v3.1.1: an R package for assessing observer coverage needs for rare species
This is the publication release of the ObsCovgTools package.
The R package ObsCovgTools provides tools for evaluating fishery observer coverage, particularly with respect to documenting and estimating rare bycatch. Current functionality includes evaluating observer coverage in terms of (1) probabilities of observing a bycatch event and of any bycatch occurring in total effort, given mean bycatch rate, dispersion index (variance to mean ratio in the bycatch rate), and total fishery effort; (2) upper confidence limit of bycatch when none was observed, given total fishery effort and dispersion index; and (3) bycatch estimation CV (coefficient of variation), given bycatch rate, dispersion index, and total fishery effort. Estimates in all cases are based directly on or simulated from the corresponding Poisson or negative binomial probability distribution.
Caveat
The current implementation of ObsCovgTools assumes observer coverage is representative of the fishery, and does not account for hierarchical sources of variance (e.g., vessel- or trip-level variation). Violating these assumptions may result in negatively biased projections of observer coverage required to meet specific objectives. Unless hierarchical sources of variance can be ruled out as potentially important, using higher-level units of effort is advised (e.g., mean bycatch per trip and number of trips, instead of mean bycatch per set and number of sets).
Shiny app
This package has been implemented as a Shiny web application, coauthored by Howard Coleman, which can be accessed at https://kacurtis.shinyapps.io/obscov/.
The latest version of the package is available at https://github.com/kacurtis/ObsCovgTools.
Please cite the following publication when you use or reference this software:
Curtis K. A., Carretta J. V. 2020. ObsCovgTools: Assessing observer coverage needed to document and estimate rare event bycatch. Fisheries Research, 225: 105493. https://doi.org/10.1016/j.fishres.2020.105493