Estimating relationship between survival and non-linear growth parameters

Hello All,

I am interested in running a model that estimates the relationship between survival (0 or 1) and parameters of a non-linear growth model (A= asymptote, k = maximum growth rate, Ti = timing of maximum growth). I’ve specified a model below that gets close to what I am looking for (I think!). However, I’d like to get the estimate between survival and parameters at the population level with the intention of eventually asking how the relationship between growth parameters and survival might vary from year to year.

Thank you for time and help in advance,


#Prior for model
prior_richards2 <- prior(normal(4, 1.0), nlpar = "A",resp='Weight') +prior(normal(5, 0.9), nlpar = "k",resp='Weight')+prior(normal(15, 2), nlpar = "Ti",resp='Weight')+prior(normal(2, 0.5), nlpar = "d",lb=1.1,ub=4,resp='Weight')+prior(normal(0.5,5), class = "Intercept",resp='Surv')

#Non-linear growth model
#Weight is repeatedly measured throughout a nestling life
Richards_mod <- bf(Weight ~ (A*100)*(1+(d-1)*exp((-(k/100)*(Age-Ti))/(d^(d/(1-d)))))^(1/(1-d)),A+k+Ti~1+(1|p|Year),d~1,nl=T)

#Survival model
#Whether or not a nestling survives to fledge from the nest. 0 or 1
Surv_mod<- bf(Surv~(1|p|Year))+bernoulli(link = 'logit')

SurvtoFledgeMod_AlphaUnfed_Year <- brm(Surv_mod+Richards_mod,data = KittChickAlphaUnfed,
  control = list(adapt_delta = 0.99,max_treedepth=15),cores=6,iter = 4000,chains = 6,thin = 10,prior=prior_richards2)

Morning and welcome,

Thanks for posting the model code. Can you share the data or better some simulated data with known parameters. This can helps folks troubleshoot what is going on.

Dear Ara,

Thank you for your reply. Attached some growth data.
Year → Year grouping level
ChickID → Nestling grouping level
Age → Age of measurement
Weight → Mass of nestling
Surv → 0 or 1 survival to fledge

share_growth_dat.csv (136.1 KB)

1 Like