cleanPacFIN: Clean raw PacFIN data

In nwfsc-assess/PacFIN.Utilities: Generate fishery composition data from PacFIN data for the NWFSC

Clean raw PacFIN data

Description

Clean raw PacFIN data to remove unsuitable samples if CLEAN = TRUE and convert units of measured quantities to work with downstream functions. Raw data are meant to be inclusive of everything from PacFIN so users can explore all that is available, but this means that raw data will ALWAYS include information that is not appropriate for use in US West Coast stock assessments.

Usage

cleanPacFIN(
  Pdata,
  keep_INPFC = lifecycle::deprecated(),
  keep_gears,
  keep_sample_type = c("M"),
  keep_sample_method = "R",
  keep_length_type,
  keep_age_method = NULL,
  keep_missing_lengths = lifecycle::deprecated(),
  keep_states = c("WA", "OR", "CA"),
  CLEAN = TRUE,
  spp = NULL,
  verbose = TRUE,
  savedir
)

Arguments

Pdata	A data frame of biological samples originating from the Pacific Fishieries Information Network (PacFIN) data warehouse, which originated in 2014. Data are pulled using sql calls, see PullBDS.PacFIN().
keep_INPFC	Deprecated. Areas are now defined using different methods.
keep_gears	A character vector including only the gear types you want to label as unique fleets. Order matters and will define fleet numbering. If the argument is missing, which is the default, then all found gear groups are maintained and ordered alphabetically. For more details see getGearGroup that lists a web link for where you can find the available gear groupings and how they link to "GRID" within your data. GRID is a legacy term from PacFIN, now identified as PACFIN_GEAR_CODE in the biological and fish ticket data, where GR is short for gear and ID is short for identification. Typical entries will include character values such as HKL, POT, TWL, where the latter is short for all non-shrimp trawls and TWS is shrimp trawls. Other gear identification codes and their definitions include DRG which is dredge gear, MSC which is all other miscellaneous gear such as diving or river trawls, NET which is all non-trawl net gear, NTW which is non-trawl gear, and TLS which is trolling gear. As a special case, MID is available for spiny dogfish to extract mid-water trawl data as a separate fleet.
keep_sample_type	A vector of character values specifying the types of samples you want to keep. The default is to keep c("M"). Available types include market (M), research (R), special request (S), and commercial on-board (C). There are additional samples without a SAMPLE_TYPE, but they are only kept if you include NA in your call. All sample types from California are assigned to M. Including commercial on-board samples is not recommended because they might also be in WCGOP data and would lead to double counting.
keep_sample_method	A vector of character values specifying the types of sampling methods you want to keep. The default is to keep "R", which refers to samples that were sampled randomly. Available types include random (R), stratified (S), systematic (N), purposive (P), and special (X). As of February 17, 2021, Washington is the only state with a sample type of "", and it was limited to two special samples of yelloweye rockfish.
keep_length_type	A vector of character values specifying the types of length samples to keep. There is no default value, though users will typically want to keep c("", "F", "A"), but should also think about using c("", "F", "A", NA). Note that types other than those listed below can be present, especially if you are dealing with a skate. A is alternate length, D is dorsal length, F is fork length, S is standard length, and T is total length.
keep_age_method	A vector of ageing methods to retain in the data. All fish aged with methods other than those listed will no longer be considered aged. A value of NULL, the default, will keep all ageing methods. However, a vector of ⁠c("B", "BB", S", "", NA, 1, 2)⁠ will keep all unaged fish and those that were aged with break and burn and surface reads. You do not really need to include such a verbose vector of values though because numbers are converted to appropriate character codes in getAge. Therefore, something like c("B", "S") would be sufficient to keep all break and burn and surface reads.
keep_missing_lengths	Deprecated. Just subset them using ⁠is.na(Pdata[, 'length']) after running ⁠cleanPacFIN' if you want to remove lengths, though there is no need because the package accommodates keeping them in.
keep_states	A vector of states that you want to keep, where each state is defined using a two-letter abbreviation, e.g., WA. The default is to keep data from all three states, keep_states = c("WA", "OR", "CA"). Add 'UNK' to the vector if you want to keep data not assigned to a state.
CLEAN	A logical value used when you want to remove data from the input data set. The default is TRUE. Where the opposite returns the original data with additional columns and reports on what would have been removed.
spp	A character string giving the species name to ensure that the methods are species specific. Leave NULL if generic methods work for your species. Currently, sablefish is the only species with species-specific code.
verbose	A logical specifying if output should be written to the screen or not. Good for testing and exploring your data but can be turned off when output indicates information that you already know. The printing of output to the screen does not affect any of the returned objects. The default is to always print to the screen, i.e., verbose = TRUE.
savedir	A file path to the directory where the results will be saved. The default is the current working directory. The path can be relative or absolute.

Details

The original fields in the returned data are left untouched, with the exception of

• SEX: modified using nwfscSurvey::codify_sex() and upon return will only include character values such that fish with an unidentified sex are now "U".

• Age: the best ages to use going forward rather than just the first age read.

CLEAN

The data are put through various tests before they are returned and the results of these tests are stored in the CLEAN column. Thus, sometimes it is informative to run cleanPacFIN(CLEAN = FALSE) and use frequency tables to inspect which groups of data will be removed from the data set when you change the code to be CLEAN = FALSE. For example, many early length compositions do not have information on the weight of fish that were sampled, and thus, there is no way to infer how much the entire sample weighed or how much the tow/trip weighed. Therefore, these data cannot be expanded and are removed using CLEAN = TRUE. Some stock assessment authors or even previous versions of this very code attempted to use adjacent years to inform weights. The number of assumptions for this was great and state representatives discouraged inferring data that did not exist.

Furthermore

The values created as new columns are for use by other functions in this package. In particular, fishyr and season are useful if there are multiple seasons (e.g., winter and summer, as in the petrale sole assessment), and the year is adjusted so that "winter" occurs in one year, rather than across two.

The fleet, fishery, and state columns are meant for use in stratifying the data according to the particulars of an assessment.

Value

The input data filtered for desired areas and record types specified, with added columns

• year: initialized from SAMPLE_YEAR
  

• fleet: initialized to 1

• fishery: initialized to 1

• season: initialized to 1. Change using getSeason

• state: initialized from SOURCE_AGID. Change using getState

• length: length in mm, where NA indicates length is not available

• lengthcm: floored cm from FORK_LENGTH when available, otherwise FISH_LENGTH

• geargroup: the gear group associated with each GRID

• weightkg: fish weight in kg from FISH_WEIGHT and FISH_WEIGHT_UNITS

Author(s)

Andi Stephens

See Also

getState, getSeason

nwfsc-assess/PacFIN.Utilities documentation built on March 19, 2024, 11:54 p.m.