Appendix E — Protocol deviation — moderators adjusted for elapsed time

E.1 Recorded change

This chapter was not part of the pre-registered protocol. It refits the primary moderators (disturbance, design, trait_type, genphen) with elapsed time added as an additive covariate.

NoteWhat changed—and what did not

Changed: for each of the four primary moderators, we fitted an additive location-scale model that adds elapsed time (log10_years, and separately log10_generations) as a second main-effect term, with no interaction. This adjusts each moderator’s estimate for the average elapsed-time difference between its levels, without letting the time slope itself vary by level.

Unchanged: the dataset, the moderator definitions, the priors (brms default priors, with the standard deviation of the known sampling-variance term fixed at 1), the diagonal sampling covariance matrix, the ref_id and phylogenetic random effects, and the MCMC settings all match the primary single-moderator models. Only the added elapsed-time term differs.

This produces eight models: four moderators × two elapsed-time scales.

These models were refitted on 2026-10-10 with the primary-model specification. This refit supersedes an earlier fit of the same eight models that used different priors and a correlated sampling covariance matrix.

E.2 Moderator–time relationships

The plots below show how elapsed time is distributed across the levels of each moderator.

Code
source(here::here("Scripts", "00_packages.R"))

dat_es <- readRDS(here::here("Rdata", "effect_sizes", "proceed_lnm_safe.rds"))

moderator_labels <- c(
  disturbance = "Disturbance context",
  design      = "Comparison design",
  trait_type  = "Trait type",
  genphen     = "Phenotypic vs. genetic"
)

theme_deviation <- function(base_size = 13) {
  ggplot2::theme_classic(base_size = base_size) +
    ggplot2::theme(
      strip.background = ggplot2::element_blank(),
      strip.text        = ggplot2::element_text(face = "bold"),
      axis.text.x       = ggplot2::element_text(angle = 30, hjust = 1),
      panel.grid.major.y = ggplot2::element_line(colour = "grey92"),
      plot.title        = ggplot2::element_text(face = "bold")
    )
}

make_time_panel <- function(dat, timevar, timelabel) {
  long_dat <- purrr::map_dfr(names(moderator_labels), function(m) {
    dat |>
      dplyr::filter(!is.na(.data[[m]]), !is.na(.data[[timevar]])) |>
      dplyr::transmute(
        moderator_label = moderator_labels[[m]],
        level = as.character(.data[[m]]),
        time_value = .data[[timevar]]
      )
  })

  ggplot2::ggplot(long_dat, ggplot2::aes(x = level, y = time_value)) +
    ggplot2::geom_boxplot(fill = "grey85", outlier.alpha = 0.35, width = 0.6) +
    ggplot2::facet_wrap(~moderator_label, scales = "free_x", nrow = 2) +
    ggplot2::labs(x = NULL, y = timelabel) +
    theme_deviation()
}
Code
make_time_panel(dat_es, "log10_years", expression(log[10]("elapsed years")))

Elapsed years (log10 scale) across the levels of each primary moderator.
Code
make_time_panel(dat_es, "log10_generations", expression(log[10]("elapsed generations")))

Elapsed generations (log10 scale) across the levels of each primary moderator.

E.3 Model specification

Models were fitted on a remote server, not while rendering this book, with Scripts/remote_fit_one_additive.R (arguments: moderator, elapsed-time variable). The script uses the same formula builder, prior builder, sampling covariance matrix and MCMC settings as the primary models: brms default priors with the standard deviation of es_id_model fixed at 1 (constant(1)), a diagonal sampling covariance matrix V (Scripts/05_phylogeny.R), 4 chains × 4,000 iterations (2,000 warmup), adapt_delta = 0.97, max_treedepth = 15, seed 123, cmdstanr backend.

Code
# From Scripts/remote_fit_one_additive.R
phylo     <- prepare_phylo_and_data(dat_model, A_full, label = label)
dat_model <- phylo$dat_model; A_mod <- phylo$A_mod; V <- phylo$V

# Main effects only, no interaction, in both the location and scale components.
formula <- build_ls_formula(paste(moderator, "+", timevar),
                            has_phylogeny = phylo$has_phylo)
priors  <- build_ls_priors(formula, dat_model, V, A = A_mod)

fit <- fit_or_read_model(model_name,
  fit_fun = function() fit_ls_model(dat_model, formula, priors, V, A_mod,
                                     mcmc_args = default_mcmc_args),
  model_dir = dir_out("models"), refit = TRUE)

E.4 Convergence diagnostics

Code
diag_tbl <- readr::read_csv(
  here::here("Rdata", "tables", "additive_diagnostics.csv"),
  show_col_types = FALSE
)
knitr::kable(
  diag_tbl,
  col.names = c("Model", "Moderator", "Elapsed-time variable", "Max Rhat",
               "Min Bulk ESS", "Min Tail ESS", "Divergent transitions",
               "Max-treedepth hits"),
  digits = 3,
  caption = "Convergence diagnostics for the eight additive (moderator + time) models."
)
Convergence diagnostics for the eight additive (moderator + time) models.
Model Moderator Elapsed-time variable Max Rhat Min Bulk ESS Min Tail ESS Divergent transitions Max-treedepth hits
disturbance_plus_log10_years disturbance log10_years 1.005 956.078 1639.076 0 0
disturbance_plus_log10_generations disturbance log10_generations 1.007 845.520 1633.319 0 0
design_plus_log10_years design log10_years 1.006 1036.886 1812.013 0 0
design_plus_log10_generations design log10_generations 1.004 1136.716 2449.895 0 0
trait_type_plus_log10_years trait_type log10_years 1.006 702.933 1203.160 0 0
trait_type_plus_log10_generations trait_type log10_generations 1.007 440.396 385.209 0 0
genphen_plus_log10_years genphen log10_years 1.005 874.342 1853.933 0 0
genphen_plus_log10_generations genphen log10_generations 1.004 695.667 1560.730 0 0
Code
crit_fail <- with(diag_tbl, max_rhat > 1.01 | min_bulk_ess < 400 |
                    min_tail_ess < 400 | n_divergent > 0)
below_crit_text <- if (!any(crit_fail)) {
  "All models meet these criteria."
} else {
  paste0(
    "Models with at least one value outside these criteria: ",
    paste(sprintf("`%s` (max Rhat %.3f, min bulk ESS %s, min tail ESS %s, %d divergent transitions)",
                  diag_tbl$model_id[crit_fail], diag_tbl$max_rhat[crit_fail],
                  round(diag_tbl$min_bulk_ess[crit_fail]),
                  round(diag_tbl$min_tail_ess[crit_fail]),
                  as.integer(diag_tbl$n_divergent[crit_fail])),
          collapse = "; "),
    "."
  )
}

Across the 8 models, the largest Rhat was 1.007, the smallest bulk ESS was 440, the smallest tail ESS was 385, and the total number of divergent transitions was 0. The convergence criteria used elsewhere in this book are max Rhat ≤ 1.01, bulk and tail ESS ≥ 400, and no divergent transitions. Models with at least one value outside these criteria: trait_type_plus_log10_generations (max Rhat 1.007, min bulk ESS 440, min tail ESS 385, 0 divergent transitions).

E.5 Model summaries

 Family: gaussian 
  Links: mu = identity; sigma = log 
Formula: yi_lnM_safe ~ disturbance + log10_years + (1 | ref_id) + (1 | gr(sp_ncbi, cov = A)) + (1 | gr(es_id_model, cov = V)) 
         sigma ~ disturbance + log10_years
   Data: structure(list(es_id = c("es0313", "es0314", "es03 (Number of observations: 7186) 
  Draws: 4 chains, each with iter = 4000; warmup = 2000; thin = 1;
         total post-warmup draws = 8000

Multilevel Hyperparameters:
~es_id_model (Number of levels: 7186) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     1.00      0.00     1.00     1.00   NA       NA       NA

~ref_id (Number of levels: 254) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     0.77      0.04     0.69     0.85 1.00     1769     3716

~sp_ncbi (Number of levels: 253) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     0.70      0.17     0.38     1.06 1.00      956     1925

Regression Coefficients:
                                         Estimate Est.Error l-95% CI u-95% CI
Intercept                                   -0.90      0.44    -1.74    -0.01
sigma_Intercept                             -0.54      0.10    -0.73    -0.35
disturbanceHunt_harv                        -0.31      0.18    -0.68     0.03
disturbanceIntroduction                     -0.11      0.20    -0.52     0.28
disturbanceLandscapechange                   0.23      0.24    -0.23     0.69
disturbanceOther                            -0.31      0.18    -0.66     0.03
disturbancePollution                         0.15      0.29    -0.41     0.71
disturbanceResponsetointroductions           0.02      0.30    -0.58     0.60
log10_years                                  0.29      0.04     0.22     0.37
sigma_disturbanceHunt_harv                   0.29      0.09     0.11     0.48
sigma_disturbanceIntroduction                0.03      0.09    -0.14     0.21
sigma_disturbanceLandscapechange            -0.54      0.16    -0.87    -0.24
sigma_disturbanceOther                       0.09      0.09    -0.08     0.28
sigma_disturbancePollution                  -0.18      0.16    -0.50     0.13
sigma_disturbanceResponsetointroductions     0.26      0.12     0.03     0.48
sigma_log10_years                           -0.10      0.03    -0.16    -0.04
                                         Rhat Bulk_ESS Tail_ESS
Intercept                                1.00     2643     3468
sigma_Intercept                          1.00     1745     3259
disturbanceHunt_harv                     1.00     1854     3064
disturbanceIntroduction                  1.00     1689     2893
disturbanceLandscapechange               1.00     1899     2964
disturbanceOther                         1.00     1852     2831
disturbancePollution                     1.00     1935     3264
disturbanceResponsetointroductions       1.00     2045     3808
log10_years                              1.00     6115     6246
sigma_disturbanceHunt_harv               1.00     1718     3329
sigma_disturbanceIntroduction            1.00     1559     3346
sigma_disturbanceLandscapechange         1.00     1014     1639
sigma_disturbanceOther                   1.00     1731     3429
sigma_disturbancePollution               1.00     1450     2826
sigma_disturbanceResponsetointroductions 1.00     2050     4065
sigma_log10_years                        1.00     3419     5415

Draws were sampled using sample(hmc). For each parameter, Bulk_ESS
and Tail_ESS are effective sample size measures, and Rhat is the potential
scale reduction factor on split chains (at convergence, Rhat = 1).
 Family: gaussian 
  Links: mu = identity; sigma = log 
Formula: yi_lnM_safe ~ disturbance + log10_generations + (1 | ref_id) + (1 | gr(sp_ncbi, cov = A)) + (1 | gr(es_id_model, cov = V)) 
         sigma ~ disturbance + log10_generations
   Data: structure(list(es_id = c("es0313", "es0314", "es03 (Number of observations: 7186) 
  Draws: 4 chains, each with iter = 4000; warmup = 2000; thin = 1;
         total post-warmup draws = 8000

Multilevel Hyperparameters:
~es_id_model (Number of levels: 7186) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     1.00      0.00     1.00     1.00   NA       NA       NA

~ref_id (Number of levels: 254) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     0.76      0.04     0.68     0.85 1.01     1759     2873

~sp_ncbi (Number of levels: 253) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     0.81      0.17     0.51     1.17 1.00      878     1871

Regression Coefficients:
                                         Estimate Est.Error l-95% CI u-95% CI
Intercept                                   -0.82      0.49    -1.76     0.16
sigma_Intercept                             -0.53      0.09    -0.71    -0.35
disturbanceHunt_harv                        -0.31      0.18    -0.65     0.04
disturbanceIntroduction                     -0.13      0.20    -0.52     0.26
disturbanceLandscapechange                   0.21      0.23    -0.24     0.68
disturbanceOther                            -0.32      0.17    -0.65     0.02
disturbancePollution                         0.12      0.29    -0.47     0.69
disturbanceResponsetointroductions           0.02      0.31    -0.60     0.61
log10_generations                            0.28      0.04     0.21     0.36
sigma_disturbanceHunt_harv                   0.23      0.10     0.03     0.42
sigma_disturbanceIntroduction                0.05      0.09    -0.12     0.23
sigma_disturbanceLandscapechange            -0.54      0.16    -0.85    -0.24
sigma_disturbanceOther                       0.06      0.09    -0.12     0.24
sigma_disturbancePollution                  -0.21      0.16    -0.53     0.09
sigma_disturbanceResponsetointroductions     0.28      0.12     0.05     0.51
sigma_log10_generations                     -0.11      0.02    -0.16    -0.07
                                         Rhat Bulk_ESS Tail_ESS
Intercept                                1.00     1942     3644
sigma_Intercept                          1.00     1408     2559
disturbanceHunt_harv                     1.00     1587     2939
disturbanceIntroduction                  1.01     1398     2585
disturbanceLandscapechange               1.00     1593     2658
disturbanceOther                         1.00     1581     2591
disturbancePollution                     1.00     1432     2308
disturbanceResponsetointroductions       1.00     1766     2880
log10_generations                        1.00     4115     5310
sigma_disturbanceHunt_harv               1.00     1453     2798
sigma_disturbanceIntroduction            1.01     1410     2664
sigma_disturbanceLandscapechange         1.00      846     1633
sigma_disturbanceOther                   1.00     1430     2649
sigma_disturbancePollution               1.00     1288     2446
sigma_disturbanceResponsetointroductions 1.00     1866     3321
sigma_log10_generations                  1.00     3528     4980

Draws were sampled using sample(hmc). For each parameter, Bulk_ESS
and Tail_ESS are effective sample size measures, and Rhat is the potential
scale reduction factor on split chains (at convergence, Rhat = 1).
 Family: gaussian 
  Links: mu = identity; sigma = log 
Formula: yi_lnM_safe ~ design + log10_years + (1 | ref_id) + (1 | gr(sp_ncbi, cov = A)) + (1 | gr(es_id_model, cov = V)) 
         sigma ~ design + log10_years
   Data: structure(list(es_id = c("es0313", "es0314", "es03 (Number of observations: 7186) 
  Draws: 4 chains, each with iter = 4000; warmup = 2000; thin = 1;
         total post-warmup draws = 8000

Multilevel Hyperparameters:
~es_id_model (Number of levels: 7186) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     1.00      0.00     1.00     1.00   NA       NA       NA

~ref_id (Number of levels: 254) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     0.76      0.04     0.68     0.84 1.00     1954     3549

~sp_ncbi (Number of levels: 253) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     0.70      0.17     0.38     1.05 1.00     1037     1812

Regression Coefficients:
                       Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS
Intercept                 -1.07      0.41    -1.87    -0.23 1.00     2994
sigma_Intercept           -0.37      0.04    -0.45    -0.30 1.00     4296
designSynchronic           0.12      0.09    -0.06     0.30 1.00     3694
log10_years                0.28      0.04     0.21     0.36 1.00     6410
sigma_designSynchronic    -0.13      0.03    -0.20    -0.07 1.00     3570
sigma_log10_years         -0.10      0.03    -0.16    -0.04 1.00     3427
                       Tail_ESS
Intercept                  3829
sigma_Intercept            5924
designSynchronic           4843
log10_years                6121
sigma_designSynchronic     4937
sigma_log10_years          5573

Draws were sampled using sample(hmc). For each parameter, Bulk_ESS
and Tail_ESS are effective sample size measures, and Rhat is the potential
scale reduction factor on split chains (at convergence, Rhat = 1).
 Family: gaussian 
  Links: mu = identity; sigma = log 
Formula: yi_lnM_safe ~ design + log10_generations + (1 | ref_id) + (1 | gr(sp_ncbi, cov = A)) + (1 | gr(es_id_model, cov = V)) 
         sigma ~ design + log10_generations
   Data: structure(list(es_id = c("es0313", "es0314", "es03 (Number of observations: 7186) 
  Draws: 4 chains, each with iter = 4000; warmup = 2000; thin = 1;
         total post-warmup draws = 8000

Multilevel Hyperparameters:
~es_id_model (Number of levels: 7186) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     1.00      0.00     1.00     1.00   NA       NA       NA

~ref_id (Number of levels: 254) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     0.75      0.04     0.67     0.84 1.00     1837     3128

~sp_ncbi (Number of levels: 253) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     0.81      0.18     0.50     1.18 1.00     1137     2450

Regression Coefficients:
                        Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS
Intercept                  -0.98      0.46    -1.89    -0.06 1.00     2559
sigma_Intercept            -0.40      0.03    -0.45    -0.35 1.00     4768
designSynchronic            0.11      0.09    -0.07     0.29 1.00     3057
log10_generations           0.27      0.04     0.20     0.35 1.00     5834
sigma_designSynchronic     -0.06      0.04    -0.13     0.02 1.00     2847
sigma_log10_generations    -0.12      0.02    -0.17    -0.08 1.00     2808
                        Tail_ESS
Intercept                   4032
sigma_Intercept             6151
designSynchronic            5090
log10_generations           5965
sigma_designSynchronic      4844
sigma_log10_generations     4968

Draws were sampled using sample(hmc). For each parameter, Bulk_ESS
and Tail_ESS are effective sample size measures, and Rhat is the potential
scale reduction factor on split chains (at convergence, Rhat = 1).
 Family: gaussian 
  Links: mu = identity; sigma = log 
Formula: yi_lnM_safe ~ trait_type + log10_years + (1 | ref_id) + (1 | gr(sp_ncbi, cov = A)) + (1 | gr(es_id_model, cov = V)) 
         sigma ~ trait_type + log10_years
   Data: structure(list(es_id = c("es0313", "es0314", "es03 (Number of observations: 7186) 
  Draws: 4 chains, each with iter = 4000; warmup = 2000; thin = 1;
         total post-warmup draws = 8000

Multilevel Hyperparameters:
~es_id_model (Number of levels: 7186) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     1.00      0.00     1.00     1.00   NA       NA       NA

~ref_id (Number of levels: 254) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     0.75      0.04     0.67     0.84 1.00     1592     3624

~sp_ncbi (Number of levels: 253) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     0.83      0.17     0.53     1.19 1.00      703     1698

Regression Coefficients:
                                Estimate Est.Error l-95% CI u-95% CI Rhat
Intercept                          -1.78      0.60    -2.97    -0.59 1.00
sigma_Intercept                    -0.24      0.25    -0.72     0.27 1.00
trait_typegrowth                    0.82      0.40     0.06     1.60 1.00
trait_typeotherLH                   0.91      0.40     0.15     1.68 1.00
trait_typeothermorphology           0.79      0.39     0.02     1.55 1.00
trait_typephenology                 0.83      0.42     0.03     1.63 1.00
trait_typephysio                    0.69      0.40    -0.06     1.46 1.00
trait_typeresponse                  1.17      0.43     0.34     1.98 1.00
trait_typesize                      0.84      0.39     0.09     1.61 1.00
log10_years                         0.28      0.04     0.21     0.36 1.00
sigma_trait_typegrowth             -0.98      0.30    -1.59    -0.43 1.00
sigma_trait_typeotherLH            -0.14      0.25    -0.64     0.34 1.00
sigma_trait_typeothermorphology    -0.50      0.25    -0.99    -0.02 1.00
sigma_trait_typephenology          -0.38      0.28    -0.94     0.15 1.01
sigma_trait_typephysio             -0.65      0.26    -1.17    -0.14 1.00
sigma_trait_typeresponse           -0.51      0.38    -1.30     0.20 1.00
sigma_trait_typesize               -0.05      0.25    -0.55     0.43 1.00
sigma_log10_years                  -0.10      0.03    -0.16    -0.05 1.00
                                Bulk_ESS Tail_ESS
Intercept                           1424     2647
sigma_Intercept                     1180     2549
trait_typegrowth                    1014     1653
trait_typeotherLH                   1001     1774
trait_typeothermorphology            993     1769
trait_typephenology                 1036     1740
trait_typephysio                    1003     1675
trait_typeresponse                  1106     2092
trait_typesize                       991     1701
log10_years                         4580     5212
sigma_trait_typegrowth               765     1203
sigma_trait_typeotherLH             1213     2646
sigma_trait_typeothermorphology     1164     2581
sigma_trait_typephenology           1224     2759
sigma_trait_typephysio              1194     2565
sigma_trait_typeresponse            1196     1814
sigma_trait_typesize                1179     2643
sigma_log10_years                   3283     5645

Draws were sampled using sample(hmc). For each parameter, Bulk_ESS
and Tail_ESS are effective sample size measures, and Rhat is the potential
scale reduction factor on split chains (at convergence, Rhat = 1).
 Family: gaussian 
  Links: mu = identity; sigma = log 
Formula: yi_lnM_safe ~ trait_type + log10_generations + (1 | ref_id) + (1 | gr(sp_ncbi, cov = A)) + (1 | gr(es_id_model, cov = V)) 
         sigma ~ trait_type + log10_generations
   Data: structure(list(es_id = c("es0313", "es0314", "es03 (Number of observations: 7186) 
  Draws: 4 chains, each with iter = 4000; warmup = 2000; thin = 1;
         total post-warmup draws = 8000

Multilevel Hyperparameters:
~es_id_model (Number of levels: 7186) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     1.00      0.00     1.00     1.00   NA       NA       NA

~ref_id (Number of levels: 254) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     0.75      0.04     0.66     0.83 1.00     2694     3712

~sp_ncbi (Number of levels: 253) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     0.87      0.16     0.57     1.22 1.00     1285     2082

Regression Coefficients:
                                Estimate Est.Error l-95% CI u-95% CI Rhat
Intercept                          -1.75      0.63    -2.97    -0.52 1.00
sigma_Intercept                    -0.31      0.25    -0.80     0.19 1.00
trait_typegrowth                    0.88      0.40     0.09     1.64 1.00
trait_typeotherLH                   0.96      0.40     0.17     1.72 1.00
trait_typeothermorphology           0.83      0.39     0.04     1.58 1.00
trait_typephenology                 0.88      0.42     0.02     1.70 1.00
trait_typephysio                    0.75      0.40    -0.05     1.51 1.00
trait_typeresponse                  1.22      0.43     0.34     2.04 1.00
trait_typesize                      0.89      0.40     0.10     1.64 1.00
log10_generations                   0.27      0.04     0.20     0.34 1.00
sigma_trait_typegrowth             -0.95      0.33    -1.64    -0.35 1.01
sigma_trait_typeotherLH            -0.11      0.25    -0.61     0.38 1.00
sigma_trait_typeothermorphology    -0.46      0.25    -0.96     0.04 1.00
sigma_trait_typephenology          -0.37      0.28    -0.95     0.19 1.00
sigma_trait_typephysio             -0.65      0.27    -1.19    -0.13 1.00
sigma_trait_typeresponse           -0.50      0.37    -1.26     0.20 1.00
sigma_trait_typesize               -0.04      0.25    -0.54     0.44 1.00
sigma_log10_generations            -0.07      0.02    -0.11    -0.03 1.00
                                Bulk_ESS Tail_ESS
Intercept                           2559     3275
sigma_Intercept                     2333     3328
trait_typegrowth                    1938     3204
trait_typeotherLH                   1925     3113
trait_typeothermorphology           1942     3166
trait_typephenology                 2036     3316
trait_typephysio                    1917     3304
trait_typeresponse                  2174     3269
trait_typesize                      1939     3262
log10_generations                   7414     6211
sigma_trait_typegrowth               440      385
sigma_trait_typeotherLH             2287     3309
sigma_trait_typeothermorphology     2334     3533
sigma_trait_typephenology           2194     3296
sigma_trait_typephysio              1799     3149
sigma_trait_typeresponse            1928     3060
sigma_trait_typesize                2312     3308
sigma_log10_generations             3747     5622

Draws were sampled using sample(hmc). For each parameter, Bulk_ESS
and Tail_ESS are effective sample size measures, and Rhat is the potential
scale reduction factor on split chains (at convergence, Rhat = 1).
 Family: gaussian 
  Links: mu = identity; sigma = log 
Formula: yi_lnM_safe ~ genphen + log10_years + (1 | ref_id) + (1 | gr(sp_ncbi, cov = A)) + (1 | gr(es_id_model, cov = V)) 
         sigma ~ genphen + log10_years
   Data: structure(list(es_id = c("es0313", "es0314", "es03 (Number of observations: 7186) 
  Draws: 4 chains, each with iter = 4000; warmup = 2000; thin = 1;
         total post-warmup draws = 8000

Multilevel Hyperparameters:
~es_id_model (Number of levels: 7186) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     1.00      0.00     1.00     1.00   NA       NA       NA

~ref_id (Number of levels: 254) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     0.76      0.04     0.68     0.85 1.00     2204     3447

~sp_ncbi (Number of levels: 253) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     0.70      0.17     0.38     1.05 1.01      874     1854

Regression Coefficients:
                        Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS
Intercept                  -0.88      0.40    -1.69    -0.07 1.00     3733
sigma_Intercept            -0.63      0.05    -0.73    -0.54 1.00     3384
genphenPhenotypic          -0.19      0.07    -0.34    -0.05 1.00     4701
log10_years                 0.29      0.04     0.21     0.36 1.00     6989
sigma_genphenPhenotypic     0.30      0.03     0.24     0.36 1.00     2592
sigma_log10_years          -0.12      0.03    -0.17    -0.07 1.00     3968
                        Tail_ESS
Intercept                   4405
sigma_Intercept             5410
genphenPhenotypic           5592
log10_years                 6713
sigma_genphenPhenotypic     4772
sigma_log10_years           5513

Draws were sampled using sample(hmc). For each parameter, Bulk_ESS
and Tail_ESS are effective sample size measures, and Rhat is the potential
scale reduction factor on split chains (at convergence, Rhat = 1).
 Family: gaussian 
  Links: mu = identity; sigma = log 
Formula: yi_lnM_safe ~ genphen + log10_generations + (1 | ref_id) + (1 | gr(sp_ncbi, cov = A)) + (1 | gr(es_id_model, cov = V)) 
         sigma ~ genphen + log10_generations
   Data: structure(list(es_id = c("es0313", "es0314", "es03 (Number of observations: 7186) 
  Draws: 4 chains, each with iter = 4000; warmup = 2000; thin = 1;
         total post-warmup draws = 8000

Multilevel Hyperparameters:
~es_id_model (Number of levels: 7186) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     1.00      0.00     1.00     1.00   NA       NA       NA

~ref_id (Number of levels: 254) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     0.75      0.04     0.67     0.84 1.00     1790     3182

~sp_ncbi (Number of levels: 253) 
              Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
sd(Intercept)     0.77      0.18     0.44     1.16 1.00      696     1561

Regression Coefficients:
                        Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS
Intercept                  -0.80      0.45    -1.70     0.07 1.00     2344
sigma_Intercept            -0.65      0.04    -0.73    -0.58 1.00     2535
genphenPhenotypic          -0.18      0.07    -0.32    -0.05 1.00     3876
log10_generations           0.27      0.04     0.19     0.34 1.00     5482
sigma_genphenPhenotypic     0.28      0.03     0.22     0.34 1.00     2269
sigma_log10_generations    -0.11      0.02    -0.15    -0.08 1.00     3514
                        Tail_ESS
Intercept                   3700
sigma_Intercept             5111
genphenPhenotypic           5931
log10_generations           6301
sigma_genphenPhenotypic     4482
sigma_log10_generations     4978

Draws were sampled using sample(hmc). For each parameter, Bulk_ESS
and Tail_ESS are effective sample size measures, and Rhat is the potential
scale reduction factor on split chains (at convergence, Rhat = 1).

E.6 Orchard-style figures

Each figure shows the focal moderator’s location (top) and scale (bottom) estimates from its additive model, with elapsed time held at its dataset mean. Points and beeswarm markers are on the same scale as the primary single-moderator orchard figures (see the model chapters); only the added time term differs.

Code
knitr::include_graphics(here::here("Rdata", "figures", "publication",
                                   "orchard_additive",
                                   "disturbance_plus_log10_years_orchard_combined.png"))

Code
knitr::include_graphics(here::here("Rdata", "figures", "publication",
                                   "orchard_additive",
                                   "disturbance_plus_log10_generations_orchard_combined.png"))

Code
knitr::include_graphics(here::here("Rdata", "figures", "publication",
                                   "orchard_additive",
                                   "design_plus_log10_years_orchard_combined.png"))

Code
knitr::include_graphics(here::here("Rdata", "figures", "publication",
                                   "orchard_additive",
                                   "design_plus_log10_generations_orchard_combined.png"))

Code
knitr::include_graphics(here::here("Rdata", "figures", "publication",
                                   "orchard_additive",
                                   "trait_type_plus_log10_years_orchard_combined.png"))

Code
knitr::include_graphics(here::here("Rdata", "figures", "publication",
                                   "orchard_additive",
                                   "trait_type_plus_log10_generations_orchard_combined.png"))

Code
knitr::include_graphics(here::here("Rdata", "figures", "publication",
                                   "orchard_additive",
                                   "genphen_plus_log10_years_orchard_combined.png"))

Code
knitr::include_graphics(here::here("Rdata", "figures", "publication",
                                   "orchard_additive",
                                   "genphen_plus_log10_generations_orchard_combined.png"))