add_mpm_error {mpmsim}

Add sampling error to matrix population models (MPMs) based on expected values of transition rates and sample sizes

Description

Produces a list of matrix population models based on expected values in the transition matrix and sample size. The expected values are provided in lists of two submatrices: mat_U for the growth/development and survival transitions and mat_F for the fecundity transitions. The output mat_U values are simulated based on expected probabilities, assuming a binomial process with a sample size defined by sample_size. The output mat_F values are simulated using a Poisson process with a sample size defined by sample_size.Thus users can expect that large sample sizes will result in simulated matrices that match closely with the expectations, while simulated matrices with small sample sizes will be more variable.

Usage

add_mpm_error(mat_U, mat_F, sample_size, split = TRUE, by_type = TRUE)

Arguments

mat_U

A list of U submatrices, or a single U submatrix.

mat_F

A list of F submatrices, or a single F submatrix.

sample_size

either (1) a single matrix of sample sizes for each element of every MPM, (2) a list of two named matrices ("mat_F_ss", "mat_U_ss") containing sample sizes for the survival and fertility submatrices of every MPM or (3) a single value applied to the every element of every matrix.

split

logical, whether to split the output into survival and fecundity matrices or not. Defaults to TRUE.

by_type

A logical indicating whether the matrices should be returned in a list by type (A, U, F, C). If split is FALSE, then by_type must also be FALSE. Defaults to TRUE.

Details

if any sample_size input is 0, it is assumed that the estimate for the element(s) concerned is known without error.

Value

list of matrices of survival and fecundity if split = TRUE, otherwise a single matrix of the sum of survival and fecundity.

Author(s)

Owen Jones [e-mail]

See Also

Other errors: calculate_errors(), compute_ci()

Other errors: calculate_errors(), compute_ci()

Examples

set.seed(42) # set seed for repeatability

# First generate a set of MPMs
mpm_set <- generate_mpm_set(n = 5, n_stages = 5, fecundity = c(
  0, 0, 4, 8, 10
), archetype = 4, split = TRUE, by_type = TRUE, as_compadre = FALSE)

# Now apply sampling error to this set
add_mpm_error(
  mat_U = mpm_set$U_list, mat_F = mpm_set$F_list, sample_size =
    50
)

# Also works with a single matrix.
mats <- make_leslie_mpm(
  survival = c(0.1, 0.2, 0.5),
  fertility = c(0, 1.2, 2.4),
  n_stages = 3, split = TRUE
)

# Sample size is a single value
add_mpm_error(mat_U = mats$mat_U, mat_F = mats$mat_F, sample_size = 20)

# Sample size is a list of two matrices
# here with a sample size of 20 for reproduction and 10 for growth/survival.
mpm_set <- generate_mpm_set(
  n = 5, n_stages = 3, fecundity = c(0, 2, 4),
  archetype = 4, split = TRUE, by_type = TRUE,
  as_compadre = FALSE
)

ssMats <- list(
  "mat_F_ss" = matrix(20, nrow = 3, ncol = 3),
  "mat_U_ss" = matrix(10, nrow = 3, ncol = 3)
)

# Add sampling error to the matrix models
output <- add_mpm_error(
  mat_U = mpm_set$U_list, mat_F = mpm_set$F_list,
  sample_size = ssMats
)

# Examine the outputs
names(output)
output