Skip to contents

Simulate catch counts from a distributional family

Source: R/simulate-creel-data.R
simulate_creel_catch.Rd

Generates catch observations from one of three distributional families: Negative Binomial ("negbin"), zero-inflated lognormal / Delta ("delta"), or Poisson. Intended for distributional sensitivity analysis, power checks, and Petrere-style estimator comparisons.

Usage

simulate_creel_catch(
  n,
  effort = 1,
  family = c("negbin", "delta", "poisson"),
  mu = 5,
  size = 0.5,
  p_zero = 0.4,
  sigma = 1,
  var_structure = c("constant", "proportional", "squared"),
  seed = NULL
)

Arguments

n

Integer. Number of catch observations to generate.

effort

Numeric vector of length n or scalar. Effort values (e.g. angler-hours) for each observation. Used to scale catch under var_structure = "proportional" or "squared".

family

Character. Distribution family: "negbin" (default), "delta", or "poisson".

mu

Numeric. Mean catch rate (catch per unit effort). Default 5.

size

Numeric. Negative Binomial dispersion parameter (NB size). Ignored when family = "poisson". Default 0.5.

p_zero

Numeric in [0, 1). Zero-inflation probability for family = "delta". Default 0.40.

sigma

Numeric. Log-scale standard deviation for family = "delta" (positive values only). Default 1.0.

var_structure

Character. Error variance structure. "constant" (default): variance independent of effort; "proportional": \(Var \propto f\); "squared": \(Var \propto f^2\).

seed

Integer or NULL. Random seed. Default NULL.

Value

Integer vector of length n containing simulated catch counts.

Details

Delta distribution (Petrere et al. 2010, Table 1): A mixture of a point mass at zero (probability p_zero) and a lognormal distribution for positive values (parameters mu and sigma on the log scale). Closely matches empirical creel catch distributions.

Variance structures (Petrere et al. 2010):

  • constant: \(\varepsilon_i \sim N(0, \sigma^2)\)

  • proportional: \(\varepsilon_i \sim N(0, \sigma^2 f_i)\)

  • squared: \(\varepsilon_i \sim N(0, \sigma^2 f_i^2)\)

References

Petrere, M. Jr., Giacomini, H.C. & De Marco, P. Jr. (2010). Catch-per-unit-effort: which estimator is best? Braz. J. Biol. 70: 483–491. doi:10.1590/S1519-69842010005000010

See also

simulate_creel_data

Other "Simulation": day_length(), simulate_creel_data()

Examples

set.seed(1)
# NB catch (default)
catch_nb <- simulate_creel_catch(n = 200, effort = 3.0, mu = 5, size = 0.5)
mean(catch_nb); var(catch_nb)
#> [1] 5.425
#> [1] 54.60741

# Delta distribution (zero-inflated lognormal)
catch_d <- simulate_creel_catch(
  n = 500, effort = 3.0, family = "delta",
  p_zero = 0.45, mu = 1.6, sigma = 0.8
)
mean(catch_d == 0)  # ~0.45
#> [1] 0.504

# Proportional variance structure
effort <- rgamma(100, shape = 2.5, rate = 0.57)
catch_prop <- simulate_creel_catch(
  n = 100, effort = effort, mu = 4,
  var_structure = "proportional"
)