Skip to contents

Model demographic rates from survival and recruitment surveys

Source: R/estimateBayesianRates.R
estimateBayesianRates.Rd

Fit Bayesian models for survival and recruitment based on boreal caribou survey data. If disturbance data is provided a Beta model is fitted with disturbance covariates and priors informed by national demographic-disturbance relationships. If disturbance data is NULL bboutools logistic models are fit with bboutools::bb_fit_survival() and bboutools::bb_fit_recruitment(). In both cases the fitted the models are used to estimate recruitment and survival rates and return summary tables.

Usage

estimateBayesianRates(
  surv_data,
  recruit_data,
  N0 = NA,
  disturbance = NULL,
  priors = NULL,
  shiny_progress = FALSE,
  return_mcmc = FALSE,
  i18n = NULL,
  niters = formals(bboutools::bb_fit_survival)$niters,
  nthin = formals(bboutools::bb_fit_survival)$nthin,
  ...
)

Arguments

surv_data

dataframe. Survival data in bboudata format.

recruit_data

dataframe. Recruitment data in bboudata format.

N0

number or dataframe. Optional. Initial population size(s). If NA (default) then population growth rate is \(\lambda_t=S_t*(1+cR_t)/s\). If a data frame N0 column is required, and PopulationName column is required if there is more than one row. Additional (optional) variation columns will be used by addN0Variation().

disturbance

dataframe. Optional. If provided, fit a Beta model that includes disturbance covariates.

priors

list. Optional. If disturbance is NULL, this should be list(priors_survival=c(...),priors_recruitment=c(...)); see bboutools::bb_priors_survival() and bboutools::bb_priors_recruitment() for details. If disturbance is not NULL, see betaNationalPriors() for details.

shiny_progress

logical. Should shiny progress bar be updated. Only set to TRUE if using in an app.

return_mcmc

boolean. If TRUE return fitted survival and recruitment models. Default FALSE.

niters

integer. The number of iterations per chain after thinning and burn-in.

nthin

integer. The number of the thinning rate.

...

Other parameters passed on to bboutools::bb_fit_survival() and bboutools::bb_fit_recruitment().

Value

If return_mcmc FALSE a data.frame containing mean survival and recruitment estimates, standard deviation, upper and lower credible intervals, and interannual variability distribution parameters for each population projected by the Bayesian model, as well as a summary of the amount of input data.

If return_mcmc is TRUE then a list with elements:

  • parTab: the data.frame returned when return_mcmc FALSE

  • parList: a list of data.frames Rbar, Sbar, Siv and Riv. Sbar and Rbar contain yearly expected survival and recruitment estimates, standard deviation, and upper and lower credible intervals for each population. Estimates will only vary by year if disturbance was provided. Siv and Riv contains interannual variability distribution parameters for each population.

  • surv_fit and recruit_fit: fitted model objects containing samples and input data

Details

See vignette("compare-bayesian-models") for more details on the differences between the two models.

Interpretation of Years

Throughout the package a year is treated as the calendar year only if it is in a table that also contains a month column. In all other contexts year is treated as the caribou year which is the annual demographic cycle of caribou, measured from one calving season to the next. By default, April is set as the start of the caribou year.

See also

Caribou demography functions: addMissingYears(), addN0Variation(), bayesianScenariosWorkflow(), bayesianTrajectoryWorkflow(), betaNationalPriors(), caribouPopGrowth(), compareTrajectories(), compositionBiasCorrection(), convertTrajectories(), dataFromSheets(), demographicProjectionApp(), estimateNationalRate(), getCaribouYear(), getNationalCoefficients(), getScenarioDefaults(), plotCompareTrajectories(), plotSurvivalSeries(), plotTrajectories(), popGrowthTableJohnsonECCC, simulateObservations(), trajectoriesFromBayesian(), trajectoriesFromNational(), trajectoriesFromSummary(), trajectoriesFromSummaryForApp()

Examples

s_data <- rbind(bboudata::bbousurv_a, bboudata::bbousurv_b)
r_data <- rbind(bboudata::bbourecruit_a, bboudata::bbourecruit_b)
estimateBayesianRates(s_data, r_data, N0 = 500)
#>   PopulationName     R_bar       R_sd R_iv_mean R_iv_shape R_bar_lower
#> 1              A 0.1892745 0.08127592 0.3153711   26.45534   0.1661514
#> 2              B 0.2030830 0.10232056 0.3153711   26.45534   0.1717109
#>   R_bar_upper     S_bar      S_sd S_iv_mean S_iv_shape S_bar_lower S_bar_upper
#> 1   0.2141778 0.8842859 0.2450002 0.5007685   12.93896   0.8259776   0.9271648
#> 2   0.2381331 0.9085433 0.2951537 0.5007685   12.93896   0.8527330   0.9474408
#>    N0 nCollarYears nSurvYears nCowsAllYears nRecruitYears
#> 1 500          900         31          2353            27
#> 2 500          519         18          2001            15