Package 'FLfse'

Title: Fisheries Sampling Evaluation In FLR
Description: Fisheries Sampling Evaluation in FLR. A package to interface betwen the SAM stock assessments and FLR, and aid the inclusion of SAM into Management Strategy Evaluation (MSE), including the set up of operating models mimicking SAM.
Authors: Simon H. Fischer [aut, cre], Sven Kupschus [aut]
Maintainer: Simon Fischer <[email protected]>
License: GPL-3
Version: 1.0.2
Built: 2024-10-27 03:03:40 UTC
Source: https://github.com/flr/FLfse

Help Index


North Sea cod

Description

Cod (Gadus morhua) in Subarea 4, Division 7.d, and Subdivision 20 (North Sea, eastern English Channel, Skagerrak). Assessment input data for North Sea cod as used by ICES WGNSSK 2018 and 2019. Includes the stock, the indices and the SAM configuration. See the example(s) below for how to run the assessment.

Usage

cod4_stk

cod4_stk_2019

cod4_stk_2019

cod4_idx

cod4_idx_2019

cod4_conf_sam

Format

An object of class FLStock of dimension 6 x 56 x 1 x 1 x 1 x 1.

An object of class FLStock of dimension 6 x 57 x 1 x 1 x 1 x 1.

An object of class FLStock of dimension 6 x 57 x 1 x 1 x 1 x 1.

An object of class FLIndices of length 2.

An object of class FLIndices of length 2.

An object of class list of length 20.

Source

https://www.stockassessment.org

Examples

# Replicate the 2018 assessment:
fit <- FLR_SAM(stk = cod4_stk,
               idx = cod4_idx,
               conf = cod4_conf_sam)

# Replicate the 2019 assessment:
fit <- FLR_SAM(stk = cod4_stk_2019,
               idx = cod4_idx_2019,
               conf = cod4_conf_sam)

Run SAM with FLR objects.

Description

This function runs a SAM assessment using FLR objects as input. Stock and fishery data is extracted from an object of class FLStock, survey indices from FLIndex or FLIndices. Additional model configurations can be passed to SAM.

Usage

FLR_SAM(
  stk,
  idx,
  conf = NULL,
  conf_full = FALSE,
  par_ini = NULL,
  DoParallel = FALSE,
  NA_rm = TRUE,
  idx_weight = FALSE,
  ...
)

## S4 method for signature 'FLStock,FLIndices'
FLR_SAM(
  stk,
  idx,
  conf = NULL,
  conf_full = FALSE,
  par_ini = NULL,
  DoParallel = FALSE,
  NA_rm = TRUE,
  idx_weight = FALSE,
  ...
)

## S4 method for signature 'FLStock,FLIndex'
FLR_SAM(
  stk,
  idx,
  conf = NULL,
  conf_full = FALSE,
  par_ini = NULL,
  DoParallel = FALSE,
  NA_rm = TRUE,
  idx_weight = FALSE,
  ...
)

Arguments

stk

Object of class FLStock with stock and fishery data.

idx

Object of class FLIndices or FLIndex object with survey index time series.

conf

Optional configurations passed to SAM. Defaults to NULL. If provided, should be a list.

conf_full

Use provided configuration object in full without ANY checking (see Details for more information).

par_ini

Optional starting parameters for SAM. See details for more information.

DoParallel

Optional, defaults to FALSE. If set to TRUE, will perform iterations of stock in parallel. See Details below for description.

NA_rm

Remove trailing years with NAs, defaults to TRUE.

idx_weight

Use index weights (index variance)? Defaults to FALSE. If required, should be slot of idx.

...

Additional arguments passed to sam.fit(), e.g. newtonsteps

Details

Stock and fishery data is extracted from stk, survey indices from idx. Stock data used are:

  • catch numbers at age (catch.n slot from stk)

  • maturity ogive (mat slot from stk)

  • stock weight at age (stock.wt slot from stk)

  • catch weight at age (catch.wt slot from stk)

  • discards weigth at age (discards.wt slot from stk)

  • landings weight at age (landings.wt slot from stk)

  • natural mortality at age (m slot from stk)

  • proportion of fishing mortality before spawning at age (harvest.spwn slot from stk)

  • proportion of natural mortality before spawning at age (catch.n slot from stk)

  • landing fraction at age (calculated as landings numbers at age divided by the catch numbers at age)

Trailing years without data in stk are removed, unless turned of by setting NA_rm = FALSE.

Survey indices are extracted from idx, using the index slot(s). SSB indices can be used. To define an index as SSB index, the first (age) dimension of the index slot of idx has to be of length 1 and the name of this dimension can be either missing (NA), non-numeric (e.g. "ssb") or -1.

Additional configurations can be passed as a list to SAM with the conf argument. If argument conf_full is set to TRUE, the configuration is passed straight on to SAM without any checking. If argument conf_full is set to FALSE (default), then FLR_SAM first generates a default model configuration with stockassessment's setup.sam.data and additional configurations available in conf replace default configurations. For details about possible configurations and format see 'help("defcon", package = "stockassessment") and https://github.com/fishfollower/SAM.

The function can handle input objects with multiple iterations (iter dimension in stk and idx). If multiple iterations are provided for stk but not for idx, idx will be inflated, and vice versa. If the assessment fails for some iterations, the error messages are returned for these iterations. If argument DoParallel is set to TRUE, the individual iterations are processed in parallel. This uses parallel computing provided by the package DoParallel. The parallel workers need to be set up and registered before calling the function. See ?DoParallel.

Argument par_ini allows the provision of initial parameter values for SAM and can speed up the model. Either a single set of parameters can be supplied and they are recycled if neccessary. Alternatively, a list of initial parameters can be supplied, one for each iteration of the stock/index. If the dimensions of initial values for numbers/fishing mortality at age differ from the data, redundant years are automatically removed and if years are missing, the values from the last provided year are recycled.

The default console output generated by SAM is not printed but saved. It is stored as an attribute of the fit and can be accessed with attr(fit, "messages").

Tagging data can be provided in the usual SAM format and should be stored as an attribute of attr(catch.n(stk), "recap"). Additional tagging configurations can be supplied as a list with the attribute attr(catch.n(stk), "recap_conf"), e.g. list(map = list(logitRecapturePhi = factor(c(1, 1)))), which are then passed on to sam.fit().

Weights for the catch numbers can be supplied as an attribute of attr(catch.n(stk), "weight") and should be formatted as FLQuant objects.

Value

An object of class sam (for single iteration) or sam_list (list of sam objects for multiple iterations) with the model results.

Warning

This methods requires the stockassessment package and all its dependencies to be installed. For details how to obtain stockassessment, see https://github.com/fishfollower/SAM/.

Examples

# fit SAM to North Sea cod
fit <- FLR_SAM(stk = cod4_stk, idx = cod4_idx)

# use WGNSSK 2017 configuration
fit <- FLR_SAM(stk = cod4_stk, idx = cod4_idx, conf = cod4_conf_sam)

# fit SAM to Irish Sea plaice
fit <- FLR_SAM(stk = ple7a_stk, idx = ple7a_idx, conf = ple7a_conf_sam)

Run SPiCT with FLR objects.

Description

This function extracts the catch from an FLStock object and the survey indices/index from a FLIndices or FLIndex object and runs a SPiCT (Surplus Production in Continuous Time) stock assessment.

Usage

FLR_SPiCT(stk, idx, conf = NULL)

## S4 method for signature 'FLStock,FLIndices'
FLR_SPiCT(stk, idx, conf = NULL)

## S4 method for signature 'FLStock,FLIndex'
FLR_SPiCT(stk, idx, conf = NULL)

Arguments

stk

Object of class FLStock with catch time series.

idx

Object of class FLIndices or FLIndex object with survey index time series.

conf

Optional configurations passed to SPiCT. Should be a list.

Details

The catch time series is obtained from the catch slot of stk.

The survey index/indices are obtained from the index slot(s) of idx. If the slot(s) contains an age age structure, the sum over all ages is used.

Additional configurations can be passed as a list to SPiCT with the conf argument. They are passed directly to SPiCT (fit.spict) without checking. Any configurations accepted by (fit.spict) can be used.

Value

An object of class spictcls with the model results.

Warning

This methods requires the spict package and all its dependencies to be installed. For details how to obtain spict, see https://github.com/mawp/spict/.

Examples

# fit SPiCT to Irish Sea plaice
fit <- FLR_SPiCT(stk = ple7a_stk, idx = ple7a_idx)
fit

# pass additional configuration, set time step to 1 per year
conf <- list(dteuler = 1)
fit <- FLR_SPiCT(stk = ple7a_stk, idx = ple7a_idx, conf = conf)
fit

Get parameter estimates from SAM model fit.

Description

This function extracts the parameter estimates from a SAM model fit. These are useful e.g. as initial values in subsequent model fits and can improve model convergence/computing time.

Usage

getpars(fit)

## S4 method for signature 'sam'
getpars(fit)

## S4 method for signature 'sam_list'
getpars(fit)

Arguments

fit

A single SAM model fit of class sam or a list of fits of class sam_lst.

Value

A list with the model parameters of the SAM or a list of them in case of several supplied models

Examples

### fit SAM to North Sea cod
fit <- FLR_SAM(stk = cod4_stk, idx = cod4_idx, conf = cod4_conf_sam)

### extract parameters
pars <- getpars(fit)

### use them as starting values
fit2 <- FLR_SAM(stk = cod4_stk, idx = cod4_idx, conf = cod4_conf_sam,
                par_ini = pars)

North Sea haddock

Description

Haddock (Melanogrammus aeglefinus) in Subarea 4, Division 6.a, and Subdivision 20 (North Sea, West of Scotland, Skagerrak). Assessment input data for North Sea haddock as used by ICES WGNSSK 2018. Includes the stock, the indices and the SAM configuration. See the example below for how to run the assessment.

Usage

had4_stk

had4_idx

had4_conf_sam

Format

An object of class FLStock of dimension 9 x 54 x 1 x 1 x 1 x 1.

An object of class FLIndices of length 2.

An object of class list of length 20.

Source

https://www.stockassessment.org

Examples

# Replicate the 2018 assessment:
fit <- FLR_SAM(stk = had4_stk,
               idx = had4_idx,
               conf = had4_stk)

North Sea herring

Description

Herring (Clupea harengus) in Subarea 4 and divisions 3.a and 7.d, autumn spawners (North Sea, Skagerrak and Kattegat, eastern English Channel). Assessment input data for North Sea herring as used by ICES HAWG 2019. Includes the stock, the indices and the SAM configuration.

Usage

her4_stk

her4_idx

her4_conf_sam

Format

An object of class FLStock of dimension 9 x 73 x 1 x 1 x 1 x 1.

An object of class FLIndices of length 8.

An object of class list of length 21.

Details

Please note that the herring SAM assessment requires a different version of the stockassessment package. For details, see the readme at https://github.com/shfischer/FLfse

Source

https://github.com/ices-eg/wg_HAWG/tree/master/NSAS

Examples

# NOTE: the herring SAM assessment requires a different version of
# the stockassessment package
## Not run: 
fit <- FLR_SAM(stk = her4_stk,
               idx = her4_idx,
               conf = her4_conf_sam)

## End(Not run)

Northeast Atlantic mackerel 2019

Description

Mackerel (Scomber scombrus) in subareas 1-8 and 14, and in Division 9.a (the Northeast Atlantic and adjacent waters). Assessment input data for Northeast Atlantic mackerel as used by ICES WGWIDE 2019. Includes the stock, the indices and the SAM configuration. See the example below for how to run the assessment.

Usage

mac.27.nea_stk_2019

mac.27.nea_idx_2019

mac.27.nea_conf_2019

Format

An object of class FLStock of dimension 13 x 40 x 1 x 1 x 1 x 1.

An object of class FLIndices of length 3.

An object of class list of length 23.

Details

mac.27.nea_stk_2019 includes tagging data, stored as an attribute in attr(catch.n(mac.27.nea_stk_2019), "recap"). An additional tagging configuration is stored in attr(catch.n(mac.27.nea_stk_2019), "recap_conf"). Also, weights for the catch numbers are supplied as an attribute in attr(catch.n(mac.27.nea_stk_2019), "weight"). All additional data is passed on automatically to the SAM assessment.

Source

https://www.stockassessment.org

Examples

# The 2019 WGWIDE Northeast Atlantic mackerel assessment can be reproduced
# with:
fit <- FLR_SAM(stk = mac.27.nea_stk_2019,
               idx = mac.27.nea_idx_2019,
               conf = mac.27.nea_conf_2019)

Irish Sea plaice

Description

Plaice (Pleuronectces platessa) in Division 7.a (Irish Sea). Assessment input data for Irish Sea plaice as used by ICES WGCSE 2017 and 2019. Includes the stock, the indices and the SAM configuration. See the example below for how to run the assessment.

Usage

ple7a_stk

ple7a_idx

ple7a_stk_2019

ple7a_idx_2019

ple7a_conf_sam

Format

An object of class FLStock of dimension 8 x 36 x 1 x 1 x 1 x 1.

An object of class FLIndices of length 3.

An object of class FLStock of dimension 8 x 38 x 1 x 1 x 1 x 1.

An object of class FLIndices of length 3.

An object of class list of length 3.

Examples

# Replicate the 2019 assessment:
fit <- FLR_SAM(stk = ple7a_stk_2019,
               idx = ple7a_idx_2019,
               conf = ple7a_conf_sam)

Create replicates/iterations of SAM model fit based on variance-covariance matrix

Description

This function use the uncertainty estimated by SAM to create replicates/iterations of assessment results. The function uses the variance-covariance matrix to quantify uncertainty.

Usage

SAM_uncertainty(
  fit,
  n = 1000,
  print_screen = FALSE,
  seed = NULL,
  idx_cov = TRUE,
  catch_est = TRUE
)

## S4 method for signature 'sam'
SAM_uncertainty(
  fit,
  n = 1000,
  print_screen = FALSE,
  seed = NULL,
  idx_cov = TRUE,
  catch_est = TRUE
)

Arguments

fit

A SAM model fit object of class sam.

n

Number of replicates

print_screen

If set to TRUE, print output of TMB::sdreport to screen.

seed

Random number seed for reproducibility.

idx_cov

If set to TRUE, return covariance of survey index/indices.

catch_est

If set to TRUE, return catch estimates from SAM.

Details

The returned objects are FLQuants where the iteration dimension contains the replicates. Each replicate is internally consistent, e.g. the fishing mortality matches the stock numbers of the same replicate.

The following metrics are returned:

  • stock.n Stock numbers at age for all years, FLQuant

  • harvest Fishing mortalities at age for all years, FLQuant

  • catch.n Estimates of catch numbers at age for all years. This differs from the assessment input values. If the SAM model fit contains catch multipliers, the values returned here are corrected for this. Class FLQuant.

  • catch_sd Standard deviation of the catch numbers at age, time invariant, FLQuant

  • survey_catchability Catchability at age for all years for all survey indices, list of FLQuants

  • survey_sd Standard deviation of all surveys at age, time invariant, list of FLQuants,

  • survey_cov Covariance matrices of survey ages, one for each survey. Return object is a list of lists, the first level corresponds to the replicates, the second level to the surveys. If no covariance between ages is assumed in the SAM model, the diagonal in the covariance matrices is simply the square root of the standard deviation (in survey_sd)

  • proc_error Standard deviation of the stock numbers at age, time invariant, class FLQuant. This corresponds to the survival process error assumed/estimated, i.e. quantifies how much the actual stock numbers at age deviate from the deterministic catch equation.

Value

A list of FLQuants with the elements: stock.n, harvest, catch.n, catch_sd, survey_catchability, survey_sd, survey_cov, proc_error.

Examples

### fit SAM to North Sea cod
fit <- FLR_SAM(stk = cod4_stk, idx = cod4_idx, conf = cod4_conf_sam)

### create 2 replicates
reps <- SAM_uncertainty(fit, n = 2)

Coerce SAM output into FLStock object

Description

This function takes the output from running the SAM stockassessment and converts them into an FLStock object.

Usage

SAM2FLStock(
  object,
  stk,
  uncertainty = FALSE,
  conf_level = 95,
  stock_only = FALSE,
  catch_estimate = FALSE,
  correct_catch = FALSE,
  mat_est = FALSE,
  stock.wt_est = FALSE,
  catch.wt_est = FALSE,
  m_est = FALSE,
  spinoutyear = FALSE
)

## S4 method for signature 'sam,missing'
SAM2FLStock(
  object,
  stk,
  uncertainty = FALSE,
  conf_level = 95,
  stock_only = FALSE,
  catch_estimate = FALSE,
  correct_catch = FALSE,
  mat_est = FALSE,
  stock.wt_est = FALSE,
  catch.wt_est = FALSE,
  m_est = FALSE,
  spinoutyear = FALSE
)

## S4 method for signature 'sam_list,missing'
SAM2FLStock(
  object,
  stk,
  uncertainty = FALSE,
  conf_level = 95,
  stock_only = FALSE,
  catch_estimate = FALSE,
  correct_catch = FALSE,
  mat_est = FALSE,
  stock.wt_est = FALSE,
  catch.wt_est = FALSE,
  m_est = FALSE,
  spinoutyear = FALSE
)

## S4 method for signature 'sam,FLStock'
SAM2FLStock(
  object,
  stk,
  uncertainty = FALSE,
  conf_level = 95,
  stock_only = FALSE,
  catch_estimate = FALSE,
  correct_catch = FALSE,
  mat_est = FALSE,
  stock.wt_est = FALSE,
  catch.wt_est = FALSE,
  m_est = FALSE,
  spinoutyear = FALSE
)

## S4 method for signature 'sam_list,FLStock'
SAM2FLStock(
  object,
  stk,
  uncertainty = FALSE,
  conf_level = 95,
  stock_only = FALSE,
  catch_estimate = FALSE,
  correct_catch = FALSE,
  mat_est = FALSE,
  stock.wt_est = FALSE,
  catch.wt_est = FALSE,
  m_est = FALSE,
  spinoutyear = FALSE
)

Arguments

object

Object of class sam with the results from a SAM stock assessment run. Alternatively, object of class sam_list, i.e. a list of sam objects.

stk

Optional. Object of class FLStock, to which the assessment results are added.

uncertainty

If set to TRUE, the estimated uncertainty from SAM will be added as attribute.

conf_level

Confidence level used when uncertainty is returned. Defaults to 95 (percent).

stock_only

Logical. If set to TRUE, catch data (numbers, weights) are ignored and only stock data (numbers, SSB, etc.) are returned

catch_estimate

Logical, return the catch estimated by SAM instead of the model input?

correct_catch

Logical, correct catch with catch multiplier estimated by SAM?

mat_est

Logical, return SAM estimates for maturity?

stock.wt_est

Logical, return SAM estimates for stock weights?

catch.wt_est

Logical, return SAM estimates for catch weights?

m_est

Logical, return SAM estimates for natural mortality?

spinoutyear

Logical, return SAM estimates of biological estimates beyond last data year?

Details

SAM2FLStock returns both the input data used for running SAM (e.g. catch) and the model estimates (stock numbers and fishing mortality). By default, the returned catch is the input catch provided to SAM. However, the catch as estimated by SAM can be returned by setting catch_estimate = TRUE. Also, estimates of biological data (stock weights, catch weights, natural mortality, maturity) can be returned if requested and available from the model fit. Setting uncertainty = TRUE returns confidence intervals for catch, stock numbers and fishing mortality, saved as attributes in the corresponding slots of the FLStock output.

If an FLStock is provided as stk argument, then this is used as template. If the dimensions (years, iterations) differ between the SAM results and the provided stock template, the returned FLStock is expanded.

The object argument can either be a single SAM model fit or a list of SAM model fits (defined as class sam_list). If a list is provided, the output is an FLStock object where the different iterations correspond to the individual model fits.

The function can handle SAM model fits with multiple fleets. In the returned FLStock, the fleets are combined into a single fleet. Some functionality (e.g. uncertainty bounds) might not work for multiple fleets.

Value

An object of class FLStock.

Examples

# fit SAM to North Sea cod
fit <- FLR_SAM(stk = cod4_stk, idx = cod4_idx, conf = cod4_conf_sam)

# coerce the output into FLStock
stk <- SAM2FLStock(fit)

# get catch estimates from model
stk <- SAM2FLStock(fit, catch_estimate = TRUE)

## Not run: 
# use multi-fleet SAM model for western Baltic spring-spawning herring and
# load model fit from stockassessment.org
fit <- stockassessment::fitfromweb("WBSS_HAWG_2021")
stk <- SAM2FLStock(fit)

## End(Not run)

North Sea whiting

Description

Whiting (Merlangius merlangus) in Subarea 4 and Division 7.d (North Sea and eastern English Channel). Assessment input data for North Sea whiting as used by ICES WGNSSK 2018. Includes the stock, the indices and the SAM configuration. See the example below for how to run the assessment.

Usage

whg4_stk

whg4_idx

whg4_conf_sam

Format

An object of class FLStock of dimension 9 x 41 x 1 x 1 x 1 x 1.

An object of class FLIndices of length 2.

An object of class list of length 20.

Source

https://www.stockassessment.org

#' @examples # Replicate the 2017 assessment: fit <- FLR_SAM(stk = whg4_stk, idx = whg4_idx, conf = whg4_conf_sam)