Predicting gear width

Preamble

The ICES VMS datacall request metrics on gear width of towed gear. Routinely the gear width is obtained from ‘the Benthis’ model. In this model gear width is a function of gear type and vessel length or power (kW), the model being either linear or a power function. Here a code snipped is provided that may be of use for predicting gear width.

Data preparation

Here will use the eflalo dataset from {vmstools}

library(ramb)
library(tidyverse)
library(vmstools)  # to get access to example data
data("eflalo")
lb <- 
  eflalo |> 
  as_tibble() |> 
  mutate(metier5 = rb_met5_from6(LE_MET_level6)) |> 
  select(length = VE_LEN, kw = VE_KW, metier5, metier6 = LE_MET_level6) |> 
  distinct() |> 
  mutate(.rid = 1:n(),
         .before = length)

Coding within a flow

A simple pipe flow:

lb1 <- 
  lb |> 
  left_join(metier5_benthis_lookup,
          by = join_by(metier5)) |> 
  left_join(benthis_parameters |> select(benthis_metier:variable),
            by = join_by(benthis_metier)) |> 
  mutate(width = 
           case_when(
             model == "power" & variable == "avg_kw" ~    a * kw^b,
             model == "power" & variable == "avg_oal" ~   a * length^b,
             model == "linear" & variable == "avg_oal" ~  a * length + b,
             .default = NA))

Using a function

Using the function in {ramb}

lb2 <- 
  lb |> 
  mutate(width = rb_benthis_width(metier5, length, kw))
identical(lb1$width, lb2$width)
#> [1] TRUE

Comparision with ICES VMS datacall flow

Just a double check:

source("~/stasi/ices/ICES-VMS-and-Logbook-Data-Call/0_global_functions.R")
lb3 <- lb
lb3$width <- add_gearwidth(lb3, "metier6", "length", "kw")
identical(lb2$width / 1000, lb3$width)
#> [1] FALSE
near(lb2$width / 1000, lb3$width) |> table(useNA = "ifany")
#> 
#> TRUE <NA> 
#>  253  239