Births/individual is based on number in population x survival rate and not number in population

[cschwarz-stat-sfu-ca]

Either the documentation for F.cjs.simulated needs to be updated, or the code for F.cjs.simulate needs adjustment

Documentation says:
If births.per.indiv is a vector of length NS - 1, then round( N(j)*births.per.indiv[,j] ) births will occur between occasions j and j+1, where N(j) is the true number of individuals in the realized population at occasion j.

However, actual code (lines 70) has

births.per.indiv is a vector of rates. Multiply by current number in population

        births <- round(sum(survived) * births.per.indiv[j-1])

Earlier in the code the survived vector is the number alive at time j from those alive at time j-1 so this does not match the documentation.

For example,

n.years <- 3

p <- rep(.1, n.years) # capture probability/year
phi <- rep(.50, n.years-1) # survival probability/year

sim.data <- mra::F.cjs.simulate( super.p=p,
super.s=phi, # survival probability
N1 = 1000, # initial population size of age 3 animals
births.per.indiv = c(1,1), # number of new recruits on a per individual basis
R=1) # number of populations to generate

actual population size

sim.data[[1]]$popln.n

gives

t1 t2 t3
1000 1036 1078

I would have expected it to give
N(2) = 1000*.5 + 10001 = 1500
It computes
N(2) = 1000.5 * 1000*.5*1 = 1000 ( the 1036 is simulation error)

Carl Schwarz
[email protected]