sample_gadget: Create length distributions and sample Gadget output
In inspktrgadget/gadgetSim: Create Gadget replicates for use in simulation studies

Description Usage Arguments Details Value Examples

These functions create length distributions from Gadget StockStdPrinter output and sample that output to simulate surveys. Error can be added to simulated surveys in survey_gadget and age and length subsamples can be taken with strip_age_length_data to more closely mimic real world surveys

survey_gadget(stock_data, length_groups, survey_suitability, survey_sigma,
  variable = "number", keep_zero_counts = FALSE)

strip_age_length_data(stock_data, length_samples = 1, age_samples = 1,
  quiet = FALSE)

replicate_datasets(stock_data, length_groups, survey_suitability, survey_sigma,
  length_samples = 1, age_samples = NULL, quiet = FALSE, n = 10,
  keep_zero_counts = FALSE)

`stock_data`	A `data.frame` of stock data retrieved via `get_stock_std` or `read_gadget_stock_std`
`length_groups`	Numeric vector of length groups to distribute by. The difference in values probably should be the same as dl in Gadget model
`survey_suitability`	Numeric vector the same length as `length_groups` representing the selection probability for each length in `length_groups`
`survey_sigma`	Numeric value of multiplicative error to place on samples
`variable`	Character vector of the variable on which to distribute length groups. Default is "number". Can be "number.consumed" or "biomass.consumed"
`keep_zero_counts`	Logical. Keep year/step/area/age combinations with no individuals
`length_samples`	Numeric. If `length_samples` <= 1, that proportion of `stock_data` will be sampled for each year/step/area combination. If `length_samples > 1`, then a number of length samples equal to the integer value provided will be sampled for each year/step/area combination.
`age_samples`	Numeric. If `age_samples` <= 1, that proportion of `length_samples` will be sampled for each year/step/area combination. If `age_samples > 1`, then a number of age samples equal to the integer value provided will be sampled for each year/step/area combination from the length data already sampled. Note that if `age_samples` >= 1 then this corresponds to the proportion of length samples to also sample for age. Therefore the actual proportion of age samples relative to sampled numbers is `length_samples * age_samples`. Sampling is similar for if `length_samples` and `age_samples` > 1; however, if `length_samples` is given as an integer value then `age_samples` must be `<= length_samples`
`quiet`	Logical. If either `length_samples` or `age_samples` are equal to 1, then this will not strip any length or age data, respectively, and will issue a warning stating so. `quiet = TRUE` suppresses this warning.
`n`	Integer. The number of times to replicate the sampling procedure

Length-structured population information from Gadget is output as mean length and standard deviation for each year, step, area, and age combination. survey_gadget takes output from get_stock_std (or read_gadget_stock_std) and distributes the number for each respective combination into numbers at each length specified by length_groups, simulates surveys given the selectivity provided in survey_suitability and error given by survey_sigma. If it is desired to have a subsample of length and age data to more closely mimic real world scenarios, then the output from survey_gadget can be fed into strip_age_length_data and subsamples of length and age data will be returned according to the given proportions. The entire process can be replicated n number of times using replicate_datasets

survey_gadget returns a data.frame similar to stock_data, but disaggregated by length based on the std. dev. found in the StockStdPrinter. strip_age_length_data returns a named list of two data.frames. One for length data and one for age data. replicate_datasets returns a names list of length 3, each element containing a data.frame with nrows the number of replications times the nrow of stock_data. The first of these data.frames is for survey indices, the second for length data, and the third for age data.

path <- system.file(gad_mod_dir, package = "gadgetSim")
cod_stock_std <- get_stock_std(main = "main",
                               path = path)
lengrps <- seq(0.5, 150.5, by = 1)
suitability <- logistic_selectivity(lengrps, 0.15, 20, 0.1)
cod_samples <- survey_gadget(cod_stock_std$cod, lengrps, suitability, 0.1)
cod_comp_data <-
  strip_age_length_data(cod_samples, length_samples = 0.2, age_samples = 0.2)

# the above process can be automated n number of times
cod_comp_data <-
    replicate_datasets(cod_stock_std$cod, lengrps, suitability,
                       survey_sigma = 0.1, length_samples = 0.2,
                       age_samples = 0.2, n = 10)