In pbs-assess/gfplot: Data Extraction and Plotting for DFO PBS Groundfish Stocks

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.asp = 0.618,
  autodep = TRUE,
  cache = TRUE,
  cache.comments = FALSE
)

library(dplyr)
library(ggplot2)
library(gfplot)

Extract catch and effort data from all databases for species of interest:

species <- c("Arrowtooth Flounder", "English Sole", "Dover Sole", "Southern Rock Sole",
  "Petrale Sole") 
d <- gfplot::get_cpue_historic(species, fishing_year = FALSE, end_year = 2017)
saveRDS(d, file = "data/flatfish-cpue-data.rds", compress = FALSE)

Read the data in, clean up the column names, and create columns for the major statistical areas:

d <- readRDS("data/flatfish-cpue-data.rds")

Filter down to remove all rows of data with missing hours fished, missing total catch, or when the total catch or hours fished are 0.

d <- d %>%
  filter(!is.na(hours_fished), !is.na(total), total > 0, hours_fished > 0) %>%
  filter(!is.na(fyear))

Now let's calculate the arithmetic and geometric mean CPUE for each species in each year:

d_sum <- d %>%
  group_by(area, fyear, species_common_name) %>%
  summarise(
    catch = sum(total, na.rm = TRUE),
    sum_hours_fished = sum(hours_fished, na.rm = TRUE),
    arith_cpue = sum(total, na.rm = TRUE) / sum(hours_fished, na.rm = TRUE),
    geo_cpue = exp(mean(log(total / hours_fished), na.rm = TRUE))
  ) %>%
  ungroup()

Let's create a version that is scaled by the geometric mean for plotting to mimic past research documents:

d_scaled <- d_sum %>%
  mutate(arith_cpue = arith_cpue / exp(mean(log(arith_cpue)))) %>%
  mutate(geo_cpue = geo_cpue / exp(mean(log(geo_cpue))))

Plot of the scaled version with species in different columns:

p <- d_scaled %>%
  tidyr::gather(cpue_type, cpue_value, arith_cpue:geo_cpue) %>%
  ggplot(aes(fyear, cpue_value, linetype = cpue_type)) +
  geom_line() +
  geom_vline(xintercept = 1996, lty = 3) +
  facet_grid(area~species_common_name, scales = "free_y") +
  xlab("") +
  ylab("CPUE divided by geometric mean") +
  ylim(0, NA) +
  theme_pbs() +
  labs(linetype = "CPUE type")
p

Plot of the scaled version with CPUE type in different columns:

d_scaled %>%
  tidyr::gather(cpue_type, cpue_value, arith_cpue:geo_cpue) %>%
  ggplot(aes(fyear, cpue_value, colour = species_common_name)) +
  geom_line() +
  geom_vline(xintercept = 1996, lty = 3) +
  facet_grid(area~cpue_type, scales = "free_y") +
  xlab("") +
  ylab("CPUE divided by geometric mean") +
  ylim(0, NA) +
  theme_pbs() +
  labs(colour = "Species")

Save the unstandardized and unscaled data:

saveRDS(d_sum, file = "../data/flatfish-annual-cpue.rds", compress = FALSE)

Save the cpue plots:

png("../data/flatfish_cpue.png")
p
dev.off()

pbs-assess/gfplot documentation built on April 3, 2024, 2:10 p.m.