Denisse Fierro Arcos 2023-11-01
- Introduction
- Loading libraries
- Downloading FishMIP regional models shapefile from THREDDS
- Loading FishMIP regional models shapefile
- Plotting map
Now that we have a single shapefile with all the FishMIP regional models, we will create a map that can be used in any publication.
Here, we will show how to access and download the FishMIP regional models shapefile available from the THREDDS server.
#Connection to THREDDS
library(thredds)
#Downloading data
library(curl)## Using libcurl 7.81.0 with GnuTLS/3.7.3
#Manipulating data
library(dplyr)##
## Attaching package: 'dplyr'
## The following objects are masked from 'package:stats':
##
## filter, lag
## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, union
library(stringr)
#Spatial data
library(sf)## Linking to GEOS 3.10.2, GDAL 3.4.1, PROJ 8.2.1; sf_use_s2() is TRUE
#Base map
library(rnaturalearth)
#Plotting maps
library(ggplot2)
library(cowplot)
library(RColorBrewer)We will use the HTTPServer service to download the shapefile files
which are available as a zip folder.
Note that you only need to download shapefiles once.
#Defining domain name for THREDDS server
domain <- "http://portal.sf.utas.edu.au"
#Defining location of regional shapefile in THREDDS server
fishmip_reg_cat <- paste0(domain, "/thredds/catalog/gem/fishmip/",
"FishMIP_regions/catalog.xml")
#Getting catalog items available in server
fishmip_reg <- CatalogNode$new(fishmip_reg_cat, prefix = "thredds")
#Getting base url for HTTPServer downloads
base_http <- paste0(domain, fishmip_reg$list_services()$http[["base"]])
#Getting names of datasets available for download
ds_name <- fishmip_reg$get_dataset_names()
#Getting complete download URL
url <- paste0(base_http, fishmip_reg$get_datasets()[[ds_name]]$url)
#Downloading zip file
curl_download(url, destfile = "FishMIP_regions.zip")
#Unzipping folder - storing files in Output folder
out_folder <- "../outputs"
#If folder does not exist, create one
if(!dir.exists(out_folder)){
dir.create(out_folder, recursive = T)}
unzip("FishMIP_regions.zip", exdir = out_folder)
#Delete zip folder
unlink("../FishMIP_regions.zip")
#Removing all variables from environment
rm(list = ls())Now that we have downloaded the merged shapefiles, we can load it to R
and create a map. We will exclude the Southern Ocean region.
fishmip_reg <- file.path("/rd/gem/private/shared_resources",
"FishMIP_regional_models/FishMIP_regional_models.shp") |>
read_sf() |>
filter(region != "Southern Ocean")
#Checking results
fishmip_reg## Simple feature collection with 36 features and 3 fields
## Geometry type: MULTIPOLYGON
## Dimension: XY
## Bounding box: xmin: -180 ymin: -78.74234 xmax: 180 ymax: 83.66553
## Geodetic CRS: WGS 84
## # A tibble: 36 × 4
## region models nmbr_md geometry
## * <chr> <chr> <int> <MULTIPOLYGON [°]>
## 1 Baltic Sea EwE EwE 1 (((23.5 64.5, 23.5 64.05…
## 2 Baltic Sea Mizer Mizer 1 (((20.5847 54.93254, 20.…
## 3 Brazil NE EcoSpace 1 (((-36.94482 -4.549031, …
## 4 Central North Pacific ECOTRAN 1 (((-140 10, -180 10, -18…
## 5 Central South Pacific Mizer 1 (((-168.2779 -22.09492, …
## 6 Chatham Rise Atlantis, EwE, Mi… 3 (((172 -45.33333, 172 -4…
## 7 Cook Strait EwE 1 (((174 -42, 174 -41, 175…
## 8 East Antarctica Atlantis Atlantis 1 (((81.87166 -53.63183, 1…
## 9 East Antarctica EwE EwE 1 (((80 -64, 80 -68.00585,…
## 10 East Bass Strait EwE 1 (((150.5 -36, 150.5 -39,…
## # ℹ 26 more rows
Since we removed the Southern Ocean region, we have 36 polygons left. However, we will need to apply some changes to our datasets to create publication ready maps.
We will use a Robinson projection for our map, so we need to reproject our shapefile. A reprojection also needs to be applied to our base world map.
We will also create a an original colour map to plot the FishMIP regions
using a combination of RColorBrewer palettes. We will add this
information to the reprojected regional shapefile for plotting.
#Define Robinson projection to be applied to shapefile and basemap
rob_proj <- paste0("+proj=robin +lon_0=0 +x_0=0 +y_0=0 +datum=WGS84 +units=m ",
"+no_defs +type=crs")
#Load world base map
world <- ne_countries(scale = "medium", returnclass = "sf") |>
#Ensuring any polygons crossing international dateline are correctly plotted
st_wrap_dateline() |>
#Reprojecting
st_transform(rob_proj)
#Creating a colour palette by merging colour brewer palettes
pal <- c("#ee3377", "#364b9a", brewer.pal(12, "Paired"), brewer.pal(10, "Set3"),
"#ee7733", "#332288", "#ccddaa", "#009988", "#993d9a",
"#117733", "#997700", "#aa4499", brewer.pal(3, "Dark2"), "#ea4648")
#Reprojecting FishMIP regions
fishmip_reg_rob <- fishmip_reg |>
#Adding colour palette
mutate(fill = pal) |>
st_wrap_dateline() |>
st_transform(rob_proj)Since there is some overlap in the spatial coverage for regional models, we will create maps showing more detail in the overlapping regions.
First, we will plot all regional models. We will use this as the main map.
#Plotting
reg <- fishmip_reg_rob |>
#Plotting
ggplot()+
#Using colours of our tailored made colourmap
geom_sf(aes(fill = region), alpha = 0.5)+
geom_sf(data = world, fill = "#f9f9f5", show.legend = F)+
scale_fill_manual(values = fishmip_reg_rob$fill)+
theme_bw()+
#Move legend to bottom
theme(legend.position = "bottom", legend.title = element_blank(),
legend.margin = margin(0, 0, 0, 0),
legend.box.margin = margin(0, 0, 0, 0),
legend.text = element_text(margin = margin(r = 5)),
panel.border = element_rect(colour = NA))+
#Split legend into four columns
guides(fill = guide_legend(ncol = 4))
#Checking map so far
reg
Our initial maps shows all FishMIP regions, but does not look great yet. We will use this as a guide to identify the areas where we will create the smaller maps. We will expand the following areas:
- Europe
- Australia and New Zealand
- East Antarctica
We will use the base map above to create these smaller maps.
europe <- reg+
#We will increase transparency to see overlapping areas better
geom_sf(inherit.aes = T, aes(alpha = 0.6))+
#Remove background
theme_bw()+
#Remove legend
theme(legend.position = "none")+
#Add world base map
geom_sf(data = world, fill = "#f9f9f5")+
#Focus on Europe. Note that we use coordinates in meters because data is
#reprojected
lims(x = c(-741458, 3251458), y = c(3256289, 6725154))+
#Add a border to map so it is easily identifiable
theme(panel.border = element_rect(colour = "#0077bb", linewidth = 2),
axis.text = element_blank(), axis.ticks = element_blank(),
plot.margin = margin(0, 0, 0, 0, unit = "cm"))
#Create a shapefile with map limits
eu_box <- st_bbox(c(xmin = -751458, xmax = 3351458, ymax = 6825154,
ymin = 3156289), crs = rob_proj) |>
st_as_sfc()
#Check map of Europe
europe
Because one of the regions in New Zealand crosses the international date line, we will show this area in an unprojected map.
au_nz <- fishmip_reg |>
#Switching to 0-360 degrees longitude
st_shift_longitude() |>
ggplot()+
#We will increase transparency to see overlapping areas better
geom_sf(aes(fill = region), alpha = 0.6)+
#Add world base map
geom_sf(inherit.aes = F, data = world, fill = "#f9f9f5", show.legend = F)+
#Focus on Australia and New Zealand
lims(x = c(115, 200), y = c(-50, -23))+
scale_fill_manual(values = fishmip_reg_rob$fill)+
#Remove background
theme_bw()+
#Remove legend
theme(legend.position = "none")
#Adding Bass Strait on top so it is easier to see
#Extract Bass Strait region
Tas <- fishmip_reg |>
filter(str_detect(region, "Bass")) |>
mutate(fill = "#CAB2D6")
#Plot Bass Strait over the original AU/NZ map
au_nz <- au_nz+
geom_sf(inherit.aes = F, data = Tas, aes(fill = region), alpha = 0.5)+
scale_fill_manual(values = fishmip_reg_rob$fill)+
#Remove background
theme_bw()+
#Remove legend
theme(legend.position = "none")+
#Add a border to map so it is easily identifiable
theme(panel.border = element_rect(colour = "#999933", linewidth = 2),
axis.text = element_blank(), axis.ticks = element_blank(),
plot.margin = margin(0, 0, 0, 0, unit = "cm"))## Scale for fill is already present.
## Adding another scale for fill, which will replace the existing scale.
#Create a shapefile with map limits
au_nz_box <- st_union(st_bbox(c(xmin = 115, xmax = 180, ymax = -23, ymin = -50),
crs = 4326) |> st_as_sfc(),
st_bbox(c(xmin = -180, xmax = 200, ymax = -23, ymin = -50),
crs = 4326) |> st_as_sfc()) |>
#Projecting to Robinson
st_transform(rob_proj)
#Check map
au_nz
so <- reg+
#We will increase transparency to see overlapping areas better
geom_sf(inherit.aes = T, aes(alpha = 0.5))+
#Remove background
theme_bw()+
#Remove legend
theme(legend.position = "none")+
#Add world base map
geom_sf(data = world, fill = "#f9f9f5")+
#Focus on Australia and New Zealand
lims(x = c(1900000, 8800000), y = c(-7525154, -4712577))+
#Add a border to map so it is easily identifiable
theme(panel.border = element_rect(colour = "#332288", linewidth = 2),
axis.text = element_blank(), axis.ticks = element_blank(),
plot.margin = margin(0, 0, 0, 0, unit = "cm"))
#Create a shapefile with map limits
so_box <- st_bbox(c(xmin = 1800000, xmax = 8900000, ymax = -4612577, ymin = -7625154),
crs = rob_proj) |>
st_as_sfc()
#Check map
so
main <- reg+
geom_sf(data = eu_box, color = "#0077bb", fill = NA, linewidth = 0.5)+
geom_sf(data = au_nz_box, color = "#999933", fill = NA, linewidth = 0.5)+
geom_sf(data = so_box, color = "#332288", fill = NA, linewidth = 0.5)
main
#Merging main map and insets
final_map <- ggdraw(main)+
draw_plot(au_nz, x = .805, y = 0.675, width = 0.185, height = 0.3)+
draw_plot(so, x = .815, y = 0.38, width = 0.175, height = 0.30)+
draw_plot(europe, x = .75, y = 0.04, width = 0.3, height = 0.3)
final_map <- final_map+
theme(plot.margin = margin(l = -5, unit = "cm"))
#Checking final result
final_map
The map above may not look great in your screen, but once it is saved to your local machine, it will plot correctly. However, you can change the code here to fit it to your needs.
The final map can be saved to your local machine. We will save it as vector image so it still looks good even when expanded. You can also find a copy of this image in the FishMIP THREDDS server.
#Saving final map
ggsave("../outputs/FishMIP_regional_models_insets.pdf", final_map,
device = "pdf", width = 25, height = 15, units = "cm")
#Saving main map with just regions and no insets
ggsave("../outputs/FishMIP_regional_models.pdf", reg, device = "pdf",
width = 9)