rPIBKC: Package to run the PIBKC stock assessment

surveyAveraging.RE

R Documentation

Smooth/interpolate survey data using a random effects/kalman filter model (RE).

Description

Function to smooth/interpolate survey data using a random effects/kalman filter model (RE).

Usage

surveyAveraging.RE(
  maxYr,
  srvData,
  type = "biomass",
  sex = "male",
  category = "mature",
  pdfType = "lognormal",
  ci = 0.95,
  modelPath = getPath2REM(),
  verbose = FALSE,
  showPlot = FALSE
)

Arguments

`maxYr`	max year for output (typically assessment year)
`srvData`	raw survey data dataframe
`type`	data type ('abundance' or 'biomass') to average
`sex`	sex ('male' or 'female') to average
`category`	category ('immature','mature', or 'legal') to average
`pdfType`	distribution for CIs
`ci`	confidence interval for CIs
`modelPath`	path to ADMB RE model used for survey averaging
`verbose`	flag (T/F) to print intermediate output
`showPlot`	flag (T/F) to plot results

Details

This function uses an ADMB random effects model (originally developed by Jim Ianelli and subsequently modified by William Stockhausen) to smooth/interpolate survey data.

Smoothing is done using a Kalman Filter/Random Effects model written in ADMB (C++) code. The single estimated parameter is the ln-scale process error variance for annual changes in survey abundance/biomass modeled as a random walk process. The estimated time series is output.

Value

list with a dataframe ('dfr') and another list ('lst') as elements. T he dataframe is the smoothed survey data, with columns

year = survey year
type = 'RE'
value = averaged or predicted value
lci = lower confidence interval
uci = upper confidence interval

The list contains other results from the model optimization, including

objFun = the final objective function value
maxGrad = the max gradient
sdrepSdLam = the estimated process error standard deviation, on the arithmetic scale
sdrepSdLam.sd = the standard deviation of the estimated process error standard deviation, on the arithmetic scale
and others