Here we will explore commercial and recreational catch data as estimated by the Sea Around Us Project. This data set includes reported data for fisheries worldwide as well as reconstructed estimates for unreported catches. For each country, project scientists assemble available data on catch and use additional sociological and fishing data plus expert information and opinion, to produce their best estimates of total catch from all fishing sectors and for all species globally.
The high seas data are accessed in the Basic Search High seas section of the site. The individual regions can be clicked on (e.g., Atlantic, Western Central) and then the data are downloaded using the Download Data button.
The current version of data in use is version 50.1, accessed June 2024 (the data files can be found on github in the SEFSC-dolphin-analyses/data /SAU/ folder). This was still the most recent version of data available as of September 2025.
First we read in the high seas catches for the Western Central Atlantic and Northwest Atlantic zones. These are equivalent to FAO areas 21 and 31.
# clear workspace
rm(list = ls())
if(!require("pals")) install.packages("pals")
library(pals)
# find data files
hslis <- dir("data/SAU/")[grep("HighSeas", dir("data/SAU/"))]
# identify areas 21 and 31
hslis[1]; hslis[3] [1] "SAU HighSeas 21 v50-1.csv"[1] "SAU HighSeas 31 v50-1.csv"#download data
h1 <- read.csv(paste0("data/SAU/", hslis[1]))
h3 <- read.csv(paste0("data/SAU/", hslis[3]))
# confirm correct areas
#unique(h1$area_name)
#unique(h3$area_name)
hs <- rbind(h1, h3) We extract the data for dolphinfish and look at the characteristics of the catches. Most of the catches are in the Western Central Atlantic, and the primary fishing entities are Venezuela and Saint Lucia. The fishing sector is all industrial, and the vast majority of the catch is reported landings, with only a tiny fraction of discards or unreported landings listed in the database. The gear type is primarily long line.
# check species name and filter out dolphinfish
unique(hs$common_name)[grep("dolphin", unique(hs$common_name))][1] "Common dolphinfish"hs <- hs[which(hs$common_name == "Common dolphinfish"), ]
# look at the categories within each column of data
apply(hs[1:13], 2, table)$area_name
Atlantic, Northwest Atlantic, Western Central
113 160
$area_type
high_seas
273
$year
1985 1986 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
1 1 1 1 1 1 1 6 4 1 5 10 5 8 10 12
2013 2014 2015 2016 2017 2018 2019
24 22 30 33 33 29 34
$scientific_name
Coryphaena hippurus
273
$common_name
Common dolphinfish
273
$functional_group
Large pelagics (>=90 cm)
273
$commercial_group
Perch-likes
273
$fishing_entity
Barbados Belize
6 3
Brazil Canada
19 18
Guyana Panama
1 1
Saint Lucia Saint Vincent & the Grenadines
5 28
Spain Suriname
80 1
Trinidad & Tobago USA
14 74
Venezuela
23
$fishing_sector
Industrial
273
$catch_type
Discards Landings
5 268
$reporting_status
Reported Unreported
253 20
$gear_type
artisanal fishing gear bottom trawl gillnet
7 9 11
hand lines longline mixed gear
17 187 1
pole and line pots or traps unknown by source
19 9 13
$end_use_type
Direct human consumption Fishmeal and fish oil
5 194 37
Other
37 # convert from metric tonnes to pounds
hs$pounds <- hs$tonnes * 1000 * 2.20462
# produce barplot of key data fields
par(mar = c(3, 20, 3, 1), mfrow = c(3, 2), mex = 0.5)
barplot(tapply(hs$pounds/10^6, hs$area_name, sum, na.rm = T), las = 2, xlab = "millions of pounds", main = "Area name", horiz = T)
#barplot(tapply(hs$pounds/10^6, hs$area_type, sum, na.rm = T), las = 2, ylab = "millions of pounds", main = "Area type")
#barplot(tapply(hs$pounds/10^6, hs$year, sum, na.rm = T), las = 2, ylab = "millions of pounds", main = "Year")
barplot(tapply(hs$pounds/10^6, hs$fishing_entity, sum, na.rm = T), las = 2, xlab = "millions of pounds", main = "Fishing entity", horiz = T)
barplot(tapply(hs$pounds/10^6, hs$fishing_sector, sum, na.rm = T), las = 2, xlab = "millions of pounds", main = "Fishing sector", horiz = T)
barplot(tapply(hs$pounds/10^6, hs$catch_type, sum, na.rm = T), las = 2, xlab = "millions of pounds", main = "Catch type", horiz = T)
barplot(tapply(hs$pounds/10^6, hs$reporting_status, sum, na.rm = T), las = 2, xlab = "millions of pounds", main = "Reporting status", horiz = T)
barplot(tapply(hs$pounds/10^6, hs$gear_type, sum, na.rm = T), las = 2, xlab = "millions of pounds", main = "Gear type", horiz = T)
round(tapply(hs$pounds, hs$catch_type, sum, na.rm = T) / sum(hs$pounds) * 100, 2) # << 1% are discardsDiscards Landings
0.03 99.97 round(tapply(hs$pounds, hs$reporting_status, sum, na.rm = T) / sum(hs$pounds) * 100, 2) # << 1% are unreported Reported Unreported
99.97 0.03 We plot the annual landings by fishing entity to look at trends over time.
hs$fishing_entity <- as.factor(hs$fishing_entity)
hs$fishing_entity <- droplevels(hs$fishing_entity)
par(mar = c(5, 5, 3, 1), mfrow = c(1, 1))
tab <- tapply(hs$pounds, list(hs$year, hs$fishing_entity), sum, na.rm = T)
matplot(as.numeric(rownames(tab)), tab/10^6, type = "l", col = glasbey(ncol(tab)), lty = 1, lwd = 2, xlab = "year", ylab = "total landings (millions of pounds)",
main = "High seas Western Central Atlantic dolphin landings")
legend("topleft", colnames(tab), col = glasbey(ncol(tab)), lty = 1, lwd = 2, cex = 0.8, bty = "n")
Finally we output the subset of high seas dolphin landings for the two areas. Because we have already accounted for the USA landings in our national reporting and are aiming to only summarize international landings here, we filter out the USA landings from the Sea Around Us database.
hs <- hs[which(hs$fishing_entity != "USA"), ]
hs$year <- factor(hs$year, levels = 1950:2019)
tab <- tapply(hs$pounds, list(hs$year, hs$area_name), sum, na.rm = T)
tab[is.na(tab)] <- 0
tail(tab, 40) Atlantic, Northwest Atlantic, Western Central
1980 0.000 0.000
1981 0.000 0.000
1982 0.000 0.000
1983 0.000 0.000
1984 0.000 0.000
1985 0.000 14037.129
1986 0.000 8748.399
1987 0.000 0.000
1988 0.000 0.000
1989 0.000 0.000
1990 0.000 0.000
1991 0.000 0.000
1992 0.000 0.000
1993 0.000 0.000
1994 0.000 0.000
1995 0.000 0.000
1996 0.000 0.000
1997 0.000 0.000
1998 0.000 0.000
1999 0.000 43168.296
2000 0.000 30284.627
2001 0.000 164272.223
2002 0.000 22147.477
2003 0.000 137103.181
2004 0.000 66058.258
2005 0.000 115171.863
2006 0.000 170522.101
2007 14687.527 330838.620
2008 28552.636 340863.719
2009 18719.121 662720.661
2010 4510.085 232093.837
2011 21589.424 300498.224
2012 72398.812 1059210.392
2013 34741.604 2471791.396
2014 27922.953 1247462.983
2015 90313.638 710059.125
2016 15319.433 1094893.891
2017 34806.933 1374259.739
2018 29328.358 885814.754
2019 21870.621 632747.157# write summary to merge with EEZ data
write.csv(tab, file = "data/SAU/high_seas_by_year.csv")The international landings are only reported as annual landings and do not have any month or season associated with them. However, the MSE operating model requires seasonal landings. The majority of the international landings are caught by longline, so we will assume that the fleet distribution of the U.S. pelagic longline is representative of the seasonality of the international high seas fleets. Now we parse the annual landings according to the seasonality of the U.S. fleet.
At the time of analysis, the SAU database only contained landings to 2019, so we interpolated the 2019 values to cover years 2020 - 2022.
# read in the percentage of the landings by area and year-quarter combinations from the U.S. PLL
load("data/outputs/per_PLLcatch_by_area_yearquarter.RData")
tab <- tab[which(rownames(tab) >= 1997), ] # PLL data are only available from 1997 on
#tab[nrow(tab), ]
tab <- rbind(tab, tab[nrow(tab), ], tab[nrow(tab), ], tab[nrow(tab), ]) # interpolate 2019 landings for 2020-2022
tail(tab) Atlantic, Northwest Atlantic, Western Central
2017 34806.93 1374259.7
2018 29328.36 885814.8
2019 21870.62 632747.2
21870.62 632747.2
21870.62 632747.2
21870.62 632747.2NED <- percatch[, , 1] # create matrices for final landings data
NCA <- percatch[, , 1]
for (i in 1:ncol(percatch)) {
NED[, i] <- percatch[, i, 7] * tab[i, 1] # these are the percentages for the NED (Atlantic Northwest)
NCA[, i] <- percatch[, i, 1] * tab[i, 2] # these are the percentages for the NCA (Western Central Atlantic)
}
# plot the parsed out data by year-quarter
yrs <- sort(rep(as.numeric(colnames(NCA)), 4))
qrt <- rep(1:4, length(yrs)/4)
matplot(yrs + qrt/4, cbind(matrix(NED), matrix(NCA))/10^6, lty = 1, lwd = 2,
col = 2:3, type = "l", main = "International high seas landings by quarter",
xlab = "", ylab = "total landings (millions of pounds)")
legend("topleft", c("NED", "NCA"), lwd = 2, col = 2:3, bty = "n")
# save NED data in this format because it needs to be combined with Canadian EEZ landings
save(NED, file = "data/outputs/NED_year_quarter.RData")Finally, we format the data in the format for input into the operating model. We only include NCA here, as the NED high seas landings must be combined with international landings in territorial waters of Canada and France in the NED region.
# format for operating model
flt <- rep("Intl", length(yrs))
are <- rep("NCA", length(yrs))
findat <- data.frame(cbind(yrs, qrt, flt, are, matrix(NCA)))
names(findat) <- c("Year", "Quarter", "Fleet", "Area", "Catch_lbs")
findat$Catch_lbs <- as.numeric(findat$Catch_lbs)
head(findat) Year Quarter Fleet Area Catch_lbs
1 1997 1 Intl NCA 0
2 1997 2 Intl NCA 0
3 1997 3 Intl NCA 0
4 1997 4 Intl NCA 0
5 1998 1 Intl NCA 0
6 1998 2 Intl NCA 0#plot(findat$Catch_lbs, type = "l")
#plot(findat$Catch_lbs ~ factor(findat$Area))
#plot(findat$Catch_lbs ~ factor(findat$Quarter))
#plot(findat$Catch_lbs ~ factor(findat$Year))
#apply(findat, 2, table)
#barplot(tapply(findat$Catch_lbs, list(findat$Quarter, findat$Area), sum, na.rm = T), beside = T, col = rainbow(4))
write.csv(findat, file = "data/FINAL_files/NCA_Intl_TomFormat.csv", row.names = FALSE)