OHI Science | Citation policy
[REFERENCE RMD FILE: https://cdn.rawgit.com/OHI-Science/ohiprep/master/globalprep/np/v2017/targetharvest_dataprep.html]
This analysis converts FAO capture production data into the OHI 2018 targeted harvest pressure data.
One more year of data
The species2group.csv file was udated: South America Sea Lion is not a cetacean therefore was assigned the pinniped order and excluded from the target list.
http://www.fao.org/fishery/statistics/software/fishstatj/en#downlApp
Release date: March 2018 FAO Global Capture Production Quantity 1950_2016
Downloaded: Aug 1 2018
Description: Quantity (tonnes) of fisheries capture for each county, species, year.
Time range: 1950-2016
# load libraries, set directories
library(ohicore) #devtools::install_github('ohi-science/ohicore@dev')
library(dplyr)
library(stringr)
library(tidyr)
library(ggplot2)
library(plotly)
## comment out when knitting
#setwd("globalprep/prs_targetedharvest/v2018")
### Load FAO-specific user-defined functions
source('../../../src/R/fao_fxn.R') # function for cleaning FAO files
source('../../../src/R/common.R') # directory locationsThis includes the FAO capture production data and a list of the “target” species.
## FAO capture production data
fis_fao_csv <- read.csv(file.path(dir_M, 'git-annex/globalprep/_raw_data/FAO_capture/d2018/Global_capture_production_Quantity_1950-2016.csv'))
# species list
sp2grp <- read.csv('raw/species2group.csv') %>%
dplyr::filter(incl_excl == 'include') %>%
dplyr::select(target, species); head(sp2grp)m <- fis_fao_csv %>%
dplyr::rename(country = Country..Country.,
species = Species..ASFIS.species.,
area = Fishing.area..FAO.major.fishing.area.,
Unit = Unit..Unit.) %>%
dplyr::select(-Unit)
m <- m %>%
tidyr::gather("year", "value", -(1:3)) %>%
dplyr::mutate(year = gsub("X", "", year)) %>%
fao_clean_data()
m <- m %>%
dplyr::mutate(species = as.character(species)) %>%
dplyr::mutate(species = ifelse(stringr::str_detect(species, "Henslow.*s swimming crab"), "Henslow's swimming crab", species))This analysis only includes target species. The warning messages need to be checked and, if necessary, changes should be made to the raw/species2group.csv
# check for discrepancies in species list
## seals are no longer included (so these errors can be ignored):
spgroups <- sort(as.character(unique(m$species)))
groups <- c('turtle', 'seal', 'whale', 'sea lion', 'dolphin', 'porpoise')
for (group in groups) { #group='dolphin'
possibles <- spgroups[grep(group, spgroups)]
d_missing_l <- setdiff(possibles, sp2grp$species)
if (length(d_missing_l)>0){
cat(sprintf("\nMISSING in the lookup the following species in target='%s'.\n %s\n",
group, paste(d_missing_l, collapse='\n ')))
}
}
# check for species in lookup not found in data
l_missing_d <- setdiff(sp2grp$species, spgroups)
if (length(l_missing_d)>0){
cat(sprintf('\nMISSING: These species in the lookup are not found in the FAO data \n'))
print(l_missing_d)
}
## filter data to include only target species ----
m2 <- m %>%
dplyr::filter(species %in% sp2grp$species)
unique(m2$area) # confirm these are all marine# spread wide to expand years
m_w = m2 %>%
tidyr::spread(year, value) %>%
dplyr::left_join(sp2grp, by='species'); head(m_w)
# gather long by target
m_l = m_w %>%
dplyr::select(-area) %>%
tidyr::gather(year, value, -country, -species, -target, na.rm=T) %>%
dplyr::mutate(year = as.integer(as.character(year))) %>%
dplyr::arrange(country, target, year); head(m_l)
#Temporary note: data for Gabn goes oly to 2009
# explore Japan[210]
m_l %>%
dplyr::group_by(country, target, year) %>%
dplyr::summarize(value = sum(value)) %>%
dplyr::filter(country == 'Japan', target == 'cetacean', year >= 2000)
# summarize totals per region per year
m_sum = m_l %>%
dplyr::group_by(country, year) %>%
dplyr::summarize(value = sum(value, na.rm=TRUE)) %>%
dplyr::filter(value != 0) %>%
dplyr::ungroup(); head(m_sum) m_sum <- m_sum %>%
dplyr::mutate(country = as.character(country)) %>%
dplyr::mutate(country = ifelse(stringr::str_detect(country, "C.*te d'Ivoire"), "Ivory Coast", country))
### Function to convert to OHI region ID
m_sum_rgn <- name_2_rgn(df_in = m_sum,
fld_name='country',
flds_unique=c('year'))
# these are duplicates for the same region
dplyr::filter(m_sum_rgn, country %in% c("Guadeloupe", "Martinique"))
# They will be summed:
m_sum_rgn <- m_sum_rgn %>%
dplyr::group_by(rgn_id, rgn_name, year) %>%
dplyr::summarize(value = sum(value)) %>%
dplyr::ungroup()Data is rescaled by dividing by the 95th quantile of values across all regions from 2011 to 2014.
target_harvest <- m_sum_rgn %>%
dplyr::mutate(quant_95 = quantile(value[year %in% 2011:2016], 0.95, na.rm = TRUE)) %>%
dplyr::mutate(score = value / quant_95) %>%
dplyr::mutate(score = ifelse(score>1, 1, score)) %>%
dplyr::select(rgn_id, year, pressure_score = score) %>%
dplyr::arrange(rgn_id, year); head(target_harvest); summary(target_harvest)
#quant_95= 160761
# any regions that did not have a catch should have score = 0
rgns <- rgn_master %>%
dplyr::filter(rgn_typ == "eez") %>%
dplyr::select(rgn_id = rgn_id_2013) %>%
dplyr::filter(rgn_id < 255) %>%
base::unique() %>%
dplyr::arrange(rgn_id)
rgns <- expand.grid(rgn_id = rgns$rgn_id, year = min(target_harvest$year):max(target_harvest$year))
target_harvest <- rgns %>%
dplyr::left_join(target_harvest) %>%
dplyr::mutate(pressure_score = ifelse(is.na(pressure_score), 0, pressure_score)) %>%
dplyr::arrange(rgn_id); head(target_harvest); summary(target_harvest)
write.csv(target_harvest, 'output/fao_targeted.csv', row.names = FALSE)
target_harvest_gf <- target_harvest %>%
dplyr::mutate(gapfill = 0) %>%
dplyr::select(rgn_id, year, gapfill)
write.csv(target_harvest_gf, 'output/fao_targeted_gf.csv', row.names = FALSE)The data from last year and this year should be the same unless there were changes to underlying FAO data or the master species list.
In this case, all of the regions looked very similar.
new <- read.csv("output/fao_targeted.csv") %>%
filter(year==2015)
old <- read.csv("../v2017/output/fao_targeted.csv") %>%
#mutate(year== year-2) %>%
dplyr::filter(year == 2015) %>%
dplyr::select(rgn_id, year, pressure_score_old=pressure_score) %>%
dplyr::left_join(new, by=c("rgn_id", "year"))
old
plot(pressure_score ~ pressure_score_old, data=old)
abline(0, 1, col="red")
compare_plot <- ggplot(data = old, aes(x=pressure_score_old, y= pressure_score, label=rgn_id))+
geom_point()+
geom_abline(color="red")
plot(compare_plot)
ggplotly(compare_plot)