Transitioning to the new version of bycatchEstimator in 2025

For details of the current version of the code see the current user guide. This page is intended to help those who are using the previous version (version 0) from 2024 translate their old code into the new format. If you wish to keep using version 0, you can install it using the following code. This is not recommended, as the new versions has corrections and new features:

# install.packages("devtools")
devtools::install_github("ebabcock/BycatchEstimator@v0.0.9")

The improvements made in 2025 include the following.

bycatchSetup was revised to provides a report (html or pdf) with data warnings, plots and summaries to check whether the data are appropriate for bycatch estimation, replacing dataCheck. You no longer need to need to include any specifications for the model-based or design-based estimation in this function call.
bycatchDesign is a new function that does design-based estimation only. The inputs are all the specifications for design-based methods, including pooling, and outputs include a report (html or pdf) with warnings, results and a summary of the pooling. It allows for a designScenario designation so that multiple scenarios (e.g. pooling choices) can be run from the same setupObj.
bycatchFit now includes all the model specifications such as modelTry, simpleModel and complexModel, and produces a report for the model results only. It allows for a model Scenario designation so that multiple scenarios can be run from the same setupObj.
loadInputs allows the user to load all the R objects from bycatchDesign or bycatchFit for further analysis if desired.

The following shows the old bycatchSetup example, with the necessary changes in comments. The new output file includes data warnings and summaries

library(BycatchEstimator)
library(MuMIn)
setupObj<-bycatchSetup(
#  modelTry = c("TMBnbinom1"),              #Moved to bycatchFit
  obsdat = droplevels(LLSIM_BUM_Example_observer[LLSIM_BUM_Example_observer$Year>2010 &LLSIM_BUM_Example_observer$fleet==2,]),
  logdat = droplevels(LLSIM_BUM_Example_logbook[LLSIM_BUM_Example_logbook$Year>2010 & LLSIM_BUM_Example_logbook$fleet==2,]),
  yearVar = "Year",
  obsEffort = "hooks",
  logEffort = "hooks",
#  logUnsampledEffort = "unsampledEffort",  #Delete, this is no longer used
#  factorNames = c("Year","area"),          #replaced with factorVariables
  factorVariables = c("Year","area"),       #NEW, replaces factorNames
  numericVariables = NA, # NEW
  EstimateBycatch = TRUE,
#  EstimateIndex = FALSE                    #Moved to bycatchFit
  logNum = NA,
  sampleUnit = "trips",
#  complexModel = formula(y~Year+area),     #Moved to bycatchFit
#  simpleModel = formula(y~Year),           #Moved to bycatchFit
#  indexModel = formula(y~Year),            #Moved to bycatchFit
 # designMethods = c("Ratio", "Delta"),     #Moved to bycatchDesign
# designVars = c("Year","area"),          #Moved to bycatchDesign
# designPooling = TRUE,                     #Moved to bycatchDesign
# poolTypes=c("adjacent","all"),            #Moved to bycatchDesign
# pooledVar=c(NA,NA),                       #Moved to bycatchDesign
# adjacentNum=c(1,NA),                      #Moved to bycatchDesign
  baseDir = getwd(),
  runName = "LLSIMBUMtripExample",
  runDescription = "LLSIm BUM by trip",
  common = "Blue marlin",
  sp = "Makaira nigricans",
  obsCatch ="BUM",
  catchUnit = "number",
  catchType = "catch",
  reportType = "html"                       #NEW: specify html or pdf 
)

The new bycatchDesign takes setupObj as an input, as well as all the specifications for design-based estimators. There is a new pooling option “none” for variables that should never by pooled.

bycatchDesign(
setupObj = setupObj,
designScenario = "noPool",                  #NEW:Name for output files 
designMethods = c("Ratio", "Delta"),        #Moved here
designVars = c("Year","area"),              #Moved here
groupVar = "Year",                          #NEW:For plotting by a variable other than Year
designPooling = TRUE,                       #Moved here
poolTypes=c("adjacent","all"),              #Moved here
pooledVar=c(NA,NA),                         #Moved here
adjacentNum=c(1,NA),                        #Moved here
minStrataUnit = 1,                          #Moved here
reportType = "html"                         #NEW
)

The new bycatchFit takes setupObj as an input, as well as all the specifications for model-based estimators.

bycatchFit(
setupObj = setupObj, 
modelScenario = "s1",                       #NEW. Name included in output files
complexModel = formula(y~(Year+area)^2),    #Moved here
simpleModel = formula(y~Year),              #Moved here
indexModel = formula(y~Year),               #Moved here
modelTry = c("Delta-Lognormal","TMBnbinom2"),     #Moved here
randomEffects = NULL,                       #Moved here
randomEffects2 = NULL,                      #Moved here
selectCriteria = "BIC",                     
DoCrossValidation = TRUE,                   
#dredgeCrossValidation =FALSE,              #Option deleted
#ResidualTest = FALSE,                      #Option deleted
CIval = 0.05,                               
VarCalc = "Simulate",                       
includeObsCatch=FALSE,                      #Moved here
matchColumn=NULL,                           #Moved here
EstimateIndex=FALSE,                        #Moved here
useParallel = TRUE,          
nSims = 100,
baseDir = getwd(),
plotValidation = FALSE,
trueVals = NULL,
trueCols = NULL,
reportType = "html"                         #NEW
)

Some minor corrections have been made, along with new options, so the new version of the code is recommended.

Beth Babcock

2025-07-03