Calculate survey indices by age.

Usage

get_surveyidx(
  x,
  ages,
  myids,
  kvecP = rep(12 * 12, length(ages)),
  kvecZ = rep(8 * 8, length(ages)),
  gamma = 1.4,
  cutOff = 1,
  fam = "Gamma",
  useBIC = FALSE,
  nBoot = 1000,
  mc.cores = 1,
  method = "ML",
  predD = NULL,
  modelZ = NULL,
  modelP = NULL,
  knotsP = NULL,
  knotsZ = NULL,
  predfix = NULL,
  linkZ = "logit",
  CIlevel = 0.95,
  ...
)

Arguments

x: DATRASraw object
ages: vector of ages
myids: haul.ids for grid
kvecP: vector with spatial smoother max. basis dimension for each age group, strictly positive part of model
kvecZ: vector with spatial smoother max. basis dimension for each age group, presence/absence part of model (ignored for Tweedie models)
gamma: model degress of freedom inflation factor (see 'gamma' argument to mgcv::gam() )
cutOff: treat observations below this value as zero
fam: distribution, either "Gamma","LogNormal", or "Tweedie".
useBIC: use BIC for smoothness selection (overrides 'gamma' argument)
nBoot: number of bootstrap samples used for calculating index confidence intervals
mc.cores: number of cores for parallel processing
method: smoothness selection method used by 'gam'
predD: optional DATRASraw object or data.frame (or named list with such objects, one for each year with names(predD) being the years) , defaults to NULL. If not null this is used as grid.
modelZ: vector of model formula for presence/absence part, one pr. age group (ignored for Tweedie models)
modelP: vector of model formula for strictly positive responses, one pr. age group
knotsP: optional list of knots to gam, strictly positive responses
knotsZ: optional list of knots to gam, presence/absence
predfix: optional named list of extra variables (besides Gear, HaulDur, Ship, and TimeShotHour), that should be fixed during prediction step (standardized)
linkZ: link function for the grDevices::dev.new part of the model, default: "logit" (not used for Tweedie models).
CIlevel: Confidence interval level, defaults to 0.95.
...: Optional extra arguments to "gam"

Value

A survey index (list)

Details

This is based on the methods described in Berg et al. (2014): "Evaluation of alternative age-based methods for estimating relative abundance from survey data in relation to assessment models", Fisheries Research 151(2014) 91-99.

Examples

if (FALSE) {
library(DATRAS) # example data
library(sdmgamindex)
library(tidyverse)
dat <- yfs <- sdmgamindex::noaa_afsc_public_foss |> 
  dplyr::filter(srvy=="EBS" & species_code == 10210)
# Megsie todo: add example here using EBS data!
}