R/Data_make_update.R
In Blue-Matter/MSEtool: Management Strategy Evaluation Toolkit
Defines functions
updateData_MS
AddRealData_MS
optESS
AddRealData
check_Index_Fit
UpdateSlot
UpdateObs
getfifth
genSizeCompWrap
simCAL
simCAA
updateData
makeData
Documented in
genSizeCompWrap
getfifth
simCAA
simCAL
makeData <- function(Biomass, CBret, Cret, N, SSB, VBiomass, StockPars,
FleetPars, ObsPars, ImpPars, RefPoints,
SampCpars, initD, Sample_Area,
Name,
nyears,
proyears,
nsim,
nareas,
reps,
CurrentYr,
silent=FALSE,
control=list()) {
if(!silent) message("Simulating observed data")
Data <- new("Data") # create a blank DLM data object
if (reps == 1) Data <- OneRep(Data) # make stochastic variables certain for only one rep
Data <- replic8(Data, nsim) # make nsim sized slots in the DLM data object
Data@Name <- Name
Data@Year <- (CurrentYr - nyears+1):CurrentYr
# --- Observed catch ----
# Simulated observed retained catch (biomass)
Data@Cat <- ObsPars$Cobs_y[,1:nyears] * apply(CBret*Sample_Area$Catch[,,1:nyears,], c(1, 3), sum)
Data@CV_Cat <- matrix(Data@CV_Cat[,1], nrow=nsim, ncol=nyears)
# --- Index of total abundance ----
# Index of abundance from total biomass - beginning of year before fishing
# apply hyperstability / hyperdepletion
II <- (apply(Biomass*Sample_Area$BInd[,,1:nyears,], c(1, 3), sum)^ObsPars$I_beta) *
ObsPars$Ierr_y[, 1:nyears]
II <- II/apply(II, 1, mean) # normalize
Data@Ind <- II # index of total abundance
Data@CV_Ind <- matrix(Data@CV_Ind[,1], nrow=nsim, ncol=nyears)
# --- Index of spawning abundance ----
# Index of abundance from total biomass - beginning of year before fishing
# apply hyperstability / hyperdepletion
II <- (apply(SSB*Sample_Area$SBInd[,,1:nyears,], c(1, 3), sum)^ObsPars$SpI_beta) *
ObsPars$SpIerr_y[, 1:nyears]
II <- II/apply(II, 1, mean) # normalize
Data@SpInd <- II # index of spawning abundance
Data@CV_SpInd <- matrix(Data@CV_SpInd[,1], nrow=nsim, ncol=nyears)
# --- Index of vulnerable abundance ----
# Index of abundance from vulnerable biomass - beginning of year before fishing
# apply hyperstability / hyperdepletion
II <- (apply(VBiomass*Sample_Area$VInd[,,1:nyears,], c(1, 3), sum)^ObsPars$VI_beta) *
ObsPars$VIerr_y[, 1:nyears]
II <- II/apply(II, 1, mean) # normalize
Data@VInd <- II # index of vulnerable abundance
Data@CV_VInd <- matrix(Data@CV_VInd[,1], nrow=nsim, ncol=nyears)
# --- Index of recruitment ----
Data@Rec <- apply(N[, 1, , ]*Sample_Area$RecInd[,1:nyears,], c(1, 2), sum) *
ObsPars$Recerr_y[, 1:nyears]
Data@t <- rep(nyears, nsim) # number of years of data
# --- Average catch ----
Data@AvC <- apply(Data@Cat, 1, mean, na.rm=TRUE) # average catch over all years
# --- Depletion ----
# observed depletion
Depletion <- apply(SSB[,,nyears,],1,sum)/RefPoints$SSB0 # current depletion
Data@Dt <- Depletion * ObsPars$Derr_y[,nyears]
Data@Dep <- Depletion * ObsPars$Derr_y[,nyears]
# --- Life-history parameters ----
Data@vbLinf <- StockPars$Linfarray[,nyears] * ObsPars$Linfbias # observed vB Linf
Data@vbK <- StockPars$Karray[,nyears] * ObsPars$Kbias # observed vB K
Data@vbt0 <- StockPars$t0array[,nyears] + ObsPars$t0bias # observed vB t0
Data@Mort <- StockPars$Marray[,nyears] * ObsPars$Mbias # natural mortality
Data@L50 <- StockPars$L50array[,nyears] * ObsPars$lenMbias # observed length at 50% maturity
Data@L95 <- StockPars$L95array[,nyears] * ObsPars$lenMbias # observed length at 95% maturity
Data@L95[Data@L95 > 0.95 * Data@vbLinf] <- 0.95 * Data@vbLinf[Data@L95 > 0.95 * Data@vbLinf] # Set a hard limit on ratio of L95 to Linf
Data@L50[Data@L50 > 0.95 * Data@L95] <- 0.95 * Data@L95[Data@L50 > 0.95 * Data@L95] # Set a hard limit on ratio of L95 to Linf
Data@LenCV <- StockPars$LenCV # variability in length-at-age - no error at this time
Data@sigmaR <- StockPars$procsd * ObsPars$sigmaRbias # observed sigmaR -
Data@MaxAge <- StockPars$maxage # maximum age - no error - used for setting up matrices only
# Observed steepness values
hs <- StockPars$hs
Data@steep <- hs * ObsPars$hbias # observed steepness
# --- Reference points ----
# Simulate observation error in BMSY/B0
ntest <- 20 # number of trials
test <- array(RefPoints$SSBMSY_SSB0 * ObsPars$BMSY_B0bias, dim = c(nsim, ntest)) # the simulated observed BMSY_B0
indy <- array(rep(1:ntest, each = nsim), c(nsim, ntest)) # index
indy[test > max(0.9, max(RefPoints$SSBMSY_SSB0))] <- NA # interval censor
ObsPars$BMSY_B0bias <- ObsPars$BMSY_B0bias[cbind(1:nsim, apply(indy, 1, min, na.rm = T))] # sample such that BMSY_B0<90%
Data@FMSY_M <- RefPoints$FMSY_M * ObsPars$FMSY_Mbias # observed FMSY/M
Data@BMSY_B0 <- RefPoints$SSBMSY_SSB0 * ObsPars$BMSY_B0bias # observed BMSY/B0
Data@Cref <- RefPoints$MSY * ObsPars$Crefbias # Catch reference - MSY with error
Data@Bref <- RefPoints$VBMSY * ObsPars$Brefbias # Vuln biomass ref - VBMSY with error
# Generate values for reference SBMSY/SB0
# should be calculated from unfished - won't be correct if initD is set
I3 <- apply(Biomass, c(1, 3), sum)^ObsPars$I_beta # apply hyperstability / hyperdepletion
I3 <- I3/apply(I3, 1, mean) # normalize index to mean 1
if (!is.null(initD)) {
b1 <- apply(Biomass, c(1, 3), sum)
b2 <- matrix(RefPoints$BMSY, nrow=nsim, ncol=nyears)
ind <- apply(abs(b1/ b2 - 1), 1, which.min) # find years closest to BMSY
Iref <- diag(I3[1:nsim,ind]) # return the real target abundance index closest to BMSY
} else {
Iref <- apply(I3[, 1:min(5, ncol(I3))], 1, mean) * RefPoints$BMSY_B0 # return the real target abundance index corresponding to BMSY
}
Data@Iref <- Iref * ObsPars$Irefbias # index reference with error
# --- Abundance ----
# Calculate vulnerable and spawning biomass abundance --
M_array <- array(0.5*StockPars$M_ageArray[,,nyears], dim=c(nsim, StockPars$maxage+1, nareas))
A <- apply(VBiomass[, , nyears, ] * exp(-M_array), 1, sum) # Abundance (mid-year before fishing)
Asp <- apply(SSB[, , nyears, ] * exp(-M_array), 1, sum) # Spawning abundance (mid-year before fishing)
OFLreal <- A * (1-exp(-RefPoints$FMSY)) # the true simulated Over Fishing Limit
# Observed total abundance
Data@Abun <- A * ObsPars$Aerr_y[,nyears]
# observed spawning abundance
Data@SpAbun <- Asp * ObsPars$Aerr_y[,nyears]
# --- Catch-at-age ----
Cret2 <- apply(Cret * Sample_Area$CAA[,,1:nyears,], 1:3, sum)
Data@CAA <- simCAA(nsim, nyears, StockPars$maxage+1, Cret2, ObsPars$CAA_ESS, ObsPars$CAA_nsamp)
# --- Catch-at-length ----
CALdone <- FALSE
if (!is.null(control$CAL)) {
if (control$CAL == 'removals') {
vn <- apply(N*Sample_Area$CAL[,,1:nyears,], c(1,2,3), sum) * FleetPars$V_real[,,1:nyears]
# numbers at age in population that would be removed
vn <- aperm(vn, c(1,3, 2))
CALdat <- simCAL(nsim, nyears, StockPars$maxage, ObsPars$CAL_ESS,
ObsPars$CAL_nsamp, StockPars$nCALbins, StockPars$CAL_binsmid, StockPars$CAL_bins,
vn, FleetPars$SLarray_real, StockPars$Linfarray,
StockPars$Karray, StockPars$t0array, StockPars$LenCV)
CALdone <- TRUE
} else {
warning('Invalid entry in OM@cpars$control$CAL (use OM@cpars$control$CAL="removals")\nSimulating CAL from retained catch')
}
}
if (!CALdone) {
vn <- apply(N*Sample_Area$CAL[,,1:nyears,], c(1,2,3), sum) * FleetPars$retA_real[,,1:nyears]
# numbers at age in population that would be retained
vn <- aperm(vn, c(1,3, 2))
CALdat <- simCAL(nsim, nyears, maxage=StockPars$maxage,
CAL_ESS=ObsPars$CAL_ESS,
CAL_nsamp=ObsPars$CAL_nsamp,
nCALbins=StockPars$nCALbins,
CAL_binsmid=StockPars$CAL_binsmid,
CAL_bins=StockPars$CAL_bins,
vn,
retL=FleetPars$retL_real,
Linfarray=StockPars$Linfarray,
Karray=StockPars$Karray,
t0array=StockPars$t0array,
LenCV=StockPars$LenCV)
}
Data@CAL_bins <- StockPars$CAL_bins
Data@CAL_mids <- StockPars$CAL_binsmid
Data@CAL <- CALdat$CAL # observed catch-at-length
Data@ML <- CALdat$ML # mean length
Data@Lc <- CALdat$Lc # modal length
Data@Lbar <- CALdat$Lbar # mean length above Lc
Data@LFC <- FleetPars$L5_y[,nyears] * ObsPars$LFCbias # length at first capture
Data@LFS <- FleetPars$LFS_y[,nyears] * ObsPars$LFSbias # length at full selection
Data@Vmaxlen <- FleetPars$Vmaxlen_y[,nyears]
# --- Previous Management Recommendations ----
Data@MPrec <- apply(CBret, c(1, 3), sum)[,nyears] # catch in last year
Data@MPeff <- rep(1, nsim) # effort in last year = 1
# --- Store OM Parameters ----
# put all the operating model parameters in one table
ind <- which(lapply(StockPars, length) == nsim)
drop_srr <- which(names(ind)=='SRRpars')
ind <- ind[-drop_srr]
stock <- as.data.frame(StockPars[ind])
stock$Fdisc <- NULL
stock$CAL_bins <- NULL
stock$CAL_binsmid <- NULL
ind <- which(lapply(FleetPars, length) == nsim)
fleet <- as.data.frame(FleetPars[ind])
ind <- which(lapply(ImpPars, length) == nsim)
imp <- as.data.frame(ImpPars[ind])
refs <- RefPoints[!names(RefPoints) %in% names(stock)]
OMtable <- data.frame(stock, fleet, imp, refs, ageM=StockPars$ageMarray[,nyears],
L5=FleetPars$L5_y[,nyears ], LFS=FleetPars$LFS_y[,nyears ],
Vmaxlen=FleetPars$Vmaxlen_y[,nyears ],
LR5=FleetPars$LR5_y[,nyears], LFR=FleetPars$LFR_y[,nyears],
Rmaxlen=FleetPars$Rmaxlen_y[,nyears],
DR=FleetPars$DR_y[,nyears], OFLreal, maxF=StockPars$maxF,
A=A, Asp=Asp, CurrentYr=CurrentYr)
OMtable <- OMtable[,order(names(OMtable))]
Data@OM <- OMtable
# --- Store Obs Parameters ----
ObsParsDF <- ObsPars
ind <- which(lapply(ObsParsDF, length)==nsim) %>% as.numeric()
ObsParsDF <- ObsParsDF[ind]
ObsTable <- as.data.frame(ObsParsDF)
ObsTable <- ObsTable[,order(names(ObsTable))]
Data@Obs <- ObsTable # put all the observation error model parameters in one table
# --- Misc ----
Data@Units <- "unitless"
Data@Ref_type <- "Simulated OFL"
Data@wla <- rep(StockPars$a, nsim)
Data@wlb <- rep(StockPars$b, nsim)
Data@nareas <- nareas
Data@Ref <- OFLreal
Data@LHYear <- CurrentYr # Last historical year is nyears (for fixed MPs)
Data@Misc <- vector("list", nsim)
Data
}
updateData <- function(Data, OM, MPCalcs, Effort, Biomass, N, Biomass_P, CB_Pret,
N_P, SSB, SSB_P, VBiomass, VBiomass_P, RefPoints,
retA_P,
retL_P, StockPars, FleetPars, ObsPars, ImpPars,
V_P,
upyrs, interval, y=2,
mm=1, Misc, RealData, Sample_Area) {
yind <- upyrs[match(y, upyrs) - 1]:(upyrs[match(y, upyrs)] - 1) # index
nyears <- OM@nyears
proyears <- OM@proyears
nsim <- OM@nsim
nareas <- StockPars$nareas
reps <- OM@reps
Data@Year <- (OM@CurrentYr - nyears+1):(OM@CurrentYr+ y - 1)
Data@t <- rep(nyears + y, nsim)
# --- Simulate catches ----
CBtemp <- CB_Pret[, , yind, , drop=FALSE] * Sample_Area$Catch[,,nyears+yind,, drop=FALSE]
CBtemp[is.na(CBtemp)] <- tiny
CBtemp[!is.finite(CBtemp)] <- tiny
yr.index <- max(which(!is.na(Data@CV_Cat[1,])))
newCV_Cat <- matrix(Data@CV_Cat[,yr.index], nrow=nsim, ncol=length(yind))
Data@CV_Cat <- cbind(Data@CV_Cat, newCV_Cat)
# --- Observed catch ----
# Simulated observed retained catch (biomass)
Cobs <- ObsPars$Cobs_y[,nyears + yind] * apply(CBtemp, c(1, 3), sum, na.rm = TRUE)
Data@Cat <- cbind(Data@Cat, Cobs)
if (!is.null(RealData) && ncol(RealData@Cat)>nyears &&
!all(is.na(RealData@Cat[1,(nyears+1):length(RealData@Cat[1,])]))) {
# update projection catches with observed catches
addYr <- min(y,ncol(RealData@Cat) - nyears)
Data@Cat[,(nyears+1):(nyears+addYr)] <- matrix(RealData@Cat[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
Data@CV_Cat[,(nyears+1):(nyears+addYr)] <- matrix(RealData@CV_Cat[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
}
# --- Index of total abundance ----
yr.ind <- max(which(!is.na(ObsPars$Ierr_y[1,1:nyears])))
I2 <- cbind(apply(Biomass*Sample_Area$BInd[,,1:nyears,], c(1, 3), sum)[,yr.ind:nyears],
apply(Biomass_P*Sample_Area$BInd[,,(nyears+1):(nyears+proyears),], c(1, 3), sum)[, 1:(y - 1)])
# standardize, apply beta & obs error
I2 <- exp(lcs(I2))^ObsPars$I_beta * ObsPars$Ierr_y[,yr.ind:(nyears + (y - 1))]
# I2 <- exp(lcs(I2)) * ObsPars$Ierr_y[,yr.ind:(nyears + (y - 1))]
year.ind <- max(which(!is.na(Data@Ind[1,1:nyears])))
scaler <- Data@Ind[,year.ind]/I2[,1]
scaler <- matrix(scaler, nrow=nsim, ncol=ncol(I2))
I2 <- I2 * scaler # convert back to historical index scale
I2 <- cbind(Data@Ind[,1:(yr.ind)], I2[,2:ncol(I2)])
Data@Ind <- I2
yr.index <- max(which(!is.na(Data@CV_Ind[1,1:nyears])))
newCV_Ind <- matrix(Data@CV_Ind[,yr.index], nrow=nsim, ncol=length(yind))
Data@CV_Ind <- cbind(Data@CV_Ind, newCV_Ind)
if (!is.null(RealData) && ncol(RealData@Ind)>nyears &&
!all(is.na(RealData@Ind[1,(nyears+1):length(RealData@Ind[1,])]))) {
# update projection index with observed index if it exists
addYr <- min(y,ncol(RealData@Ind) - nyears)
Data@Ind[,(nyears+1):(nyears+addYr)] <- matrix(RealData@Ind[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
Data@CV_Ind[,(nyears+1):(nyears+addYr)] <- matrix(RealData@CV_Ind[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
}
# --- Index of spawning abundance ----
yr.ind <- max(which(!is.na(ObsPars$SpIerr_y[1,1:nyears])))
I2 <- cbind(apply(SSB*Sample_Area$SBInd[,,1:nyears,], c(1, 3), sum)[,yr.ind:nyears],
apply(SSB_P*Sample_Area$SBInd[,,(nyears+1):(nyears+proyears),], c(1, 3), sum)[, 1:(y - 1)])
# standardize, apply beta & obs error
I2 <- exp(lcs(I2))^ObsPars$SpI_beta * ObsPars$SpIerr_y[,yr.ind:(nyears + (y - 1))]
year.ind <- max(which(!is.na(Data@SpInd[1,1:nyears])))
scaler <- Data@SpInd[,year.ind]/I2[,1]
scaler <- matrix(scaler, nrow=nsim, ncol=ncol(I2))
I2 <- I2 * scaler # convert back to historical index scale
I2 <- cbind(Data@SpInd[,1:(yr.ind)], I2[,2:ncol(I2)])
Data@SpInd <- I2
yr.index <- max(which(!is.na(Data@CV_SpInd[1,1:nyears])))
newCV_Ind <- matrix(Data@CV_SpInd[,yr.index], nrow=nsim, ncol=length(yind))
Data@CV_SpInd <- cbind(Data@CV_SpInd, newCV_Ind)
if (!is.null(RealData) && ncol(RealData@SpInd)>nyears &&
!all(is.na(RealData@SpInd[1,(nyears+1):length(RealData@SpInd[1,])]))) {
# update projection index with observed index if it exists
addYr <- min(y,ncol(RealData@SpInd) - nyears)
Data@SpInd[,(nyears+1):(nyears+addYr)] <- matrix(RealData@SpInd[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
Data@CV_SpInd[,(nyears+1):(nyears+addYr)] <- matrix(RealData@CV_SpInd[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
}
# --- Index of vulnerable abundance ----
yr.ind <- max(which(!is.na(ObsPars$VIerr_y[1,1:nyears])))
I2 <- cbind(apply(VBiomass*Sample_Area$VInd[,,1:nyears,], c(1, 3), sum)[,yr.ind:nyears],
apply(VBiomass_P*Sample_Area$VInd[,,(nyears+1):(nyears+proyears),], c(1, 3), sum)[, 1:(y - 1)])
# standardize, apply beta & obs error
I2 <- exp(lcs(I2))^ObsPars$VI_beta * ObsPars$VIerr_y[,yr.ind:(nyears + (y - 1))]
year.ind <- max(which(!is.na(Data@VInd[1,1:nyears])))
scaler <- Data@VInd[,year.ind]/I2[,1]
scaler <- matrix(scaler, nrow=nsim, ncol=ncol(I2))
I2 <- I2 * scaler # convert back to historical index scale
I2 <- cbind(Data@VInd[,1:(yr.ind)], I2[,2:ncol(I2)])
Data@VInd <- I2
yr.index <- max(which(!is.na(Data@CV_VInd[1,1:nyears])))
newCV_Ind <- matrix(Data@CV_VInd[,yr.index], nrow=nsim, ncol=length(yind))
Data@CV_VInd <- cbind(Data@CV_VInd, newCV_Ind)
if (!is.null(RealData) && ncol(RealData@VInd)>nyears &&
!all(is.na(RealData@VInd[1,(nyears+1):length(RealData@VInd[1,])]))) {
# update projection index with observed index if it exists
addYr <- min(y,ncol(RealData@VInd) - nyears)
Data@VInd[,(nyears+1):(nyears+addYr)] <- matrix(RealData@VInd[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
Data@CV_VInd[,(nyears+1):(nyears+addYr)] <- matrix(RealData@CV_VInd[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
}
# --- Update additional indices (if they exist) ----
AddIunits <- Data@AddIunits
AddIndType <- Data@AddIndType
if (length(ObsPars$AddIerr)>0) {
n.ind <- dim(ObsPars$AddIerr)[2]
AddInd <- array(NA, dim=c(nsim, n.ind, nyears+y-1))
CV_AddInd <- array(NA, dim=c(nsim, n.ind, nyears+y-1))
for (i in 1:n.ind) {
if (all(is.na(RealData@AddIndV[1, , ]))) {
Ind_V <- rep(1, Data@MaxAge+1)
} else {
Ind_V <- RealData@AddIndV[1,i, ]
}
Ind_V <- matrix(Ind_V, nrow=Data@MaxAge+1, ncol= nyears+proyears)
Ind_V <- replicate(nsim, Ind_V) %>% aperm(c(3,1,2))
nas <- which(!is.na(ObsPars$AddIerr[1,i, 1:nyears]))
if (length(nas)>0) {
yr.ind <- max(nas)
if (AddIunits[i]) { # Biomass-based index
if (AddIndType[i]==1) {
# total biomass
b1 <- apply(Biomass[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum)
b2 <- apply(Biomass_P, c(1, 2, 3), sum)
}
if (AddIndType[i]==2) {
# spawning biomass
b1 <- apply(SSB[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum)
b2 <- apply(SSB_P, c(1, 2, 3), sum)
}
if (AddIndType[i]==3) {
# vulnerable biomass
b1 <- apply(VBiomass[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum)
b2 <- apply(VBiomass_P, c(1, 2, 3), sum)
}
} else {
if (AddIndType[i]==1) {
# total stock
b1 <- apply(N[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum) # Abundance-based index
b2 <- apply(N_P, c(1, 2, 3), sum)
}
if (AddIndType[i]==2) {
# spawning stock
b1 <- apply(N[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum) * StockPars$Mat_age[,,yr.ind:nyears, drop=FALSE]
b2 <- apply(N_P, c(1, 2, 3), sum) * StockPars$Mat_age[,,(nyears+1):(nyears+proyears), drop=FALSE]
}
if (AddIndType[i]==3) {
# vuln stock
b1 <- apply(N[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum) * V_P[,,yr.ind:nyears, drop=FALSE]
b2 <- apply(N_P, c(1, 2, 3), sum) * V_P[,,(nyears+1):(nyears+proyears), drop=FALSE]
}
}
b1 <- apply(b1 * Ind_V[,,yr.ind:nyears, drop=FALSE], c(1,3), sum)
b2 <- apply(b2 * Ind_V[,,(nyears+1):(nyears+proyears), drop=FALSE], c(1,3), sum)
tempI <- cbind(b1, b2[, 1:(y - 1)])
# standardize, apply beta & obs error
tempI <- exp(lcs(tempI))^ObsPars$AddIbeta[,i] * ObsPars$AddIerr[,i,yr.ind:(nyears + (y - 1))]
year.ind <- max(which(!is.na(RealData@AddInd[1,i,])))
scaler <- RealData@AddInd[1,i,year.ind]/tempI[,1]
scaler <- matrix(scaler, nrow=nsim, ncol=ncol(tempI))
tempI <- tempI * scaler # convert back to historical index scale
AddInd[,i,] <- cbind(Data@AddInd[1:nsim,i,1:yr.ind], tempI[,2:ncol(tempI)])
yr.index <- max(which(!is.na(Data@CV_AddInd[1,i,1:nyears])))
newCV_Ind <- matrix(Data@CV_AddInd[,i,yr.index], nrow=nsim, ncol=length(yind))
CV_AddInd[,i,] <- cbind(Data@CV_AddInd[,i,], newCV_Ind)
if (!is.null(RealData) && length(RealData@AddInd[1,i,])>nyears &&
!all(is.na(RealData@AddInd[1,i,(nyears+1):length(RealData@AddInd[1,i,])]))) {
# update projection index with observed index if it exists
addYr <- min(y-1,length(RealData@AddInd[1,i,]) - nyears)
AddInd[,i,(nyears+1):(nyears+addYr)] <- matrix(RealData@AddInd[1,i,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
CV_AddInd[,i,(nyears+1):(nyears+addYr)] <- matrix(RealData@CV_AddInd[1,i,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
}
}
}
Data@AddInd <- AddInd
Data@CV_AddInd <- CV_AddInd
}
# --- Index of recruitment ----
Recobs <- ObsPars$Recerr_y[, nyears + yind] * apply(array(N_P[, 1, yind, ] *
Sample_Area$RecInd[,nyears+yind,],
c(nsim, length(yind), nareas)),
c(1, 2), sum)
Data@Rec <- cbind(Data@Rec, Recobs)
# --- Average catch ----
Data@AvC <- apply(Data@Cat, 1, mean)
# --- Depletion ----
Depletion <- apply(SSB_P[, , y, ], 1, sum)/RefPoints$SSB0
Depletion[Depletion < tiny] <- tiny
Data@Dt <- Depletion * ObsPars$Derr_y[,nyears+y]
Data@Dep <- Depletion * ObsPars$Derr_y[,nyears+y]
# --- Update life-history parameter estimates for current year ----
Data@vbLinf <- StockPars$Linfarray[,nyears+y] * ObsPars$Linfbias # observed vB Linf
Data@vbK <- StockPars$Karray[,nyears+y] * ObsPars$Kbias # observed vB K
Data@vbt0 <- StockPars$t0array[,nyears+y] + ObsPars$t0bias # observed vB t0
Data@Mort <- StockPars$Marray[,nyears+y] * ObsPars$Mbias # natural mortality
Data@L50 <- StockPars$L50array[,nyears+y] * ObsPars$lenMbias # observed length at 50% maturity
Data@L95 <- StockPars$L95array[,nyears+y] * ObsPars$lenMbias # observed length at 95% maturity
Data@L95[Data@L95 > 0.9 * Data@vbLinf] <- 0.9 * Data@vbLinf[Data@L95 > 0.9 * Data@vbLinf] # Set a hard limit on ratio of L95 to Linf
Data@L50[Data@L50 > 0.9 * Data@L95] <- 0.9 * Data@L95[Data@L50 > 0.9 * Data@L95] # Set a hard limit on ratio of L95 to Linf
# --- Abundance ----
# Calculate vulnerable and spawning biomass abundance --
M_array <- array(0.5*StockPars$M_ageArray[,,nyears+y], dim=c(nsim, StockPars$maxage+1, nareas))
A <- apply(VBiomass_P[, , y, ] * exp(-M_array), 1, sum) # Abundance (mid-year before fishing)
Asp <- apply(SSB_P[, , y, ] * exp(-M_array), 1, sum) # Spawning abundance (mid-year before fishing)
Data@Abun <- A * ObsPars$Aerr_y[,nyears+yind]
Data@SpAbun <- Asp * ObsPars$Aerr_y[,nyears+yind]
# --- Catch-at-age ----
# previous CAA
oldCAA <- Data@CAA
Data@CAA <- array(0, dim = c(nsim, nyears + y - 1, StockPars$maxage+1))
Data@CAA[, 1:(nyears + y - interval[mm] - 1), ] <- oldCAA[, 1:(nyears + y - interval[mm] - 1), ]
# update CAA
CNtemp <- retA_P[,,yind+nyears, drop=FALSE] *
apply(N_P[,,yind,, drop=FALSE]*Sample_Area$CAA[,,nyears+yind,, drop=FALSE], 1:3, sum)
CNtemp[is.na(CNtemp)] <- tiny
CNtemp[!is.finite(CNtemp)] <- tiny
CAA <- simCAA(nsim, yrs=length(yind), StockPars$maxage+1, Cret=CNtemp, ObsPars$CAA_ESS, ObsPars$CAA_nsamp)
Data@CAA[, nyears + yind, ] <- CAA
# --- Catch-at-length ----
oldCAL <- Data@CAL
Data@CAL <- array(0, dim = c(nsim, nyears + y - 1, StockPars$nCALbins))
Data@CAL[, 1:(nyears + y - interval[mm] - 1), ] <- oldCAL[, 1:(nyears + y - interval[mm] - 1), ]
CAL <- array(NA, dim = c(nsim, interval[mm], StockPars$nCALbins))
if (!all(is.na(Data@Vuln_CAL))) {
Vuln_CAL <- replicate(nsim,Data@Vuln_CAL[1,])
Vuln_CAL <- replicate(length(yind),Vuln_CAL)
Vuln_CAL <- aperm(Vuln_CAL, c(2,1,3))
VList_dat <- lapply(1:nsim, calcV,
Len_age=StockPars$Len_age[,,nyears+yind, drop=FALSE],
LatASD=StockPars$LatASD[,,nyears+yind, drop=FALSE],
SLarray=Vuln_CAL,
n_age=StockPars$maxage+1,
nyears=length(yind), proyears = 0,
CAL_binsmid=StockPars$CAL_binsmid)
V_data <- aperm(array(as.numeric(unlist(VList_dat, use.names=FALSE)), dim=c(StockPars$maxage+1, length(yind), nsim)), c(3,1,2))
vn <- (apply(N_P[,,yind,, drop=FALSE]*Sample_Area$CAL[,,(nyears+yind),, drop=FALSE], c(1,2,3), sum) * V_data)
vn <- aperm(vn, c(1,3,2))
CALdat <- simCAL(nsim, nyears=length(yind), StockPars$maxage, ObsPars$CAL_ESS,
ObsPars$CAL_nsamp, StockPars$nCALbins, StockPars$CAL_binsmid, StockPars$CAL_bins,
vn=vn, retL=Vuln_CAL,
Linfarray=StockPars$Linfarray[,nyears + yind, drop=FALSE],
Karray=StockPars$Karray[,nyears + yind, drop=FALSE],
t0array=StockPars$t0array[,nyears + yind,drop=FALSE],
LenCV=StockPars$LenCV)
} else {
vn <- (apply(N_P*Sample_Area$CAL[,,(nyears+1):(nyears+proyears),], c(1,2,3), sum) *
V_P[,,(nyears+1):(nyears+proyears)] *
retA_P[,,(nyears+1):(nyears+proyears)]) # numbers at age that would be retained
vn <- aperm(vn, c(1,3,2))
CALdat <- simCAL(nsim, nyears=length(yind), maxage=StockPars$maxage,
CAL_ESS=ObsPars$CAL_ESS,
CAL_nsamp=ObsPars$CAL_nsamp,
nCALbins=StockPars$nCALbins,
CAL_binsmid=StockPars$CAL_binsmid,
CAL_bins=StockPars$CAL_bins,
vn=vn[,yind,, drop=FALSE],
retL=retL_P[,,nyears+yind, drop=FALSE],
Linfarray=StockPars$Linfarray[,nyears + yind, drop=FALSE],
Karray=StockPars$Karray[,nyears + yind, drop=FALSE],
t0array=StockPars$t0array[,nyears + yind,drop=FALSE],
LenCV=StockPars$LenCV)
}
Data@CAL[, nyears + yind, ] <- CALdat$CAL # observed catch-at-length
Data@ML <- cbind(Data@ML, CALdat$ML) # mean length
Data@Lc <- cbind(Data@Lc, CALdat$Lc) # modal length
Data@Lbar <- cbind(Data@Lbar, CALdat$Lbar) # mean length above Lc
Data@LFC <- FleetPars$L5_y[,nyears+y] * ObsPars$LFCbias # length at first capture
Data@LFS <- FleetPars$LFS_y[,nyears+y] * ObsPars$LFSbias # length at full selection
Data@Vmaxlen <- FleetPars$Vmaxlen_y[,nyears+y]
# --- Previous Management Recommendations ----
Data@MPrec <- MPCalcs$TACrec # last MP TAC recommendation
if (length(dim(Effort)) == 5) {
Data@MPeff <- Effort[, 1,1,mm, y-1] # last recommended effort
} else {
Data@MPeff <- Effort[, mm, y-1] # last recommended effort
}
# --- Store OM Parameters ----
# put all the operating model parameters in one table
ind <- which(lapply(StockPars, length) == nsim)
drop_srr <- which(names(ind)=='SRRpars')
ind <- ind[-drop_srr]
stock <- as.data.frame(StockPars[ind])
stock$Fdisc <- NULL
stock$CAL_bins <- NULL
stock$CAL_binsmid <- NULL
ind <- which(lapply(FleetPars, length) == nsim)
fleet <- as.data.frame(FleetPars[ind])
ind <- which(lapply(ImpPars, length) == nsim)
imp <- as.data.frame(ImpPars[ind])
refs <- RefPoints# [!names(RefPoints) %in% names(stock)]
OFLreal <- A * (1-exp(-RefPoints$FMSY)) # the true simulated Over Fishing Limit
OMtable <- data.frame(stock, fleet, imp, refs, ageM=StockPars$ageMarray[,nyears+y],
L5=FleetPars$L5_y[,nyears+y], LFS=FleetPars$LFS_y[,nyears+y],
Vmaxlen=FleetPars$Vmaxlen_y[,nyears+y],
LR5=FleetPars$LR5_y[,nyears], LFR=FleetPars$LFR_y[,nyears+y],
Rmaxlen=FleetPars$Rmaxlen_y[,nyears+y],
DR=FleetPars$DR_y[,nyears+y], OFLreal, maxF=StockPars$maxF,
A=A, Asp=Asp, CurrentYr=OM@CurrentYr)
OMtable <- OMtable[,order(names(OMtable))]
Data@OM <- OMtable
Data@Misc <- Misc
if (!is.null(RealData)) {
if (length(RealData@Misc) && !is.null(names(RealData@Misc))) {
Data@Misc[names(RealData@Misc)] <- RealData@Misc[names(RealData@Misc)]
}
}
Data
}
#' Simulate Catch-at-Age Data
#'
#' CAA generated with either a multinomial or logistic normal observation model from retained catch-at-age
#' array
#'
#' @param nsim Number of simulations
#' @param yrs Number of years
#' @param n_age Number of age classes
#' @param Cret Retained Catch at age in numbers - array(sim, years, maxage+1)
#' @param CAA_ESS CAA effective sample size. If greater than 1, then this is the multinomial distribution sample size.
#' If less than 1, this is the coefficient of variation for the logistic normal distribution (see details).
#' @param CAA_nsamp CAA sample size
#' @details The logistic normal generates the catch-at-age sample by first sampling once from a multivariate normal distribution
#' with the mean vector equal to the logarithm of the proportions-at-age and the diagonal of the covariance matrix is the square of the
#' product of the CV and the log proportions (all off-diagonals are zero). The sampled vector is then converted to proportions
#' with the softmax function and expanded to numbers (CAA_nsamp). This method allows for simulating fractional values in the
#' catch-at-age matrix.
#' @return CAA array
simCAA <- function(nsim, yrs, n_age, Cret, CAA_ESS, CAA_nsamp) {
# generate CAA from retained catch-at-age
CAA <- array(NA, dim = c(nsim, yrs, n_age)) # Catch at age array
# a multinomial observation model for catch-at-age data
# logistic normal if CAA_ESS < 1
if(any(CAA_ESS < 1) && !requireNamespace("mvtnorm", quietly = TRUE)) stop("The mvtnorm package is needed.")
for (i in 1:nsim) {
for (j in 1:yrs) {
if (!sum(Cret[i, ,j])) {
CAA[i, j, ] <- 0
} else {
if(CAA_ESS[i] < 1) {
if(CAA_ESS[i] <= 0) {
CAA[i, j, ] <- 0
} else {
log_PAA <- log(Cret[i, , j]/sum(Cret[i, , j]))
vcov_PAA <- (CAA_ESS[i] * log_PAA)^2 * diag(length(log_PAA))
samp_PAA <- mvtnorm::rmvnorm(1, log_PAA, vcov_PAA) %>% ilogitm()
CAA[i, j, ] <- CAA_nsamp[i] * samp_PAA
}
} else {
CAA[i, j, ] <- ceiling(-0.5 + rmultinom(1, CAA_ESS[i], Cret[i, ,j]) * CAA_nsamp[i]/CAA_ESS[i])
}
}
}
}
CAA
}
#' Simulate Catch-at-Length Data
#'
#' Simulate CAL and calculate length-at-first capture (LFC),
#' mean length (ML), modal length (Lc), and mean length over modal length (Lbar)
#'
#' @param nsim Number of simulations
#' @param nyears Number of years
#' @param maxage Maximum age
#' @param CAL_ESS CAA effective sample size
#' @param CAL_nsamp CAA sample size
#' @param nCALbins number of CAL bins
#' @param CAL_binsmid mid-points of CAL bins
#' @param CAL_bins Boundary of CAL bins
#' @param vn Vulnerable numbers-at-age
#' @param retL Retention at length curve
#' @param Linfarray Array of Linf values by simulation and year
#' @param Karray Array of K values by simulation and year
#' @param t0array Array of t0 values by simulation and year
#' @param LenCV CV of length-at-age#'
#' @return named list with CAL array and LFC, ML, & Lc vectors
#'
simCAL <- function(nsim, nyears, maxage, CAL_ESS, CAL_nsamp, nCALbins, CAL_binsmid, CAL_bins,
vn, retL, Linfarray, Karray, t0array, LenCV) {
# a multinomial observation model for catch-at-length data
# assumed normally-distributed length-at-age truncated at 2 standard deviations from the mean
CAL <- array(NA, dim=c(nsim, nyears, nCALbins))
# Generate size comp data with variability in age
runParallel <- snowfall::sfIsRunning()
if (runParallel) {
tempSize <- snowfall::sfLapply(1:nsim, genSizeCompWrap, vn, CAL_binsmid, CAL_bins,
retL, CAL_ESS, CAL_nsamp,
Linfarray, Karray, t0array, LenCV, truncSD=2)
} else {
# vn <<- vn
# CAL_binsmid <<- CAL_binsmid
# CAL_bins <<- CAL_bins
# retL <<- retL
# CAL_ESS <<- CAL_ESS
# CAL_nsamp <<- CAL_nsamp
# Linfarray <<- Linfarray
# Karray <<- Karray
# t0array <<- t0array
# LenCV <<- LenCV
#
tempSize <- lapply(1:nsim, genSizeCompWrap, vn, CAL_binsmid, CAL_bins, retL, CAL_ESS,
CAL_nsamp,
Linfarray, Karray, t0array, LenCV, truncSD=2)
}
CAL <- aperm(array(as.numeric(unlist(tempSize, use.names=FALSE)),
dim=c(nyears, length(CAL_binsmid), nsim)), c(3,1,2))
# calculate LFC - length-at-first capture - 5th percentile
LFC <- rep(NA, nsim)
LFC <- unlist(lapply(tempSize, function(x) getfifth(x[nyears, ], CAL_binsmid)))
LFC[is.na(LFC)] <- 1
LFC[LFC<1] <- 1
# Mean Length
temp <- CAL * rep(CAL_binsmid, each = nsim * nyears)
ML <- apply(temp, 1:2, sum)/apply(CAL, 1:2, sum)
ML[!is.finite(ML)] <- 0
# Lc - modal length
Lc <- array(CAL_binsmid[apply(CAL, 1:2, which.max)], dim = c(nsim, nyears))
# Lbar - mean length above Lc
nuCAL <- CAL
for (i in 1:nsim) {
for (j in 1:nyears) {
# nuCAL[i, j, 1:match(max(1, Lc[i, j]), CAL_binsmid, nomatch=1)] <- NA
lcbin <- max(1,match(max(1, Lc[i, j]), CAL_binsmid, nomatch=1)-1)
nuCAL[i, j, 1:lcbin] <- NA
}
}
temp <- nuCAL * rep(CAL_binsmid, each = nsim * nyears)
Lbar <- apply(temp, 1:2, sum, na.rm=TRUE)/apply(nuCAL, 1:2, sum, na.rm=TRUE)
Lbar[!is.finite(Lbar)] <- 0
out <- list()
out$CAL <- CAL
out$LFC <- LFC
out$ML <- ML
out$Lc <- Lc
out$Lbar <- Lbar
out
}
#' Wrapper for C++ function to generate length composition
#'
#' And other internal related functions
#'
#' @param i Simulation number
#' @param vn Array of vulnerable numbers
#' @param CAL_binsmid Mid-points of CAL bins
#' @param CAL_bins Boundary of length bins
#' @param retL Array of retention-at-length
#' @param CAL_ESS CAL effective sample size
#' @param CAL_nsamp CAL sample size
#' @param Linfarray Matrix of Linf
#' @param Karray Matrix of K values
#' @param t0array Matrix of t0 values
#' @param LenCV Vector of LenCV
#' @param truncSD Numeric. Number of standard deviations to truncate normal d
#' distribution
#'
#' @return Generated length composition from `genSizeComp`
#'
#' @keywords internal
genSizeCompWrap <- function(i, vn, CAL_binsmid, CAL_bins, retL,
CAL_ESS, CAL_nsamp,
Linfarray, Karray, t0array,
LenCV, truncSD=2) {
VulnN <- as.matrix(vn[i,,])
if (ncol(VulnN)>1) {
VulnN <- VulnN/rowSums(VulnN) * CAL_nsamp[i] # get relative numbers at age
} else {
VulnN <- VulnN/sum(VulnN) * CAL_nsamp[i] # get relative numbers at age
}
# VulnN <- round(VulnN,0) # convert to integers
nyrs <- nrow(as.matrix(Linfarray[i,]))
if (nyrs == 1) VulnN <- t(VulnN)
retLa <- as.matrix(retL[i,,])
# assumes lengths sampled throughout years
lens <- genSizeComp(VulnN, CAL_binsmid, CAL_bins, retLa,
CAL_ESS=CAL_ESS[i], CAL_nsamp=CAL_nsamp[i],
Linfs=Linfarray[i,], Ks=Karray[i,], t0s=t0array[i,],
LenCV=LenCV[i], truncSD)
# snapshot length comp
# lens <- genSizeComp2(VulnN, CAL_binsmid, CAL_bins, retLa,
# CAL_ESS=CAL_ESS[i], CAL_nsamp=CAL_nsamp[i],
# Linfs=Linfarray[i,], Ks=Karray[i,], t0s=t0array[i,],
# LenCV=LenCV[i], truncSD)
lens[!is.finite(lens)] <- 0
lens
}
#' @describeIn genSizeCompWrap Internal function to calculate fifth percentile of size composition
#' @param lenvec Vector of lengths
getfifth <- function(lenvec, CAL_binsmid) {
temp <- rep(CAL_binsmid, lenvec)
if(sum(lenvec)==0) return(NA)
dens <- try(density(temp), silent=TRUE)
if(!methods::is(dens, "density")) return(NA)
dens$x[min(which(cumsum(dens$y/sum(dens$y)) >0.05))]
}
UpdateObs <- function(sl, obsval, OMval, RealData, SimData, msg){
RealVal <- slot(RealData, sl)[1]
SimVal <- slot(SimData, sl)
if (!is.na(RealVal) & !all(SimVal == tiny)) {
if (msg)
message_info(paste0('Updating Observation Error for `OM@cpars$Data@', sl, '`'))
# bias
if (sl == 'vbt0') {
calcBias <- RealVal-OMval
} else {
calcBias <- RealVal/OMval
}
return(calcBias)
} else {
return(obsval)
}
}
UpdateSlot <- function(sl, RealData, SimData, msg) {
RealVal <- slot(RealData, sl)[1]
SimVal <- slot(SimData, sl)
if (!is.na(RealVal) & !all(SimVal == tiny)) {
if (msg)
message_info(paste0('Using `OM@cpars$Data@', sl, '` (', RealVal, ')'))
nrep <- length(slot(SimData, sl))
return(rep(RealVal, nrep))
}
return(slot(SimData, sl))
}
check_Index_Fit <- function(Stats, index=1) {
ac_err <- sum(!is.finite(Stats$AC))
sd_err <- sum(!is.finite(Stats$SD))
if (ac_err | sd_err) {
if (is.numeric(index)) {
warning('An error occurred in calculating statistical properties of fit to Additional Index ', index,
' (possibly because there was only one observed data point). \nUsing the index observation error for slot `Ind` from `Obs` object (or possibly conditioned if `cpars$Data@Ind` was provided).\nUse `cpars$AddIerr` to manually set the observation error.')
} else {
warning('An error occurred in calculating statistical properties of fit to Index ', index,
' (possibly because there was only one observed data point). \nUsing the index observation error for slot `Ind` from `Obs` object (or possibly conditioned if `cpars$Data@Ind` was provided).\nUse `cpars$AddIerr` to manually set the observation error.')
}
return(FALSE)
}
TRUE
}
AddRealData <- function(SimData, RealData, ObsPars, StockPars, FleetPars, nsim,
nyears, proyears, SampCpars, msg, control, Sample_Area) {
Data_out <- SimData
if (msg)
message('Updating Simulated Data with Real Data from `OM@cpars$Data`')
# check last year
if (!is.na(RealData@LHYear) && !SimData@LHYear == RealData@LHYear) {
warning('`Fleet@CurrentYear` (', SimData@LHYear, ') is not the same as `OM@cpars$Data@LHYear` (', RealData@LHYear, ')')
}
# check maxage
if (!is.na(RealData@MaxAge) && !SimData@MaxAge == RealData@MaxAge) {
warning('`Stock@MaxAge` (', SimData@MaxAge, ') is not the same as `OM@cpars$Data@MaxAge` (', RealData@MaxAge, ')')
}
# check dimensions of real data CAA
if (!all(is.na(RealData@CAA)) && dim(RealData@CAA)[3] > 1) {
if (!dim(RealData@CAA)[3] == SimData@MaxAge+1)
stop('`OM@cpars$Data@CAA` has incorrect dimensions, should be `Stock@maxage+1` age-classes')
}
# update static slots
slts <- c('Name', 'Common_Name', 'Species', 'Region', 'LHYear', 'Units')
for (sl in slts)
slot(Data_out, sl) <- slot(RealData, sl)
Data_out@MaxAge <- SimData@MaxAge
if (!is.na(RealData@MPrec)) {
Data_out@MPrec <- rep(RealData@MPrec, nsim)
} else {
Data_out@MPrec <- SimData@MPrec
}
Data_out@MPeff <- SimData@MPeff
Data_out@nareas <- SimData@nareas
Data_out@LHYear <- SimData@LHYear
# ---- Update Life-history parameters ----
Data_out@Mort <- UpdateSlot('Mort', RealData, SimData, msg)
ObsPars$Mbias <- UpdateObs('Mort', ObsPars$Mbias, StockPars$Marray[, nyears],
RealData, SimData, msg)
Data_out@vbLinf <- UpdateSlot('vbLinf', RealData, SimData, msg)
ObsPars$Linfbias <- UpdateObs('vbLinf', ObsPars$Linfbias, StockPars$Linfarray[, nyears],
RealData, SimData, msg)
Data_out@vbK <- UpdateSlot('vbK', RealData, SimData, msg)
ObsPars$Kbias <- UpdateObs('vbK', ObsPars$Kbias, StockPars$Karray[, nyears],
RealData, SimData, msg)
Data_out@vbt0 <- UpdateSlot('vbt0', RealData, SimData, msg)
ObsPars$t0bias <- UpdateObs('vbt0', ObsPars$t0bias, StockPars$t0array[, nyears],
RealData, SimData, msg)
Data_out@CV_Mort <- UpdateSlot('CV_Mort', RealData, SimData, msg)
Data_out@CV_vbLinf <- UpdateSlot('CV_vbLinf', RealData, SimData, msg)
Data_out@CV_vbK <- UpdateSlot('CV_vbK', RealData, SimData, msg)
Data_out@CV_vbt0 <- UpdateSlot('CV_vbt0', RealData, SimData, msg)
Data_out@wla <- UpdateSlot('wla', RealData, SimData, msg)
Data_out@wlb <- UpdateSlot('wlb', RealData, SimData, msg)
Data_out@CV_wla <- UpdateSlot('CV_wla', RealData, SimData, msg)
Data_out@CV_wlb <- UpdateSlot('CV_wlb', RealData, SimData, msg)
Data_out@steep <- UpdateSlot('steep', RealData, SimData, msg)
Data_out@CV_steep <- UpdateSlot('CV_steep', RealData, SimData, msg)
ObsPars$hbias <- UpdateObs('steep', ObsPars$hbias, StockPars$hs,
RealData, SimData, msg)
Data_out@sigmaR <- UpdateSlot('sigmaR', RealData, SimData, msg)
Data_out@CV_sigmaR <- UpdateSlot('CV_sigmaR', RealData, SimData, msg)
# TODO ObsPars$sigmaRbias
Data_out@L50 <- UpdateSlot('L50', RealData, SimData, msg)
Data_out@CV_L50 <- UpdateSlot('CV_L50', RealData, SimData, msg)
ObsPars$lenMbias <- UpdateObs('L50', ObsPars$lenMbias, StockPars$L50,
RealData, SimData, msg)
Data_out@L95 <- UpdateSlot('L95', RealData, SimData, msg)
Data_out@LenCV <- UpdateSlot('LenCV', RealData, SimData, msg)
Data_out@LFC <- UpdateSlot('LFC', RealData, SimData, msg)
Data_out@CV_LFC <- UpdateSlot('CV_LFC', RealData, SimData, msg)
ObsPars$LFCbias <- UpdateObs('LFC', ObsPars$LFCbias, FleetPars$L5_y[,nyears],
RealData, SimData, msg)
Data_out@LFS <- UpdateSlot('LFS', RealData, SimData, msg)
Data_out@CV_LFS <- UpdateSlot('CV_LFS', RealData, SimData, msg)
ObsPars$LFSbias <- UpdateObs('LFS', ObsPars$LFSbias, FleetPars$LFS_y[,nyears],
RealData, SimData, msg)
# Data_out@Vmaxlen <- UpdateSlot('Vmaxlen', RealData, SimData, msg)
if (length(RealData@Year)>1) {
# check years
if (all(RealData@Year !=SimData@Year)) {
stop('`OM$cpars$Data@Year` does not match `SimData@Year`. \nAre `Fleet@nyears` and `Fleet@CurrentYr` correct?')
}
Data_out@Year <- RealData@Year
}
# ---- Update Catch ----
if (!all(is.na(RealData@Cat[1,]))) {
if (msg)
message('Updating Simulated Catch from `OM@cpars$Data@Cat`')
Data_out@Cat <- matrix(RealData@Cat[1,1:nyears], nrow=nsim, ncol=nyears, byrow=TRUE)
Data_out@CV_Cat <- matrix(RealData@CV_Cat[1,1:nyears], nrow=nsim, ncol=nyears, byrow=TRUE)
simcatch <- apply(StockPars$CBret, c(1,3), sum)
simcatch[simcatch==0] <- tiny
Cbias <- matrix(apply(Data_out@Cat, 1, mean)/apply(simcatch, 1, mean),
nrow=nsim, ncol=nyears+proyears)
Cerr <- Data_out@Cat/(simcatch*Cbias[,1:nyears])
t1<- Cerr[,max(nyears-10, 1):nyears]/apply(Cerr[,max(nyears-10, 1):nyears],1,mean)
SDs <- apply(log(t1), 1, sd)
Cerr_proj <- matrix(NA, nsim, proyears)
nas <- which(is.na(SDs))
Cerr_proj[nas,] <- tiny # no catch
if (!all(is.na(SDs))) {
for (i in (1:nsim)[!is.na(SDs)]) {
Cerr_proj[i,] <- exp(rnorm(proyears, -((SDs[i]^2)/2), SDs[i]))
}
}
Cerr <- cbind(Cerr, Cerr_proj)
if(!is.null(SampCpars$Cobs_y)) {
if (msg) message_info('Catch observation error detected in cpars (`cpars$Cobs_y`). Not updating catch obs error')
} else {
if(!is.null(SampCpars$Cbias)) {
if (msg) message_info('Catch Cbias detected in cpars (`cpars$Cbias`). Not updating catch bias')
} else {
if (msg) message_info('Updating Catch bias from `OM@cpars$Data@Cat`')
ObsPars$Cbias <- Cbias[,1]
}
if(!is.null(SampCpars$Cerr_y)) {
if (msg) message_info('Catch variability detected in cpars (`cpars$Cerr_y`). Not updating catch error')
} else {
if (msg) message_info('Updating Catch variability from `OM@cpars$Data@Cat`')
ObsPars$Cerr_y <- Cerr
}
if (msg) message_info('Updating catch observation error from `OM@cpars$Data@Cat`')
ObsPars$Cobs_y <- Cerr * Cbias
}
}
# ---- Update Effort -----
# TODO
# ---- Update Index (total biomass) ----
if (!all(is.na(RealData@Ind[1,]))) { # Index exists
fit_ind <- Fit_Index(ind_slot='Ind', indcv_slot="CV_Ind", Data_out,
RealData, StockPars, ObsPars, SampCpars, nsim, nyears, proyears, msg=msg)
Data_out <- fit_ind$Data_out
ObsPars <- fit_ind$ObsPars
}
# ---- Update Index (spawning biomass) ----
if (!all(is.na(RealData@SpInd[1,]))) { # Index exists
fit_ind <- Fit_Index(ind_slot='SpInd', indcv_slot="CV_SpInd", Data_out,
RealData, StockPars, ObsPars, SampCpars, nsim, nyears, proyears, msg=msg)
Data_out <- fit_ind$Data_out
ObsPars <- fit_ind$ObsPars
}
# ---- Index (vulnerable biomass) ----
if (!all(is.na(RealData@VInd[1,]))) { # Index exists
fit_ind <- Fit_Index(ind_slot='VInd', indcv_slot="CV_VInd", Data_out,
RealData, StockPars, ObsPars, SampCpars, nsim, nyears, proyears, msg=msg)
Data_out <- fit_ind$Data_out
ObsPars <- fit_ind$ObsPars
}
# ---- Add Additional Indices ----
if (!all(is.na(RealData@AddInd))) {
if (msg)
message('Adding Additional Indices to Simulated Data from `OM@cpars$Data@AddInd`')
n.ind <- dim(RealData@AddInd)[2]
Data_out@AddInd <- Data_out@CV_AddInd <- array(NA, dim=c(nsim, n.ind, nyears))
fitbeta <- fitIerr <- TRUE
if (!is.null(SampCpars$AddIbeta)) {
if (any(dim(SampCpars$AddIbeta) != c(nsim, n.ind)))
stop("cpars$AddIbeta must be dimensions c(nsim, n.ind)")
if (msg) message_info('cpars$AddIbeta detected. Not updating beta for additional indices')
ObsPars$AddIbeta <- SampCpars$AddIbeta
fitbeta <- FALSE
} else {
if (msg) message_info('Updating beta for additional indices from real data')
ObsPars$AddIbeta <- matrix(NA, nsim, n.ind)
}
if (!is.null(SampCpars$AddIerr)) {
if (any(dim(SampCpars$AddIerr) != c(nsim, n.ind, nyears+proyears)))
stop("cpars$AddIerr must be dimensions c(nsim, n.ind, nyears+proyears)")
if (msg) message_info('cpars$AddIerr detected. Not updating observation variability for additional indices')
ObsPars$AddIerr <- SampCpars$AddIerr
fitIerr <- FALSE
} else {
if (msg) message_info('Updating observation variability (AddIerr) for additional indices from real data')
ObsPars$AddIerr <- array(NA, dim=c(nsim, n.ind, nyears+proyears))
}
if (!is.null(SampCpars$AddIunits) && length(SampCpars$AddIunits) != n.ind)
stop("cpars$AddIunits must be length n.ind")
AddIunits <- RealData@AddIunits
if (all(is.na(AddIunits))) AddIunits <- rep(1, n.ind)
if (!is.null(SampCpars$AddIunits)) AddIunits <- SampCpars$AddIunits
Data_out@AddIunits <- AddIunits
AddIndType <- RealData@AddIndType
if (all(is.na(AddIndType))) AddIndType <- rep(1, n.ind)
Data_out@AddIndType <- AddIndType
ObsPars$AddInd_Stat <- vector(mode = "list", length = n.ind)
UnitsTab <- data.frame(n=1:0, units=c('biomass', 'numbers'))
TypeTab <- data.frame(n=1:3, type=c('total', 'spawning', 'vuln.'))
Data_out@AddIndV <- array(NA, dim=c(nsim, n.ind, Data_out@MaxAge+1))
# loop over additional indices
for (i in 1:n.ind) {
units <- UnitsTab$units[match(AddIunits[i], UnitsTab$n)]
type <- TypeTab$type[match(AddIndType[i], TypeTab$n)]
if(msg) message_info("Additional index", i, '-', type, 'stock', paste0('(', units, ')'))
nyrs <- min(length(RealData@AddInd[1,i,]), nyears)
ind <- RealData@AddInd[1,i,1:nyrs]
cv_ind <- RealData@CV_AddInd[1,i,1:nyrs]
if (nyrs < nyears) {
ind <- c(ind, rep(NA,nyears-nyrs))
cv_ind <- c(cv_ind, rep(NA,nyears-nyrs))
}
Data_out@AddInd[,i,] <- matrix(ind, nrow=nsim, ncol=nyears, byrow=TRUE)
Data_out@CV_AddInd[,i,] <- matrix(cv_ind, nrow=nsim, ncol=nyears, byrow=TRUE)
if (all(is.na(RealData@AddIndV[1,, ]))) {
# no vulnerability-at-age included
Ind_V <- rep(1, Data_out@MaxAge+1)
} else {
Ind_V <- RealData@AddIndV[1,i, ]
}
Data_out@AddIndV[,i,] <- t(replicate(nsim,Ind_V))
if (!all(is.na(ind))) {
# check dimensions
if (!length(Ind_V) == Data_out@MaxAge+1)
stop('Vulnerability-at-age for additional index ', i, ' is not length `maxage`+1' )
# calculate simulated index
if (AddIunits[i]) {
if (AddIndType[i]==1) SimIndex <- apply(StockPars$Biomass, c(1, 2, 3), sum) # Total Biomass-based index
if (AddIndType[i]==2) SimIndex <- apply(StockPars$SSB, c(1, 2, 3), sum) # Spawning Biomass-based index
if (AddIndType[i]==3) SimIndex <- apply(StockPars$VBiomass, c(1, 2, 3), sum) # vuln Biomass-based index
} else {
if (AddIndType[i]==1) SimIndex <- apply(StockPars$N, c(1, 2, 3), sum) # Total Abundance-based index
if (AddIndType[i]==2) SimIndex <- apply(StockPars$N, c(1, 2, 3), sum) * StockPars$Mat_age[,,1:nyears] # Spawning abundance-based index
if (AddIndType[i]==3) SimIndex <- apply(StockPars$N, c(1, 2, 3), sum) * FleetPars$V_real[,,1:nyears] # Vulnerable abundance-based index
}
Ind_V <- matrix(Ind_V, nrow=SimData@MaxAge+1, ncol= nyears)
Ind_V <- replicate(nsim, Ind_V) %>% aperm(., c(3,1,2))
# apply vulnerability-at-age and sum over age-classes
SimIndex <- apply(SimIndex*Ind_V, c(1,3), sum)
# Fit to observed index and generate residuals for projections
if (fitbeta) {
beta <- rep(NA, nsim)
} else {
beta <- ObsPars$AddIbeta[,i]
}
# Calculate residuals (with or without estimated beta)
Res_List <- lapply(1:nsim, function(x) Calc_Residuals(sim.index=SimIndex[x,],
obs.ind=ind,
beta=beta[x]))
lResids_Hist <- do.call('rbind', lapply(Res_List, '[[', 1))
if (fitbeta)
ObsPars$AddIbeta[,i] <- as.vector(do.call('cbind', lapply(Res_List, '[[', 2)))
if (msg & fitbeta)
message_info(paste0('Estimated beta for Additional Index ', i, ':'),
paste0(range(round(ObsPars$AddIbeta[,i],2)), collapse = "-"),
"Use `cpars$AddIbeta` to override")
# Calculate statistics
Stats_List <- lapply(1:nsim, function(x) Calc_Stats(lResids_Hist[x,]))
Stats <- do.call('rbind', Stats_List)
check_Index <- check_Index_Fit(Stats, i)
if (check_Index) {
# Generate residuals for projections
Resid_Hist <- exp(lResids_Hist) # historical residuals in normal space
Resid_Proj <- Gen_Residuals(Stats, nsim, proyears)
if (fitIerr) ObsPars$AddIerr[,i, ] <- cbind(Resid_Hist, Resid_Proj)
ObsPars$AddInd_Stat[[i]] <- Stats[,1:2] # index fit statistics
} else {
ObsPars$AddIerr[,i, ] <- ObsPars$Ierr_y
ObsPars$AddInd_Stat[[i]] <- Stats[,1:2] # index fit statistics
}
}
}
}
# ---- Update Recruitment ----
if (!all(is.na(RealData@Rec))) {
if (msg)
message('Updating Simulated Recruitment Data from `OM@cpars$Data@Rec`')
Data_out@Rec <- matrix(RealData@Rec[1,1:nyears], nrow=nsim, ncol=nyears, byrow=TRUE)
dd <- dim(RealData@CV_Rec)
if (dd[2]<nyears) {
RealData@CV_Rec <- matrix(RealData@CV_Rec[1,1], nrow=nsim, ncol=nyears, byrow=TRUE)
}
Data_out@CV_Rec <- matrix(RealData@CV_Rec[1,1:nyears], nrow=nsim, ncol=nyears, byrow=TRUE)
# Calculate Error
Rec <- apply(StockPars$N[, 1, , ], c(1, 2), sum) # simulated recruitment
Recbias <- matrix(apply(Data_out@Rec, 1, mean)/apply(Rec, 1, mean),
nrow=nsim, ncol=nyears+proyears)
Rec_err <- Data_out@Rec/(Rec*Recbias[,1:nyears])
t1 <- Rec_err[,max(nyears-10, 1):nyears]/apply(Rec_err[,max(nyears-10, 1):nyears],1,mean) # last 10 years used for projections
SDs <- apply(log(t1), 1, sd)
Rec_err_proj <- matrix(NA, nsim, proyears)
for (i in 1:nsim) {
Rec_err_proj[i,] <- exp(rnorm(proyears, -((SDs[i]^2)/2), SDs[i]))
}
Rec_err_proj <- Rec_err_proj * Recbias[,(nyears+1):(nyears+proyears)]
Rec_err <- cbind(Rec_err, Rec_err_proj)
ObsPars$Recerr_y <- Rec_err * Recbias
ObsPars$Recsd <- SDs
ObsPars$Recbias <- Recbias
}
# ---- Update CAA ----
if (!all(is.na(RealData@CAA)) & !all(RealData@CAA ==0)) {
if (msg)
message('Updating Simulated Catch-at-Age Data from `OM@cpars$Data@CAA`. Note: CAA_ESS is currently NOT updated')
Data_out@CAA <- aperm(replicate(nsim, RealData@CAA[1,1:nyears,]),c(3,1,2))
Data_out@Vuln_CAA <- RealData@Vuln_CAA
# Get average sample size
nsamp <- ceiling(mean(apply(RealData@CAA[1,1:nyears,], 1, sum)))
ObsPars$CAA_nsamp <- rep(nsamp, nsim)
}
# ---- Update CAL ----
if (!all(is.na(RealData@CAL)) & !all(RealData@CAL ==0)) {
# check length bins
if (!all(RealData@CAL_bins %in% StockPars$CAL_bins)) {
warning('cpars$Data@CAL_bins cannot be matched with Simulated Data@CAL_bins. Add cpars$Data@CAL_bins to cpars$CAL_bins. cpars$Data@CAL are NOT being used')
} else {
if (msg)
message('Updating Simulated Catch-at-Length Data, Obs@CAL_nsamp, and Obs@CAL_ESS from `OM@cpars$Data@CAL`')
# match length bins
ind <- match(RealData@CAL_mids, StockPars$CAL_binsmid)
dd <- dim(Data_out@CAL)
Data_out@CAL <- array(0, dim=dd)
CAL <- Data_out@CAL[1,,]
CAL[,ind] <- RealData@CAL[1,,] # replace with real CAL
CAL <- aperm(replicate(nsim, CAL[1:nyears,]),c(3,1,2))
Data_out@CAL <- CAL
Data_out@Vuln_CAL <- RealData@Vuln_CAL
# Get average sample size
nsamp <- ceiling(mean(apply(RealData@CAL[1,1:nyears,], 1, sum, na.rm=TRUE), na.rm=TRUE))
ObsPars$CAL_nsamp <- rep(nsamp, nsim)
# calculate effective sample size for projections
# vuln N-at-age in last year with length data
nas <- !apply(CAL[1,,], 2, is.na)
yr.ind <- max(which(apply(nas, 1, prod)>0))
if (!is.null(control$CAL) && control$CAL == 'removals') {
vn <- apply(StockPars$N[,,yr.ind,, drop=FALSE], c(1,2,3), sum) *
FleetPars$V_real[,,yr.ind, drop=FALSE]
# numbers at age in population that would be removed
vn <- aperm(vn, c(1,3, 2))
doopt <- optimise(optESS, c(10, 10000), vn, StockPars,
FleetPars$SLarray_real[1,,yr.ind, drop=FALSE], yr.ind, ObsPars, Data_out, CAL)
ObsPars$CAL_ESS <- rep(doopt$minimum, nsim)
} else {
vn <- apply(StockPars$N[,,yr.ind,, drop=FALSE], c(1,2,3), sum) *
FleetPars$retA_real[,,yr.ind, drop=FALSE]
# numbers at age in population that would be removed
vn <- aperm(vn, c(1,3, 2))
doopt <- optimise(optESS, c(10, 10000), vn, StockPars,
FleetPars$retL_real[1,,yr.ind, drop=FALSE], yr.ind, ObsPars, Data_out, CAL)
ObsPars$CAL_ESS <- rep(doopt$minimum, nsim)
}
}
}
# ---- Depletion ----
Data_out@Dep <- UpdateSlot('Dep', RealData, SimData, msg)
Data_out@CV_Dep <- UpdateSlot('CV_Dep', RealData, SimData, msg)
ObsPars$Dbias <- UpdateObs('Dep', ObsPars$Dbias, StockPars$Depletion,
RealData, SimData, msg)
# ---- Index Reference -----
Data_out@Iref <- UpdateSlot('Iref', RealData, SimData, msg)
Data_out@CV_Iref <- UpdateSlot('CV_Iref', RealData, SimData, msg)
# ObsPars$Irefbias <- rep(NA, nsim) # not calculated
# ---- Misc ----
if (length(RealData@Misc) && !is.null(names(RealData@Misc))) {
Data_out@Misc[names(RealData@Misc)] <- RealData@Misc[names(RealData@Misc)]
if (msg)
message('Adding named list from `OM@cpars$Data@Misc`')
}
NotUpdated <- function(RealData, sl, msg) {
if (!all(is.na(slot(RealData, sl)))) {
if (msg)
message_info(paste0('Data detected in `OM@cpars$Data@', sl, '` but is NOT being used.'))
}
}
NotUpdated(RealData, 'ML', msg)
NotUpdated(RealData, 'Lc', msg)
NotUpdated(RealData, 'Lbar', msg)
NotUpdated(RealData, 'Abun', msg)
NotUpdated(RealData, 'SpAbun', msg)
NotUpdated(RealData, 'FMSY_M', msg)
NotUpdated(RealData, 'BMSY_B0', msg)
NotUpdated(RealData, 'Cref', msg)
NotUpdated(RealData, 'Bref', msg)
NotUpdated(RealData, 'AvC', msg)
NotUpdated(RealData, 'Dt', msg)
NotUpdated(RealData, 'Ref', msg)
list(Data=Data_out, ObsPars=ObsPars)
}
optESS <- function(CAL_ESS, vn, StockPars, sel_ret, yr.ind, ObsPars, Data_out, CAL) {
set.seed(101)
tempSize <- genSizeCompWrap(1, vn, StockPars$CAL_binsmid, StockPars$CAL_bins,
sel_ret, CAL_ESS,
ObsPars$CAL_nsamp,
StockPars$Linfarray[1, yr.ind, drop=FALSE],
StockPars$Karray[1, yr.ind, drop=FALSE],
StockPars$t0array[1, yr.ind, drop=FALSE],StockPars$LenCV, truncSD=2)
tempSize <- tempSize/sum(tempSize)
obsSize <- Data_out@CAL[1,yr.ind,]/sum(Data_out@CAL[1,yr.ind,])
sum((obsSize-tempSize)^2)
}
AddRealData_MS <- function(SimData,
RealData,
StockPars,
FleetPars,
ObsPars,
SampCpars,
map.stocks,
CBret,
VBF,
p,
f,
nsim,
nyears,
proyears,
msg=TRUE,
control) {
Data_out <- SimData
if (msg)
message('Updating Simulated Data with Real Data from `OM@cpars$Data`')
# check last year
if (!is.na(RealData@LHYear) && !SimData@LHYear == RealData@LHYear) {
warning('`Fleet@CurrentYear` (', SimData@LHYear, ') is not the same as `OM@cpars$Data@LHYear` (', RealData@LHYear, ')')
}
# check maxage
if (!is.na(RealData@MaxAge) && !SimData@MaxAge == RealData@MaxAge) {
warning('`Stock@MaxAge` (', SimData@MaxAge, ') is not the same as `OM@cpars$Data@MaxAge` (', RealData@MaxAge, ')')
}
# check dimensions of real data CAA
if (!all(is.na(RealData@CAA)) && dim(RealData@CAA)[3] > 1) {
if (!dim(RealData@CAA)[3] == SimData@MaxAge+1)
stop('`OM@cpars$Data@CAA` has incorrect dimensions, should be `Stock@maxage+1` age-classes')
}
# update static slots
slts <- c('Name', 'Common_Name', 'Species', 'Region', 'LHYear', 'Units')
for (sl in slts)
slot(Data_out, sl) <- slot(RealData, sl)
Data_out@MaxAge <- SimData@MaxAge
if (!is.na(RealData@MPrec)) {
Data_out@MPrec <- rep(RealData@MPrec, nsim)
} else {
Data_out@MPrec <- SimData@MPrec
}
Data_out@MPeff <- SimData@MPeff
Data_out@nareas <- SimData@nareas
Data_out@LHYear <- SimData@LHYear
# ---- Update Life-history parameters ----
Data_out@Mort <- UpdateSlot('Mort', RealData, SimData, msg)
ObsPars[[p]][[f]]$Mbias <- UpdateObs('Mort', ObsPars[[p]][[f]]$Mbias, StockPars[[p]]$Marray[, nyears],
RealData, SimData, msg)
Data_out@vbLinf <- UpdateSlot('vbLinf', RealData, SimData, msg)
ObsPars[[p]][[f]]$Linfbias <- UpdateObs('vbLinf', ObsPars[[p]][[f]]$Linfbias, StockPars[[p]]$Linfarray[, nyears],
RealData, SimData, msg)
Data_out@vbK <- UpdateSlot('vbK', RealData, SimData, msg)
ObsPars[[p]][[f]]$Kbias <- UpdateObs('vbK', ObsPars[[p]][[f]]$Kbias, StockPars[[p]]$Karray[, nyears],
RealData, SimData, msg)
Data_out@vbt0 <- UpdateSlot('vbt0', RealData, SimData, msg)
ObsPars[[p]][[f]]$t0bias <- UpdateObs('vbt0', ObsPars[[p]][[f]]$t0bias, StockPars[[p]]$t0array[, nyears],
RealData, SimData, msg)
Data_out@CV_Mort <- UpdateSlot('CV_Mort', RealData, SimData, msg)
Data_out@CV_vbLinf <- UpdateSlot('CV_vbLinf', RealData, SimData, msg)
Data_out@CV_vbK <- UpdateSlot('CV_vbK', RealData, SimData, msg)
Data_out@CV_vbt0 <- UpdateSlot('CV_vbt0', RealData, SimData, msg)
Data_out@wla <- UpdateSlot('wla', RealData, SimData, msg)
Data_out@wlb <- UpdateSlot('wlb', RealData, SimData, msg)
Data_out@CV_wla <- UpdateSlot('CV_wla', RealData, SimData, msg)
Data_out@CV_wlb <- UpdateSlot('CV_wlb', RealData, SimData, msg)
Data_out@steep <- UpdateSlot('steep', RealData, SimData, msg)
Data_out@CV_steep <- UpdateSlot('CV_steep', RealData, SimData, msg)
ObsPars[[p]][[f]]$hbias <- UpdateObs('steep', ObsPars[[p]][[f]]$hbias, StockPars[[p]]$hs,
RealData, SimData, msg)
Data_out@sigmaR <- UpdateSlot('sigmaR', RealData, SimData, msg)
Data_out@CV_sigmaR <- UpdateSlot('CV_sigmaR', RealData, SimData, msg)
# TODO ObsPars$sigmaRbias
Data_out@L50 <- UpdateSlot('L50', RealData, SimData, msg)
Data_out@CV_L50 <- UpdateSlot('CV_L50', RealData, SimData, msg)
ObsPars[[p]][[f]]$lenMbias <- UpdateObs('L50', ObsPars[[p]][[f]]$lenMbias, StockPars[[p]]$L50,
RealData, SimData, msg)
Data_out@L95 <- UpdateSlot('L95', RealData, SimData, msg)
Data_out@LenCV <- UpdateSlot('LenCV', RealData, SimData, msg)
Data_out@LFC <- UpdateSlot('LFC', RealData, SimData, msg)
Data_out@CV_LFC <- UpdateSlot('CV_LFC', RealData, SimData, msg)
ObsPars[[p]][[f]]$LFCbias <- UpdateObs('LFC', ObsPars[[p]][[f]]$LFCbias, FleetPars[[p]][[f]]$L5_y[,nyears],
RealData, SimData, msg)
Data_out@LFS <- UpdateSlot('LFS', RealData, SimData, msg)
Data_out@CV_LFS <- UpdateSlot('CV_LFS', RealData, SimData, msg)
ObsPars[[p]][[f]]$LFSbias <- UpdateObs('LFS', ObsPars[[p]][[f]]$LFSbias, FleetPars[[p]][[f]]$LFS_y[,nyears],
RealData, SimData, msg)
if (length(RealData@Year)>1) {
# check years
# if (!all(RealData@Year %in%SimData@Year)) {
# stop('`OM$cpars$Data@Year` does not match `SimData@Year`. \nAre `Fleet@nyears` and `Fleet@CurrentYr` correct?')
# }
Data_out@Year <- RealData@Year
}
# ---- Update Catch ----
if (!all(is.na(RealData@Cat[1,]))) {
if (msg)
message('Updating Simulated Catch from `OM@cpars$Data@Cat`')
Data_out@Cat <- matrix(RealData@Cat[1,1:nyears], nrow=nsim, ncol=nyears, byrow=TRUE)
Data_out@CV_Cat <- matrix(RealData@CV_Cat[1,1:nyears], nrow=nsim, ncol=nyears, byrow=TRUE)
simcatch <- apply(CBret[,map.stocks,f,,, ,drop=FALSE], c(1,5), sum)
simcatch[simcatch==0] <- tiny
Cbias <- matrix(apply(Data_out@Cat, 1, mean)/apply(simcatch, 1, mean),
nrow=nsim, ncol=nyears+proyears)
Cerr <- Data_out@Cat/(simcatch*Cbias[,1:nyears])
t1<- Cerr[,max(nyears-10, 1):nyears]/apply(Cerr[,max(nyears-10, 1):nyears],1,mean)
SDs <- apply(log(t1), 1, sd)
Cerr_proj <- matrix(NA, nsim, proyears)
nas <- which(is.na(SDs))
Cerr_proj[nas,] <- tiny # no catch
if (!all(is.na(SDs))) {
for (i in (1:nsim)[!is.na(SDs)]) {
Cerr_proj[i,] <- exp(rnorm(proyears, -((SDs[i]^2)/2), SDs[i]))
}
}
Cerr <- cbind(Cerr, Cerr_proj)
if(!is.null(SampCpars[[p]][[f]]$Cobs_y)) {
if (msg) message_info('Catch observation error detected in cpars (`cpars$Cobs_y`). Not updating catch obs error')
} else {
if(!is.null(SampCpars[[p]][[f]]$Cbias)) {
if (msg) message_info('Catch Cbias detected in cpars (`cpars$Cbias`). Not updating catch bias')
} else {
if (msg) message_info('Updating Catch bias from `OM@cpars$Data@Cat`')
ObsPars[[p]][[f]]$Cbias <- Cbias[,1]
}
if(!is.null(SampCpars[[p]][[f]]$Cerr_y)) {
if (msg) message_info('Catch variability detected in cpars (`cpars$Cerr_y`). Not updating catch error')
} else {
if (msg) message_info('Updating Catch variability from `OM@cpars$Data@Cat`')
ObsPars[[p]][[f]]$Cerr_y <- Cerr
}
if (msg) message_info('Updating catch observation error from `OM@cpars$Data@Cat`')
ObsPars[[p]][[f]]$Cobs_y <- Cerr * Cbias
}
}
# ---- Update Effort -----
# TODO
# ---- Update Index (total biomass) ----
if (!all(is.na(RealData@Ind[1,]))) { # Index exists
fit_ind <- Fit_Index_MS(ind_slot='Ind', indcv_slot="CV_Ind", Data_out,
RealData, StockPars, ObsPars, SampCpars, nsim, nyears, proyears,
map.stocks,
p, f,
msg=msg)
Data_out <- fit_ind$Data_out
# add future year data if they exist
n_future_years <- length(RealData@Ind[1,]) - length(Data_out@Ind[1,])
if (n_future_years>0) {
ll <- length(RealData@Ind[1,])
future_index <- matrix(RealData@Ind[1,(ll-n_future_years+1):ll], nsim, n_future_years, byrow=TRUE)
future_index_cv <- matrix(RealData@CV_Ind[1,(ll-n_future_years+1):ll], nsim, n_future_years, byrow=TRUE)
Data_out@Ind <- cbind(Data_out@Ind, future_index)
Data_out@CV_Ind <- cbind(Data_out@CV_Ind, future_index_cv)
}
ObsPars <- fit_ind$ObsPars
}
# ---- Update Index (spawning biomass) ----
if (!all(is.na(RealData@SpInd[1,]))) { # Index exists
fit_ind <- Fit_Index_MS(ind_slot='SpInd', indcv_slot="CV_SpInd", Data_out,
RealData, StockPars, ObsPars, SampCpars, nsim, nyears, proyears,
map.stocks,
p, f,
msg=msg)
Data_out <- fit_ind$Data_out
ObsPars <- fit_ind$ObsPars
}
# ---- Index (vulnerable biomass) ----
if (!all(is.na(RealData@VInd[1,]))) { # Index exists
StockPars[[p]]$VBiomass <- VBF[,,f,,,]
fit_ind <- Fit_Index_MS(ind_slot='VInd', indcv_slot="CV_VInd", Data_out,
RealData, StockPars, ObsPars, SampCpars, nsim, nyears, proyears,
map.stocks,
p, f,
msg=msg)
Data_out <- fit_ind$Data_out
ObsPars <- fit_ind$ObsPars
fit_ind$ObsPars[[1]][[1]]$VIerr_y[,1:69]
}
# ---- Add Additional Indices ----
if (!all(is.na(RealData@AddInd))) {
if (msg)
message('Adding Additional Indices to Simulated Data from `OM@cpars$Data@AddInd`')
n.ind <- dim(RealData@AddInd)[2]
Data_out@AddInd <- Data_out@CV_AddInd <- array(NA, dim=c(nsim, n.ind, nyears))
fitbeta <- fitIerr <- TRUE
if (!is.null(SampCpars[[p]][[f]]$AddIbeta)) {
if (any(dim(SampCpars[[p]][[f]]$AddIbeta) != c(nsim, n.ind)))
stop("cpars$AddIbeta must be dimensions c(nsim, n.ind)")
if (msg) message_info('cpars$AddIbeta detected. Not updating beta for additional indices')
ObsPars[[p]][[f]]$AddIbeta <- SampCpars[[p]][[f]]$AddIbeta
fitbeta <- FALSE
} else {
if (msg) message_info('Updating beta for additional indices from real data')
ObsPars[[p]][[f]]$AddIbeta <- matrix(NA, nsim, n.ind)
}
if (!is.null(SampCpars[[p]][[f]]$AddIerr)) {
if (any(dim(SampCpars[[p]][[f]]$AddIerr) != c(nsim, n.ind, nyears+proyears)))
stop("cpars$AddIerr must be dimensions c(nsim, n.ind, nyears+proyears)")
if (msg) message_info('cpars$AddIerr detected. Not updating observation variability for additional indices')
ObsPars[[p]][[f]]$AddIerr <- SampCpars[[p]][[f]]$AddIerr
fitIerr <- FALSE
} else {
if (msg) message_info('Updating observation variability (AddIerr) for additional indices from real data')
ObsPars[[p]][[f]]$AddIerr <- array(NA, dim=c(nsim, n.ind, nyears+proyears))
}
if (!is.null(SampCpars$AddIunits) && length(SampCpars$AddIunits) != n.ind)
stop("cpars$AddIunits must be length n.ind")
AddIunits <- RealData@AddIunits
if (all(is.na(AddIunits))) AddIunits <- rep(1, n.ind)
if (!is.null(SampCpars[[p]][[f]]$AddIunits)) AddIunits <- SampCpars[[p]][[f]]$AddIunits
Data_out@AddIunits <- AddIunits
AddIndType <- RealData@AddIndType
if (all(is.na(AddIndType))) AddIndType <- rep(1, n.ind)
Data_out@AddIndType <- AddIndType
ObsPars[[p]][[f]]$AddInd_Stat <- vector(mode = "list", length = n.ind)
UnitsTab <- data.frame(n=1:0, units=c('biomass', 'numbers'))
TypeTab <- data.frame(n=1:3, type=c('total', 'spawning', 'vuln.'))
Data_out@AddIndV <- array(NA, dim=c(nsim, n.ind, Data_out@MaxAge+1))
# loop over additional indices
for (i in 1:n.ind) {
units <- UnitsTab$units[match(AddIunits[i], UnitsTab$n)]
type <- TypeTab$type[match(AddIndType[i], TypeTab$n)]
if(msg) message_info("Additional index", i, '-', type, 'stock', paste0('(', units, ')'))
nyrs <- min(length(RealData@AddInd[1,i,]), nyears)
ind <- RealData@AddInd[1,i,1:nyrs]
cv_ind <- RealData@CV_AddInd[1,i,1:nyrs]
if (nyrs < nyears) {
ind <- c(ind, rep(NA,nyears-nyrs))
cv_ind <- c(cv_ind, rep(NA,nyears-nyrs))
}
Data_out@AddInd[,i,] <- matrix(ind, nrow=nsim, ncol=nyears, byrow=TRUE)
Data_out@CV_AddInd[,i,] <- matrix(cv_ind, nrow=nsim, ncol=nyears, byrow=TRUE)
if (all(is.na(RealData@AddIndV[1,, ]))) {
# no vulnerability-at-age included
Ind_V <- rep(1, Data_out@MaxAge+1)
} else {
Ind_V <- RealData@AddIndV[1,i, ]
}
Data_out@AddIndV[,i,] <- t(replicate(nsim,Ind_V))
if (!all(is.na(ind))) {
# check dimensions
if (!length(Ind_V) == Data_out@MaxAge+1)
stop('Vulnerability-at-age for additional index ', i, ' is not length `maxage`+1' )
Ind_V <- matrix(Ind_V, nrow=SimData@MaxAge+1, ncol= nyears)
Ind_V <- replicate(nsim, Ind_V) %>% aperm(., c(3,1,2))
Ind_V_list <- list()
ObsPars[[p]][[f]]$AddIV <- SampCpars[[p]][[f]]$AddIV
for (pp in map.stocks) {
if (!is.null(SampCpars[[pp]][[f]]$AddIV)) {
Ind_V_list[[pp]] <- SampCpars[[pp]][[f]]$AddIV[,,i,1:nyears]
} else {
Ind_V_list[[pp]] <- Ind_V
}
}
# calculate simulated index
if (AddIunits[i]) {
if (AddIndType[i]==1) SimIndex <- apply(StockPars[[p]]$Biomass[,map.stocks,,,,drop=FALSE], c(1, 2, 3, 4), sum) # Total Biomass-based index
} else {
if (AddIndType[i]==1) SimIndex <- apply(StockPars[[p]]$N[,map.stocks,,,,drop=FALSE], c(1, 2, 3, 4), sum) # Total Abundance-based index
}
if (AddIndType[i]!=1)
stop('Vulnerable and spawning indices not supported for multiMSE. Use `cpars$AddIV` to specify selectivity pattern')
for (pp in seq_along(map.stocks)) {
SimIndex[,pp,,] <- SimIndex[,pp,,] * Ind_V_list[[map.stocks[pp]]]
}
SimIndex <- apply(SimIndex, c(1,4), sum)
# Fit to observed index and generate residuals for projections
if (fitbeta) {
beta <- rep(NA, nsim)
} else {
beta <- ObsPars[[p]][[f]]$AddIbeta[,i]
}
# Calculate residuals (with or without estimated beta)
Res_List <- lapply(1:nsim, function(x) Calc_Residuals(sim.index=SimIndex[x,],
obs.ind=ind,
beta=beta[x]))
lResids_Hist <- do.call('rbind', lapply(Res_List, '[[', 1))
if (fitbeta)
ObsPars[[p]][[f]]$AddIbeta[,i] <- as.vector(do.call('cbind', lapply(Res_List, '[[', 2)))
if (msg & fitbeta)
message_info(paste0('Estimated beta for Additional Index ', i, ':'),
paste0(range(round(ObsPars[[p]][[f]]$AddIbeta[,i],2)), collapse = "-"),
"Use `cpars$AddIbeta` to override")
# Calculate statistics
Stats_List <- lapply(1:nsim, function(x) Calc_Stats(lResids_Hist[x,]))
Stats <- do.call('rbind', Stats_List)
check_Index <- check_Index_Fit(Stats, i)
if (check_Index) {
# Generate residuals for projections
Resid_Hist <- exp(lResids_Hist) # historical residuals in normal space
Resid_Proj <- Gen_Residuals(Stats, nsim, proyears)
if (fitIerr) ObsPars[[p]][[f]]$AddIerr[,i, ] <- cbind(Resid_Hist, Resid_Proj)
ObsPars[[p]][[f]]$AddInd_Stat[[i]] <- Stats[,1:2] # index fit statistics
} else {
ObsPars[[p]][[f]]$AddIerr[,i, ] <- ObsPars[[p]][[f]]$Ierr_y
ObsPars[[p]][[f]]$AddInd_Stat[[i]] <- Stats[,1:2] # index fit statistics
}
}
}
}
# ---- Update Recruitment ----
if (!all(is.na(RealData@Rec))) {
if (msg)
message('Updating Simulated Recruitment Data from `OM@cpars$Data@Rec`')
Data_out@Rec <- matrix(RealData@Rec[1,1:nyears], nrow=nsim, ncol=nyears, byrow=TRUE)
dd <- dim(RealData@CV_Rec)
if (dd[2]<nyears) {
RealData@CV_Rec <- matrix(RealData@CV_Rec[1,1], nrow=nsim, ncol=nyears, byrow=TRUE)
}
Data_out@CV_Rec <- matrix(RealData@CV_Rec[1,1:nyears], nrow=nsim, ncol=nyears, byrow=TRUE)
# Calculate Error
Rec <- apply(StockPars[[p]]$N[, map.stocks, 1, , ], c(1, 3), sum) # simulated recruitment
Recbias <- matrix(apply(Data_out@Rec, 1, mean)/apply(Rec, 1, mean),
nrow=nsim, ncol=nyears+proyears)
Rec_err <- Data_out@Rec/(Rec*Recbias[,1:nyears])
t1 <- Rec_err[,max(nyears-10, 1):nyears]/apply(Rec_err[,max(nyears-10, 1):nyears],1,mean) # last 10 years used for projections
SDs <- apply(log(t1), 1, sd)
Rec_err_proj <- matrix(NA, nsim, proyears)
for (i in 1:nsim) {
Rec_err_proj[i,] <- exp(rnorm(proyears, -((SDs[i]^2)/2), SDs[i]))
}
Rec_err_proj <- Rec_err_proj * Recbias[,(nyears+1):(nyears+proyears)]
Rec_err <- cbind(Rec_err, Rec_err_proj)
ObsPars[[p]][[f]]$Recerr_y <- Rec_err * Recbias
ObsPars[[p]][[f]]$Recsd <- SDs
ObsPars[[p]][[f]]$Recbias <- Recbias
}
# ---- Update CAA ----
if (!all(is.na(RealData@CAA)) & !all(RealData@CAA ==0)) {
if (msg)
message('Updating Simulated Catch-at-Age Data from `OM@cpars$Data@CAA`. Note: CAA_ESS is currently NOT updated')
Data_out@CAA <- aperm(replicate(nsim, RealData@CAA[1,1:nyears,]),c(3,1,2))
Data_out@Vuln_CAA <- RealData@Vuln_CAA
# Get average sample size
nsamp <- ceiling(mean(apply(RealData@CAA[1,1:nyears,], 1, sum)))
ObsPars[[p]][[f]]$CAA_nsamp <- rep(nsamp, nsim)
}
# ---- Update CAL ----
if (!all(is.na(RealData@CAL)) & !all(RealData@CAL ==0)) {
# check length bins
if (!all(RealData@CAL_bins %in% StockPars[[p]]$CAL_bins)) {
warning('cpars$Data@CAL_bins cannot be matched with Simulated Data@CAL_bins. Add cpars$Data@CAL_bins to cpars$CAL_bins. cpars$Data@CAL are NOT being used')
} else {
if (msg)
message('Updating Simulated Catch-at-Length Data, Obs@CAL_nsamp, and Obs@CAL_ESS from `OM@cpars$Data@CAL`')
# match length bins
ind <- match(RealData@CAL_mids, StockPars[[p]]$CAL_binsmid)
dd <- dim(Data_out@CAL)
dd_real <- dim(RealData@CAL[1,,])
dd[2] <- dd_real[1]
Data_out@CAL <- array(0, dim=dd)
CAL <- Data_out@CAL[1,,]
CAL[,ind] <- RealData@CAL[1,,] # replace with real CAL
n_dat_years <- dim(RealData@CAL[1,,])[1]
CAL <- aperm(replicate(nsim, CAL[1:n_dat_years,]),c(3,1,2))
Data_out@CAL <- CAL
Data_out@Vuln_CAL <- RealData@Vuln_CAL
# Get average sample size
nsamp <- ceiling(mean(apply(RealData@CAL[1,1:n_dat_years,], 1, sum, na.rm=TRUE), na.rm=TRUE))
ObsPars[[p]][[f]]$CAL_nsamp <- rep(nsamp, nsim)
# calculate effective sample size for projections
# vuln N-at-age in last year with length data
nas <- !apply(CAL[1,,], 2, is.na)
yr.ind <- max(which(apply(nas, 1, prod)>0))
if (!is.null(control$CAL) && control$CAL == 'removals') {
vn <- apply(StockPars[[p]]$N[,,yr.ind,, drop=FALSE], c(1,2,3), sum) *
FleetPars[[p]][[f]]$V_real[,,yr.ind, drop=FALSE]
# numbers at age in population that would be removed
vn <- aperm(vn, c(1,3, 2))
doopt <- optimise(optESS, c(10, 10000), vn, StockPars,
FleetPars$SLarray_real[1,,yr.ind, drop=FALSE], yr.ind, ObsPars, Data_out, CAL)
ObsPars$CAL_ESS <- rep(doopt$minimum, nsim)
} else {
vn <- apply(StockPars[[p]]$N[,map.stocks,,yr.ind,, drop=FALSE], c(1,3,4), sum) *
FleetPars[[p]][[f]]$retA_real[,,yr.ind, drop=FALSE]
# numbers at age in population that would be removed
vn <- aperm(vn, c(1,3, 2))
doopt <- optimise(optESS, c(10, 10000), vn, StockPars[[p]],
FleetPars[[p]][[f]]$retL_real[1,,yr.ind, drop=FALSE], yr.ind, ObsPars[[p]][[f]], Data_out, CAL)
ObsPars[[p]][[f]]$CAL_ESS <- rep(doopt$minimum, nsim)
}
}
}
# ---- Depletion ----
Data_out@Dep <- UpdateSlot('Dep', RealData, SimData, msg)
Data_out@CV_Dep <- UpdateSlot('CV_Dep', RealData, SimData, msg)
ObsPars[[p]][[f]]$Dbias <- UpdateObs('Dep', ObsPars$Dbias, StockPars$Depletion,
RealData, SimData, msg)
# ---- Index Reference -----
Data_out@Iref <- UpdateSlot('Iref', RealData, SimData, msg)
Data_out@CV_Iref <- UpdateSlot('CV_Iref', RealData, SimData, msg)
# ObsPars$Irefbias <- rep(NA, nsim) # not calculated
# ---- Misc ----
if (length(RealData@Misc) && !is.null(names(RealData@Misc))) {
Data_out@Misc[names(RealData@Misc)] <- RealData@Misc[names(RealData@Misc)]
if (msg)
message('Adding named list from `OM@cpars$Data@Misc`')
}
NotUpdated <- function(RealData, sl, msg) {
if (!all(is.na(slot(RealData, sl)))) {
if (msg)
message_info(paste0('Data detected in `OM@cpars$Data@', sl, '` but is NOT being used.'))
}
}
NotUpdated(RealData, 'ML', msg)
NotUpdated(RealData, 'Lc', msg)
NotUpdated(RealData, 'Lbar', msg)
NotUpdated(RealData, 'Abun', msg)
NotUpdated(RealData, 'SpAbun', msg)
NotUpdated(RealData, 'FMSY_M', msg)
NotUpdated(RealData, 'BMSY_B0', msg)
NotUpdated(RealData, 'Cref', msg)
NotUpdated(RealData, 'Bref', msg)
NotUpdated(RealData, 'AvC', msg)
NotUpdated(RealData, 'Dt', msg)
NotUpdated(RealData, 'Ref', msg)
list(Data=Data_out, ObsPars=ObsPars)
}
updateData_MS <- function(Data, OM, MPCalcs, Effort, Biomass, N, Biomass_P, CB_Pret,
N_P, SSB, SSB_P, VBiomass, VBiomass_P,
StockPars, FleetPars, ObsPars, ImpPars,
upyrs, interval, y=2, mm=1, Misc, RealData,
p, f, map.stocks) {
yind <- upyrs[match(y, upyrs) - 1]:(upyrs[match(y, upyrs)] - 1) # index
nyears <- OM@nyears
proyears <- OM@proyears
nsim <- OM@nsim
nareas <- StockPars[[p]]$nareas
reps <- OM@reps
Data@Year <- (OM@CurrentYr - nyears+1):(OM@CurrentYr+ y - 1)
Data@t <- rep(nyears + y, nsim)
# --- Simulate catches ----
CBtemp <- CB_Pret[, map.stocks,f, , yind, , drop=FALSE]
CBtemp[is.na(CBtemp)] <- tiny
CBtemp[!is.finite(CBtemp)] <- tiny
yr.index <- max(which(!is.na(Data@CV_Cat[1,])))
newCV_Cat <- matrix(Data@CV_Cat[,yr.index], nrow=nsim, ncol=length(yind))
Data@CV_Cat <- cbind(Data@CV_Cat, newCV_Cat)
# --- Observed catch ----
# Simulated observed retained catch (biomass)
Cobs <- ObsPars[[p]][[f]]$Cobs_y[,nyears + yind] * apply(CBtemp, c(1,5), sum, na.rm = TRUE)
Data@Cat <- cbind(Data@Cat, Cobs)
if (!is.null(RealData) && ncol(RealData@Cat)>nyears &&
!all(is.na(RealData@Cat[1,(nyears+1):length(RealData@Cat[1,])]))) {
# update projection catches with observed catches
addYr <- min(y,ncol(RealData@Cat) - nyears)
Data@Cat[,(nyears+1):(nyears+addYr)] <- matrix(RealData@Cat[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
Data@CV_Cat[,(nyears+1):(nyears+addYr)] <- matrix(RealData@CV_Cat[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
}
# --- Index of total abundance ----
yr.ind <- max(which(!is.na(ObsPars[[p]][[f]]$Ierr_y[1,1:nyears])))
I2 <- cbind(apply(Biomass[,map.stocks,,,,drop=FALSE], c(1, 4), sum)[,yr.ind:nyears],
apply(Biomass_P[,map.stocks,,,,drop=FALSE], c(1, 4), sum)[, 1:(y - 1)])
# standardize, apply beta & obs error
I2 <- exp(lcs(I2))^ObsPars[[p]][[f]]$I_beta * ObsPars[[p]][[f]]$Ierr_y[,yr.ind:(nyears + (y - 1))]
# I2 <- exp(lcs(I2)) * ObsPars$Ierr_y[,yr.ind:(nyears + (y - 1))]
if (sum(Data@Ind[1,1:nyears], na.rm = TRUE)) {
year.ind <- max(which(!is.na(Data@Ind[1,1:nyears])))
scaler <- Data@Ind[,year.ind]/I2[,1]
} else {
scaler <- rep(1, nsim)
}
scaler <- matrix(scaler, nrow=nsim, ncol=ncol(I2))
I2 <- I2 * scaler # convert back to historical index scale
I2 <- cbind(Data@Ind[,1:(yr.ind)], I2[,2:ncol(I2)])
Data@Ind <- I2
yr.index <- max(which(!is.na(Data@CV_Ind[1,1:nyears])))
newCV_Ind <- matrix(Data@CV_Ind[,yr.index], nrow=nsim, ncol=length(yind))
Data@CV_Ind <- cbind(Data@CV_Ind, newCV_Ind)
if (!is.null(RealData) && ncol(RealData@Ind)>nyears &&
!all(is.na(RealData@Ind[1,(nyears+1):length(RealData@Ind[1,])]))) {
# update projection index with observed index if it exists
dd <- length(RealData@Ind[1,])
p_yr_ind <- (nyears+1):(nyears+y-1)
p_yr_ind <- p_yr_ind[p_yr_ind<=dd]
p_ind <- matrix(RealData@Ind[1,p_yr_ind],
nrow=nsim, ncol=length(p_yr_ind), byrow=TRUE)
p_cv <- matrix(RealData@CV_Ind[1,p_yr_ind],
nrow=nsim, ncol=length(p_yr_ind), byrow=TRUE)
Data@Ind[,p_yr_ind] <- p_ind
Data@CV_Ind[,p_yr_ind] <- p_cv
}
# --- Index of spawning abundance ----
yr.ind <- max(which(!is.na(ObsPars[[p]][[f]]$SpIerr_y[1,1:nyears])))
I2 <- cbind(apply(SSB[,map.stocks,,,,drop=FALSE], c(1, 4), sum)[,yr.ind:nyears],
apply(SSB_P[,map.stocks,,,,drop=FALSE], c(1, 4), sum)[, 1:(y - 1)])
# standardize, apply beta & obs error
I2 <- exp(lcs(I2))^ObsPars[[p]][[f]]$SpI_beta * ObsPars[[p]][[f]]$SpIerr_y[,yr.ind:(nyears + (y - 1))]
if (sum(Data@SpInd[1,1:nyears], na.rm = TRUE)) {
year.ind <- max(which(!is.na(Data@SpInd[1,1:nyears])))
scaler <- Data@SpInd[,year.ind]/I2[,1]
} else {
scaler <- rep(1, nsim)
}
scaler <- matrix(scaler, nrow=nsim, ncol=ncol(I2))
I2 <- I2 * scaler # convert back to historical index scale
I2 <- cbind(Data@SpInd[,1:(yr.ind)], I2[,2:ncol(I2)])
Data@SpInd <- I2
yr.index <- max(which(!is.na(Data@CV_SpInd[1,1:nyears])))
newCV_Ind <- matrix(Data@CV_SpInd[,yr.index], nrow=nsim, ncol=length(yind))
Data@CV_SpInd <- cbind(Data@CV_SpInd, newCV_Ind)
if (!is.null(RealData) && ncol(RealData@SpInd)>nyears &&
!all(is.na(RealData@SpInd[1,(nyears+1):length(RealData@SpInd[1,])]))) {
# update projection index with observed index if it exists
addYr <- min(y,ncol(RealData@SpInd) - nyears)
Data@SpInd[,(nyears+1):(nyears+addYr)] <- matrix(RealData@SpInd[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
Data@CV_SpInd[,(nyears+1):(nyears+addYr)] <- matrix(RealData@CV_SpInd[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
}
# --- Index of vulnerable abundance ----
I2 <- cbind(apply(VBiomass[,map.stocks,f,,,,drop=FALSE], c(1, 5), sum)[,yr.ind:nyears],
apply(VBiomass_P[,map.stocks,f,,,,drop=FALSE], c(1, 5), sum)[, 1:(y - 1)])
# standardize, apply beta & obs error
I2 <- exp(lcs(I2))^ObsPars[[p]][[f]]$VI_beta * ObsPars[[p]][[f]]$VIerr_y[,yr.ind:(nyears + (y - 1))]
if (sum(Data@VInd[1,1:nyears], na.rm = TRUE)) {
year.ind <- max(which(!is.na(Data@VInd[1,1:nyears])))
scaler <- Data@VInd[,year.ind]/I2[,1]
} else {
scaler <- rep(1, nsim)
}
scaler <- matrix(scaler, nrow=nsim, ncol=ncol(I2))
I2 <- I2 * scaler # convert back to historical index scale
I2 <- cbind(Data@VInd[,1:(yr.ind)], I2[,2:ncol(I2)])
Data@VInd <- I2
yr.index <- max(which(!is.na(Data@CV_VInd[1,1:nyears])))
newCV_Ind <- matrix(Data@CV_VInd[,yr.index], nrow=nsim, ncol=length(yind))
Data@CV_VInd <- cbind(Data@CV_VInd, newCV_Ind)
if (!is.null(RealData) && ncol(RealData@VInd)>nyears &&
!all(is.na(RealData@VInd[1,(nyears+1):length(RealData@VInd[1,])]))) {
# update projection index with observed index if it exists
addYr <- min(y,ncol(RealData@VInd) - nyears)
Data@VInd[,(nyears+1):(nyears+addYr)] <- matrix(RealData@VInd[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
Data@CV_VInd[,(nyears+1):(nyears+addYr)] <- matrix(RealData@CV_VInd[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
}
# --- Update additional indices (if they exist) ----
AddIunits <- Data@AddIunits
AddIndType <- Data@AddIndType
if (length(ObsPars[[p]][[f]]$AddIerr)>0) {
n.ind <- dim(ObsPars[[p]][[f]]$AddIerr)[2]
AddInd <- array(NA, dim=c(nsim, n.ind, nyears+y-1))
CV_AddInd <- array(NA, dim=c(nsim, n.ind, nyears+y-1))
for (i in 1:n.ind) {
if (all(is.na(RealData@AddIndV[1, , ]))) {
Ind_V <- rep(1, Data@MaxAge+1)
} else {
Ind_V <- RealData@AddIndV[1,i, ]
}
Ind_V <- matrix(Ind_V, nrow=Data@MaxAge+1, ncol= nyears+proyears)
Ind_V <- replicate(nsim, Ind_V) %>% aperm(c(3,1,2))
Ind_V_list <- list()
for (pp in map.stocks) {
if (!is.null(ObsPars[[pp]][[f]]$AddIV)) {
Ind_V_list[[pp]] <- ObsPars[[pp]][[f]]$AddIV[,,i,]
} else {
Ind_V_list[[pp]] <- Ind_V
}
}
nas <- which(!is.na(ObsPars[[p]][[f]]$AddIerr[1,i, 1:nyears]))
if (length(nas)>0) {
yr.ind <- max(nas)
if (AddIunits[i]) { # Biomass-based index
if (AddIndType[i]==1) {
# total biomass
b1 <- apply(Biomass[,,,yr.ind:nyears, ,drop=FALSE], c(1:4), sum)
b2 <- apply(Biomass_P, c(1:4), sum)
}
#
# if (AddIndType[i]==2) {
# # spawning biomass
# b1 <- apply(SSB[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum)
# b2 <- apply(SSB_P, c(1, 2, 3), sum)
# }
# if (AddIndType[i]==3) {
# # vulnerable biomass
# b1 <- apply(VBiomass[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum)
# b2 <- apply(VBiomass_P, c(1, 2, 3), sum)
# }
} else {
if (AddIndType[i]==1) {
# total stock
b1 <- apply(N[,,,yr.ind:nyears,,drop=FALSE], 1:4, sum) # Abundance-based index
b2 <- apply(N_P, 1:4, sum)
}
# if (AddIndType[i]==2) {
# # spawning stock
# b1 <- apply(N[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum) * StockPars$Mat_age[,,yr.ind:nyears, drop=FALSE]
# b2 <- apply(N_P, c(1, 2, 3), sum) * StockPars$Mat_age[,,(nyears+1):(nyears+proyears), drop=FALSE]
# }
# if (AddIndType[i]==3) {
# # vuln stock
# b1 <- apply(N[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum) * V_P[,,yr.ind:nyears, drop=FALSE]
# b2 <- apply(N_P, c(1, 2, 3), sum) * V_P[,,(nyears+1):(nyears+proyears), drop=FALSE]
# }
}
for (pp in map.stocks) {
histV <- array(NA, dim=dim(b1[,pp,,, drop=FALSE]))
histV[,1,,] <- Ind_V_list[[pp]][,,yr.ind:nyears, drop=FALSE]
b1[,pp,,] <- b1[,pp,,, drop=FALSE] * histV
b2[,pp,,] <- b2[,pp,,] * Ind_V_list[[pp]][,,(nyears+1):(nyears+proyears), drop=FALSE]
}
b1 <- apply(b1, c(1,4), sum)
b2 <- apply(b2, c(1,4), sum)
tempI <- cbind(b1, b2[, 1:(y - 1)])
# standardize, apply beta & obs error
tempI <- exp(lcs(tempI))^ObsPars[[p]][[f]]$AddIbeta[,i] * ObsPars[[p]][[f]]$AddIerr[,i,yr.ind:(nyears + (y - 1))]
year.ind <- max(which(!is.na(RealData@AddInd[1,i,])))
scaler <- RealData@AddInd[1,i,year.ind]/tempI[,1]
scaler <- matrix(scaler, nrow=nsim, ncol=ncol(tempI))
tempI <- tempI * scaler # convert back to historical index scale
AddInd[,i,] <- cbind(Data@AddInd[1:nsim,i,1:yr.ind], tempI[,2:ncol(tempI)])
yr.index <- max(which(!is.na(Data@CV_AddInd[1,i,1:nyears])))
newCV_Ind <- matrix(Data@CV_AddInd[,i,yr.index], nrow=nsim, ncol=length(yind))
CV_AddInd[,i,] <- cbind(Data@CV_AddInd[,i,], newCV_Ind)
if (!is.null(RealData) && length(RealData@AddInd[1,i,])>nyears &&
!all(is.na(RealData@AddInd[1,i,(nyears+1):length(RealData@AddInd[1,i,])]))) {
# update projection index with observed index if it exists
addYr <- min(y-1,length(RealData@AddInd[1,i,]) - nyears)
AddInd[,i,(nyears+1):(nyears+addYr)] <- matrix(RealData@AddInd[1,i,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
CV_AddInd[,i,(nyears+1):(nyears+addYr)] <- matrix(RealData@CV_AddInd[1,i,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
}
}
}
Data@AddInd <- AddInd
Data@CV_AddInd <- CV_AddInd
}
# --- Index of recruitment ----
Recobs <- ObsPars[[p]][[f]]$Recerr_y[, nyears + yind] * apply(N_P[, p, 1, yind, , drop=FALSE], c(1, 4), sum)
Data@Rec <- cbind(Data@Rec, Recobs)
# --- Average catch ----
Data@AvC <- apply(Data@Cat, 1, mean)
# --- Depletion ----
Depletion <- apply(SSB_P[,p, , y-1, ], 1, sum)/StockPars[[p]]$ReferencePoints$ReferencePoints$SSB0
Depletion[Depletion < tiny] <- tiny
Data@Dt <- Depletion * ObsPars[[p]][[f]]$Derr_y[,nyears+y]
Data@Dep <- Depletion * ObsPars[[p]][[f]]$Derr_y[,nyears+y]
# --- Update life-history parameter estimates for current year ----
Data@vbLinf <- StockPars[[p]]$Linfarray[,nyears+y] * ObsPars[[p]][[f]]$Linfbias # observed vB Linf
Data@vbK <- StockPars[[p]]$Karray[,nyears+y] * ObsPars[[p]][[f]]$Kbias # observed vB K
Data@vbt0 <- StockPars[[p]]$t0array[,nyears+y] + ObsPars[[p]][[f]]$t0bias # observed vB t0
Data@Mort <- StockPars[[p]]$Marray[,nyears+y] * ObsPars[[p]][[f]]$Mbias # natural mortality
Data@L50 <- StockPars[[p]]$L50array[,nyears+y] * ObsPars[[p]][[f]]$lenMbias # observed length at 50% maturity
Data@L95 <- StockPars[[p]]$L95array[,nyears+y] * ObsPars[[p]][[f]]$lenMbias # observed length at 95% maturity
Data@L95[Data@L95 > 0.9 * Data@vbLinf] <- 0.9 * Data@vbLinf[Data@L95 > 0.9 * Data@vbLinf] # Set a hard limit on ratio of L95 to Linf
Data@L50[Data@L50 > 0.9 * Data@L95] <- 0.9 * Data@L95[Data@L50 > 0.9 * Data@L95] # Set a hard limit on ratio of L95 to Linf
# --- Abundance ----
# Calculate vulnerable and spawning biomass abundance --
M_array <- array(0.5*StockPars[[p]]$M_ageArray[,,nyears+y], dim=c(nsim, StockPars[[p]]$maxage+1, nareas))
A <- apply(VBiomass_P[, p,f, , y, ] * exp(-M_array), 1, sum) # Abundance (mid-year before fishing)
Asp <- apply(SSB_P[, p,, y, ] * exp(-M_array), 1, sum) # Spawning abundance (mid-year before fishing)
Data@Abun <- A * ObsPars[[p]][[f]]$Aerr_y[,nyears+yind]
Data@SpAbun <- Asp * ObsPars[[p]][[f]]$Aerr_y[,nyears+yind]
# --- Catch-at-age ----
# previous CAA
oldCAA <- Data@CAA
Data@CAA <- array(0, dim = c(nsim, nyears + y - 1, StockPars[[p]]$maxage+1))
Data@CAA[, 1:(nyears + y - interval[mm] - 1), ] <- oldCAA[, 1:(nyears + y - interval[mm] - 1), ]
# update CAA
CNtemp <- FleetPars[[p]][[f]]$retA_P[,,yind+nyears, drop=FALSE] *
apply(N_P[,p,,yind,, drop=FALSE], c(1,3,4), sum)
CNtemp[is.na(CNtemp)] <- tiny
CNtemp[!is.finite(CNtemp)] <- tiny
CAA <- simCAA(nsim, yrs=length(yind), StockPars[[p]]$maxage+1, Cret=CNtemp, ObsPars[[p]][[f]]$CAA_ESS, ObsPars[[p]][[f]]$CAA_nsamp)
Data@CAA[, nyears + yind, ] <- CAA
# --- Catch-at-length ----
oldCAL <- Data@CAL
Data@CAL <- array(0, dim = c(nsim, nyears + y - 1, StockPars[[p]]$nCALbins))
Data@CAL[, 1:(nyears + y - interval[mm] - 1), ] <- oldCAL[, 1:(nyears + y - interval[mm] - 1), ]
CAL <- array(NA, dim = c(nsim, interval[mm], StockPars[[p]]$nCALbins))
if (!all(is.na(Data@Vuln_CAL))) {
Vuln_CAL <- replicate(nsim,Data@Vuln_CAL[1,])
Vuln_CAL <- replicate(length(yind),Vuln_CAL)
Vuln_CAL <- aperm(Vuln_CAL, c(2,1,3))
VList_dat <- lapply(1:nsim, calcV,
Len_age=StockPars[[p]]$Len_age[,,nyears+yind, drop=FALSE],
LatASD=StockPars[[p]]$LatASD[,,nyears+yind, drop=FALSE],
SLarray=Vuln_CAL,
n_age=StockPars[[p]]$maxage+1,
nyears=length(yind), proyears = 0,
CAL_binsmid=StockPars[[p]]$CAL_binsmid)
V_data <- aperm(array(as.numeric(unlist(VList_dat, use.names=FALSE)), dim=c(StockPars[[p]]$maxage+1, length(yind), nsim)), c(3,1,2))
vn <- (apply(N_P[,p,,yind,, drop=FALSE], c(1,3,4), sum) * V_data)
vn <- aperm(vn, c(1,3,2))
CALdat <- simCAL(nsim, nyears=length(yind), StockPars[[p]]$maxage, ObsPars[[p]][[f]]$CAL_ESS,
ObsPars[[p]][[f]]$CAL_nsamp, StockPars[[p]]$nCALbins, StockPars[[p]]$CAL_binsmid,
StockPars[[p]]$CAL_bins,
vn=vn, retL=Vuln_CAL,
Linfarray=StockPars[[p]]$Linfarray[,nyears + yind, drop=FALSE],
Karray=StockPars[[p]]$Karray[,nyears + yind, drop=FALSE],
t0array=StockPars[[p]]$t0array[,nyears + yind,drop=FALSE],
LenCV=StockPars[[p]]$LenCV)
} else {
vn <- (apply(N_P[,p,,,, drop=FALSE], c(1,3,4), sum) * FleetPars[[p]][[f]]$retA_P_real[,,(nyears+1):(nyears+proyears)]) # numbers at age that would be retained
vn <- aperm(vn, c(1,3,2))
vn <- vn[,yind, ,drop=FALSE]
CALdat <- simCAL(nsim, nyears=length(yind), StockPars[[p]]$maxage, ObsPars[[p]][[f]]$CAL_ESS,
ObsPars[[p]][[f]]$CAL_nsamp, StockPars[[p]]$nCALbins, StockPars[[p]]$CAL_binsmid,
StockPars[[p]]$CAL_bins,
vn=vn, retL=FleetPars[[p]][[f]]$retL_P[,,nyears+yind, drop=FALSE],
Linfarray=StockPars[[p]]$Linfarray[,nyears + yind, drop=FALSE],
Karray=StockPars[[p]]$Karray[,nyears + yind, drop=FALSE],
t0array=StockPars[[p]]$t0array[,nyears + yind,drop=FALSE],
LenCV=StockPars[[p]]$LenCV)
}
Data@CAL[, nyears + yind, ] <- CALdat$CAL # observed catch-at-length
Data@ML <- cbind(Data@ML, CALdat$ML) # mean length
Data@Lc <- cbind(Data@Lc, CALdat$Lc) # modal length
Data@Lbar <- cbind(Data@Lbar, CALdat$Lbar) # mean length above Lc
Data@LFC <- FleetPars[[p]][[f]]$L5_y[,nyears+y] * ObsPars[[p]][[f]]$LFCbias # length at first capture
Data@LFS <- FleetPars[[p]][[f]]$LFS_y[,nyears+y] * ObsPars[[p]][[f]]$LFSbias # length at full selection
Data@Vmaxlen <- FleetPars[[p]][[f]]$Vmaxlen_y[,nyears+y]
# --- Previous Management Recommendations ----
Data@MPrec <- MPCalcs$TACrec # last MP TAC recommendation
if (length(dim(Effort)) == 5) {
Data@MPeff <- Effort[, 1,1,mm, y-1] # last recommended effort
} else {
Data@MPeff <- Effort[, mm, y-1] # last recommended effort
}
# --- Store OM Parameters ----
# put all the operating model parameters in one table
ind <- which(lapply(StockPars[[p]], length) == nsim)
drop_srr <- which(names(ind)=='SRRpars')
ind <- ind[-drop_srr]
stock <- as.data.frame(StockPars[[p]][ind])
stock$Fdisc <- NULL
stock$CAL_bins <- NULL
stock$CAL_binsmid <- NULL
ind <- which(lapply(FleetPars[[p]][[f]], length) == nsim)
fleet <- as.data.frame(FleetPars[[p]][[f]][ind])
ind <- which(lapply(ImpPars[[p]][[f]], length) == nsim)
imp <- as.data.frame(ImpPars[[p]][[f]][ind])
RefPoints <- StockPars[[p]]$ReferencePoints$ReferencePoints
refs <- RefPoints# [!names(RefPoints) %in% names(stock)]
OFLreal <- A * (1-exp(-RefPoints$FMSY)) # the true simulated Over Fishing Limit
OMtable <- data.frame(stock, fleet, imp, refs,
ageM=StockPars[[p]]$ageMarray[,nyears+y],
L5=FleetPars[[p]][[f]]$L5_y[,nyears+y],
LFS=FleetPars[[p]][[f]]$LFS_y[,nyears+y],
Vmaxlen=FleetPars[[p]][[f]]$Vmaxlen_y[,nyears+y],
LR5=FleetPars[[p]][[f]]$LR5_y[,nyears],
LFR=FleetPars[[p]][[f]]$LFR_y[,nyears+y],
Rmaxlen=FleetPars[[p]][[f]]$Rmaxlen_y[,nyears+y],
DR=FleetPars[[p]][[f]]$DR_y[,nyears+y], OFLreal, maxF=StockPars[[p]]$maxF,
A=A, Asp=Asp, CurrentYr=OM@CurrentYr)
OMtable <- OMtable[,order(names(OMtable))]
Data@OM <- OMtable
Data@Misc <- Misc
if (!is.null(RealData)) {
if (length(RealData@Misc) && !is.null(names(RealData@Misc))) {
Data@Misc[names(RealData@Misc)] <- RealData@Misc[names(RealData@Misc)]
}
}
Data
}
Blue-Matter/MSEtool documentation built on Nov. 7, 2024, 11:38 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions updateData_MS AddRealData_MS optESS AddRealData check_Index_Fit UpdateSlot UpdateObs getfifth genSizeCompWrap simCAL simCAA updateData makeData
Documented in genSizeCompWrap getfifth simCAA simCAL
makeData <- function(Biomass, CBret, Cret, N, SSB, VBiomass, StockPars,
FleetPars, ObsPars, ImpPars, RefPoints,
SampCpars, initD, Sample_Area,
Name,
nyears,
proyears,
nsim,
nareas,
reps,
CurrentYr,
silent=FALSE,
control=list()) {
if(!silent) message("Simulating observed data")
Data <- new("Data") # create a blank DLM data object
if (reps == 1) Data <- OneRep(Data) # make stochastic variables certain for only one rep
Data <- replic8(Data, nsim) # make nsim sized slots in the DLM data object
Data@Name <- Name
Data@Year <- (CurrentYr - nyears+1):CurrentYr
# --- Observed catch ----
# Simulated observed retained catch (biomass)
Data@Cat <- ObsPars$Cobs_y[,1:nyears] * apply(CBret*Sample_Area$Catch[,,1:nyears,], c(1, 3), sum)
Data@CV_Cat <- matrix(Data@CV_Cat[,1], nrow=nsim, ncol=nyears)
# --- Index of total abundance ----
# Index of abundance from total biomass - beginning of year before fishing
# apply hyperstability / hyperdepletion
II <- (apply(Biomass*Sample_Area$BInd[,,1:nyears,], c(1, 3), sum)^ObsPars$I_beta) *
ObsPars$Ierr_y[, 1:nyears]
II <- II/apply(II, 1, mean) # normalize
Data@Ind <- II # index of total abundance
Data@CV_Ind <- matrix(Data@CV_Ind[,1], nrow=nsim, ncol=nyears)
# --- Index of spawning abundance ----
# Index of abundance from total biomass - beginning of year before fishing
# apply hyperstability / hyperdepletion
II <- (apply(SSB*Sample_Area$SBInd[,,1:nyears,], c(1, 3), sum)^ObsPars$SpI_beta) *
ObsPars$SpIerr_y[, 1:nyears]
II <- II/apply(II, 1, mean) # normalize
Data@SpInd <- II # index of spawning abundance
Data@CV_SpInd <- matrix(Data@CV_SpInd[,1], nrow=nsim, ncol=nyears)
# --- Index of vulnerable abundance ----
# Index of abundance from vulnerable biomass - beginning of year before fishing
# apply hyperstability / hyperdepletion
II <- (apply(VBiomass*Sample_Area$VInd[,,1:nyears,], c(1, 3), sum)^ObsPars$VI_beta) *
ObsPars$VIerr_y[, 1:nyears]
II <- II/apply(II, 1, mean) # normalize
Data@VInd <- II # index of vulnerable abundance
Data@CV_VInd <- matrix(Data@CV_VInd[,1], nrow=nsim, ncol=nyears)
# --- Index of recruitment ----
Data@Rec <- apply(N[, 1, , ]*Sample_Area$RecInd[,1:nyears,], c(1, 2), sum) *
ObsPars$Recerr_y[, 1:nyears]
Data@t <- rep(nyears, nsim) # number of years of data
# --- Average catch ----
Data@AvC <- apply(Data@Cat, 1, mean, na.rm=TRUE) # average catch over all years
# --- Depletion ----
# observed depletion
Depletion <- apply(SSB[,,nyears,],1,sum)/RefPoints$SSB0 # current depletion
Data@Dt <- Depletion * ObsPars$Derr_y[,nyears]
Data@Dep <- Depletion * ObsPars$Derr_y[,nyears]
# --- Life-history parameters ----
Data@vbLinf <- StockPars$Linfarray[,nyears] * ObsPars$Linfbias # observed vB Linf
Data@vbK <- StockPars$Karray[,nyears] * ObsPars$Kbias # observed vB K
Data@vbt0 <- StockPars$t0array[,nyears] + ObsPars$t0bias # observed vB t0
Data@Mort <- StockPars$Marray[,nyears] * ObsPars$Mbias # natural mortality
Data@L50 <- StockPars$L50array[,nyears] * ObsPars$lenMbias # observed length at 50% maturity
Data@L95 <- StockPars$L95array[,nyears] * ObsPars$lenMbias # observed length at 95% maturity
Data@L95[Data@L95 > 0.95 * Data@vbLinf] <- 0.95 * Data@vbLinf[Data@L95 > 0.95 * Data@vbLinf] # Set a hard limit on ratio of L95 to Linf
Data@L50[Data@L50 > 0.95 * Data@L95] <- 0.95 * Data@L95[Data@L50 > 0.95 * Data@L95] # Set a hard limit on ratio of L95 to Linf
Data@LenCV <- StockPars$LenCV # variability in length-at-age - no error at this time
Data@sigmaR <- StockPars$procsd * ObsPars$sigmaRbias # observed sigmaR -
Data@MaxAge <- StockPars$maxage # maximum age - no error - used for setting up matrices only
# Observed steepness values
hs <- StockPars$hs
Data@steep <- hs * ObsPars$hbias # observed steepness
# --- Reference points ----
# Simulate observation error in BMSY/B0
ntest <- 20 # number of trials
test <- array(RefPoints$SSBMSY_SSB0 * ObsPars$BMSY_B0bias, dim = c(nsim, ntest)) # the simulated observed BMSY_B0
indy <- array(rep(1:ntest, each = nsim), c(nsim, ntest)) # index
indy[test > max(0.9, max(RefPoints$SSBMSY_SSB0))] <- NA # interval censor
ObsPars$BMSY_B0bias <- ObsPars$BMSY_B0bias[cbind(1:nsim, apply(indy, 1, min, na.rm = T))] # sample such that BMSY_B0<90%
Data@FMSY_M <- RefPoints$FMSY_M * ObsPars$FMSY_Mbias # observed FMSY/M
Data@BMSY_B0 <- RefPoints$SSBMSY_SSB0 * ObsPars$BMSY_B0bias # observed BMSY/B0
Data@Cref <- RefPoints$MSY * ObsPars$Crefbias # Catch reference - MSY with error
Data@Bref <- RefPoints$VBMSY * ObsPars$Brefbias # Vuln biomass ref - VBMSY with error
# Generate values for reference SBMSY/SB0
# should be calculated from unfished - won't be correct if initD is set
I3 <- apply(Biomass, c(1, 3), sum)^ObsPars$I_beta # apply hyperstability / hyperdepletion
I3 <- I3/apply(I3, 1, mean) # normalize index to mean 1
if (!is.null(initD)) {
b1 <- apply(Biomass, c(1, 3), sum)
b2 <- matrix(RefPoints$BMSY, nrow=nsim, ncol=nyears)
ind <- apply(abs(b1/ b2 - 1), 1, which.min) # find years closest to BMSY
Iref <- diag(I3[1:nsim,ind]) # return the real target abundance index closest to BMSY
} else {
Iref <- apply(I3[, 1:min(5, ncol(I3))], 1, mean) * RefPoints$BMSY_B0 # return the real target abundance index corresponding to BMSY
}
Data@Iref <- Iref * ObsPars$Irefbias # index reference with error
# --- Abundance ----
# Calculate vulnerable and spawning biomass abundance --
M_array <- array(0.5*StockPars$M_ageArray[,,nyears], dim=c(nsim, StockPars$maxage+1, nareas))
A <- apply(VBiomass[, , nyears, ] * exp(-M_array), 1, sum) # Abundance (mid-year before fishing)
Asp <- apply(SSB[, , nyears, ] * exp(-M_array), 1, sum) # Spawning abundance (mid-year before fishing)
OFLreal <- A * (1-exp(-RefPoints$FMSY)) # the true simulated Over Fishing Limit
# Observed total abundance
Data@Abun <- A * ObsPars$Aerr_y[,nyears]
# observed spawning abundance
Data@SpAbun <- Asp * ObsPars$Aerr_y[,nyears]
# --- Catch-at-age ----
Cret2 <- apply(Cret * Sample_Area$CAA[,,1:nyears,], 1:3, sum)
Data@CAA <- simCAA(nsim, nyears, StockPars$maxage+1, Cret2, ObsPars$CAA_ESS, ObsPars$CAA_nsamp)
# --- Catch-at-length ----
CALdone <- FALSE
if (!is.null(control$CAL)) {
if (control$CAL == 'removals') {
vn <- apply(N*Sample_Area$CAL[,,1:nyears,], c(1,2,3), sum) * FleetPars$V_real[,,1:nyears]
# numbers at age in population that would be removed
vn <- aperm(vn, c(1,3, 2))
CALdat <- simCAL(nsim, nyears, StockPars$maxage, ObsPars$CAL_ESS,
ObsPars$CAL_nsamp, StockPars$nCALbins, StockPars$CAL_binsmid, StockPars$CAL_bins,
vn, FleetPars$SLarray_real, StockPars$Linfarray,
StockPars$Karray, StockPars$t0array, StockPars$LenCV)
CALdone <- TRUE
} else {
warning('Invalid entry in OM@cpars$control$CAL (use OM@cpars$control$CAL="removals")\nSimulating CAL from retained catch')
}
}
if (!CALdone) {
vn <- apply(N*Sample_Area$CAL[,,1:nyears,], c(1,2,3), sum) * FleetPars$retA_real[,,1:nyears]
# numbers at age in population that would be retained
vn <- aperm(vn, c(1,3, 2))
CALdat <- simCAL(nsim, nyears, maxage=StockPars$maxage,
CAL_ESS=ObsPars$CAL_ESS,
CAL_nsamp=ObsPars$CAL_nsamp,
nCALbins=StockPars$nCALbins,
CAL_binsmid=StockPars$CAL_binsmid,
CAL_bins=StockPars$CAL_bins,
vn,
retL=FleetPars$retL_real,
Linfarray=StockPars$Linfarray,
Karray=StockPars$Karray,
t0array=StockPars$t0array,
LenCV=StockPars$LenCV)
}
Data@CAL_bins <- StockPars$CAL_bins
Data@CAL_mids <- StockPars$CAL_binsmid
Data@CAL <- CALdat$CAL # observed catch-at-length
Data@ML <- CALdat$ML # mean length
Data@Lc <- CALdat$Lc # modal length
Data@Lbar <- CALdat$Lbar # mean length above Lc
Data@LFC <- FleetPars$L5_y[,nyears] * ObsPars$LFCbias # length at first capture
Data@LFS <- FleetPars$LFS_y[,nyears] * ObsPars$LFSbias # length at full selection
Data@Vmaxlen <- FleetPars$Vmaxlen_y[,nyears]
# --- Previous Management Recommendations ----
Data@MPrec <- apply(CBret, c(1, 3), sum)[,nyears] # catch in last year
Data@MPeff <- rep(1, nsim) # effort in last year = 1
# --- Store OM Parameters ----
# put all the operating model parameters in one table
ind <- which(lapply(StockPars, length) == nsim)
drop_srr <- which(names(ind)=='SRRpars')
ind <- ind[-drop_srr]
stock <- as.data.frame(StockPars[ind])
stock$Fdisc <- NULL
stock$CAL_bins <- NULL
stock$CAL_binsmid <- NULL
ind <- which(lapply(FleetPars, length) == nsim)
fleet <- as.data.frame(FleetPars[ind])
ind <- which(lapply(ImpPars, length) == nsim)
imp <- as.data.frame(ImpPars[ind])
refs <- RefPoints[!names(RefPoints) %in% names(stock)]
OMtable <- data.frame(stock, fleet, imp, refs, ageM=StockPars$ageMarray[,nyears],
L5=FleetPars$L5_y[,nyears ], LFS=FleetPars$LFS_y[,nyears ],
Vmaxlen=FleetPars$Vmaxlen_y[,nyears ],
LR5=FleetPars$LR5_y[,nyears], LFR=FleetPars$LFR_y[,nyears],
Rmaxlen=FleetPars$Rmaxlen_y[,nyears],
DR=FleetPars$DR_y[,nyears], OFLreal, maxF=StockPars$maxF,
A=A, Asp=Asp, CurrentYr=CurrentYr)
OMtable <- OMtable[,order(names(OMtable))]
Data@OM <- OMtable
# --- Store Obs Parameters ----
ObsParsDF <- ObsPars
ind <- which(lapply(ObsParsDF, length)==nsim) %>% as.numeric()
ObsParsDF <- ObsParsDF[ind]
ObsTable <- as.data.frame(ObsParsDF)
ObsTable <- ObsTable[,order(names(ObsTable))]
Data@Obs <- ObsTable # put all the observation error model parameters in one table
# --- Misc ----
Data@Units <- "unitless"
Data@Ref_type <- "Simulated OFL"
Data@wla <- rep(StockPars$a, nsim)
Data@wlb <- rep(StockPars$b, nsim)
Data@nareas <- nareas
Data@Ref <- OFLreal
Data@LHYear <- CurrentYr # Last historical year is nyears (for fixed MPs)
Data@Misc <- vector("list", nsim)
Data
}
updateData <- function(Data, OM, MPCalcs, Effort, Biomass, N, Biomass_P, CB_Pret,
N_P, SSB, SSB_P, VBiomass, VBiomass_P, RefPoints,
retA_P,
retL_P, StockPars, FleetPars, ObsPars, ImpPars,
V_P,
upyrs, interval, y=2,
mm=1, Misc, RealData, Sample_Area) {
yind <- upyrs[match(y, upyrs) - 1]:(upyrs[match(y, upyrs)] - 1) # index
nyears <- OM@nyears
proyears <- OM@proyears
nsim <- OM@nsim
nareas <- StockPars$nareas
reps <- OM@reps
Data@Year <- (OM@CurrentYr - nyears+1):(OM@CurrentYr+ y - 1)
Data@t <- rep(nyears + y, nsim)
# --- Simulate catches ----
CBtemp <- CB_Pret[, , yind, , drop=FALSE] * Sample_Area$Catch[,,nyears+yind,, drop=FALSE]
CBtemp[is.na(CBtemp)] <- tiny
CBtemp[!is.finite(CBtemp)] <- tiny
yr.index <- max(which(!is.na(Data@CV_Cat[1,])))
newCV_Cat <- matrix(Data@CV_Cat[,yr.index], nrow=nsim, ncol=length(yind))
Data@CV_Cat <- cbind(Data@CV_Cat, newCV_Cat)
# --- Observed catch ----
# Simulated observed retained catch (biomass)
Cobs <- ObsPars$Cobs_y[,nyears + yind] * apply(CBtemp, c(1, 3), sum, na.rm = TRUE)
Data@Cat <- cbind(Data@Cat, Cobs)
if (!is.null(RealData) && ncol(RealData@Cat)>nyears &&
!all(is.na(RealData@Cat[1,(nyears+1):length(RealData@Cat[1,])]))) {
# update projection catches with observed catches
addYr <- min(y,ncol(RealData@Cat) - nyears)
Data@Cat[,(nyears+1):(nyears+addYr)] <- matrix(RealData@Cat[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
Data@CV_Cat[,(nyears+1):(nyears+addYr)] <- matrix(RealData@CV_Cat[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
}
# --- Index of total abundance ----
yr.ind <- max(which(!is.na(ObsPars$Ierr_y[1,1:nyears])))
I2 <- cbind(apply(Biomass*Sample_Area$BInd[,,1:nyears,], c(1, 3), sum)[,yr.ind:nyears],
apply(Biomass_P*Sample_Area$BInd[,,(nyears+1):(nyears+proyears),], c(1, 3), sum)[, 1:(y - 1)])
# standardize, apply beta & obs error
I2 <- exp(lcs(I2))^ObsPars$I_beta * ObsPars$Ierr_y[,yr.ind:(nyears + (y - 1))]
# I2 <- exp(lcs(I2)) * ObsPars$Ierr_y[,yr.ind:(nyears + (y - 1))]
year.ind <- max(which(!is.na(Data@Ind[1,1:nyears])))
scaler <- Data@Ind[,year.ind]/I2[,1]
scaler <- matrix(scaler, nrow=nsim, ncol=ncol(I2))
I2 <- I2 * scaler # convert back to historical index scale
I2 <- cbind(Data@Ind[,1:(yr.ind)], I2[,2:ncol(I2)])
Data@Ind <- I2
yr.index <- max(which(!is.na(Data@CV_Ind[1,1:nyears])))
newCV_Ind <- matrix(Data@CV_Ind[,yr.index], nrow=nsim, ncol=length(yind))
Data@CV_Ind <- cbind(Data@CV_Ind, newCV_Ind)
if (!is.null(RealData) && ncol(RealData@Ind)>nyears &&
!all(is.na(RealData@Ind[1,(nyears+1):length(RealData@Ind[1,])]))) {
# update projection index with observed index if it exists
addYr <- min(y,ncol(RealData@Ind) - nyears)
Data@Ind[,(nyears+1):(nyears+addYr)] <- matrix(RealData@Ind[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
Data@CV_Ind[,(nyears+1):(nyears+addYr)] <- matrix(RealData@CV_Ind[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
}
# --- Index of spawning abundance ----
yr.ind <- max(which(!is.na(ObsPars$SpIerr_y[1,1:nyears])))
I2 <- cbind(apply(SSB*Sample_Area$SBInd[,,1:nyears,], c(1, 3), sum)[,yr.ind:nyears],
apply(SSB_P*Sample_Area$SBInd[,,(nyears+1):(nyears+proyears),], c(1, 3), sum)[, 1:(y - 1)])
# standardize, apply beta & obs error
I2 <- exp(lcs(I2))^ObsPars$SpI_beta * ObsPars$SpIerr_y[,yr.ind:(nyears + (y - 1))]
year.ind <- max(which(!is.na(Data@SpInd[1,1:nyears])))
scaler <- Data@SpInd[,year.ind]/I2[,1]
scaler <- matrix(scaler, nrow=nsim, ncol=ncol(I2))
I2 <- I2 * scaler # convert back to historical index scale
I2 <- cbind(Data@SpInd[,1:(yr.ind)], I2[,2:ncol(I2)])
Data@SpInd <- I2
yr.index <- max(which(!is.na(Data@CV_SpInd[1,1:nyears])))
newCV_Ind <- matrix(Data@CV_SpInd[,yr.index], nrow=nsim, ncol=length(yind))
Data@CV_SpInd <- cbind(Data@CV_SpInd, newCV_Ind)
if (!is.null(RealData) && ncol(RealData@SpInd)>nyears &&
!all(is.na(RealData@SpInd[1,(nyears+1):length(RealData@SpInd[1,])]))) {
# update projection index with observed index if it exists
addYr <- min(y,ncol(RealData@SpInd) - nyears)
Data@SpInd[,(nyears+1):(nyears+addYr)] <- matrix(RealData@SpInd[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
Data@CV_SpInd[,(nyears+1):(nyears+addYr)] <- matrix(RealData@CV_SpInd[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
}
# --- Index of vulnerable abundance ----
yr.ind <- max(which(!is.na(ObsPars$VIerr_y[1,1:nyears])))
I2 <- cbind(apply(VBiomass*Sample_Area$VInd[,,1:nyears,], c(1, 3), sum)[,yr.ind:nyears],
apply(VBiomass_P*Sample_Area$VInd[,,(nyears+1):(nyears+proyears),], c(1, 3), sum)[, 1:(y - 1)])
# standardize, apply beta & obs error
I2 <- exp(lcs(I2))^ObsPars$VI_beta * ObsPars$VIerr_y[,yr.ind:(nyears + (y - 1))]
year.ind <- max(which(!is.na(Data@VInd[1,1:nyears])))
scaler <- Data@VInd[,year.ind]/I2[,1]
scaler <- matrix(scaler, nrow=nsim, ncol=ncol(I2))
I2 <- I2 * scaler # convert back to historical index scale
I2 <- cbind(Data@VInd[,1:(yr.ind)], I2[,2:ncol(I2)])
Data@VInd <- I2
yr.index <- max(which(!is.na(Data@CV_VInd[1,1:nyears])))
newCV_Ind <- matrix(Data@CV_VInd[,yr.index], nrow=nsim, ncol=length(yind))
Data@CV_VInd <- cbind(Data@CV_VInd, newCV_Ind)
if (!is.null(RealData) && ncol(RealData@VInd)>nyears &&
!all(is.na(RealData@VInd[1,(nyears+1):length(RealData@VInd[1,])]))) {
# update projection index with observed index if it exists
addYr <- min(y,ncol(RealData@VInd) - nyears)
Data@VInd[,(nyears+1):(nyears+addYr)] <- matrix(RealData@VInd[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
Data@CV_VInd[,(nyears+1):(nyears+addYr)] <- matrix(RealData@CV_VInd[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
}
# --- Update additional indices (if they exist) ----
AddIunits <- Data@AddIunits
AddIndType <- Data@AddIndType
if (length(ObsPars$AddIerr)>0) {
n.ind <- dim(ObsPars$AddIerr)[2]
AddInd <- array(NA, dim=c(nsim, n.ind, nyears+y-1))
CV_AddInd <- array(NA, dim=c(nsim, n.ind, nyears+y-1))
for (i in 1:n.ind) {
if (all(is.na(RealData@AddIndV[1, , ]))) {
Ind_V <- rep(1, Data@MaxAge+1)
} else {
Ind_V <- RealData@AddIndV[1,i, ]
}
Ind_V <- matrix(Ind_V, nrow=Data@MaxAge+1, ncol= nyears+proyears)
Ind_V <- replicate(nsim, Ind_V) %>% aperm(c(3,1,2))
nas <- which(!is.na(ObsPars$AddIerr[1,i, 1:nyears]))
if (length(nas)>0) {
yr.ind <- max(nas)
if (AddIunits[i]) { # Biomass-based index
if (AddIndType[i]==1) {
# total biomass
b1 <- apply(Biomass[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum)
b2 <- apply(Biomass_P, c(1, 2, 3), sum)
}
if (AddIndType[i]==2) {
# spawning biomass
b1 <- apply(SSB[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum)
b2 <- apply(SSB_P, c(1, 2, 3), sum)
}
if (AddIndType[i]==3) {
# vulnerable biomass
b1 <- apply(VBiomass[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum)
b2 <- apply(VBiomass_P, c(1, 2, 3), sum)
}
} else {
if (AddIndType[i]==1) {
# total stock
b1 <- apply(N[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum) # Abundance-based index
b2 <- apply(N_P, c(1, 2, 3), sum)
}
if (AddIndType[i]==2) {
# spawning stock
b1 <- apply(N[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum) * StockPars$Mat_age[,,yr.ind:nyears, drop=FALSE]
b2 <- apply(N_P, c(1, 2, 3), sum) * StockPars$Mat_age[,,(nyears+1):(nyears+proyears), drop=FALSE]
}
if (AddIndType[i]==3) {
# vuln stock
b1 <- apply(N[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum) * V_P[,,yr.ind:nyears, drop=FALSE]
b2 <- apply(N_P, c(1, 2, 3), sum) * V_P[,,(nyears+1):(nyears+proyears), drop=FALSE]
}
}
b1 <- apply(b1 * Ind_V[,,yr.ind:nyears, drop=FALSE], c(1,3), sum)
b2 <- apply(b2 * Ind_V[,,(nyears+1):(nyears+proyears), drop=FALSE], c(1,3), sum)
tempI <- cbind(b1, b2[, 1:(y - 1)])
# standardize, apply beta & obs error
tempI <- exp(lcs(tempI))^ObsPars$AddIbeta[,i] * ObsPars$AddIerr[,i,yr.ind:(nyears + (y - 1))]
year.ind <- max(which(!is.na(RealData@AddInd[1,i,])))
scaler <- RealData@AddInd[1,i,year.ind]/tempI[,1]
scaler <- matrix(scaler, nrow=nsim, ncol=ncol(tempI))
tempI <- tempI * scaler # convert back to historical index scale
AddInd[,i,] <- cbind(Data@AddInd[1:nsim,i,1:yr.ind], tempI[,2:ncol(tempI)])
yr.index <- max(which(!is.na(Data@CV_AddInd[1,i,1:nyears])))
newCV_Ind <- matrix(Data@CV_AddInd[,i,yr.index], nrow=nsim, ncol=length(yind))
CV_AddInd[,i,] <- cbind(Data@CV_AddInd[,i,], newCV_Ind)
if (!is.null(RealData) && length(RealData@AddInd[1,i,])>nyears &&
!all(is.na(RealData@AddInd[1,i,(nyears+1):length(RealData@AddInd[1,i,])]))) {
# update projection index with observed index if it exists
addYr <- min(y-1,length(RealData@AddInd[1,i,]) - nyears)
AddInd[,i,(nyears+1):(nyears+addYr)] <- matrix(RealData@AddInd[1,i,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
CV_AddInd[,i,(nyears+1):(nyears+addYr)] <- matrix(RealData@CV_AddInd[1,i,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
}
}
}
Data@AddInd <- AddInd
Data@CV_AddInd <- CV_AddInd
}
# --- Index of recruitment ----
Recobs <- ObsPars$Recerr_y[, nyears + yind] * apply(array(N_P[, 1, yind, ] *
Sample_Area$RecInd[,nyears+yind,],
c(nsim, length(yind), nareas)),
c(1, 2), sum)
Data@Rec <- cbind(Data@Rec, Recobs)
# --- Average catch ----
Data@AvC <- apply(Data@Cat, 1, mean)
# --- Depletion ----
Depletion <- apply(SSB_P[, , y, ], 1, sum)/RefPoints$SSB0
Depletion[Depletion < tiny] <- tiny
Data@Dt <- Depletion * ObsPars$Derr_y[,nyears+y]
Data@Dep <- Depletion * ObsPars$Derr_y[,nyears+y]
# --- Update life-history parameter estimates for current year ----
Data@vbLinf <- StockPars$Linfarray[,nyears+y] * ObsPars$Linfbias # observed vB Linf
Data@vbK <- StockPars$Karray[,nyears+y] * ObsPars$Kbias # observed vB K
Data@vbt0 <- StockPars$t0array[,nyears+y] + ObsPars$t0bias # observed vB t0
Data@Mort <- StockPars$Marray[,nyears+y] * ObsPars$Mbias # natural mortality
Data@L50 <- StockPars$L50array[,nyears+y] * ObsPars$lenMbias # observed length at 50% maturity
Data@L95 <- StockPars$L95array[,nyears+y] * ObsPars$lenMbias # observed length at 95% maturity
Data@L95[Data@L95 > 0.9 * Data@vbLinf] <- 0.9 * Data@vbLinf[Data@L95 > 0.9 * Data@vbLinf] # Set a hard limit on ratio of L95 to Linf
Data@L50[Data@L50 > 0.9 * Data@L95] <- 0.9 * Data@L95[Data@L50 > 0.9 * Data@L95] # Set a hard limit on ratio of L95 to Linf
# --- Abundance ----
# Calculate vulnerable and spawning biomass abundance --
M_array <- array(0.5*StockPars$M_ageArray[,,nyears+y], dim=c(nsim, StockPars$maxage+1, nareas))
A <- apply(VBiomass_P[, , y, ] * exp(-M_array), 1, sum) # Abundance (mid-year before fishing)
Asp <- apply(SSB_P[, , y, ] * exp(-M_array), 1, sum) # Spawning abundance (mid-year before fishing)
Data@Abun <- A * ObsPars$Aerr_y[,nyears+yind]
Data@SpAbun <- Asp * ObsPars$Aerr_y[,nyears+yind]
# --- Catch-at-age ----
# previous CAA
oldCAA <- Data@CAA
Data@CAA <- array(0, dim = c(nsim, nyears + y - 1, StockPars$maxage+1))
Data@CAA[, 1:(nyears + y - interval[mm] - 1), ] <- oldCAA[, 1:(nyears + y - interval[mm] - 1), ]
# update CAA
CNtemp <- retA_P[,,yind+nyears, drop=FALSE] *
apply(N_P[,,yind,, drop=FALSE]*Sample_Area$CAA[,,nyears+yind,, drop=FALSE], 1:3, sum)
CNtemp[is.na(CNtemp)] <- tiny
CNtemp[!is.finite(CNtemp)] <- tiny
CAA <- simCAA(nsim, yrs=length(yind), StockPars$maxage+1, Cret=CNtemp, ObsPars$CAA_ESS, ObsPars$CAA_nsamp)
Data@CAA[, nyears + yind, ] <- CAA
# --- Catch-at-length ----
oldCAL <- Data@CAL
Data@CAL <- array(0, dim = c(nsim, nyears + y - 1, StockPars$nCALbins))
Data@CAL[, 1:(nyears + y - interval[mm] - 1), ] <- oldCAL[, 1:(nyears + y - interval[mm] - 1), ]
CAL <- array(NA, dim = c(nsim, interval[mm], StockPars$nCALbins))
if (!all(is.na(Data@Vuln_CAL))) {
Vuln_CAL <- replicate(nsim,Data@Vuln_CAL[1,])
Vuln_CAL <- replicate(length(yind),Vuln_CAL)
Vuln_CAL <- aperm(Vuln_CAL, c(2,1,3))
VList_dat <- lapply(1:nsim, calcV,
Len_age=StockPars$Len_age[,,nyears+yind, drop=FALSE],
LatASD=StockPars$LatASD[,,nyears+yind, drop=FALSE],
SLarray=Vuln_CAL,
n_age=StockPars$maxage+1,
nyears=length(yind), proyears = 0,
CAL_binsmid=StockPars$CAL_binsmid)
V_data <- aperm(array(as.numeric(unlist(VList_dat, use.names=FALSE)), dim=c(StockPars$maxage+1, length(yind), nsim)), c(3,1,2))
vn <- (apply(N_P[,,yind,, drop=FALSE]*Sample_Area$CAL[,,(nyears+yind),, drop=FALSE], c(1,2,3), sum) * V_data)
vn <- aperm(vn, c(1,3,2))
CALdat <- simCAL(nsim, nyears=length(yind), StockPars$maxage, ObsPars$CAL_ESS,
ObsPars$CAL_nsamp, StockPars$nCALbins, StockPars$CAL_binsmid, StockPars$CAL_bins,
vn=vn, retL=Vuln_CAL,
Linfarray=StockPars$Linfarray[,nyears + yind, drop=FALSE],
Karray=StockPars$Karray[,nyears + yind, drop=FALSE],
t0array=StockPars$t0array[,nyears + yind,drop=FALSE],
LenCV=StockPars$LenCV)
} else {
vn <- (apply(N_P*Sample_Area$CAL[,,(nyears+1):(nyears+proyears),], c(1,2,3), sum) *
V_P[,,(nyears+1):(nyears+proyears)] *
retA_P[,,(nyears+1):(nyears+proyears)]) # numbers at age that would be retained
vn <- aperm(vn, c(1,3,2))
CALdat <- simCAL(nsim, nyears=length(yind), maxage=StockPars$maxage,
CAL_ESS=ObsPars$CAL_ESS,
CAL_nsamp=ObsPars$CAL_nsamp,
nCALbins=StockPars$nCALbins,
CAL_binsmid=StockPars$CAL_binsmid,
CAL_bins=StockPars$CAL_bins,
vn=vn[,yind,, drop=FALSE],
retL=retL_P[,,nyears+yind, drop=FALSE],
Linfarray=StockPars$Linfarray[,nyears + yind, drop=FALSE],
Karray=StockPars$Karray[,nyears + yind, drop=FALSE],
t0array=StockPars$t0array[,nyears + yind,drop=FALSE],
LenCV=StockPars$LenCV)
}
Data@CAL[, nyears + yind, ] <- CALdat$CAL # observed catch-at-length
Data@ML <- cbind(Data@ML, CALdat$ML) # mean length
Data@Lc <- cbind(Data@Lc, CALdat$Lc) # modal length
Data@Lbar <- cbind(Data@Lbar, CALdat$Lbar) # mean length above Lc
Data@LFC <- FleetPars$L5_y[,nyears+y] * ObsPars$LFCbias # length at first capture
Data@LFS <- FleetPars$LFS_y[,nyears+y] * ObsPars$LFSbias # length at full selection
Data@Vmaxlen <- FleetPars$Vmaxlen_y[,nyears+y]
# --- Previous Management Recommendations ----
Data@MPrec <- MPCalcs$TACrec # last MP TAC recommendation
if (length(dim(Effort)) == 5) {
Data@MPeff <- Effort[, 1,1,mm, y-1] # last recommended effort
} else {
Data@MPeff <- Effort[, mm, y-1] # last recommended effort
}
# --- Store OM Parameters ----
# put all the operating model parameters in one table
ind <- which(lapply(StockPars, length) == nsim)
drop_srr <- which(names(ind)=='SRRpars')
ind <- ind[-drop_srr]
stock <- as.data.frame(StockPars[ind])
stock$Fdisc <- NULL
stock$CAL_bins <- NULL
stock$CAL_binsmid <- NULL
ind <- which(lapply(FleetPars, length) == nsim)
fleet <- as.data.frame(FleetPars[ind])
ind <- which(lapply(ImpPars, length) == nsim)
imp <- as.data.frame(ImpPars[ind])
refs <- RefPoints# [!names(RefPoints) %in% names(stock)]
OFLreal <- A * (1-exp(-RefPoints$FMSY)) # the true simulated Over Fishing Limit
OMtable <- data.frame(stock, fleet, imp, refs, ageM=StockPars$ageMarray[,nyears+y],
L5=FleetPars$L5_y[,nyears+y], LFS=FleetPars$LFS_y[,nyears+y],
Vmaxlen=FleetPars$Vmaxlen_y[,nyears+y],
LR5=FleetPars$LR5_y[,nyears], LFR=FleetPars$LFR_y[,nyears+y],
Rmaxlen=FleetPars$Rmaxlen_y[,nyears+y],
DR=FleetPars$DR_y[,nyears+y], OFLreal, maxF=StockPars$maxF,
A=A, Asp=Asp, CurrentYr=OM@CurrentYr)
OMtable <- OMtable[,order(names(OMtable))]
Data@OM <- OMtable
Data@Misc <- Misc
if (!is.null(RealData)) {
if (length(RealData@Misc) && !is.null(names(RealData@Misc))) {
Data@Misc[names(RealData@Misc)] <- RealData@Misc[names(RealData@Misc)]
}
}
Data
}
#' Simulate Catch-at-Age Data
#'
#' CAA generated with either a multinomial or logistic normal observation model from retained catch-at-age
#' array
#'
#' @param nsim Number of simulations
#' @param yrs Number of years
#' @param n_age Number of age classes
#' @param Cret Retained Catch at age in numbers - array(sim, years, maxage+1)
#' @param CAA_ESS CAA effective sample size. If greater than 1, then this is the multinomial distribution sample size.
#' If less than 1, this is the coefficient of variation for the logistic normal distribution (see details).
#' @param CAA_nsamp CAA sample size
#' @details The logistic normal generates the catch-at-age sample by first sampling once from a multivariate normal distribution
#' with the mean vector equal to the logarithm of the proportions-at-age and the diagonal of the covariance matrix is the square of the
#' product of the CV and the log proportions (all off-diagonals are zero). The sampled vector is then converted to proportions
#' with the softmax function and expanded to numbers (CAA_nsamp). This method allows for simulating fractional values in the
#' catch-at-age matrix.
#' @return CAA array
simCAA <- function(nsim, yrs, n_age, Cret, CAA_ESS, CAA_nsamp) {
# generate CAA from retained catch-at-age
CAA <- array(NA, dim = c(nsim, yrs, n_age)) # Catch at age array
# a multinomial observation model for catch-at-age data
# logistic normal if CAA_ESS < 1
if(any(CAA_ESS < 1) && !requireNamespace("mvtnorm", quietly = TRUE)) stop("The mvtnorm package is needed.")
for (i in 1:nsim) {
for (j in 1:yrs) {
if (!sum(Cret[i, ,j])) {
CAA[i, j, ] <- 0
} else {
if(CAA_ESS[i] < 1) {
if(CAA_ESS[i] <= 0) {
CAA[i, j, ] <- 0
} else {
log_PAA <- log(Cret[i, , j]/sum(Cret[i, , j]))
vcov_PAA <- (CAA_ESS[i] * log_PAA)^2 * diag(length(log_PAA))
samp_PAA <- mvtnorm::rmvnorm(1, log_PAA, vcov_PAA) %>% ilogitm()
CAA[i, j, ] <- CAA_nsamp[i] * samp_PAA
}
} else {
CAA[i, j, ] <- ceiling(-0.5 + rmultinom(1, CAA_ESS[i], Cret[i, ,j]) * CAA_nsamp[i]/CAA_ESS[i])
}
}
}
}
CAA
}
#' Simulate Catch-at-Length Data
#'
#' Simulate CAL and calculate length-at-first capture (LFC),
#' mean length (ML), modal length (Lc), and mean length over modal length (Lbar)
#'
#' @param nsim Number of simulations
#' @param nyears Number of years
#' @param maxage Maximum age
#' @param CAL_ESS CAA effective sample size
#' @param CAL_nsamp CAA sample size
#' @param nCALbins number of CAL bins
#' @param CAL_binsmid mid-points of CAL bins
#' @param CAL_bins Boundary of CAL bins
#' @param vn Vulnerable numbers-at-age
#' @param retL Retention at length curve
#' @param Linfarray Array of Linf values by simulation and year
#' @param Karray Array of K values by simulation and year
#' @param t0array Array of t0 values by simulation and year
#' @param LenCV CV of length-at-age#'
#' @return named list with CAL array and LFC, ML, & Lc vectors
#'
simCAL <- function(nsim, nyears, maxage, CAL_ESS, CAL_nsamp, nCALbins, CAL_binsmid, CAL_bins,
vn, retL, Linfarray, Karray, t0array, LenCV) {
# a multinomial observation model for catch-at-length data
# assumed normally-distributed length-at-age truncated at 2 standard deviations from the mean
CAL <- array(NA, dim=c(nsim, nyears, nCALbins))
# Generate size comp data with variability in age
runParallel <- snowfall::sfIsRunning()
if (runParallel) {
tempSize <- snowfall::sfLapply(1:nsim, genSizeCompWrap, vn, CAL_binsmid, CAL_bins,
retL, CAL_ESS, CAL_nsamp,
Linfarray, Karray, t0array, LenCV, truncSD=2)
} else {
# vn <<- vn
# CAL_binsmid <<- CAL_binsmid
# CAL_bins <<- CAL_bins
# retL <<- retL
# CAL_ESS <<- CAL_ESS
# CAL_nsamp <<- CAL_nsamp
# Linfarray <<- Linfarray
# Karray <<- Karray
# t0array <<- t0array
# LenCV <<- LenCV
#
tempSize <- lapply(1:nsim, genSizeCompWrap, vn, CAL_binsmid, CAL_bins, retL, CAL_ESS,
CAL_nsamp,
Linfarray, Karray, t0array, LenCV, truncSD=2)
}
CAL <- aperm(array(as.numeric(unlist(tempSize, use.names=FALSE)),
dim=c(nyears, length(CAL_binsmid), nsim)), c(3,1,2))
# calculate LFC - length-at-first capture - 5th percentile
LFC <- rep(NA, nsim)
LFC <- unlist(lapply(tempSize, function(x) getfifth(x[nyears, ], CAL_binsmid)))
LFC[is.na(LFC)] <- 1
LFC[LFC<1] <- 1
# Mean Length
temp <- CAL * rep(CAL_binsmid, each = nsim * nyears)
ML <- apply(temp, 1:2, sum)/apply(CAL, 1:2, sum)
ML[!is.finite(ML)] <- 0
# Lc - modal length
Lc <- array(CAL_binsmid[apply(CAL, 1:2, which.max)], dim = c(nsim, nyears))
# Lbar - mean length above Lc
nuCAL <- CAL
for (i in 1:nsim) {
for (j in 1:nyears) {
# nuCAL[i, j, 1:match(max(1, Lc[i, j]), CAL_binsmid, nomatch=1)] <- NA
lcbin <- max(1,match(max(1, Lc[i, j]), CAL_binsmid, nomatch=1)-1)
nuCAL[i, j, 1:lcbin] <- NA
}
}
temp <- nuCAL * rep(CAL_binsmid, each = nsim * nyears)
Lbar <- apply(temp, 1:2, sum, na.rm=TRUE)/apply(nuCAL, 1:2, sum, na.rm=TRUE)
Lbar[!is.finite(Lbar)] <- 0
out <- list()
out$CAL <- CAL
out$LFC <- LFC
out$ML <- ML
out$Lc <- Lc
out$Lbar <- Lbar
out
}
#' Wrapper for C++ function to generate length composition
#'
#' And other internal related functions
#'
#' @param i Simulation number
#' @param vn Array of vulnerable numbers
#' @param CAL_binsmid Mid-points of CAL bins
#' @param CAL_bins Boundary of length bins
#' @param retL Array of retention-at-length
#' @param CAL_ESS CAL effective sample size
#' @param CAL_nsamp CAL sample size
#' @param Linfarray Matrix of Linf
#' @param Karray Matrix of K values
#' @param t0array Matrix of t0 values
#' @param LenCV Vector of LenCV
#' @param truncSD Numeric. Number of standard deviations to truncate normal d
#' distribution
#'
#' @return Generated length composition from `genSizeComp`
#'
#' @keywords internal
genSizeCompWrap <- function(i, vn, CAL_binsmid, CAL_bins, retL,
CAL_ESS, CAL_nsamp,
Linfarray, Karray, t0array,
LenCV, truncSD=2) {
VulnN <- as.matrix(vn[i,,])
if (ncol(VulnN)>1) {
VulnN <- VulnN/rowSums(VulnN) * CAL_nsamp[i] # get relative numbers at age
} else {
VulnN <- VulnN/sum(VulnN) * CAL_nsamp[i] # get relative numbers at age
}
# VulnN <- round(VulnN,0) # convert to integers
nyrs <- nrow(as.matrix(Linfarray[i,]))
if (nyrs == 1) VulnN <- t(VulnN)
retLa <- as.matrix(retL[i,,])
# assumes lengths sampled throughout years
lens <- genSizeComp(VulnN, CAL_binsmid, CAL_bins, retLa,
CAL_ESS=CAL_ESS[i], CAL_nsamp=CAL_nsamp[i],
Linfs=Linfarray[i,], Ks=Karray[i,], t0s=t0array[i,],
LenCV=LenCV[i], truncSD)
# snapshot length comp
# lens <- genSizeComp2(VulnN, CAL_binsmid, CAL_bins, retLa,
# CAL_ESS=CAL_ESS[i], CAL_nsamp=CAL_nsamp[i],
# Linfs=Linfarray[i,], Ks=Karray[i,], t0s=t0array[i,],
# LenCV=LenCV[i], truncSD)
lens[!is.finite(lens)] <- 0
lens
}
#' @describeIn genSizeCompWrap Internal function to calculate fifth percentile of size composition
#' @param lenvec Vector of lengths
getfifth <- function(lenvec, CAL_binsmid) {
temp <- rep(CAL_binsmid, lenvec)
if(sum(lenvec)==0) return(NA)
dens <- try(density(temp), silent=TRUE)
if(!methods::is(dens, "density")) return(NA)
dens$x[min(which(cumsum(dens$y/sum(dens$y)) >0.05))]
}
UpdateObs <- function(sl, obsval, OMval, RealData, SimData, msg){
RealVal <- slot(RealData, sl)[1]
SimVal <- slot(SimData, sl)
if (!is.na(RealVal) & !all(SimVal == tiny)) {
if (msg)
message_info(paste0('Updating Observation Error for `OM@cpars$Data@', sl, '`'))
# bias
if (sl == 'vbt0') {
calcBias <- RealVal-OMval
} else {
calcBias <- RealVal/OMval
}
return(calcBias)
} else {
return(obsval)
}
}
UpdateSlot <- function(sl, RealData, SimData, msg) {
RealVal <- slot(RealData, sl)[1]
SimVal <- slot(SimData, sl)
if (!is.na(RealVal) & !all(SimVal == tiny)) {
if (msg)
message_info(paste0('Using `OM@cpars$Data@', sl, '` (', RealVal, ')'))
nrep <- length(slot(SimData, sl))
return(rep(RealVal, nrep))
}
return(slot(SimData, sl))
}
check_Index_Fit <- function(Stats, index=1) {
ac_err <- sum(!is.finite(Stats$AC))
sd_err <- sum(!is.finite(Stats$SD))
if (ac_err | sd_err) {
if (is.numeric(index)) {
warning('An error occurred in calculating statistical properties of fit to Additional Index ', index,
' (possibly because there was only one observed data point). \nUsing the index observation error for slot `Ind` from `Obs` object (or possibly conditioned if `cpars$Data@Ind` was provided).\nUse `cpars$AddIerr` to manually set the observation error.')
} else {
warning('An error occurred in calculating statistical properties of fit to Index ', index,
' (possibly because there was only one observed data point). \nUsing the index observation error for slot `Ind` from `Obs` object (or possibly conditioned if `cpars$Data@Ind` was provided).\nUse `cpars$AddIerr` to manually set the observation error.')
}
return(FALSE)
}
TRUE
}
AddRealData <- function(SimData, RealData, ObsPars, StockPars, FleetPars, nsim,
nyears, proyears, SampCpars, msg, control, Sample_Area) {
Data_out <- SimData
if (msg)
message('Updating Simulated Data with Real Data from `OM@cpars$Data`')
# check last year
if (!is.na(RealData@LHYear) && !SimData@LHYear == RealData@LHYear) {
warning('`Fleet@CurrentYear` (', SimData@LHYear, ') is not the same as `OM@cpars$Data@LHYear` (', RealData@LHYear, ')')
}
# check maxage
if (!is.na(RealData@MaxAge) && !SimData@MaxAge == RealData@MaxAge) {
warning('`Stock@MaxAge` (', SimData@MaxAge, ') is not the same as `OM@cpars$Data@MaxAge` (', RealData@MaxAge, ')')
}
# check dimensions of real data CAA
if (!all(is.na(RealData@CAA)) && dim(RealData@CAA)[3] > 1) {
if (!dim(RealData@CAA)[3] == SimData@MaxAge+1)
stop('`OM@cpars$Data@CAA` has incorrect dimensions, should be `Stock@maxage+1` age-classes')
}
# update static slots
slts <- c('Name', 'Common_Name', 'Species', 'Region', 'LHYear', 'Units')
for (sl in slts)
slot(Data_out, sl) <- slot(RealData, sl)
Data_out@MaxAge <- SimData@MaxAge
if (!is.na(RealData@MPrec)) {
Data_out@MPrec <- rep(RealData@MPrec, nsim)
} else {
Data_out@MPrec <- SimData@MPrec
}
Data_out@MPeff <- SimData@MPeff
Data_out@nareas <- SimData@nareas
Data_out@LHYear <- SimData@LHYear
# ---- Update Life-history parameters ----
Data_out@Mort <- UpdateSlot('Mort', RealData, SimData, msg)
ObsPars$Mbias <- UpdateObs('Mort', ObsPars$Mbias, StockPars$Marray[, nyears],
RealData, SimData, msg)
Data_out@vbLinf <- UpdateSlot('vbLinf', RealData, SimData, msg)
ObsPars$Linfbias <- UpdateObs('vbLinf', ObsPars$Linfbias, StockPars$Linfarray[, nyears],
RealData, SimData, msg)
Data_out@vbK <- UpdateSlot('vbK', RealData, SimData, msg)
ObsPars$Kbias <- UpdateObs('vbK', ObsPars$Kbias, StockPars$Karray[, nyears],
RealData, SimData, msg)
Data_out@vbt0 <- UpdateSlot('vbt0', RealData, SimData, msg)
ObsPars$t0bias <- UpdateObs('vbt0', ObsPars$t0bias, StockPars$t0array[, nyears],
RealData, SimData, msg)
Data_out@CV_Mort <- UpdateSlot('CV_Mort', RealData, SimData, msg)
Data_out@CV_vbLinf <- UpdateSlot('CV_vbLinf', RealData, SimData, msg)
Data_out@CV_vbK <- UpdateSlot('CV_vbK', RealData, SimData, msg)
Data_out@CV_vbt0 <- UpdateSlot('CV_vbt0', RealData, SimData, msg)
Data_out@wla <- UpdateSlot('wla', RealData, SimData, msg)
Data_out@wlb <- UpdateSlot('wlb', RealData, SimData, msg)
Data_out@CV_wla <- UpdateSlot('CV_wla', RealData, SimData, msg)
Data_out@CV_wlb <- UpdateSlot('CV_wlb', RealData, SimData, msg)
Data_out@steep <- UpdateSlot('steep', RealData, SimData, msg)
Data_out@CV_steep <- UpdateSlot('CV_steep', RealData, SimData, msg)
ObsPars$hbias <- UpdateObs('steep', ObsPars$hbias, StockPars$hs,
RealData, SimData, msg)
Data_out@sigmaR <- UpdateSlot('sigmaR', RealData, SimData, msg)
Data_out@CV_sigmaR <- UpdateSlot('CV_sigmaR', RealData, SimData, msg)
# TODO ObsPars$sigmaRbias
Data_out@L50 <- UpdateSlot('L50', RealData, SimData, msg)
Data_out@CV_L50 <- UpdateSlot('CV_L50', RealData, SimData, msg)
ObsPars$lenMbias <- UpdateObs('L50', ObsPars$lenMbias, StockPars$L50,
RealData, SimData, msg)
Data_out@L95 <- UpdateSlot('L95', RealData, SimData, msg)
Data_out@LenCV <- UpdateSlot('LenCV', RealData, SimData, msg)
Data_out@LFC <- UpdateSlot('LFC', RealData, SimData, msg)
Data_out@CV_LFC <- UpdateSlot('CV_LFC', RealData, SimData, msg)
ObsPars$LFCbias <- UpdateObs('LFC', ObsPars$LFCbias, FleetPars$L5_y[,nyears],
RealData, SimData, msg)
Data_out@LFS <- UpdateSlot('LFS', RealData, SimData, msg)
Data_out@CV_LFS <- UpdateSlot('CV_LFS', RealData, SimData, msg)
ObsPars$LFSbias <- UpdateObs('LFS', ObsPars$LFSbias, FleetPars$LFS_y[,nyears],
RealData, SimData, msg)
# Data_out@Vmaxlen <- UpdateSlot('Vmaxlen', RealData, SimData, msg)
if (length(RealData@Year)>1) {
# check years
if (all(RealData@Year !=SimData@Year)) {
stop('`OM$cpars$Data@Year` does not match `SimData@Year`. \nAre `Fleet@nyears` and `Fleet@CurrentYr` correct?')
}
Data_out@Year <- RealData@Year
}
# ---- Update Catch ----
if (!all(is.na(RealData@Cat[1,]))) {
if (msg)
message('Updating Simulated Catch from `OM@cpars$Data@Cat`')
Data_out@Cat <- matrix(RealData@Cat[1,1:nyears], nrow=nsim, ncol=nyears, byrow=TRUE)
Data_out@CV_Cat <- matrix(RealData@CV_Cat[1,1:nyears], nrow=nsim, ncol=nyears, byrow=TRUE)
simcatch <- apply(StockPars$CBret, c(1,3), sum)
simcatch[simcatch==0] <- tiny
Cbias <- matrix(apply(Data_out@Cat, 1, mean)/apply(simcatch, 1, mean),
nrow=nsim, ncol=nyears+proyears)
Cerr <- Data_out@Cat/(simcatch*Cbias[,1:nyears])
t1<- Cerr[,max(nyears-10, 1):nyears]/apply(Cerr[,max(nyears-10, 1):nyears],1,mean)
SDs <- apply(log(t1), 1, sd)
Cerr_proj <- matrix(NA, nsim, proyears)
nas <- which(is.na(SDs))
Cerr_proj[nas,] <- tiny # no catch
if (!all(is.na(SDs))) {
for (i in (1:nsim)[!is.na(SDs)]) {
Cerr_proj[i,] <- exp(rnorm(proyears, -((SDs[i]^2)/2), SDs[i]))
}
}
Cerr <- cbind(Cerr, Cerr_proj)
if(!is.null(SampCpars$Cobs_y)) {
if (msg) message_info('Catch observation error detected in cpars (`cpars$Cobs_y`). Not updating catch obs error')
} else {
if(!is.null(SampCpars$Cbias)) {
if (msg) message_info('Catch Cbias detected in cpars (`cpars$Cbias`). Not updating catch bias')
} else {
if (msg) message_info('Updating Catch bias from `OM@cpars$Data@Cat`')
ObsPars$Cbias <- Cbias[,1]
}
if(!is.null(SampCpars$Cerr_y)) {
if (msg) message_info('Catch variability detected in cpars (`cpars$Cerr_y`). Not updating catch error')
} else {
if (msg) message_info('Updating Catch variability from `OM@cpars$Data@Cat`')
ObsPars$Cerr_y <- Cerr
}
if (msg) message_info('Updating catch observation error from `OM@cpars$Data@Cat`')
ObsPars$Cobs_y <- Cerr * Cbias
}
}
# ---- Update Effort -----
# TODO
# ---- Update Index (total biomass) ----
if (!all(is.na(RealData@Ind[1,]))) { # Index exists
fit_ind <- Fit_Index(ind_slot='Ind', indcv_slot="CV_Ind", Data_out,
RealData, StockPars, ObsPars, SampCpars, nsim, nyears, proyears, msg=msg)
Data_out <- fit_ind$Data_out
ObsPars <- fit_ind$ObsPars
}
# ---- Update Index (spawning biomass) ----
if (!all(is.na(RealData@SpInd[1,]))) { # Index exists
fit_ind <- Fit_Index(ind_slot='SpInd', indcv_slot="CV_SpInd", Data_out,
RealData, StockPars, ObsPars, SampCpars, nsim, nyears, proyears, msg=msg)
Data_out <- fit_ind$Data_out
ObsPars <- fit_ind$ObsPars
}
# ---- Index (vulnerable biomass) ----
if (!all(is.na(RealData@VInd[1,]))) { # Index exists
fit_ind <- Fit_Index(ind_slot='VInd', indcv_slot="CV_VInd", Data_out,
RealData, StockPars, ObsPars, SampCpars, nsim, nyears, proyears, msg=msg)
Data_out <- fit_ind$Data_out
ObsPars <- fit_ind$ObsPars
}
# ---- Add Additional Indices ----
if (!all(is.na(RealData@AddInd))) {
if (msg)
message('Adding Additional Indices to Simulated Data from `OM@cpars$Data@AddInd`')
n.ind <- dim(RealData@AddInd)[2]
Data_out@AddInd <- Data_out@CV_AddInd <- array(NA, dim=c(nsim, n.ind, nyears))
fitbeta <- fitIerr <- TRUE
if (!is.null(SampCpars$AddIbeta)) {
if (any(dim(SampCpars$AddIbeta) != c(nsim, n.ind)))
stop("cpars$AddIbeta must be dimensions c(nsim, n.ind)")
if (msg) message_info('cpars$AddIbeta detected. Not updating beta for additional indices')
ObsPars$AddIbeta <- SampCpars$AddIbeta
fitbeta <- FALSE
} else {
if (msg) message_info('Updating beta for additional indices from real data')
ObsPars$AddIbeta <- matrix(NA, nsim, n.ind)
}
if (!is.null(SampCpars$AddIerr)) {
if (any(dim(SampCpars$AddIerr) != c(nsim, n.ind, nyears+proyears)))
stop("cpars$AddIerr must be dimensions c(nsim, n.ind, nyears+proyears)")
if (msg) message_info('cpars$AddIerr detected. Not updating observation variability for additional indices')
ObsPars$AddIerr <- SampCpars$AddIerr
fitIerr <- FALSE
} else {
if (msg) message_info('Updating observation variability (AddIerr) for additional indices from real data')
ObsPars$AddIerr <- array(NA, dim=c(nsim, n.ind, nyears+proyears))
}
if (!is.null(SampCpars$AddIunits) && length(SampCpars$AddIunits) != n.ind)
stop("cpars$AddIunits must be length n.ind")
AddIunits <- RealData@AddIunits
if (all(is.na(AddIunits))) AddIunits <- rep(1, n.ind)
if (!is.null(SampCpars$AddIunits)) AddIunits <- SampCpars$AddIunits
Data_out@AddIunits <- AddIunits
AddIndType <- RealData@AddIndType
if (all(is.na(AddIndType))) AddIndType <- rep(1, n.ind)
Data_out@AddIndType <- AddIndType
ObsPars$AddInd_Stat <- vector(mode = "list", length = n.ind)
UnitsTab <- data.frame(n=1:0, units=c('biomass', 'numbers'))
TypeTab <- data.frame(n=1:3, type=c('total', 'spawning', 'vuln.'))
Data_out@AddIndV <- array(NA, dim=c(nsim, n.ind, Data_out@MaxAge+1))
# loop over additional indices
for (i in 1:n.ind) {
units <- UnitsTab$units[match(AddIunits[i], UnitsTab$n)]
type <- TypeTab$type[match(AddIndType[i], TypeTab$n)]
if(msg) message_info("Additional index", i, '-', type, 'stock', paste0('(', units, ')'))
nyrs <- min(length(RealData@AddInd[1,i,]), nyears)
ind <- RealData@AddInd[1,i,1:nyrs]
cv_ind <- RealData@CV_AddInd[1,i,1:nyrs]
if (nyrs < nyears) {
ind <- c(ind, rep(NA,nyears-nyrs))
cv_ind <- c(cv_ind, rep(NA,nyears-nyrs))
}
Data_out@AddInd[,i,] <- matrix(ind, nrow=nsim, ncol=nyears, byrow=TRUE)
Data_out@CV_AddInd[,i,] <- matrix(cv_ind, nrow=nsim, ncol=nyears, byrow=TRUE)
if (all(is.na(RealData@AddIndV[1,, ]))) {
# no vulnerability-at-age included
Ind_V <- rep(1, Data_out@MaxAge+1)
} else {
Ind_V <- RealData@AddIndV[1,i, ]
}
Data_out@AddIndV[,i,] <- t(replicate(nsim,Ind_V))
if (!all(is.na(ind))) {
# check dimensions
if (!length(Ind_V) == Data_out@MaxAge+1)
stop('Vulnerability-at-age for additional index ', i, ' is not length `maxage`+1' )
# calculate simulated index
if (AddIunits[i]) {
if (AddIndType[i]==1) SimIndex <- apply(StockPars$Biomass, c(1, 2, 3), sum) # Total Biomass-based index
if (AddIndType[i]==2) SimIndex <- apply(StockPars$SSB, c(1, 2, 3), sum) # Spawning Biomass-based index
if (AddIndType[i]==3) SimIndex <- apply(StockPars$VBiomass, c(1, 2, 3), sum) # vuln Biomass-based index
} else {
if (AddIndType[i]==1) SimIndex <- apply(StockPars$N, c(1, 2, 3), sum) # Total Abundance-based index
if (AddIndType[i]==2) SimIndex <- apply(StockPars$N, c(1, 2, 3), sum) * StockPars$Mat_age[,,1:nyears] # Spawning abundance-based index
if (AddIndType[i]==3) SimIndex <- apply(StockPars$N, c(1, 2, 3), sum) * FleetPars$V_real[,,1:nyears] # Vulnerable abundance-based index
}
Ind_V <- matrix(Ind_V, nrow=SimData@MaxAge+1, ncol= nyears)
Ind_V <- replicate(nsim, Ind_V) %>% aperm(., c(3,1,2))
# apply vulnerability-at-age and sum over age-classes
SimIndex <- apply(SimIndex*Ind_V, c(1,3), sum)
# Fit to observed index and generate residuals for projections
if (fitbeta) {
beta <- rep(NA, nsim)
} else {
beta <- ObsPars$AddIbeta[,i]
}
# Calculate residuals (with or without estimated beta)
Res_List <- lapply(1:nsim, function(x) Calc_Residuals(sim.index=SimIndex[x,],
obs.ind=ind,
beta=beta[x]))
lResids_Hist <- do.call('rbind', lapply(Res_List, '[[', 1))
if (fitbeta)
ObsPars$AddIbeta[,i] <- as.vector(do.call('cbind', lapply(Res_List, '[[', 2)))
if (msg & fitbeta)
message_info(paste0('Estimated beta for Additional Index ', i, ':'),
paste0(range(round(ObsPars$AddIbeta[,i],2)), collapse = "-"),
"Use `cpars$AddIbeta` to override")
# Calculate statistics
Stats_List <- lapply(1:nsim, function(x) Calc_Stats(lResids_Hist[x,]))
Stats <- do.call('rbind', Stats_List)
check_Index <- check_Index_Fit(Stats, i)
if (check_Index) {
# Generate residuals for projections
Resid_Hist <- exp(lResids_Hist) # historical residuals in normal space
Resid_Proj <- Gen_Residuals(Stats, nsim, proyears)
if (fitIerr) ObsPars$AddIerr[,i, ] <- cbind(Resid_Hist, Resid_Proj)
ObsPars$AddInd_Stat[[i]] <- Stats[,1:2] # index fit statistics
} else {
ObsPars$AddIerr[,i, ] <- ObsPars$Ierr_y
ObsPars$AddInd_Stat[[i]] <- Stats[,1:2] # index fit statistics
}
}
}
}
# ---- Update Recruitment ----
if (!all(is.na(RealData@Rec))) {
if (msg)
message('Updating Simulated Recruitment Data from `OM@cpars$Data@Rec`')
Data_out@Rec <- matrix(RealData@Rec[1,1:nyears], nrow=nsim, ncol=nyears, byrow=TRUE)
dd <- dim(RealData@CV_Rec)
if (dd[2]<nyears) {
RealData@CV_Rec <- matrix(RealData@CV_Rec[1,1], nrow=nsim, ncol=nyears, byrow=TRUE)
}
Data_out@CV_Rec <- matrix(RealData@CV_Rec[1,1:nyears], nrow=nsim, ncol=nyears, byrow=TRUE)
# Calculate Error
Rec <- apply(StockPars$N[, 1, , ], c(1, 2), sum) # simulated recruitment
Recbias <- matrix(apply(Data_out@Rec, 1, mean)/apply(Rec, 1, mean),
nrow=nsim, ncol=nyears+proyears)
Rec_err <- Data_out@Rec/(Rec*Recbias[,1:nyears])
t1 <- Rec_err[,max(nyears-10, 1):nyears]/apply(Rec_err[,max(nyears-10, 1):nyears],1,mean) # last 10 years used for projections
SDs <- apply(log(t1), 1, sd)
Rec_err_proj <- matrix(NA, nsim, proyears)
for (i in 1:nsim) {
Rec_err_proj[i,] <- exp(rnorm(proyears, -((SDs[i]^2)/2), SDs[i]))
}
Rec_err_proj <- Rec_err_proj * Recbias[,(nyears+1):(nyears+proyears)]
Rec_err <- cbind(Rec_err, Rec_err_proj)
ObsPars$Recerr_y <- Rec_err * Recbias
ObsPars$Recsd <- SDs
ObsPars$Recbias <- Recbias
}
# ---- Update CAA ----
if (!all(is.na(RealData@CAA)) & !all(RealData@CAA ==0)) {
if (msg)
message('Updating Simulated Catch-at-Age Data from `OM@cpars$Data@CAA`. Note: CAA_ESS is currently NOT updated')
Data_out@CAA <- aperm(replicate(nsim, RealData@CAA[1,1:nyears,]),c(3,1,2))
Data_out@Vuln_CAA <- RealData@Vuln_CAA
# Get average sample size
nsamp <- ceiling(mean(apply(RealData@CAA[1,1:nyears,], 1, sum)))
ObsPars$CAA_nsamp <- rep(nsamp, nsim)
}
# ---- Update CAL ----
if (!all(is.na(RealData@CAL)) & !all(RealData@CAL ==0)) {
# check length bins
if (!all(RealData@CAL_bins %in% StockPars$CAL_bins)) {
warning('cpars$Data@CAL_bins cannot be matched with Simulated Data@CAL_bins. Add cpars$Data@CAL_bins to cpars$CAL_bins. cpars$Data@CAL are NOT being used')
} else {
if (msg)
message('Updating Simulated Catch-at-Length Data, Obs@CAL_nsamp, and Obs@CAL_ESS from `OM@cpars$Data@CAL`')
# match length bins
ind <- match(RealData@CAL_mids, StockPars$CAL_binsmid)
dd <- dim(Data_out@CAL)
Data_out@CAL <- array(0, dim=dd)
CAL <- Data_out@CAL[1,,]
CAL[,ind] <- RealData@CAL[1,,] # replace with real CAL
CAL <- aperm(replicate(nsim, CAL[1:nyears,]),c(3,1,2))
Data_out@CAL <- CAL
Data_out@Vuln_CAL <- RealData@Vuln_CAL
# Get average sample size
nsamp <- ceiling(mean(apply(RealData@CAL[1,1:nyears,], 1, sum, na.rm=TRUE), na.rm=TRUE))
ObsPars$CAL_nsamp <- rep(nsamp, nsim)
# calculate effective sample size for projections
# vuln N-at-age in last year with length data
nas <- !apply(CAL[1,,], 2, is.na)
yr.ind <- max(which(apply(nas, 1, prod)>0))
if (!is.null(control$CAL) && control$CAL == 'removals') {
vn <- apply(StockPars$N[,,yr.ind,, drop=FALSE], c(1,2,3), sum) *
FleetPars$V_real[,,yr.ind, drop=FALSE]
# numbers at age in population that would be removed
vn <- aperm(vn, c(1,3, 2))
doopt <- optimise(optESS, c(10, 10000), vn, StockPars,
FleetPars$SLarray_real[1,,yr.ind, drop=FALSE], yr.ind, ObsPars, Data_out, CAL)
ObsPars$CAL_ESS <- rep(doopt$minimum, nsim)
} else {
vn <- apply(StockPars$N[,,yr.ind,, drop=FALSE], c(1,2,3), sum) *
FleetPars$retA_real[,,yr.ind, drop=FALSE]
# numbers at age in population that would be removed
vn <- aperm(vn, c(1,3, 2))
doopt <- optimise(optESS, c(10, 10000), vn, StockPars,
FleetPars$retL_real[1,,yr.ind, drop=FALSE], yr.ind, ObsPars, Data_out, CAL)
ObsPars$CAL_ESS <- rep(doopt$minimum, nsim)
}
}
}
# ---- Depletion ----
Data_out@Dep <- UpdateSlot('Dep', RealData, SimData, msg)
Data_out@CV_Dep <- UpdateSlot('CV_Dep', RealData, SimData, msg)
ObsPars$Dbias <- UpdateObs('Dep', ObsPars$Dbias, StockPars$Depletion,
RealData, SimData, msg)
# ---- Index Reference -----
Data_out@Iref <- UpdateSlot('Iref', RealData, SimData, msg)
Data_out@CV_Iref <- UpdateSlot('CV_Iref', RealData, SimData, msg)
# ObsPars$Irefbias <- rep(NA, nsim) # not calculated
# ---- Misc ----
if (length(RealData@Misc) && !is.null(names(RealData@Misc))) {
Data_out@Misc[names(RealData@Misc)] <- RealData@Misc[names(RealData@Misc)]
if (msg)
message('Adding named list from `OM@cpars$Data@Misc`')
}
NotUpdated <- function(RealData, sl, msg) {
if (!all(is.na(slot(RealData, sl)))) {
if (msg)
message_info(paste0('Data detected in `OM@cpars$Data@', sl, '` but is NOT being used.'))
}
}
NotUpdated(RealData, 'ML', msg)
NotUpdated(RealData, 'Lc', msg)
NotUpdated(RealData, 'Lbar', msg)
NotUpdated(RealData, 'Abun', msg)
NotUpdated(RealData, 'SpAbun', msg)
NotUpdated(RealData, 'FMSY_M', msg)
NotUpdated(RealData, 'BMSY_B0', msg)
NotUpdated(RealData, 'Cref', msg)
NotUpdated(RealData, 'Bref', msg)
NotUpdated(RealData, 'AvC', msg)
NotUpdated(RealData, 'Dt', msg)
NotUpdated(RealData, 'Ref', msg)
list(Data=Data_out, ObsPars=ObsPars)
}
optESS <- function(CAL_ESS, vn, StockPars, sel_ret, yr.ind, ObsPars, Data_out, CAL) {
set.seed(101)
tempSize <- genSizeCompWrap(1, vn, StockPars$CAL_binsmid, StockPars$CAL_bins,
sel_ret, CAL_ESS,
ObsPars$CAL_nsamp,
StockPars$Linfarray[1, yr.ind, drop=FALSE],
StockPars$Karray[1, yr.ind, drop=FALSE],
StockPars$t0array[1, yr.ind, drop=FALSE],StockPars$LenCV, truncSD=2)
tempSize <- tempSize/sum(tempSize)
obsSize <- Data_out@CAL[1,yr.ind,]/sum(Data_out@CAL[1,yr.ind,])
sum((obsSize-tempSize)^2)
}
AddRealData_MS <- function(SimData,
RealData,
StockPars,
FleetPars,
ObsPars,
SampCpars,
map.stocks,
CBret,
VBF,
p,
f,
nsim,
nyears,
proyears,
msg=TRUE,
control) {
Data_out <- SimData
if (msg)
message('Updating Simulated Data with Real Data from `OM@cpars$Data`')
# check last year
if (!is.na(RealData@LHYear) && !SimData@LHYear == RealData@LHYear) {
warning('`Fleet@CurrentYear` (', SimData@LHYear, ') is not the same as `OM@cpars$Data@LHYear` (', RealData@LHYear, ')')
}
# check maxage
if (!is.na(RealData@MaxAge) && !SimData@MaxAge == RealData@MaxAge) {
warning('`Stock@MaxAge` (', SimData@MaxAge, ') is not the same as `OM@cpars$Data@MaxAge` (', RealData@MaxAge, ')')
}
# check dimensions of real data CAA
if (!all(is.na(RealData@CAA)) && dim(RealData@CAA)[3] > 1) {
if (!dim(RealData@CAA)[3] == SimData@MaxAge+1)
stop('`OM@cpars$Data@CAA` has incorrect dimensions, should be `Stock@maxage+1` age-classes')
}
# update static slots
slts <- c('Name', 'Common_Name', 'Species', 'Region', 'LHYear', 'Units')
for (sl in slts)
slot(Data_out, sl) <- slot(RealData, sl)
Data_out@MaxAge <- SimData@MaxAge
if (!is.na(RealData@MPrec)) {
Data_out@MPrec <- rep(RealData@MPrec, nsim)
} else {
Data_out@MPrec <- SimData@MPrec
}
Data_out@MPeff <- SimData@MPeff
Data_out@nareas <- SimData@nareas
Data_out@LHYear <- SimData@LHYear
# ---- Update Life-history parameters ----
Data_out@Mort <- UpdateSlot('Mort', RealData, SimData, msg)
ObsPars[[p]][[f]]$Mbias <- UpdateObs('Mort', ObsPars[[p]][[f]]$Mbias, StockPars[[p]]$Marray[, nyears],
RealData, SimData, msg)
Data_out@vbLinf <- UpdateSlot('vbLinf', RealData, SimData, msg)
ObsPars[[p]][[f]]$Linfbias <- UpdateObs('vbLinf', ObsPars[[p]][[f]]$Linfbias, StockPars[[p]]$Linfarray[, nyears],
RealData, SimData, msg)
Data_out@vbK <- UpdateSlot('vbK', RealData, SimData, msg)
ObsPars[[p]][[f]]$Kbias <- UpdateObs('vbK', ObsPars[[p]][[f]]$Kbias, StockPars[[p]]$Karray[, nyears],
RealData, SimData, msg)
Data_out@vbt0 <- UpdateSlot('vbt0', RealData, SimData, msg)
ObsPars[[p]][[f]]$t0bias <- UpdateObs('vbt0', ObsPars[[p]][[f]]$t0bias, StockPars[[p]]$t0array[, nyears],
RealData, SimData, msg)
Data_out@CV_Mort <- UpdateSlot('CV_Mort', RealData, SimData, msg)
Data_out@CV_vbLinf <- UpdateSlot('CV_vbLinf', RealData, SimData, msg)
Data_out@CV_vbK <- UpdateSlot('CV_vbK', RealData, SimData, msg)
Data_out@CV_vbt0 <- UpdateSlot('CV_vbt0', RealData, SimData, msg)
Data_out@wla <- UpdateSlot('wla', RealData, SimData, msg)
Data_out@wlb <- UpdateSlot('wlb', RealData, SimData, msg)
Data_out@CV_wla <- UpdateSlot('CV_wla', RealData, SimData, msg)
Data_out@CV_wlb <- UpdateSlot('CV_wlb', RealData, SimData, msg)
Data_out@steep <- UpdateSlot('steep', RealData, SimData, msg)
Data_out@CV_steep <- UpdateSlot('CV_steep', RealData, SimData, msg)
ObsPars[[p]][[f]]$hbias <- UpdateObs('steep', ObsPars[[p]][[f]]$hbias, StockPars[[p]]$hs,
RealData, SimData, msg)
Data_out@sigmaR <- UpdateSlot('sigmaR', RealData, SimData, msg)
Data_out@CV_sigmaR <- UpdateSlot('CV_sigmaR', RealData, SimData, msg)
# TODO ObsPars$sigmaRbias
Data_out@L50 <- UpdateSlot('L50', RealData, SimData, msg)
Data_out@CV_L50 <- UpdateSlot('CV_L50', RealData, SimData, msg)
ObsPars[[p]][[f]]$lenMbias <- UpdateObs('L50', ObsPars[[p]][[f]]$lenMbias, StockPars[[p]]$L50,
RealData, SimData, msg)
Data_out@L95 <- UpdateSlot('L95', RealData, SimData, msg)
Data_out@LenCV <- UpdateSlot('LenCV', RealData, SimData, msg)
Data_out@LFC <- UpdateSlot('LFC', RealData, SimData, msg)
Data_out@CV_LFC <- UpdateSlot('CV_LFC', RealData, SimData, msg)
ObsPars[[p]][[f]]$LFCbias <- UpdateObs('LFC', ObsPars[[p]][[f]]$LFCbias, FleetPars[[p]][[f]]$L5_y[,nyears],
RealData, SimData, msg)
Data_out@LFS <- UpdateSlot('LFS', RealData, SimData, msg)
Data_out@CV_LFS <- UpdateSlot('CV_LFS', RealData, SimData, msg)
ObsPars[[p]][[f]]$LFSbias <- UpdateObs('LFS', ObsPars[[p]][[f]]$LFSbias, FleetPars[[p]][[f]]$LFS_y[,nyears],
RealData, SimData, msg)
if (length(RealData@Year)>1) {
# check years
# if (!all(RealData@Year %in%SimData@Year)) {
# stop('`OM$cpars$Data@Year` does not match `SimData@Year`. \nAre `Fleet@nyears` and `Fleet@CurrentYr` correct?')
# }
Data_out@Year <- RealData@Year
}
# ---- Update Catch ----
if (!all(is.na(RealData@Cat[1,]))) {
if (msg)
message('Updating Simulated Catch from `OM@cpars$Data@Cat`')
Data_out@Cat <- matrix(RealData@Cat[1,1:nyears], nrow=nsim, ncol=nyears, byrow=TRUE)
Data_out@CV_Cat <- matrix(RealData@CV_Cat[1,1:nyears], nrow=nsim, ncol=nyears, byrow=TRUE)
simcatch <- apply(CBret[,map.stocks,f,,, ,drop=FALSE], c(1,5), sum)
simcatch[simcatch==0] <- tiny
Cbias <- matrix(apply(Data_out@Cat, 1, mean)/apply(simcatch, 1, mean),
nrow=nsim, ncol=nyears+proyears)
Cerr <- Data_out@Cat/(simcatch*Cbias[,1:nyears])
t1<- Cerr[,max(nyears-10, 1):nyears]/apply(Cerr[,max(nyears-10, 1):nyears],1,mean)
SDs <- apply(log(t1), 1, sd)
Cerr_proj <- matrix(NA, nsim, proyears)
nas <- which(is.na(SDs))
Cerr_proj[nas,] <- tiny # no catch
if (!all(is.na(SDs))) {
for (i in (1:nsim)[!is.na(SDs)]) {
Cerr_proj[i,] <- exp(rnorm(proyears, -((SDs[i]^2)/2), SDs[i]))
}
}
Cerr <- cbind(Cerr, Cerr_proj)
if(!is.null(SampCpars[[p]][[f]]$Cobs_y)) {
if (msg) message_info('Catch observation error detected in cpars (`cpars$Cobs_y`). Not updating catch obs error')
} else {
if(!is.null(SampCpars[[p]][[f]]$Cbias)) {
if (msg) message_info('Catch Cbias detected in cpars (`cpars$Cbias`). Not updating catch bias')
} else {
if (msg) message_info('Updating Catch bias from `OM@cpars$Data@Cat`')
ObsPars[[p]][[f]]$Cbias <- Cbias[,1]
}
if(!is.null(SampCpars[[p]][[f]]$Cerr_y)) {
if (msg) message_info('Catch variability detected in cpars (`cpars$Cerr_y`). Not updating catch error')
} else {
if (msg) message_info('Updating Catch variability from `OM@cpars$Data@Cat`')
ObsPars[[p]][[f]]$Cerr_y <- Cerr
}
if (msg) message_info('Updating catch observation error from `OM@cpars$Data@Cat`')
ObsPars[[p]][[f]]$Cobs_y <- Cerr * Cbias
}
}
# ---- Update Effort -----
# TODO
# ---- Update Index (total biomass) ----
if (!all(is.na(RealData@Ind[1,]))) { # Index exists
fit_ind <- Fit_Index_MS(ind_slot='Ind', indcv_slot="CV_Ind", Data_out,
RealData, StockPars, ObsPars, SampCpars, nsim, nyears, proyears,
map.stocks,
p, f,
msg=msg)
Data_out <- fit_ind$Data_out
# add future year data if they exist
n_future_years <- length(RealData@Ind[1,]) - length(Data_out@Ind[1,])
if (n_future_years>0) {
ll <- length(RealData@Ind[1,])
future_index <- matrix(RealData@Ind[1,(ll-n_future_years+1):ll], nsim, n_future_years, byrow=TRUE)
future_index_cv <- matrix(RealData@CV_Ind[1,(ll-n_future_years+1):ll], nsim, n_future_years, byrow=TRUE)
Data_out@Ind <- cbind(Data_out@Ind, future_index)
Data_out@CV_Ind <- cbind(Data_out@CV_Ind, future_index_cv)
}
ObsPars <- fit_ind$ObsPars
}
# ---- Update Index (spawning biomass) ----
if (!all(is.na(RealData@SpInd[1,]))) { # Index exists
fit_ind <- Fit_Index_MS(ind_slot='SpInd', indcv_slot="CV_SpInd", Data_out,
RealData, StockPars, ObsPars, SampCpars, nsim, nyears, proyears,
map.stocks,
p, f,
msg=msg)
Data_out <- fit_ind$Data_out
ObsPars <- fit_ind$ObsPars
}
# ---- Index (vulnerable biomass) ----
if (!all(is.na(RealData@VInd[1,]))) { # Index exists
StockPars[[p]]$VBiomass <- VBF[,,f,,,]
fit_ind <- Fit_Index_MS(ind_slot='VInd', indcv_slot="CV_VInd", Data_out,
RealData, StockPars, ObsPars, SampCpars, nsim, nyears, proyears,
map.stocks,
p, f,
msg=msg)
Data_out <- fit_ind$Data_out
ObsPars <- fit_ind$ObsPars
fit_ind$ObsPars[[1]][[1]]$VIerr_y[,1:69]
}
# ---- Add Additional Indices ----
if (!all(is.na(RealData@AddInd))) {
if (msg)
message('Adding Additional Indices to Simulated Data from `OM@cpars$Data@AddInd`')
n.ind <- dim(RealData@AddInd)[2]
Data_out@AddInd <- Data_out@CV_AddInd <- array(NA, dim=c(nsim, n.ind, nyears))
fitbeta <- fitIerr <- TRUE
if (!is.null(SampCpars[[p]][[f]]$AddIbeta)) {
if (any(dim(SampCpars[[p]][[f]]$AddIbeta) != c(nsim, n.ind)))
stop("cpars$AddIbeta must be dimensions c(nsim, n.ind)")
if (msg) message_info('cpars$AddIbeta detected. Not updating beta for additional indices')
ObsPars[[p]][[f]]$AddIbeta <- SampCpars[[p]][[f]]$AddIbeta
fitbeta <- FALSE
} else {
if (msg) message_info('Updating beta for additional indices from real data')
ObsPars[[p]][[f]]$AddIbeta <- matrix(NA, nsim, n.ind)
}
if (!is.null(SampCpars[[p]][[f]]$AddIerr)) {
if (any(dim(SampCpars[[p]][[f]]$AddIerr) != c(nsim, n.ind, nyears+proyears)))
stop("cpars$AddIerr must be dimensions c(nsim, n.ind, nyears+proyears)")
if (msg) message_info('cpars$AddIerr detected. Not updating observation variability for additional indices')
ObsPars[[p]][[f]]$AddIerr <- SampCpars[[p]][[f]]$AddIerr
fitIerr <- FALSE
} else {
if (msg) message_info('Updating observation variability (AddIerr) for additional indices from real data')
ObsPars[[p]][[f]]$AddIerr <- array(NA, dim=c(nsim, n.ind, nyears+proyears))
}
if (!is.null(SampCpars$AddIunits) && length(SampCpars$AddIunits) != n.ind)
stop("cpars$AddIunits must be length n.ind")
AddIunits <- RealData@AddIunits
if (all(is.na(AddIunits))) AddIunits <- rep(1, n.ind)
if (!is.null(SampCpars[[p]][[f]]$AddIunits)) AddIunits <- SampCpars[[p]][[f]]$AddIunits
Data_out@AddIunits <- AddIunits
AddIndType <- RealData@AddIndType
if (all(is.na(AddIndType))) AddIndType <- rep(1, n.ind)
Data_out@AddIndType <- AddIndType
ObsPars[[p]][[f]]$AddInd_Stat <- vector(mode = "list", length = n.ind)
UnitsTab <- data.frame(n=1:0, units=c('biomass', 'numbers'))
TypeTab <- data.frame(n=1:3, type=c('total', 'spawning', 'vuln.'))
Data_out@AddIndV <- array(NA, dim=c(nsim, n.ind, Data_out@MaxAge+1))
# loop over additional indices
for (i in 1:n.ind) {
units <- UnitsTab$units[match(AddIunits[i], UnitsTab$n)]
type <- TypeTab$type[match(AddIndType[i], TypeTab$n)]
if(msg) message_info("Additional index", i, '-', type, 'stock', paste0('(', units, ')'))
nyrs <- min(length(RealData@AddInd[1,i,]), nyears)
ind <- RealData@AddInd[1,i,1:nyrs]
cv_ind <- RealData@CV_AddInd[1,i,1:nyrs]
if (nyrs < nyears) {
ind <- c(ind, rep(NA,nyears-nyrs))
cv_ind <- c(cv_ind, rep(NA,nyears-nyrs))
}
Data_out@AddInd[,i,] <- matrix(ind, nrow=nsim, ncol=nyears, byrow=TRUE)
Data_out@CV_AddInd[,i,] <- matrix(cv_ind, nrow=nsim, ncol=nyears, byrow=TRUE)
if (all(is.na(RealData@AddIndV[1,, ]))) {
# no vulnerability-at-age included
Ind_V <- rep(1, Data_out@MaxAge+1)
} else {
Ind_V <- RealData@AddIndV[1,i, ]
}
Data_out@AddIndV[,i,] <- t(replicate(nsim,Ind_V))
if (!all(is.na(ind))) {
# check dimensions
if (!length(Ind_V) == Data_out@MaxAge+1)
stop('Vulnerability-at-age for additional index ', i, ' is not length `maxage`+1' )
Ind_V <- matrix(Ind_V, nrow=SimData@MaxAge+1, ncol= nyears)
Ind_V <- replicate(nsim, Ind_V) %>% aperm(., c(3,1,2))
Ind_V_list <- list()
ObsPars[[p]][[f]]$AddIV <- SampCpars[[p]][[f]]$AddIV
for (pp in map.stocks) {
if (!is.null(SampCpars[[pp]][[f]]$AddIV)) {
Ind_V_list[[pp]] <- SampCpars[[pp]][[f]]$AddIV[,,i,1:nyears]
} else {
Ind_V_list[[pp]] <- Ind_V
}
}
# calculate simulated index
if (AddIunits[i]) {
if (AddIndType[i]==1) SimIndex <- apply(StockPars[[p]]$Biomass[,map.stocks,,,,drop=FALSE], c(1, 2, 3, 4), sum) # Total Biomass-based index
} else {
if (AddIndType[i]==1) SimIndex <- apply(StockPars[[p]]$N[,map.stocks,,,,drop=FALSE], c(1, 2, 3, 4), sum) # Total Abundance-based index
}
if (AddIndType[i]!=1)
stop('Vulnerable and spawning indices not supported for multiMSE. Use `cpars$AddIV` to specify selectivity pattern')
for (pp in seq_along(map.stocks)) {
SimIndex[,pp,,] <- SimIndex[,pp,,] * Ind_V_list[[map.stocks[pp]]]
}
SimIndex <- apply(SimIndex, c(1,4), sum)
# Fit to observed index and generate residuals for projections
if (fitbeta) {
beta <- rep(NA, nsim)
} else {
beta <- ObsPars[[p]][[f]]$AddIbeta[,i]
}
# Calculate residuals (with or without estimated beta)
Res_List <- lapply(1:nsim, function(x) Calc_Residuals(sim.index=SimIndex[x,],
obs.ind=ind,
beta=beta[x]))
lResids_Hist <- do.call('rbind', lapply(Res_List, '[[', 1))
if (fitbeta)
ObsPars[[p]][[f]]$AddIbeta[,i] <- as.vector(do.call('cbind', lapply(Res_List, '[[', 2)))
if (msg & fitbeta)
message_info(paste0('Estimated beta for Additional Index ', i, ':'),
paste0(range(round(ObsPars[[p]][[f]]$AddIbeta[,i],2)), collapse = "-"),
"Use `cpars$AddIbeta` to override")
# Calculate statistics
Stats_List <- lapply(1:nsim, function(x) Calc_Stats(lResids_Hist[x,]))
Stats <- do.call('rbind', Stats_List)
check_Index <- check_Index_Fit(Stats, i)
if (check_Index) {
# Generate residuals for projections
Resid_Hist <- exp(lResids_Hist) # historical residuals in normal space
Resid_Proj <- Gen_Residuals(Stats, nsim, proyears)
if (fitIerr) ObsPars[[p]][[f]]$AddIerr[,i, ] <- cbind(Resid_Hist, Resid_Proj)
ObsPars[[p]][[f]]$AddInd_Stat[[i]] <- Stats[,1:2] # index fit statistics
} else {
ObsPars[[p]][[f]]$AddIerr[,i, ] <- ObsPars[[p]][[f]]$Ierr_y
ObsPars[[p]][[f]]$AddInd_Stat[[i]] <- Stats[,1:2] # index fit statistics
}
}
}
}
# ---- Update Recruitment ----
if (!all(is.na(RealData@Rec))) {
if (msg)
message('Updating Simulated Recruitment Data from `OM@cpars$Data@Rec`')
Data_out@Rec <- matrix(RealData@Rec[1,1:nyears], nrow=nsim, ncol=nyears, byrow=TRUE)
dd <- dim(RealData@CV_Rec)
if (dd[2]<nyears) {
RealData@CV_Rec <- matrix(RealData@CV_Rec[1,1], nrow=nsim, ncol=nyears, byrow=TRUE)
}
Data_out@CV_Rec <- matrix(RealData@CV_Rec[1,1:nyears], nrow=nsim, ncol=nyears, byrow=TRUE)
# Calculate Error
Rec <- apply(StockPars[[p]]$N[, map.stocks, 1, , ], c(1, 3), sum) # simulated recruitment
Recbias <- matrix(apply(Data_out@Rec, 1, mean)/apply(Rec, 1, mean),
nrow=nsim, ncol=nyears+proyears)
Rec_err <- Data_out@Rec/(Rec*Recbias[,1:nyears])
t1 <- Rec_err[,max(nyears-10, 1):nyears]/apply(Rec_err[,max(nyears-10, 1):nyears],1,mean) # last 10 years used for projections
SDs <- apply(log(t1), 1, sd)
Rec_err_proj <- matrix(NA, nsim, proyears)
for (i in 1:nsim) {
Rec_err_proj[i,] <- exp(rnorm(proyears, -((SDs[i]^2)/2), SDs[i]))
}
Rec_err_proj <- Rec_err_proj * Recbias[,(nyears+1):(nyears+proyears)]
Rec_err <- cbind(Rec_err, Rec_err_proj)
ObsPars[[p]][[f]]$Recerr_y <- Rec_err * Recbias
ObsPars[[p]][[f]]$Recsd <- SDs
ObsPars[[p]][[f]]$Recbias <- Recbias
}
# ---- Update CAA ----
if (!all(is.na(RealData@CAA)) & !all(RealData@CAA ==0)) {
if (msg)
message('Updating Simulated Catch-at-Age Data from `OM@cpars$Data@CAA`. Note: CAA_ESS is currently NOT updated')
Data_out@CAA <- aperm(replicate(nsim, RealData@CAA[1,1:nyears,]),c(3,1,2))
Data_out@Vuln_CAA <- RealData@Vuln_CAA
# Get average sample size
nsamp <- ceiling(mean(apply(RealData@CAA[1,1:nyears,], 1, sum)))
ObsPars[[p]][[f]]$CAA_nsamp <- rep(nsamp, nsim)
}
# ---- Update CAL ----
if (!all(is.na(RealData@CAL)) & !all(RealData@CAL ==0)) {
# check length bins
if (!all(RealData@CAL_bins %in% StockPars[[p]]$CAL_bins)) {
warning('cpars$Data@CAL_bins cannot be matched with Simulated Data@CAL_bins. Add cpars$Data@CAL_bins to cpars$CAL_bins. cpars$Data@CAL are NOT being used')
} else {
if (msg)
message('Updating Simulated Catch-at-Length Data, Obs@CAL_nsamp, and Obs@CAL_ESS from `OM@cpars$Data@CAL`')
# match length bins
ind <- match(RealData@CAL_mids, StockPars[[p]]$CAL_binsmid)
dd <- dim(Data_out@CAL)
dd_real <- dim(RealData@CAL[1,,])
dd[2] <- dd_real[1]
Data_out@CAL <- array(0, dim=dd)
CAL <- Data_out@CAL[1,,]
CAL[,ind] <- RealData@CAL[1,,] # replace with real CAL
n_dat_years <- dim(RealData@CAL[1,,])[1]
CAL <- aperm(replicate(nsim, CAL[1:n_dat_years,]),c(3,1,2))
Data_out@CAL <- CAL
Data_out@Vuln_CAL <- RealData@Vuln_CAL
# Get average sample size
nsamp <- ceiling(mean(apply(RealData@CAL[1,1:n_dat_years,], 1, sum, na.rm=TRUE), na.rm=TRUE))
ObsPars[[p]][[f]]$CAL_nsamp <- rep(nsamp, nsim)
# calculate effective sample size for projections
# vuln N-at-age in last year with length data
nas <- !apply(CAL[1,,], 2, is.na)
yr.ind <- max(which(apply(nas, 1, prod)>0))
if (!is.null(control$CAL) && control$CAL == 'removals') {
vn <- apply(StockPars[[p]]$N[,,yr.ind,, drop=FALSE], c(1,2,3), sum) *
FleetPars[[p]][[f]]$V_real[,,yr.ind, drop=FALSE]
# numbers at age in population that would be removed
vn <- aperm(vn, c(1,3, 2))
doopt <- optimise(optESS, c(10, 10000), vn, StockPars,
FleetPars$SLarray_real[1,,yr.ind, drop=FALSE], yr.ind, ObsPars, Data_out, CAL)
ObsPars$CAL_ESS <- rep(doopt$minimum, nsim)
} else {
vn <- apply(StockPars[[p]]$N[,map.stocks,,yr.ind,, drop=FALSE], c(1,3,4), sum) *
FleetPars[[p]][[f]]$retA_real[,,yr.ind, drop=FALSE]
# numbers at age in population that would be removed
vn <- aperm(vn, c(1,3, 2))
doopt <- optimise(optESS, c(10, 10000), vn, StockPars[[p]],
FleetPars[[p]][[f]]$retL_real[1,,yr.ind, drop=FALSE], yr.ind, ObsPars[[p]][[f]], Data_out, CAL)
ObsPars[[p]][[f]]$CAL_ESS <- rep(doopt$minimum, nsim)
}
}
}
# ---- Depletion ----
Data_out@Dep <- UpdateSlot('Dep', RealData, SimData, msg)
Data_out@CV_Dep <- UpdateSlot('CV_Dep', RealData, SimData, msg)
ObsPars[[p]][[f]]$Dbias <- UpdateObs('Dep', ObsPars$Dbias, StockPars$Depletion,
RealData, SimData, msg)
# ---- Index Reference -----
Data_out@Iref <- UpdateSlot('Iref', RealData, SimData, msg)
Data_out@CV_Iref <- UpdateSlot('CV_Iref', RealData, SimData, msg)
# ObsPars$Irefbias <- rep(NA, nsim) # not calculated
# ---- Misc ----
if (length(RealData@Misc) && !is.null(names(RealData@Misc))) {
Data_out@Misc[names(RealData@Misc)] <- RealData@Misc[names(RealData@Misc)]
if (msg)
message('Adding named list from `OM@cpars$Data@Misc`')
}
NotUpdated <- function(RealData, sl, msg) {
if (!all(is.na(slot(RealData, sl)))) {
if (msg)
message_info(paste0('Data detected in `OM@cpars$Data@', sl, '` but is NOT being used.'))
}
}
NotUpdated(RealData, 'ML', msg)
NotUpdated(RealData, 'Lc', msg)
NotUpdated(RealData, 'Lbar', msg)
NotUpdated(RealData, 'Abun', msg)
NotUpdated(RealData, 'SpAbun', msg)
NotUpdated(RealData, 'FMSY_M', msg)
NotUpdated(RealData, 'BMSY_B0', msg)
NotUpdated(RealData, 'Cref', msg)
NotUpdated(RealData, 'Bref', msg)
NotUpdated(RealData, 'AvC', msg)
NotUpdated(RealData, 'Dt', msg)
NotUpdated(RealData, 'Ref', msg)
list(Data=Data_out, ObsPars=ObsPars)
}
updateData_MS <- function(Data, OM, MPCalcs, Effort, Biomass, N, Biomass_P, CB_Pret,
N_P, SSB, SSB_P, VBiomass, VBiomass_P,
StockPars, FleetPars, ObsPars, ImpPars,
upyrs, interval, y=2, mm=1, Misc, RealData,
p, f, map.stocks) {
yind <- upyrs[match(y, upyrs) - 1]:(upyrs[match(y, upyrs)] - 1) # index
nyears <- OM@nyears
proyears <- OM@proyears
nsim <- OM@nsim
nareas <- StockPars[[p]]$nareas
reps <- OM@reps
Data@Year <- (OM@CurrentYr - nyears+1):(OM@CurrentYr+ y - 1)
Data@t <- rep(nyears + y, nsim)
# --- Simulate catches ----
CBtemp <- CB_Pret[, map.stocks,f, , yind, , drop=FALSE]
CBtemp[is.na(CBtemp)] <- tiny
CBtemp[!is.finite(CBtemp)] <- tiny
yr.index <- max(which(!is.na(Data@CV_Cat[1,])))
newCV_Cat <- matrix(Data@CV_Cat[,yr.index], nrow=nsim, ncol=length(yind))
Data@CV_Cat <- cbind(Data@CV_Cat, newCV_Cat)
# --- Observed catch ----
# Simulated observed retained catch (biomass)
Cobs <- ObsPars[[p]][[f]]$Cobs_y[,nyears + yind] * apply(CBtemp, c(1,5), sum, na.rm = TRUE)
Data@Cat <- cbind(Data@Cat, Cobs)
if (!is.null(RealData) && ncol(RealData@Cat)>nyears &&
!all(is.na(RealData@Cat[1,(nyears+1):length(RealData@Cat[1,])]))) {
# update projection catches with observed catches
addYr <- min(y,ncol(RealData@Cat) - nyears)
Data@Cat[,(nyears+1):(nyears+addYr)] <- matrix(RealData@Cat[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
Data@CV_Cat[,(nyears+1):(nyears+addYr)] <- matrix(RealData@CV_Cat[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
}
# --- Index of total abundance ----
yr.ind <- max(which(!is.na(ObsPars[[p]][[f]]$Ierr_y[1,1:nyears])))
I2 <- cbind(apply(Biomass[,map.stocks,,,,drop=FALSE], c(1, 4), sum)[,yr.ind:nyears],
apply(Biomass_P[,map.stocks,,,,drop=FALSE], c(1, 4), sum)[, 1:(y - 1)])
# standardize, apply beta & obs error
I2 <- exp(lcs(I2))^ObsPars[[p]][[f]]$I_beta * ObsPars[[p]][[f]]$Ierr_y[,yr.ind:(nyears + (y - 1))]
# I2 <- exp(lcs(I2)) * ObsPars$Ierr_y[,yr.ind:(nyears + (y - 1))]
if (sum(Data@Ind[1,1:nyears], na.rm = TRUE)) {
year.ind <- max(which(!is.na(Data@Ind[1,1:nyears])))
scaler <- Data@Ind[,year.ind]/I2[,1]
} else {
scaler <- rep(1, nsim)
}
scaler <- matrix(scaler, nrow=nsim, ncol=ncol(I2))
I2 <- I2 * scaler # convert back to historical index scale
I2 <- cbind(Data@Ind[,1:(yr.ind)], I2[,2:ncol(I2)])
Data@Ind <- I2
yr.index <- max(which(!is.na(Data@CV_Ind[1,1:nyears])))
newCV_Ind <- matrix(Data@CV_Ind[,yr.index], nrow=nsim, ncol=length(yind))
Data@CV_Ind <- cbind(Data@CV_Ind, newCV_Ind)
if (!is.null(RealData) && ncol(RealData@Ind)>nyears &&
!all(is.na(RealData@Ind[1,(nyears+1):length(RealData@Ind[1,])]))) {
# update projection index with observed index if it exists
dd <- length(RealData@Ind[1,])
p_yr_ind <- (nyears+1):(nyears+y-1)
p_yr_ind <- p_yr_ind[p_yr_ind<=dd]
p_ind <- matrix(RealData@Ind[1,p_yr_ind],
nrow=nsim, ncol=length(p_yr_ind), byrow=TRUE)
p_cv <- matrix(RealData@CV_Ind[1,p_yr_ind],
nrow=nsim, ncol=length(p_yr_ind), byrow=TRUE)
Data@Ind[,p_yr_ind] <- p_ind
Data@CV_Ind[,p_yr_ind] <- p_cv
}
# --- Index of spawning abundance ----
yr.ind <- max(which(!is.na(ObsPars[[p]][[f]]$SpIerr_y[1,1:nyears])))
I2 <- cbind(apply(SSB[,map.stocks,,,,drop=FALSE], c(1, 4), sum)[,yr.ind:nyears],
apply(SSB_P[,map.stocks,,,,drop=FALSE], c(1, 4), sum)[, 1:(y - 1)])
# standardize, apply beta & obs error
I2 <- exp(lcs(I2))^ObsPars[[p]][[f]]$SpI_beta * ObsPars[[p]][[f]]$SpIerr_y[,yr.ind:(nyears + (y - 1))]
if (sum(Data@SpInd[1,1:nyears], na.rm = TRUE)) {
year.ind <- max(which(!is.na(Data@SpInd[1,1:nyears])))
scaler <- Data@SpInd[,year.ind]/I2[,1]
} else {
scaler <- rep(1, nsim)
}
scaler <- matrix(scaler, nrow=nsim, ncol=ncol(I2))
I2 <- I2 * scaler # convert back to historical index scale
I2 <- cbind(Data@SpInd[,1:(yr.ind)], I2[,2:ncol(I2)])
Data@SpInd <- I2
yr.index <- max(which(!is.na(Data@CV_SpInd[1,1:nyears])))
newCV_Ind <- matrix(Data@CV_SpInd[,yr.index], nrow=nsim, ncol=length(yind))
Data@CV_SpInd <- cbind(Data@CV_SpInd, newCV_Ind)
if (!is.null(RealData) && ncol(RealData@SpInd)>nyears &&
!all(is.na(RealData@SpInd[1,(nyears+1):length(RealData@SpInd[1,])]))) {
# update projection index with observed index if it exists
addYr <- min(y,ncol(RealData@SpInd) - nyears)
Data@SpInd[,(nyears+1):(nyears+addYr)] <- matrix(RealData@SpInd[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
Data@CV_SpInd[,(nyears+1):(nyears+addYr)] <- matrix(RealData@CV_SpInd[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
}
# --- Index of vulnerable abundance ----
I2 <- cbind(apply(VBiomass[,map.stocks,f,,,,drop=FALSE], c(1, 5), sum)[,yr.ind:nyears],
apply(VBiomass_P[,map.stocks,f,,,,drop=FALSE], c(1, 5), sum)[, 1:(y - 1)])
# standardize, apply beta & obs error
I2 <- exp(lcs(I2))^ObsPars[[p]][[f]]$VI_beta * ObsPars[[p]][[f]]$VIerr_y[,yr.ind:(nyears + (y - 1))]
if (sum(Data@VInd[1,1:nyears], na.rm = TRUE)) {
year.ind <- max(which(!is.na(Data@VInd[1,1:nyears])))
scaler <- Data@VInd[,year.ind]/I2[,1]
} else {
scaler <- rep(1, nsim)
}
scaler <- matrix(scaler, nrow=nsim, ncol=ncol(I2))
I2 <- I2 * scaler # convert back to historical index scale
I2 <- cbind(Data@VInd[,1:(yr.ind)], I2[,2:ncol(I2)])
Data@VInd <- I2
yr.index <- max(which(!is.na(Data@CV_VInd[1,1:nyears])))
newCV_Ind <- matrix(Data@CV_VInd[,yr.index], nrow=nsim, ncol=length(yind))
Data@CV_VInd <- cbind(Data@CV_VInd, newCV_Ind)
if (!is.null(RealData) && ncol(RealData@VInd)>nyears &&
!all(is.na(RealData@VInd[1,(nyears+1):length(RealData@VInd[1,])]))) {
# update projection index with observed index if it exists
addYr <- min(y,ncol(RealData@VInd) - nyears)
Data@VInd[,(nyears+1):(nyears+addYr)] <- matrix(RealData@VInd[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
Data@CV_VInd[,(nyears+1):(nyears+addYr)] <- matrix(RealData@CV_VInd[1,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
}
# --- Update additional indices (if they exist) ----
AddIunits <- Data@AddIunits
AddIndType <- Data@AddIndType
if (length(ObsPars[[p]][[f]]$AddIerr)>0) {
n.ind <- dim(ObsPars[[p]][[f]]$AddIerr)[2]
AddInd <- array(NA, dim=c(nsim, n.ind, nyears+y-1))
CV_AddInd <- array(NA, dim=c(nsim, n.ind, nyears+y-1))
for (i in 1:n.ind) {
if (all(is.na(RealData@AddIndV[1, , ]))) {
Ind_V <- rep(1, Data@MaxAge+1)
} else {
Ind_V <- RealData@AddIndV[1,i, ]
}
Ind_V <- matrix(Ind_V, nrow=Data@MaxAge+1, ncol= nyears+proyears)
Ind_V <- replicate(nsim, Ind_V) %>% aperm(c(3,1,2))
Ind_V_list <- list()
for (pp in map.stocks) {
if (!is.null(ObsPars[[pp]][[f]]$AddIV)) {
Ind_V_list[[pp]] <- ObsPars[[pp]][[f]]$AddIV[,,i,]
} else {
Ind_V_list[[pp]] <- Ind_V
}
}
nas <- which(!is.na(ObsPars[[p]][[f]]$AddIerr[1,i, 1:nyears]))
if (length(nas)>0) {
yr.ind <- max(nas)
if (AddIunits[i]) { # Biomass-based index
if (AddIndType[i]==1) {
# total biomass
b1 <- apply(Biomass[,,,yr.ind:nyears, ,drop=FALSE], c(1:4), sum)
b2 <- apply(Biomass_P, c(1:4), sum)
}
#
# if (AddIndType[i]==2) {
# # spawning biomass
# b1 <- apply(SSB[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum)
# b2 <- apply(SSB_P, c(1, 2, 3), sum)
# }
# if (AddIndType[i]==3) {
# # vulnerable biomass
# b1 <- apply(VBiomass[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum)
# b2 <- apply(VBiomass_P, c(1, 2, 3), sum)
# }
} else {
if (AddIndType[i]==1) {
# total stock
b1 <- apply(N[,,,yr.ind:nyears,,drop=FALSE], 1:4, sum) # Abundance-based index
b2 <- apply(N_P, 1:4, sum)
}
# if (AddIndType[i]==2) {
# # spawning stock
# b1 <- apply(N[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum) * StockPars$Mat_age[,,yr.ind:nyears, drop=FALSE]
# b2 <- apply(N_P, c(1, 2, 3), sum) * StockPars$Mat_age[,,(nyears+1):(nyears+proyears), drop=FALSE]
# }
# if (AddIndType[i]==3) {
# # vuln stock
# b1 <- apply(N[,,yr.ind:nyears,, drop=FALSE], c(1, 2, 3), sum) * V_P[,,yr.ind:nyears, drop=FALSE]
# b2 <- apply(N_P, c(1, 2, 3), sum) * V_P[,,(nyears+1):(nyears+proyears), drop=FALSE]
# }
}
for (pp in map.stocks) {
histV <- array(NA, dim=dim(b1[,pp,,, drop=FALSE]))
histV[,1,,] <- Ind_V_list[[pp]][,,yr.ind:nyears, drop=FALSE]
b1[,pp,,] <- b1[,pp,,, drop=FALSE] * histV
b2[,pp,,] <- b2[,pp,,] * Ind_V_list[[pp]][,,(nyears+1):(nyears+proyears), drop=FALSE]
}
b1 <- apply(b1, c(1,4), sum)
b2 <- apply(b2, c(1,4), sum)
tempI <- cbind(b1, b2[, 1:(y - 1)])
# standardize, apply beta & obs error
tempI <- exp(lcs(tempI))^ObsPars[[p]][[f]]$AddIbeta[,i] * ObsPars[[p]][[f]]$AddIerr[,i,yr.ind:(nyears + (y - 1))]
year.ind <- max(which(!is.na(RealData@AddInd[1,i,])))
scaler <- RealData@AddInd[1,i,year.ind]/tempI[,1]
scaler <- matrix(scaler, nrow=nsim, ncol=ncol(tempI))
tempI <- tempI * scaler # convert back to historical index scale
AddInd[,i,] <- cbind(Data@AddInd[1:nsim,i,1:yr.ind], tempI[,2:ncol(tempI)])
yr.index <- max(which(!is.na(Data@CV_AddInd[1,i,1:nyears])))
newCV_Ind <- matrix(Data@CV_AddInd[,i,yr.index], nrow=nsim, ncol=length(yind))
CV_AddInd[,i,] <- cbind(Data@CV_AddInd[,i,], newCV_Ind)
if (!is.null(RealData) && length(RealData@AddInd[1,i,])>nyears &&
!all(is.na(RealData@AddInd[1,i,(nyears+1):length(RealData@AddInd[1,i,])]))) {
# update projection index with observed index if it exists
addYr <- min(y-1,length(RealData@AddInd[1,i,]) - nyears)
AddInd[,i,(nyears+1):(nyears+addYr)] <- matrix(RealData@AddInd[1,i,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
CV_AddInd[,i,(nyears+1):(nyears+addYr)] <- matrix(RealData@CV_AddInd[1,i,(nyears+1):(nyears+addYr)],
nrow=nsim, ncol=addYr, byrow=TRUE)
}
}
}
Data@AddInd <- AddInd
Data@CV_AddInd <- CV_AddInd
}
# --- Index of recruitment ----
Recobs <- ObsPars[[p]][[f]]$Recerr_y[, nyears + yind] * apply(N_P[, p, 1, yind, , drop=FALSE], c(1, 4), sum)
Data@Rec <- cbind(Data@Rec, Recobs)
# --- Average catch ----
Data@AvC <- apply(Data@Cat, 1, mean)
# --- Depletion ----
Depletion <- apply(SSB_P[,p, , y-1, ], 1, sum)/StockPars[[p]]$ReferencePoints$ReferencePoints$SSB0
Depletion[Depletion < tiny] <- tiny
Data@Dt <- Depletion * ObsPars[[p]][[f]]$Derr_y[,nyears+y]
Data@Dep <- Depletion * ObsPars[[p]][[f]]$Derr_y[,nyears+y]
# --- Update life-history parameter estimates for current year ----
Data@vbLinf <- StockPars[[p]]$Linfarray[,nyears+y] * ObsPars[[p]][[f]]$Linfbias # observed vB Linf
Data@vbK <- StockPars[[p]]$Karray[,nyears+y] * ObsPars[[p]][[f]]$Kbias # observed vB K
Data@vbt0 <- StockPars[[p]]$t0array[,nyears+y] + ObsPars[[p]][[f]]$t0bias # observed vB t0
Data@Mort <- StockPars[[p]]$Marray[,nyears+y] * ObsPars[[p]][[f]]$Mbias # natural mortality
Data@L50 <- StockPars[[p]]$L50array[,nyears+y] * ObsPars[[p]][[f]]$lenMbias # observed length at 50% maturity
Data@L95 <- StockPars[[p]]$L95array[,nyears+y] * ObsPars[[p]][[f]]$lenMbias # observed length at 95% maturity
Data@L95[Data@L95 > 0.9 * Data@vbLinf] <- 0.9 * Data@vbLinf[Data@L95 > 0.9 * Data@vbLinf] # Set a hard limit on ratio of L95 to Linf
Data@L50[Data@L50 > 0.9 * Data@L95] <- 0.9 * Data@L95[Data@L50 > 0.9 * Data@L95] # Set a hard limit on ratio of L95 to Linf
# --- Abundance ----
# Calculate vulnerable and spawning biomass abundance --
M_array <- array(0.5*StockPars[[p]]$M_ageArray[,,nyears+y], dim=c(nsim, StockPars[[p]]$maxage+1, nareas))
A <- apply(VBiomass_P[, p,f, , y, ] * exp(-M_array), 1, sum) # Abundance (mid-year before fishing)
Asp <- apply(SSB_P[, p,, y, ] * exp(-M_array), 1, sum) # Spawning abundance (mid-year before fishing)
Data@Abun <- A * ObsPars[[p]][[f]]$Aerr_y[,nyears+yind]
Data@SpAbun <- Asp * ObsPars[[p]][[f]]$Aerr_y[,nyears+yind]
# --- Catch-at-age ----
# previous CAA
oldCAA <- Data@CAA
Data@CAA <- array(0, dim = c(nsim, nyears + y - 1, StockPars[[p]]$maxage+1))
Data@CAA[, 1:(nyears + y - interval[mm] - 1), ] <- oldCAA[, 1:(nyears + y - interval[mm] - 1), ]
# update CAA
CNtemp <- FleetPars[[p]][[f]]$retA_P[,,yind+nyears, drop=FALSE] *
apply(N_P[,p,,yind,, drop=FALSE], c(1,3,4), sum)
CNtemp[is.na(CNtemp)] <- tiny
CNtemp[!is.finite(CNtemp)] <- tiny
CAA <- simCAA(nsim, yrs=length(yind), StockPars[[p]]$maxage+1, Cret=CNtemp, ObsPars[[p]][[f]]$CAA_ESS, ObsPars[[p]][[f]]$CAA_nsamp)
Data@CAA[, nyears + yind, ] <- CAA
# --- Catch-at-length ----
oldCAL <- Data@CAL
Data@CAL <- array(0, dim = c(nsim, nyears + y - 1, StockPars[[p]]$nCALbins))
Data@CAL[, 1:(nyears + y - interval[mm] - 1), ] <- oldCAL[, 1:(nyears + y - interval[mm] - 1), ]
CAL <- array(NA, dim = c(nsim, interval[mm], StockPars[[p]]$nCALbins))
if (!all(is.na(Data@Vuln_CAL))) {
Vuln_CAL <- replicate(nsim,Data@Vuln_CAL[1,])
Vuln_CAL <- replicate(length(yind),Vuln_CAL)
Vuln_CAL <- aperm(Vuln_CAL, c(2,1,3))
VList_dat <- lapply(1:nsim, calcV,
Len_age=StockPars[[p]]$Len_age[,,nyears+yind, drop=FALSE],
LatASD=StockPars[[p]]$LatASD[,,nyears+yind, drop=FALSE],
SLarray=Vuln_CAL,
n_age=StockPars[[p]]$maxage+1,
nyears=length(yind), proyears = 0,
CAL_binsmid=StockPars[[p]]$CAL_binsmid)
V_data <- aperm(array(as.numeric(unlist(VList_dat, use.names=FALSE)), dim=c(StockPars[[p]]$maxage+1, length(yind), nsim)), c(3,1,2))
vn <- (apply(N_P[,p,,yind,, drop=FALSE], c(1,3,4), sum) * V_data)
vn <- aperm(vn, c(1,3,2))
CALdat <- simCAL(nsim, nyears=length(yind), StockPars[[p]]$maxage, ObsPars[[p]][[f]]$CAL_ESS,
ObsPars[[p]][[f]]$CAL_nsamp, StockPars[[p]]$nCALbins, StockPars[[p]]$CAL_binsmid,
StockPars[[p]]$CAL_bins,
vn=vn, retL=Vuln_CAL,
Linfarray=StockPars[[p]]$Linfarray[,nyears + yind, drop=FALSE],
Karray=StockPars[[p]]$Karray[,nyears + yind, drop=FALSE],
t0array=StockPars[[p]]$t0array[,nyears + yind,drop=FALSE],
LenCV=StockPars[[p]]$LenCV)
} else {
vn <- (apply(N_P[,p,,,, drop=FALSE], c(1,3,4), sum) * FleetPars[[p]][[f]]$retA_P_real[,,(nyears+1):(nyears+proyears)]) # numbers at age that would be retained
vn <- aperm(vn, c(1,3,2))
vn <- vn[,yind, ,drop=FALSE]
CALdat <- simCAL(nsim, nyears=length(yind), StockPars[[p]]$maxage, ObsPars[[p]][[f]]$CAL_ESS,
ObsPars[[p]][[f]]$CAL_nsamp, StockPars[[p]]$nCALbins, StockPars[[p]]$CAL_binsmid,
StockPars[[p]]$CAL_bins,
vn=vn, retL=FleetPars[[p]][[f]]$retL_P[,,nyears+yind, drop=FALSE],
Linfarray=StockPars[[p]]$Linfarray[,nyears + yind, drop=FALSE],
Karray=StockPars[[p]]$Karray[,nyears + yind, drop=FALSE],
t0array=StockPars[[p]]$t0array[,nyears + yind,drop=FALSE],
LenCV=StockPars[[p]]$LenCV)
}
Data@CAL[, nyears + yind, ] <- CALdat$CAL # observed catch-at-length
Data@ML <- cbind(Data@ML, CALdat$ML) # mean length
Data@Lc <- cbind(Data@Lc, CALdat$Lc) # modal length
Data@Lbar <- cbind(Data@Lbar, CALdat$Lbar) # mean length above Lc
Data@LFC <- FleetPars[[p]][[f]]$L5_y[,nyears+y] * ObsPars[[p]][[f]]$LFCbias # length at first capture
Data@LFS <- FleetPars[[p]][[f]]$LFS_y[,nyears+y] * ObsPars[[p]][[f]]$LFSbias # length at full selection
Data@Vmaxlen <- FleetPars[[p]][[f]]$Vmaxlen_y[,nyears+y]
# --- Previous Management Recommendations ----
Data@MPrec <- MPCalcs$TACrec # last MP TAC recommendation
if (length(dim(Effort)) == 5) {
Data@MPeff <- Effort[, 1,1,mm, y-1] # last recommended effort
} else {
Data@MPeff <- Effort[, mm, y-1] # last recommended effort
}
# --- Store OM Parameters ----
# put all the operating model parameters in one table
ind <- which(lapply(StockPars[[p]], length) == nsim)
drop_srr <- which(names(ind)=='SRRpars')
ind <- ind[-drop_srr]
stock <- as.data.frame(StockPars[[p]][ind])
stock$Fdisc <- NULL
stock$CAL_bins <- NULL
stock$CAL_binsmid <- NULL
ind <- which(lapply(FleetPars[[p]][[f]], length) == nsim)
fleet <- as.data.frame(FleetPars[[p]][[f]][ind])
ind <- which(lapply(ImpPars[[p]][[f]], length) == nsim)
imp <- as.data.frame(ImpPars[[p]][[f]][ind])
RefPoints <- StockPars[[p]]$ReferencePoints$ReferencePoints
refs <- RefPoints# [!names(RefPoints) %in% names(stock)]
OFLreal <- A * (1-exp(-RefPoints$FMSY)) # the true simulated Over Fishing Limit
OMtable <- data.frame(stock, fleet, imp, refs,
ageM=StockPars[[p]]$ageMarray[,nyears+y],
L5=FleetPars[[p]][[f]]$L5_y[,nyears+y],
LFS=FleetPars[[p]][[f]]$LFS_y[,nyears+y],
Vmaxlen=FleetPars[[p]][[f]]$Vmaxlen_y[,nyears+y],
LR5=FleetPars[[p]][[f]]$LR5_y[,nyears],
LFR=FleetPars[[p]][[f]]$LFR_y[,nyears+y],
Rmaxlen=FleetPars[[p]][[f]]$Rmaxlen_y[,nyears+y],
DR=FleetPars[[p]][[f]]$DR_y[,nyears+y], OFLreal, maxF=StockPars[[p]]$maxF,
A=A, Asp=Asp, CurrentYr=OM@CurrentYr)
OMtable <- OMtable[,order(names(OMtable))]
Data@OM <- OMtable
Data@Misc <- Misc
if (!is.null(RealData)) {
if (length(RealData@Misc) && !is.null(names(RealData@Misc))) {
Data@Misc[names(RealData@Misc)] <- RealData@Misc[names(RealData@Misc)]
}
}
Data
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.