations. Critically, changes and fluctuations in selection can impact both population demography and evolutionary change. Therefore, predicting the impacts of climate change depends on understanding the magnitude and variation in selection on traits across different life stages and environments. Long-term experiments in wild settings are a great opportunity to determine the impact of environmental conditions on selection. Here we examined variability in the strength of selection on size traits of nestling black-legged kittiwakes (Rissa tridactyla) in a 25-year study including a food supplementation experiment on Middleton Island in the Gulf of Alaska. Using mixed effect models, we examined the annual variability of stage-specific and resource-specific selection gradients across 25 years. We found that (a) larger and heavier hatchlings were the most likely to survive during early ontogeny, (b) non-food supplemented younger nestlings in a brood experienced the strongest selection, and (c) warmer conditions increased the magnitude of selection on nestling mass and affected non-food supplemented and second-hatched nestlings the most. Our results suggested that variable resource dynamics likely caused some of the changes in selection from year to year and that warming conditions increased the strength of selection on subarctic seabird growth. However, our experimental manipulation revealed that local environmental heterogeneity could buffer the selection expected from broader climatic changes. Consequently, understanding the interactive effects of local conditions and general changes in climate seems likely to improve our ability to predict future selection gradients.


Fitness Functions
As noted in the main text we contrasted five different fitness function shapes at each survival stage.Following de Villemereuil et al., (2020) we used a logit link function for our flat, linear, and planar fitness functions (eqs. 1a, 1b, & 1c) but used an exponential for the Gaussian or bivariate Gaussian fitness peak and a logit link for the   (maximum fitness) parameter (eqs 1d &1e).In each of the below equations the survival probability is modelled as a function of either one trait (mass or wing length; ).or two traits (mass & wing length;  1 &  2 ).The below equations correspond to the fitness functions described in the text () =  (𝑎) 1a () = ( + ) 1b ( 1 ,  2 ) = ( +  1  1 +  2  2 ) 1c ] where  is the intercept (Eq.1a, b, c),  is the selection differential (Eq.1b) or gradient (Eq.1c),   is the maximum fitness,  is the optimal mass or wing length (Eqs.1d, e),  is the width of the fitness function (Eqs.1d, e), and  is the correlation between mass and wing length (Eq.1e).95% credible interval (l-95% CI), and upper 95% credible interval (u-95% CI) are reported for the posterior distribution of that parameter.Rhat is an indicate of convergence of chains and if it is considerably greater than 1 the chains have not converged.The Bulk effective sampling size (Bulk_ESS) is a diagnostic for sampling efficiency of the bulk of the posterior and the tail effective sampling size (Tail_ESS) is a diagnostic for sampling efficiency of the tails of the posterior(see Bürkner, 2017;Vehtari et al., 2017).Parameters under the "Population-Level" heading indicate fitness function parameters estimated across the entire dataset.Parameters under the "Group Level" heading indicate variation (standard deviation) in annual mean deviations from the population level parameters.The different parameters estimated are the maximum fitness Wmax, the trait specific optimum θ, the correlation between wing length and mass ρ, and the width of the fitness function for each trait ω.The suffices indicate the treatment and nestling rank level that a parameter was estimated at relative to the first-hatched food supplemented nestling group.For example, the optimal mass for a non-food supplemented first-hatched nestling would be sum of θ_Mass_First-hatched_Fed and θ_ Mass_Treatmentunfed.If there is no suffix for a parameter the same parameter was assumed for all food treatment and rank groupings.95% credible interval (l-95% CI), and upper 95% credible interval (u-95% CI) are reported for the posterior distribution of that parameter.Rhat is an indicate of convergence of chains and if it is considerably greater than 1 the chains have not converged.The Bulk effective sampling size (Bulk_ESS) is a diagnostic for sampling efficiency of the bulk of the posterior and the tail effective sampling size (Tail_ESS) is a diagnostic for sampling efficiency of the tails of the posterior(see Bürkner, 2017;Vehtari et al., 2017).Parameters under the "Population-Level" heading indicate fitness function parameters estimated across the entire dataset.Parameters under the "Group Level" heading indicate variation (standard deviation) in annual mean deviations from the population level parameters.The different parameters estimated are the maximum fitness Wmax, the trait specific optimum θ, the correlation between wing length and mass ρ, and the width of the fitness function for each trait ω.The suffices indicate the treatment and nestling rank level that a parameter was estimated at relative to the first-hatched food supplemented nestling group.For example, the optimal mass for a non-food supplemented first-hatched nestling would be sum of θ_Mass_First-hatched_Fed and θ_ Mass_Treatmentunfed.If there is no suffix for a parameter the same parameter was assumed for all food treatment and rank groupings.An "AirWin", "SSTWin", "pdo", or "pdo_lagged" suffix indicates the estimate of the linear relationship between a function parameter and our chosen air temperature, sea-surface temperature, Pacific Decadal Oscillation, or lagged Pacific Decadal Oscillation parameters.95% credible interval (l-95% CI), and upper 95% credible interval (u-95% CI) are reported for the posterior distribution of that parameter.Rhat is an indicate of convergence of chains and if it is considerably greater than 1 the chains have not converged.The Bulk effective sampling size (Bulk_ESS) is a diagnostic for sampling efficiency of the bulk of the posterior and the tail effective sampling size (Tail_ESS) is a diagnostic for sampling efficiency of the tails of the posterior(see Bürkner, 2017;Vehtari et al., 2017).Parameters under the "Population-Level" heading indicate fitness function parameters estimated across the entire dataset.Parameters under the "Group Level" heading indicate variation (standard deviation) in annual mean deviations from the population level parameters.The different parameters estimated are the fitness intercept , and the selection gradient ."_Mass" or "_Wing" following the parameter indicate the estimate for mass or wing length, respectively.The suffixes indicate the treatment and nestling rank level that a parameter was estimated at relative to the first-hatched food supplemented nestling group.For example, the selection gradeintl on mass for a non-food supplemented first-hatched nestling would be sum of _Mass_First-hatched_Fed and _Mass_Treatmentunfed.If there is no suffix for a parameter, the same parameter value was estimated for all food treatment and rank groupings.A "AirWin", "SSTWin", "pdo", or "pdo_lagged" suffix indicates the estimate of the linear relationship between a function parameter and our chosen air temperature, sea-surface temperature, Pacific Decadal Oscillation, or lagged Pacific Decadal Oscillation parameters.95% credible interval (l-95% CI), and upper 95% credible interval (u-95% CI) are reported for the posterior distribution of that parameter.Rhat is an indicate of convergence of chains and if it is considerably greater than 1 the chains have not converged.The Bulk effective sampling size (Bulk_ESS) is a diagnostic for sampling efficiency of the bulk of the posterior and the tail effective sampling size (Tail_ESS) is a diagnostic for sampling efficiency of the tails of the posterior(see Bürkner, 2017;Vehtari et al., 2017).Parameters under the "Population-Level" heading indicate fitness function parameters estimated across the entire dataset.Parameters under the "Group Level" heading indicate variation (standard deviation) in annual mean deviations from the population level parameters.The different parameters estimated are the fitness intercept , and the selection differential .The suffixes indicate the treatment and nestling rank level that a parameter was estimated at relative to the first-hatched food supplemented nestling group.For example, the selection differential on mass for a non-food supplemented first-hatched nestling would be sum of _First-hatched_Fed and _Treatmentunfed.If there is no suffix for a parameter the same parameter value was estimated for all food treatment and rank groupings.A "AirWin", "SSTWin", "pdo", or "pdo_lagged" suffix indicates the estimate of the linear relationship between a function parameter and our chosen air temperature, seasurface temperature, Pacific Decadal Oscillation, or lagged Pacific Decadal Oscillation parameters.

Age 35 Annual Fluctuations
Group

Table 1 : ΔLOOIC values for fitness functions evaluated for survival from age 0 to age 5 as a function of mass and/or wing length
. ΔLOOIC values are the differences between LOOIC values of the best model (lowest LOOIC value) and LOOIC value for a given model.Models with a ΔLOOIC <5 are bolded.

Table 2 : ΔLOOIC values for fitness functions evaluated for survival from age 5 to age 10 as a function of mass and/or wing length
. ΔLOOIC values are the differences between LOOIC values of the best model (lowest LOOIC value) and LOOIC value for a given model.Models with a ΔLOOIC <5 are bolded.

Table 3 : ΔLOOIC values for fitness functions evaluated for survival from age 10 to age 15 as a function of mass and/or wing length
. ΔLOOIC values are the differences between LOOIC values of the best model (lowest LOOIC value) and LOOIC value for a given model.Models with a ΔLOOIC <5 are bolded.

Table 4 : ΔLOOIC values for fitness functions evaluated for survival from age 15 to age 20 as a function of mass and/or wing length
. ΔLOOIC values are the differences between LOOIC values of the best model (lowest LOOIC value) and LOOIC value for a given model.Models with a ΔLOOIC <5 are bolded.

Table 5 : ΔLOOIC values for fitness functions evaluated for survival from age 20 to age 25 as a function of mass and/or wing length
. ΔLOOIC values are the differences between LOOIC values of the best model (lowest LOOIC value) and LOOIC value for a given model.Models with a ΔLOOIC <5 are bolded.

Table 6 : ΔLOOIC values for fitness functions evaluated for survival from age 25 to age 30 as a function of mass and/or wing length
. ΔLOOIC values are the differences between LOOIC values of the best model (lowest LOOIC value) and LOOIC value for a given model.Models with a ΔLOOIC <5 are bolded.

Table 7 : ΔLOOIC values for fitness functions evaluated for survival from age 30 to age 35 as a function of mass and/or wing length
. ΔLOOIC values are the differences between LOOIC values of the best model (lowest LOOIC value) and LOOIC value for a given model.Models with a ΔLOOIC <5 are bolded.

Table 8 : ΔLOOIC values for fitness functions evaluated for survival from age 35 to age 40 as a function of mass and/or wing length
. ΔLOOIC values are the differences between LOOIC values of the best model (lowest LOOIC value) and LOOIC value for a given model.Models with a ΔLOOIC <5 are bolded.

Table 9 : ΔLOOIC values for fitness functions evaluated for survival from age 40 to fledging as a function of mass and/or wing length
. ΔLOOIC values are the differences between LOOIC values of the best model (lowest LOOIC value) and LOOIC value for a given model.Models with a ΔLOOIC <5 are bolded.

Supplementary Table 10: Top fitness function for survival from age 0 to age 5 based on ΔLOOIC comparisons.
For each model parameter the mean (Estimate), standard deviation (Est.Error), lower

Supplementary Table 12: Top fitness function for survival from age 10 to age 15 based on ΔLOOIC comparisons. For
parameters estimated across the entire dataset.Parameters under the "Group Level" heading indicate variation (standard deviation) in annual mean deviations from the population level parameters.The different parameters estimated are the fitness intercept , and the selection gradient ."_Mass" or "_Wing" following the parameter indicate the estimate for mass or wing length, respectively.The suffixes indicate the treatment and nestling rank level that a parameter was estimated at relative to the first-hatched food supplemented nestling group.For example, the selection differential on mass for a non-food supplemented first-hatched nestling would be sum of _First-hatched_Fed and _Treatmentunfed.If there is no suffix for a parameter, the same parameter value was estimated for all food treatment and rank groupings.

Table 13 : Top fitness function for survival from age 15 to age 20 based on ΔLOOIC comparisons.
Bürkner, 2017;Vehtari et al., 2017)Estimate), standard deviation (Est.Error), lower 95% credible interval (l-95% CI), and upper 95% credible interval (u-95% CI) are reported for the posterior distribution of that parameter.Rhat is an indicate of convergence of chains and if it is considerably greater than 1 the chains have not converged.The Bulk effective sampling size (Bulk_ESS) is a diagnostic for sampling efficiency of the bulk of the posterior and the tail effective sampling size (Tail_ESS) is a diagnostic for sampling efficiency of the tails of the posterior(seeBürkner, 2017;Vehtari et al., 2017).Parameters under the "Population-Level" heading indicate fitness function parameters estimated across the entire dataset.Parameters under the "Group Level" heading indicate variation (standard deviation) in annual mean deviations from the population level parameters.The different parameters estimated are the fitness intercept , and the selection gradient ."_Mass" or "_Wing" following the parameter indicate the estimate for mass or wing length, respectively.The suffixes indicate the treatment and nestling rank level that a parameter was estimated at relative to the first-hatched food supplemented nestling group.For example, the selection differential on mass for a non-food supplemented first-hatched nestling would be sum of _First-hatched_Fed and _Treatmentunfed.If there is no suffix for a parameter, the same parameter value was estimated for all food treatment and rank groupings.

Table 14 : Top fitness function for survival from age 20 to age 25 based on ΔLOOIC comparisons.
Bürkner, 2017;Vehtari et al., 2017)Estimate), standard deviation (Est.Error), lower 95% credible interval (l-95% CI), and upper 95% credible interval (u-95% CI) are reported for the posterior distribution of that parameter.Rhat is an indicate of convergence of chains and if it is considerably greater than 1 the chains have not converged.The Bulk effective sampling size (Bulk_ESS) is a diagnostic for sampling efficiency of the bulk of the posterior and the tail effective sampling size (Tail_ESS) is a diagnostic for sampling efficiency of the tails of the posterior(seeBürkner, 2017;Vehtari et al., 2017).Parameters under the "Population-Level" heading indicate fitness function parameters estimated across the entire dataset.Parameters under the "Group Level" heading indicate variation (standard deviation) in annual mean deviations from the population level parameters.The different parameters estimated are the fitness intercept , and the selection differential .The suffixes indicate the treatment and nestling rank level that a parameter was estimated at relative to the first-hatched food supplemented nestling group.For example, the selection differential on mass for a non-food supplemented first-hatched nestling would be sum of _First-hatched_Fed and _Treatmentunfed.If there is no suffix for a parameter, the same parameter value was estimated for all food treatment and rank groupings.

Table 15 : Top fitness function for survival from age 25 to age 30 based on ΔLOOIC comparisons.
Bürkner, 2017;Vehtari et al., 2017)Estimate), standard deviation (Est.Error), lower 95% credible interval (l-95% CI), and upper 95% credible interval (u-95% CI) are reported for the posterior distribution of that parameter.Rhat is an indicate of convergence of chains and if it is considerably greater than 1 the chains have not converged.The Bulk effective sampling size (Bulk_ESS) is a diagnostic for sampling efficiency of the bulk of the posterior and the tail effective sampling size (Tail_ESS) is a diagnostic for sampling efficiency of the tails of the posterior(seeBürkner, 2017;Vehtari et al., 2017).Parameters under the "Population-Level" heading indicate fitness function parameters estimated across the entire dataset.Parameters under the "Group Level" heading indicate variation (standard deviation) in annual mean deviations from the population level parameters.The different parameters estimated are the fitness intercept , and the selection differential .The suffixes indicate the treatment and nestling rank level that a parameter was estimated at relative to the first-hatched food supplemented nestling group.For example, the selection differential on mass for a non-food supplemented first-hatched nestling would be sum of _First-hatched_Fed and _Treatmentunfed.If there is no suffix for a parameter, the same parameter value was estimated for all food treatment and rank groupings.

Table 16 : Top fitness function for survival from age 30 to age 35 based on ΔLOOIC comparisons.
Bürkner, 2017;Vehtari et al., 2017)Estimate), standard deviation (Est.Error), lower 95% credible interval (l-95% CI), and upper 95% credible interval (u-95% CI) are reported for the posterior distribution of that parameter.Rhat is an indicate of convergence of chains and if it is considerably greater than 1 the chains have not converged.The Bulk effective sampling size (Bulk_ESS) is a diagnostic for sampling efficiency of the bulk of the posterior and the tail effective sampling size (Tail_ESS) is a diagnostic for sampling efficiency of the tails of the posterior(seeBürkner, 2017;Vehtari et al., 2017).Parameters under the "Population-Level" heading indicate fitness function parameters estimated across the entire dataset.Parameters under the "Group Level" heading indicate variation (standard deviation) in annual mean deviations from the population level parameters.The different parameters estimated are the fitness intercept , and the selection differential .The suffixes indicate the treatment and nestling rank level that a parameter was estimated at relative to the first-hatched food supplemented nestling group.For example, the selection differential on mass for a non-food supplemented first-hatched nestling would be sum of _First-hatched_Fed and _Treatmentunfed.If there is no suffix for a parameter, the same parameter value was estimated for all food treatment and rank groupings.

Table 17 : Top fitness function for survival from age 35 to age 40 based on ΔLOOIC comparisons.
Bürkner, 2017;Vehtari et al., 2017)Estimate), standard deviation (Est.Error), lower 95% credible interval (l-95% CI), and upper 95% credible interval (u-95% CI) are reported for the posterior distribution of that parameter.Rhat is an indicate of convergence of chains and if it is considerably greater than 1 the chains have not converged.The Bulk effective sampling size (Bulk_ESS) is a diagnostic for sampling efficiency of the bulk of the posterior and the tail effective sampling size (Tail_ESS) is a diagnostic for sampling efficiency of the tails of the posterior(seeBürkner, 2017;Vehtari et al., 2017).Parameters under the "Population-Level" heading indicate fitness function parameters estimated across the entire dataset.Parameters under the "Group Level" heading indicate variation (standard deviation) in annual mean deviations from the population level parameters.The different parameters estimated are the fitness intercept , and the selection differential .The suffixes indicate the treatment and nestling rank level that a parameter was estimated at relative to the first-hatched food supplemented nestling group.For example, the selection differential on mass for a non-food supplemented first-hatched nestling would be sum of _First-hatched_Fed and _Treatmentunfed.If there is no suffix for a parameter, the same parameter value was estimated for all food treatment and rank groupings

Table 18 : Top fitness function for survival from age 40 to age fledging based on ΔLOOIC comparisons.
Bürkner, 2017;Vehtari et al., 2017)Estimate), standard deviation (Est.Error), lower 95% credible interval (l-95% CI), and upper 95% credible interval (u-95% CI) are reported for the posterior distribution of that parameter.Rhat is an indicate of convergence of chains and if it is considerably greater than 1 the chains have not converged.The Bulk effective sampling size (Bulk_ESS) is a diagnostic for sampling efficiency of the bulk of the posterior and the tail effective sampling size (Tail_ESS) is a diagnostic for sampling efficiency of the tails of the posterior(seeBürkner, 2017;Vehtari et al., 2017).Parameters under the "Population-Level" heading indicate fitness function parameters estimated across the entire dataset.Parameters under the "Group Level" heading indicate variation (standard deviation) in annual mean deviations from the population level parameters.The different parameters estimated are the fitness intercept , and the selection differential .The suffixes indicate the treatment and nestling rank level that a parameter was estimated at relative to the first-hatched food supplemented nestling group.For example, the selection differential on mass for a non-food supplemented first-hatched nestling would be sum of _First-hatched_Fed and _Treatmentunfed.If there is no suffix for a parameter, the same parameter value was estimated for all food treatment and rank groupings

Supplementary Table 29: Top fitness function for survival from age 0 to age 5 based on ΔLOOIC comparisons.
For each model parameter the mean (Estimate), standard deviation (Est.Error), lower

Table 30 : Top fitness function for survival from age 5 to age 10 based on ΔLOOIC comparisons.
Bürkner, 2017;Vehtari et al., 2017)Estimate), standard deviation (Est.Error), lower 95% credible interval (l-95% CI), and upper 95% credible interval (u-95% CI) are reported for the posterior distribution of that parameter.Rhat is an indicate of convergence of chains and if it is considerably greater than 1 the chains have not converged.The Bulk effective sampling size (Bulk_ESS) is a diagnostic for sampling efficiency of the bulk of the posterior and the tail effective sampling size (Tail_ESS) is a diagnostic for sampling efficiency of the tails of the posterior(seeBürkner, 2017;Vehtari et al., 2017).Parameters under the "Population-Level" heading indicate fitness function parameters estimated across the entire dataset.Parameters under the "Group Level" heading indicate variation (standard deviation) in annual mean deviations from the population level parameters.The different parameters estimated are the fitness intercept , and the selection differential .The suffixes indicate the treatment and nestling rank level that a parameter was estimated at relative to the first-hatched food supplemented nestling group.For example, the selection differential on mass for a non-food supplemented first-hatched nestling would be sum of _First-hatched_Fed and _Treatmentunfed.If there is no suffix for a parameter the same parameter value was estimated for all food treatment and rank groupings.A "AirWin", "SSTWin", "pdo", or "pdo_lagged" suffix indicates the estimate of the linear relationship between a function parameter and our chosen air temperature, seasurface temperature, Pacific Decadal Oscillation, or lagged Pacific Decadal Oscillation parameters.

Table 31 : Top fitness function for survival from age 10 to age 15 based on ΔLOOIC comparisons.
For each model parameter the mean (Estimate), standard deviation (Est.Error), lower

Table 32 : Top fitness function for survival from age 15 to age 20 based on ΔLOOIC comparisons.
Bürkner, 2017;Vehtari et al., 2017)Estimate), standard deviation (Est.Error), lower 95% credible interval (l-95% CI), and upper 95% credible interval (u-95% CI) are reported for the posterior distribution of that parameter.Rhat is an indicate of convergence of chains and if it is considerably greater than 1 the chains have not converged.The Bulk effective sampling size (Bulk_ESS) is a diagnostic for sampling efficiency of the bulk of the posterior and the tail effective sampling size (Tail_ESS) is a diagnostic for sampling efficiency of the tails of the posterior(seeBürkner, 2017;Vehtari et al., 2017).Parameters under the "Population-Level" heading indicate fitness function parameters estimated across the entire dataset.Parameters under the "Group Level" heading indicate variation (standard deviation) in annual mean deviations from the population level parameters.The different parameters estimated are the fitness intercept , and the selection gradient ."_Mass" or "_Wing" following the parameter indicate the estimate for mass or wing length, respectively.The suffixes indicate the treatment and nestling rank level that a parameter was estimated at relative to the first-hatched food supplemented nestling group.For example, the selection gradient on mass for a non-food supplemented first-hatched nestling would be sum of _Mass_First-hatched_Fed and _Mass_Treatmentunfed.If there is no suffix for a parameter, the same parameter value was estimated for all food treatment and rank groupings.A "AirWin", "SSTWin", "pdo", or "pdo_lagged" suffix indicates the estimate of the linear relationship between a function parameter and our chosen air temperature, sea-surface temperature, Pacific Decadal Oscillation, or lagged Pacific Decadal Oscillation parameters.

Table 33 : Top fitness function for survival from age 20 to age 25 based on ΔLOOIC comparisons.
Bürkner, 2017;Vehtari et al., 2017)Estimate), standard deviation (Est.Error), lower 95% credible interval (l-95% CI), and upper 95% credible interval (u-95% CI) are reported for the posterior distribution of that parameter.Rhat is an indicate of convergence of chains and if it is considerably greater than 1 the chains have not converged.The Bulk effective sampling size (Bulk_ESS) is a diagnostic for sampling efficiency of the bulk of the posterior and the tail effective sampling size (Tail_ESS) is a diagnostic for sampling efficiency of the tails of the posterior(seeBürkner, 2017;Vehtari et al., 2017).Parameters under the "Population-Level" heading indicate fitness function parameters estimated across the entire dataset.Parameters under the "Group Level" heading indicate variation (standard deviation) in annual mean deviations from the population level parameters.The different parameters estimated are the fitness intercept , and the selection gradient ."_Mass" or "_Wing" following the parameter indicate the estimate for mass or wing length, respectively.The suffixes indicate the treatment and nestling rank level that a parameter was estimated at relative to the first-hatched food supplemented nestling group.For example, the selection differential on mass for a non-food supplemented first-hatched nestling would be sum of _First-hatched_Fed and _Treatmentunfed.If there is no suffix for a parameter, the same parameter value was estimated for all food treatment and rank groupings.A "AirWin", "SSTWin", "pdo", or "pdo_lagged" suffix indicates the estimate of the linear relationship between a function parameter and our chosen air temperature, seasurface temperature, Pacific Decadal Oscillation, or lagged Pacific Decadal Oscillation parameters.

Table 34 : Top fitness function for survival from age 25 to age 30 based on ΔLOOIC comparisons.
Bürkner, 2017;Vehtari et al., 2017)Estimate), standard deviation (Est.Error), lower 95% credible interval (l-95% CI), and upper 95% credible interval (u-95% CI) are reported for the posterior distribution of that parameter.Rhat is an indicate of convergence of chains and if it is considerably greater than 1 the chains have not converged.The Bulk effective sampling size (Bulk_ESS) is a diagnostic for sampling efficiency of the bulk of the posterior and the tail effective sampling size (Tail_ESS) is a diagnostic for sampling efficiency of the tails of the posterior(seeBürkner, 2017;Vehtari et al., 2017).Parameters under the "Population-Level" heading indicate fitness function parameters estimated across the entire dataset.Parameters under the "Group Level" heading indicate variation (standard deviation) in annual mean deviations from the population level parameters.The different parameters estimated are the fitness intercept , and the selection differential .The suffixes indicate the treatment and nestling rank level that a parameter was estimated at relative to the first-hatched food supplemented nestling group.For example, the selection differential on mass for a non-food supplemented first-hatched nestling would be sum of _First-hatched_Fed and _Treatmentunfed.If there is no suffix for a parameter the same parameter value was estimated for all food treatment and rank groupings.A "AirWin", "SSTWin", "pdo", or "pdo_lagged" suffix indicates the estimate of the linear relationship between a function parameter and our chosen air temperature, seasurface temperature, Pacific Decadal Oscillation, or lagged Pacific Decadal Oscillation parameters.

Table 35 : Top fitness function for survival from age 30 to age 35 based on ΔLOOIC comparisons.
Bürkner, 2017;Vehtari et al., 2017)Estimate), standard deviation (Est.Error), lower 95% credible interval (l-95% CI), and upper 95% credible interval (u-95% CI) are reported for the posterior distribution of that parameter.Rhat is an indicate of convergence of chains and if it is considerably greater than 1 the chains have not converged.The Bulk effective sampling size (Bulk_ESS) is a diagnostic for sampling efficiency of the bulk of the posterior and the tail effective sampling size (Tail_ESS) is a diagnostic for sampling efficiency of the tails of the posterior(seeBürkner, 2017;Vehtari et al., 2017).Parameters under the "Population-Level" heading indicate fitness function parameters estimated across the entire dataset.Parameters under the "Group Level" heading indicate variation (standard deviation) in annual mean deviations from the population level parameters.The different parameters estimated are the fitness intercept , and the selection differential .The suffixes indicate the treatment and nestling rank level that a parameter was estimated at relative to the first-hatched food supplemented nestling group.For example, the selection differential on mass for a non-food supplemented first-hatched nestling would be sum of _First-hatched_Fed and _Treatmentunfed.If there is no suffix for a parameter the same parameter value was estimated for all food treatment and rank groupings.A "AirWin", "SSTWin", "pdo", or "pdo_lagged" suffix indicates the estimate of the linear relationship between a function parameter and our chosen air temperature, seasurface temperature, Pacific Decadal Oscillation, or lagged Pacific Decadal Oscillation parameters.

Table 36 : Top fitness function for survival from age 35 to age 40 based on ΔLOOIC comparisons.
For each model parameter the mean (Estimate), standard deviation (Est.Error), lower