CatDynBSD {CatDyn}

Calculate Annual Biomass and its Standard Error

Description

Using results of a model fit by CatDynFit, calculate annual biomass considering initial abundance, mean weight, abundance inputs, fishery removals, and natural mortality, and then computes its standard error using the delta method.

Usage

CatDynBSD(x, method, multi, mbw.sd)

Arguments

x

One object of class catdyn when the time step is the month (multi-annual version) or a list of objects of class catdyn when the time step is the week or the day (intra-annual versions). The number of years when the time step is the month or the number of separate catdyn objects when the time step is the week or the day must be 15 or higher.

method

Character or character vector. The numerical method used to fit the model when the time step is the month or a character vector with all the methods used to fit the several models when the time step is the week or the day.

multi

Logical. Wheter there is a single multi-annual model (monthly time step), or many intra-annual models (daily or weekly time step)

mbw.sd

Numeric. A vector of length 12 with the standard deviation or standard error of the mean weight in kg per month when the time step is the month, or a data.frame with three columns when the time step is the week or the day: year, mean weight in kg, and standard deviation or standard error of mean weight in kg. In the latter case the number of rows must equal the number of years (>14), also the length of the list x.

Details

The main purpose of this function is to obtain annual biomass estimates to be passed as input information for the fit of a population dynamics of the surplus production kind in a hierarchical inference framework. Thus is carries over most of the uncertainty in the original catch, effort and mean weight data, to inform the population dynamics model. The limit of 15 years of data as a minimum for the use of this function is set so that the fit of the population dynamics model has sufficient information to estimate its parameters.

When the time step is the month, this function will calculate the biomass and its standard error for every month in the time series of data. When the time step is the week or the day it will calculate the biomass and its standard error at the start of the season, only one value per year.

Value

When the time step is the month, a data.frame with columns for year, month, time step, mean weight (kg), standard deviation or standard error of mean weight, abundance, standard error of abundance, biomass (tons), and standard error of biomass (tons). The data.frame has as many rows as time steps (minimum of 180 months)

When the time step is the week or the day, a data.frame with the year, mean weight (kg), standard deviation or standard error of mean weight, abundance, standard error of abundance, biomass (tons), and standard error of biomass (tons). The data.frame has as many rows as years (minimum of 15 years).

Note

This function makes extensive use of the delta method for carrying over the original uncertainty. It uses the asymptotic standard errors of parameters in the object of class catdyn and their correlation matrix, along with standard deviation or standard error of mean weight per time step, to calculate the standard error of biomass.

In case a model output (object of class catdyn) is selected which has not produced standard errors for all parameters that are involved in the calculation of biomass (natural mortality, initial abundance, and magnitude of each input/output pulses of abundance) then the missing standard errors are replaced by imputed standard errors computed as the estimate for which the standard error is missing times the mean coefficient of variation across all parameters which did get standard errors. In this manner the degree of statistical uncertainty is preserved and all standard errors are available to use the delta method to calculate the standard error of biomass.

Author(s)

Ruben H. Roa-Ureta (ORCID ID 0000-0002-9620-5224)

References

Roa-Ureta, R. H. et al. 2015. Fisheries Research 171 (Special Issue), 59-67.