Check the parameter name in MTI function #88
Description
The result of MTI function is not same with Ecopath windows version.
There are some parameters names not update, like Bimomass (BB in MTI)
please update it.
Below is a update MTI function I used.
MTI_new<- function (Rpath, Rpath.params, increase = T) {
Group <- Type <- V1 <- NULL
x <- copy(Rpath.params)
y <- copy(Rpath)
nfleets <- length(which(x$model$Type == 3))
ndetritus <- length(which(x$model$Type == 2))
ngroups <- length(x$model$Group)
fleetnames <- x$model$Group[which(x$model$Type == 3)]
detritusnames <- x$model$Group[which(x$model$Type == 2)]
livingnames <- x$model$Group[which(x$model$Type < 2)]
allnames <- x$model$Group
DC <- as.data.table(x$diet)
DC <- DC[Group != "Import"]
DC.types <- x$model[Type < 2, Type]
DC.colsum <- DC[, colSums(.SD, na.rm = T), .SDcols = livingnames]
DC.check <- round(DC.colsum + DC.types, 3)
for (ipred in 1:length(livingnames)) {
if (DC.check[ipred] != 1) {
DC[, :=(ipred + 1, .SD/DC.colsum[ipred]),
.SDcols = livingnames[ipred]]
}
}
fleetrows <- as.data.table(matrix(rep(0, nfleets * ncol(DC)),
nfleets, ncol(DC)))
fleetrows[, :=(V1, fleetnames)]
DC <- rbindlist(list(DC, fleetrows), use.names = F)
detcols <- as.data.table(matrix(rep(0, ndetritus * ngroups),
ngroups, ndetritus))
setnames(detcols, paste0("V", seq_along(detritusnames)),
detritusnames)
DC <- cbind(DC, detcols)
totcatch <- y$Landings + y$Discards
totcatch.sum <- colSums(totcatch)
for (ifleet in 1:ncol(totcatch)) {
fleet.prop <- as.data.table(totcatch[, ifleet]/totcatch.sum[ifleet])
setnames(fleet.prop, "V1", fleetnames[ifleet])
if (ifleet == 1) {
fleetcols <- fleet.prop
}
else {
fleetcols <- cbind(fleetcols, fleet.prop)
}
}
DC <- cbind(DC, fleetcols)
DCi.order <- c("Group", DC$Group)
setcolorder(DC, DCi.order)
DC[is.na(DC)] <- 0
DC[, :=(Group, NULL)]
bio <- y$Biomass
BQB <- bio * y$QB
Tij <- DC * BQB[col(as.matrix(DC))]
Tij <- Tij[, which(!names(Tij) %in% fleetnames), with = F]
Tij <- cbind(Tij, totcatch)
Tim <- rowSums(Tij)
FCij <- c()
for (iprey in 1:nrow(Tij)) {
fij <- Tij[iprey, ]/Tim[iprey]
FCij <- rbind(FCij, fij)
}
FCji <- t(FCij)
FCji[is.na(FCji)] <- 0
net.impact <- as.matrix(DC - FCji)
identity.matrix <- diag(ncol(net.impact))
MTI <- MASS::ginv(identity.matrix - net.impact) - identity.matrix
return(MTI)
}