NEWS.md

In tcarruth/MSEtool: Management Strategy Evaluation Toolkit

The latest release of the MSEtool package is available on CRAN.

MSEtool 2.0.1

• Markdown reports gradients and CVs of parameter estimates.
• SS2MOM function imports SS assessments and builds a multi-sex, multi-fleet MOM.
• VPA2OM function helps build on OM from arrays of F-at-age, Abundance-at-age, and Catch-at-age

MSEtool 2.0.0

SRA_scope

• data$I_type is now obsolete to remove redundancy with argument s_selectivity. Use s_selectivity to specify the selectivity of surveys. If data$I_type is detected, the code will attempt to update s_selectivity.
• data$MS (mean size) is now used instead of data$ML (fishery mean lengths). data$MS can also be mean weights, specify with data$MS_type to be either "length" or "weight". The likelihood of data$MS uses a normal distribution with constant CV specified in data$MS_cv (default = 0.2).
• The function is now a generic. The data argument can now be either a list (preferred) or a DLMtool Data S4 object.

SS2Data

• Update function to fill in new Data slots: CAL_mids and Vmaxlen.
• Both SS2Data and SS2OM imports all indices into Data@AddInd.

Other

• Update calculation of Pearson residual of age/length comps in bubble residual plots.

MSEtool 1.7.0

SRA_scope

• Allow for OM conditioning by resampling the Hessian matrix from a single model fit.
• Projection period rec devs are calculated with autocorrelation from the last estimated historical rec dev.
• More error checks if there are issues with the predicted catch in the model.
• Update AddInd calcs for DLMtool 5.4.4.
• Constants, including fixed standard deviations, are removed from the normal and multinomial likelihoods.

Other

• Rename model parameters that correspond with magnitude and estimated on the log-scale. By default, these estimated parameters are also rescaled internally to be within an order of magnitude of other parameters. As such, they are no longer the log of the corresponding population parameter, e.g., R0.
• Delay-difference and delay-differential models accommodate alternative initial depletion values in the first year of the model instead of the default unfished assumption. Along with the surplus production model, these models also accommodate multiple indices in the Data object.

MSEtool 1.6.0

• Time-varying fleet selectivity in SRA_scope can be set up with blocks. Unique blocks are defined and then assigned to fleet and year. New vignettes and updated help files for SRA_scope describe the set up in the function call.
• An implemention of Simple Stock Synthesis (function SSS, catch-only method with fixed depletion assumption) is now added to the package.
• The retrospective function is now available for spict assessments.
• The delay-difference models (DD_TMB and DD_SS) can now be conditioned on catch (previous versions only allowed conditioning on effort). The default is now to condition on catch, which is standard practice.

MSEtool 1.5.0

SP

• SP and SP_SS now support multiple indices in the model, using Data@AddInd and Data@CV_AddInd. These assessments still support Data@Ind but a custom wrapper function is still needed to use either Data@SpInd or Data@VInd.

SRA_scope

• Option to remove plusgroup in SRA_scope has been added.
• When conditioned on catch, SRA_scope can now solve F iteratively using Newton's method (argument condition = "catch2"). F as independently estimated parameters is still available with argument condition = "catch".
• Added a retrospective generic function for SRA objects.
• An ageing error matrix can now be provided to the model.
• Improvements in the backend for reporting (plotting) and error messaging.
• Additional function arguments for adjusting control parameters for optimization (in function nlminb), generating replicates by resampling from the covariance matrix, and filtering non-converged simulation replicates.

MSEtool 1.4.3

Updates to SRA_scope

• Bugs in (1) likelihood weights, and (2) indexing for initial biomass calcs have been fixed.
• Absolute indices, i.e., indices where q = 1, age-specific indices, and abundance/biomass-based indices can now be accommodated.
• The lognormal likelihood for composition data is now available as an alternative to the multinomial.
• When conditioned on catch, fishing mortality is estimated as deviations from the F in the middle of the time series (instead of all independent parameters) for more stability.
• Markdown reporting now includes likelihood weights. Slight changes in color schemes are included for consistency among plots.
• compare_SRA is a function that compares output and fits from multiple SRA objects with identical model structures in slot SRA@mean_fit but different data weightings, omissions, multipliers, etc.

MSEtool 1.4.1

• Fixed error in Solaris build.

MSEtool 1.4.0

Updates to SRA_scope

• Revised function arguments. Data inputs are in a single list in order to keep function calls tidy. The function should be backwards compatible for the most part.
• Mean weight-at-age is now used to calculate biomass and catch in order to match calculations in DLMtool::runMSE. Depletion calculations also match those in DLMtool::runMSE.
• Survey selectivity can now be estimated if age or length compositions for the survey are provided. See help file for setup.
• New plots have been added to the markdown report, include those that compare the outputs from the SRA and the updated operating model.
• A vignette for SRA_scope has now been added.

Other

• Extensive revisions to SS2OM have been added. The function also generates a markdown report to compare operating model output to Stock Synthesis outputs, e.g., recruitment, catch, spawning biomass time series.
• A log-Jacobian transform has been added for the r prior in SP and SP_SS (surplus production models). This is needed because FMSY is estimated rate parameter rather than r. By default, the minimum CV on the r-prior is 0.1 to allow the model to update r. It is assumed n is fixed in the model.
• Re-organize TMB files.

MSEtool 1.3.0

Updates to SRA_scope

• Defaults for SRA_scope are now more robust (set maximum F in model, higher std. dev. for likelihood of mean lengths).
• Users can choose to use SRA_scope conditioned on either observed catch or observed effort.
• SRA_scope returns an S4 object of class SRA with a plot() method that generates a markdown report of model fits.

Assessment models

• A prior for r is now possible with SP and SP_SS using life history information (priors in natural mortality and steepness, as well as maturity/weight at age). To use this feature, set argument use_r_prior = TRUE.
• Default process error standard deviation for SP_SS is reduced to 0.1.
• cDD and cDD_SS are more robust when catch is very, very small (F is set to 0). This is important for management procedures that shut down fishing.
• Minor updates to simplify TMB code.
• Minor revision to make_MP.

Other edits

• Vignette links are now available through the MSEtool help page. Type ?MSEtool into the console.
• Minor fixes to multiMSE.
• SS2OM now has an option for selecting male or female life history parameters.

MSEtool 1.2.1

• Fixed error in Solaris build.

MSEtool 1.2.0

For the new features described below, DLMtool version 5.3.1 is recommended.

multiMSE

• The initial release for multi-fleet and multi-stock operating models and MSEs are released in this version, with multiMSE being the core function. The multiMSE vignette will be quite useful and can be accessed at browseVignettes("MSEtool").

Assessment models

Quite a few additions and changes have been made to the Assessment models. See the help manual and vignettes for descriptions of these new Assessment functions.

• The continuous delay-differential model with deterministic and stochastic recruitment (cDD and cDD_SS, respectively) have been added as new Assessment models to the package. The continuous formulation should be more stable in high F situations.
• A virtual population analysis VPA model has also been added to the package.
• The surplus production model SP assumes continuous production and estimates continuous F's, similar to ASPIC. This formulation will be more stable in high F situations. The Fox model can be implemented by setting the production function exponent n = 1.
• A wrapper function for spict (state-space surplus production model) has been written and is available in the DLMextra package (located on Github). While reporting functions are available in MSEtool, the output of the wrapper function can still be used with the diagnostic functions in the spict package.
• SCA and SCA2 estimate annual F's and include a likelihood function for the catch. In previous versions, SCA matched the predicted catch to observed catch. This feature has been transfered over to the SCA_Pope function.
• Summary of assessment results can be obtained by the plot function which now generates a markdown report. This will be useful for diagnosing model fits and evaluating parameter estimates.

Other edits and additions, including:

• A scoping function SRA_scope fits an assessment model to catch, indices, and age/length comps to inform historical effort, recruitment deviations, and depletion for data-moderate operating models. Multiple fits are done based on the different life history parameters assumed in the operating model. This function is intended to be an alternative to DLMtool::StochasticSRA.
• profile and retrospective functions for profiling the likelihood and retrospective analyses, respectively, of assessment models are now improved.
• The compare_models function has been added to compare time series estimates, e.g. B/BMSY and F/FMSY, among different assessment models.
• Manual starting values (the start argument) for parameters of assessment models can be expressions and subsequently evaluated in the assessment function. This can be very helpful when passing starting values in the make_MP function.
• The CASAL2OM function can be used to generate an operating model from CASAL assessments.
• The SS2OM and SS2Data functions are updated for the latest versions of r4ss on Github.
• More options are provided to increase flexibility for ramped harvest control rules (HCR_ramp).

MSEtool 1.1.0

New additions, including:

• Re-parameterization of dome selectivity in SCA so that estimated parameters are age-based after transformation.
• Additional argument in SCA for lognormal distribution for age comps.
• A more efficient method is used to report convergence diagnostics of assessment models when running in closed-loop simulation.

Minor edits, including:

• By default, steepness is now fixed in the SCA and SCA2 assessment functions.
• By default, nine data-rich MPs are now included in the package. See the help documentation: ?Data-rich-MP
• A generic function for ramped harvest control rules (HCR_ramp) is now included. Users can input the desired limit and target reference points.
• make_MP adds dependencies to the MP so that DLMtool::Required returns the appropriate dependencies. Dependencies are dynamic based on the configuration of the assessment model. For example, Data@steep is a dependency for a SCA-based model only if steepness is fixed.

MSEtool 1.0.0

• Initial CRAN release.
• Assessment models: Delay-Difference (DD_TMB, DD_SS); Surplus Production (SP, SP_SS); Statistical Catch-at-Age (SCA, SCA2)
• Harvest control rules: HCR_MSY, HCR40_10, HCR60_20
• simmov function for multiple-area movement models (age-independent)
• Functions for converting Stock Synthesis and iSCAM assessments to OM and Data objects (classes inherited from DLMtool)

tcarruth/MSEtool documentation built on Oct. 19, 2020, 6:09 a.m.