pseudo.pred {QFASA}

Generate a pseudo predator by sampling with replacement from prey database.

Description

Generates a single pseudo predator by sampling with replacement from prey database. To generate a sample of pseudo predators, please refer to example code.

Usage

pseudo.pred(diet, preybase, cal.vec, fat.vec, preysize = 2)

Arguments

diet

the "true" or "desired" diet of the pseudo predator with prey species in alphabetical order (i.e.in the order of table(preyFAs[,2])). A compositional vector of proportions that sums to one with length equal to the number of prey species.

preybase

prey database from which to generate the pseudo predator. First column must provide the species name.

cal.vec

vector of calibration coefficients whose length is the same as the number of fatty acids in prey database.

fat.vec

vector of fat content whose length is the same as the number of species.

preysize

number of prey to sample from prey database. If preysize=1, then one prey is selected from each species. Otherwise, a sample of n_k signatures (where n_k is sample size for species k) is obtained by sampling with replacement.

Details

The default is to re-sample all of the prey signatures within each species (that is, preysize=2). Alternatively, one prey may be randomly selected from each species yielding potentially more variable pseudo-predators. For details on simulating realistic predators signatures, see Bromaghin, J. (2015) Simulating realistic predator signatures in quantitative fatty acid signature analysis, Ecological Informatics, 30, 68-71.

Value

A simulated predator FA signature.

Examples

data(preyFAs)

# Generating a sample of 10 pseudo predators each with "true" diet being
# (1/11,1/11,...1/11), no calibration effect and no fat content.  The QFASA diet estimate
# is then computed for each pseudo predator.

# Note: To incorporate calibration and fat content in a simulation study,
# one set of calibration and fat content is generally used to simulate the pseudo predator
# and another is used to estimate the diet.

set.seed(11)
p.mat <- matrix(rep(NA,10*11),nrow=10)
for (i in 1: 10) {
    my.seal <- pseudo.pred(rep(1/11,11),
                            preyFAs[,-c(1,3)],
                            rep(1,ncol(preyFAs[,-c(1,3)])-1),
                            rep(1,11))
     p.mat[i,] <- p.QFASA(my.seal,
                          MEANmeth(preyFAs[,-c(1,3)]),
                          rep(1,length(my.seal)),
                          2,
                          ext.fa=colnames(preyFAs[,-c(1:3)]))$`Diet Estimates`
 }

# Can verify that average diet estimate of the 10 pseudo predators is close to
# "true" diet.

 colnames(p.mat) <- as.vector(rownames(MEANmeth(preyFAs[,-c(1,3)])))
 round(apply(p.mat,2,mean),3)