echam

install

git clone --recurse-submodules https://github.com/chrisdane/echam

run

Go into the repo and create an example namelist:

cd echam
cat <<EOF >> example_namelist.post.r
# load defaults
repopath <- getwd() # change here if necessary
repopath <- normalizePath(repopath, mustWork=T) # error if not found
source(paste0(repopath, "/namelist.general.post.r"))

# my setting
models <- "CanESM5" # ocean: nemo3.4.1: native: orca1 tripolar; cdo: curvilinear; fldint ok
datapaths <- "/mnt/lustre02/work/ik1017/CMIP6/data/CMIP6/CMIP/CCCma/CanESM5/piControl/r1i1p1f1/Omon/fgco2/gn/v20190429"
fpatterns <- "fgco2_Omon_CanESM5_piControl_r1i1p1f1_gn_<YYYY_from>01-<YYYY_to>12.nc"
fvarnames <- "fgco2"
prefixes <- "CanESM5_piControl_r1i1p1f1"
modes <- "fldint"
froms <- 6150
tos <- 6200

# run
source(paste0(repopath, "/post_echam.r"))
EOF

Install or load R if necessary:

module load r # or R

Running

source("example_namelist.post.r")

in an active R session or

nohup Rscript example_namelist.post.r > example.log 2>&1 &

in background with will calculate fldint from years 6150 to 6200 of variable fgco2 of all files found with pattern fpatterns in directory datapaths and save it as nc file named "prefixes....nc".

notes

Allowed special patterns in fpatterns are printed in log. modes can be anything from cdo. The mode "fldint" works as soon as cdo gridarea does not return an error.

Adding

cdo_after_calcs <- list(c("setunit,\"PgC yr-1\"", "mulc,365.25", "mulc,86400", "divc,1e12")) # kgC s-1 --> PgC yr-1

to the namelist adds the given cdo commands to the calculation.

Several settings can be used in one call via

# my settings
models <- c("model1", "model2")
datapaths <- c("path1", "path2")
fpatterns <- c("fpattern1", "fpattern2")
fvarnames <- c("variable1", "variable2")
prefixes <- c("prefix1", "prefix2")
modes <- c("mode1", "mode2")
froms <- c(from1, from2)
tos <- c(to1, to2)

in the namelist.

Check the already existing namelist.post.r for examples.