script to process Copernicus Multi Observation Global Ocean 3D Temperature Salinity

multiobs_glo_phy_nc_to_csv.R

# =======================================================================================================================
# Antonio Olinto Avila-da-Silva, Instituto de Pesca, Brasil
# version 2024/06/04
# https://gist.github.com/aolinto/3828be642d06857788500d51f114eb49
# script to process Multi Observation Global Ocean 3D Temperature Salinity Height Geostrophic Current and MLD
# product MULTIOBS_GLO_PHY_TSUV_3D_MYNRT_015_012
# https://doi.org/10.48670/moi-00052
# nc file downloaded from
# https://data.marine.copernicus.eu/product/MULTIOBS_GLO_PHY_TSUV_3D_MYNRT_015_012/services
# dataset-armor-3d-rep-monthly
# using the phyton script copernicus_mod_glo_phy.py
# https://gist.github.com/aolinto/d14d72711a648a34ab5b31a8c26d6360
# the script will open nc file, organize data in a data matrix and write it into a single csv file
# =======================================================================================================================

# load libraries
# --------------

# install.packages(c("ncdf4"))
# ncdf4 needs libnetcdf-dev netcdf-bin in Linux
library(ncdf4)

# set workspace
# --------------

# list and remove objects from workspace
ls()
rm(list = ls())
options("width"=226) # 100,226

# set working directory
setwd("...")   # indicate the path to the nc file

# import nc file
# --------------

list.files(pattern = "^dataset-armor.*\\.nc$")
nc.file <- "dataset-armor-3d-rep-monthly_mlotst-so-to_....nc"
nc.data <- nc_open(nc.file)
nc.data

# check ncfile information
# ------------------------

# list of the nc variable names
var.names <- attributes(nc.data$var)$names
var.names

# variables info
ncatt_get(nc.data,var.names[1])
ncatt_get(nc.data,var.names[2])
ncatt_get(nc.data,var.names[3])

# list of the nc dimensions names
attributes(nc.data$dim)
nc.data$dim$longitude$len
nc.data$dim$latitude$len
nc.data$dim$depth$len
nc.data$dim$time$len

# dimensions info
ncatt_get(nc.data,"longitude")
ncvar_get(nc.data,"longitude")

ncatt_get(nc.data,"latitude")
ncvar_get(nc.data,"latitude")

ncatt_get(nc.data,"depth")
ncvar_get(nc.data,"depth")

ncatt_get(nc.data,"time")
ncvar_get(nc.data,"time")
as.Date(ncvar_get(nc.data,"time")/24, origin="1950-01-01")

# dimentions levels (depth, latitude, longitude, time)
longitude <- ncvar_get(nc.data,"longitude")
latitude <- ncvar_get(nc.data,"latitude")
depth <- ncvar_get(nc.data,"depth")
time <- ncvar_get(nc.data,"time")
date <- as.Date(time/24, origin="1950-01-01")

# dataframe per variable
#. mixed layer depth
arr.mlotst <- ncvar_get(nc.data,"mlotst",raw_datavals=F)
dim(arr.mlotst)
dimnames(arr.mlotst) <- list(lon = longitude, lat = latitude, time = time)
indices.mlotst <- expand.grid(lon = longitude, lat = latitude, time = time)
val.mlotst <- c(arr.mlotst)
length(val.mlotst)
dm.mlotst <- cbind(indices.mlotst, value = val.mlotst)
dim(dm.mlotst)
dm.mlotst$depth <- NA_integer_
dm.mlotst$cod_variable <- 1 # mlotst
dm.mlotst$cod_source <- 1
dm.mlotst$date <- as.Date(dm.mlotst$time/24, origin="1950-01-01")
dm.mlotst_final <- dm.mlotst[,c("cod_source","date","lon","lat","depth","cod_variable","value")]
dm.mlotst_final <- dm.mlotst_final[order(dm.mlotst_final$date,dm.mlotst_final$lon,dm.mlotst_final$lat),]
summary(dm.mlotst_final)
dm.mlotst_final <- dm.mlotst_final[!is.na(dm.mlotst_final$value ),]
head(dm.mlotst_final)
tail(dm.mlotst_final)

#. salinity
arr.so <- ncvar_get(nc.data,"so",raw_datavals=F)
dim(arr.so)
dimnames(arr.so) <- list(lon = longitude, lat = latitude, depth = depth, time = time)
indices.so <- expand.grid(lon = longitude, lat = latitude, depth = depth, time = time)
val.so <- c(arr.so)
length(val.so)
dm.so <- cbind(indices.so, value = val.so)
dim(dm.so)
dm.so$cod_variable <- 2 # so
dm.so$cod_source <- 1
dm.so$date <- as.Date(dm.so$time/24, origin="1950-01-01")
dm.so_final <- dm.so[,c("cod_source","date","lon","lat","depth","cod_variable","value")]
dm.so_final <- dm.so_final[order(dm.so_final$date,dm.so_final$lon,dm.so_final$lat),]
summary(dm.so_final)
dm.so_final <- dm.so_final[!is.na(dm.so_final$value ),]
head(dm.so_final)
tail(dm.so_final)

#. temperature
arr.to <- ncvar_get(nc.data,"to",raw_datavals=F)
dim(arr.to)
dimnames(arr.to) <- list(lon = longitude, lat = latitude, depth = depth, time = time)
indices.to <- expand.grid(lon = longitude, lat = latitude, depth = depth, time = time)
val.to <- c(arr.to)
length(val.to)
dm.to <- cbind(indices.to, value = val.to)
dim(dm.to)
dm.to$cod_variable <- 3 # to
dm.to$cod_source <- 1
dm.to$date <- as.Date(dm.to$time/24, origin="1950-01-01")
dm.to_final <- dm.to[,c("cod_source","date","lon","lat","depth","cod_variable","value")]
dm.to_final <- dm.to_final[order(dm.to_final$date,dm.to_final$lon,dm.to_final$lat),]
summary(dm.to_final)
dm.to_final <- dm.to_final[!is.na(dm.to_final$value ),]
head(dm.to_final)
tail(dm.to_final)

# merge data
dm.armor_3d_rep_monthly <- rbind(dm.mlotst_final,dm.so_final,dm.to_final)
summary(dm.armor_3d_rep_monthly)
dim(dm.armor_3d_rep_monthly)

# save csv
write.csv(dm.armor_3d_rep_monthly,file=paste0(substr(nc.file,1,nchar(nc.file)-3),".csv"),row.names = FALSE)

# close
nc_close(nc.data)