rfesom

With this R tool you can read/post-process/plot FESOM version 1.x (Danilov et al. 2004; Wang et al. 2014) data.

Please note that the tool was only tested on linux so far and bugs do exist!

Table of Contents

Install
Get R
Demo
- Demo1
- Demo2
- Demo3
- Demo4
Modify the runscript
Help
Contribute
References
Appendix: Available variables

Install

Clone this repo with

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

Get R

Download and install R or load it via module:

$ module load r # you can check the available modules with 'module avail'

Demo

Go to cd rfesom and run either in an active R session with

$ R # from terminal
source("runscripts/demo.run.r") # from within R

or from terminal with

$ Rcript runscripts/demo.run.r

$ nohup Rscript runscripts/demo.run.r > demo.run.log 2>&1 &

so that the program runs in background and you can close the connection when running a long job.

Demo1

Saves fesoms mesh resolution in km on regular coordinates as netcdf and saves a spatial plot:

$ ncdump -h demo__resolutionkm_ltm_area_lsea_rectangular_regular_dx0.250_dy0.250.nc 
dimensions:
        lon = 96 ;
        lat = 48 ;
variables:
        double resolutionkm(lat, lon) ;
               resolutionkm:units = "km" ;
               resolutionkm:_FillValue = NaN ;

Demo2

Saves 1) 1948 mean, 2) standard deviation (sd) and 3) 12 (monthly) records of ssh on regular coordinates as netcdf and saves a spatial plot of the temporal mean and sd fields.

$ ncdump -h demo__monthly_ssh_ltm_area_Jan-Dec_1948_mean_lsea_rectangular_regular_dx0.250_dy0.250.nc
dimensions:
        lon = 96 ;
        lat = 48 ;
        double ssh(lat, lon) ;
               ssh:units = "m" ;
               ssh:_FillValue = NaN ;
        double ssh_sd(lat, lon) ;
               ssh_sd:units = "m" ;
               ssh_sd:_FillValue = NaN ;
$ ncdump -h demo__monthly_ssh_transient_area_Jan-Dec_1948_mean_lsea_rectangular_regular_dx0.250_dy0.250.nc 
dimensions:
        time = 12 ;
        lon = 96 ;
        lat = 48 ;
variables:
        double ssh(time, lat, lon) ;
               ssh:units = "m" ;
               ssh:_FillValue = NaN ;
               ssh:long_name = "Sea Surface Height" ;

Demo3

Saves potential temperature field mean in area "lsea" averaged between 0 and 100 meters depth as netcdf and saves a spatial plot of the temporal and depth mean.

$ ncdump -h demo__monthly_temp_transient_mean_Jan-Dec_1948_mean_0-100m_lsea_rectangular.nc
dimensions:
        time = 12 ;
variables:
        double temp_mean(time) ;
               temp_mean:units = "degC" ;
               temp_mean:_FillValue = NaN ;
               temp_mean:long_name = "Potential Temperature" ;

Demo4

Saves potential temperature field mean in area "lsea" as a function of depth as netcdf and saves a spatial plot of the temporal and depth mean.

$ ncdump -h demo__monthly_temp_transient_depth_Jan-Dec_1948_mean_0-3600m_lsea_rectangular.nc
dimensions:
        time = 12 ;
        depth = 31 ;
variables:
        double temp_depth(depth, time) ;
               temp_depth:units = "degC" ;
               temp_depth:_FillValue = NaN ;
               temp_depth:long_name = "Potential Temperature" ;

Modify the runscript

Different postprocessing options for different variables (see appendix here and namelist.var.r) are available, see namelist.config.r: todo: insert table

Using this tool with your own modified runscript works best if

you make a copy of runscripts/myrunscript.r into e.g. rfesom or whereever you want
by that, any future git pulls will not change your individualized runscript

Help

R syntax basics

any help text for a function can be obtained within R by ?<functionname>, e.g. ?sum
R counts from 1, not zero
index syntax is [], not (). So mat2[1,2] yields the 1st row and 2nd column element of the 2d-array mat2 and mat3[1:2,,c(4,6,8)] all entries of the 2nd dimension of the 3d-array mat3 in the 1st and 2nd row and the 4th, 6th and 8th entries of the 3rd dimension.
booleans TRUE and FALSE can be abbreviated with T and F
variable assignment symbol is <-, e.g. a <- 1 (a = 1 works as well)
'not equal' condition is !=
By default, subsetting in R removes redundant dimensions, i.e. the effect of matlabs squeeze() is applied automatically. For example

a <- array(1:6, c(3,2)); message("a:"); a; message("dim of 2d-array:"); dim(a); message("dim of 1d-subset of 2d-array:"); dim(a[1,]); message("length of 2d-array:"); length(a); message("length of 1d-subset of 2d-array:"); length(a[1,])
a:
     [,1] [,2]
[1,]    1    4
[2,]    2    5
[3,]    3    6
dim of 2d-array:
[1] 3 2
dim of 1d-subset of 2d-array:
NULL
length of 2d-array:
[1] 6
length of 1d-subset of 2d-array:
[1] 2

However, rfesom uses the 'matlab' way of subsetting:

a <- array(1:6, c(3,2)); message("a:"); a; message("dim of 2d-array:"); dim(a); message("dim of 1d-subset of 2d-array:"); dim(a[1,]); message("length of 2d-array:"); length(a); message("length of 1d-subset of 2d-array:"); length(a[1,])
a:
     [,1] [,2]
[1,]    1    4
[2,]    2    5
[3,]    3    6
dim of 2d-array:
[1] 3 2
dim of 1d-subset of 2d-array:
[1] 1 2
length of 2d-array:
[1] 6
length of 1d-subset of 2d-array:
[1] 2

For further information check the drop argument of the the [ function: ?`[`.

elements from list or data.frame objects can be accessed via the $ notation:

list1 <- list()
list1$arr1 <- array(1:6, c(2, 3))
list1$char1 <- c("a", "b", "c")
list1$[tabtab]
st1$arr1   list1$char1

Installing new R packages (= libraries)

Within R, you can install the ncdf4 package with

install.packages("ncdf4")

or with

install.packages("ncdf4", lib="/my/own/package/directory")

if you want to set a path where the package should be installed. The default package installation path is the first entry of .libPaths(), i.e. by default the argument lib=.libPaths()[1].

Missing header or config files

Some packages may need special header files or libraries that can be provided in an active R session as follows:

package: rgdal

gdal_path <- "/sw/rhel6-x64/gdal-2.1.3-gcc48"
proj4_path <- "/sw/rhel6-x64/graphics/proj4-4.9.3-gcc48"
pkg_libs <- paste0("-Wl,-rpath,", gdal_path, "/lib:", proj4_path, "/lib")
install.packages("rgdal", 
                 configure.args=paste0("--with-gdal-config=", gdal_path, "/bin/gdal-config 
                                        --with-proj-include=", proj4_path, "/include 
                                        --with-proj-lib=", proj4_path, "/lib 
                                        PKG_LIBS=", pkg_libs))

package: proj4

gdal_path <- "/sw/rhel6-x64/gdal-2.1.3-gcc48"
proj4_path <- "/sw/rhel6-x64/graphics/proj4-4.9.3-gcc48"
pkg_libs <- paste0("-Wl,-rpath,", gdal_path, "/lib")
install.packages("proj4", 
                 configure.args=paste0("--with-proj-include=", proj4_path, "/include 
                                        --with-proj-lib=", proj4_path, "/lib 
                                        PKG_LIBS=", pkg_libs))

pakage: rgeos

export LD_LIBRARY_PATH=/sw/rhel6-x64/geos-3.6.1-gcc48/lib/:$LD_LIBRARY_PATH

package: units

export LD_LIBRARY_PATH=/sw/rhel6-x64/util/udunits-2.2.26-gcc64/lib:$LD_LIBRARY_PATH

Unable to load shared library

Package installation from source requires the same compiler that was used for building R. On a supercomputer, this sometimes raises a problem if compilers and/or R were loaded via the default module load gcc r command, which loads the current default version numbers which are not neccesarily compatible. Then, running the command given above for installing a package, a typical error looks like

libstdc++.so.6: version `CXXABI_1.3.8' not found

A solution to this is as follows. First, identify the R executable. Within R, run

file.path(R.home(), "bin", "exec", "R")
[1] "/sw/rhel6-x64/r/r-3.5.3-gcc48/lib64/R/bin/exec/R"

Then, in the shell, type

$ ldd /sw/rhel6-x64/r/r-3.5.3-gcc48/lib64/R/bin/exec/R
        linux-vdso.so.1 =>  (0x00007ffd249ee000)
        libR.so => not found
        libRblas.so => not found
        libgomp.so.1 => /sw/rhel6-x64/gcc/gcc-4.8.2/lib64/libgomp.so.1 (0x00002b5263631000)
        libpthread.so.0 => /lib64/libpthread.so.0 (0x00002b5263840000)
        libc.so.6 => /lib64/libc.so.6 (0x00002b5263a5d000)
        librt.so.1 => /lib64/librt.so.1 (0x00002b5263df1000)
        /lib64/ld-linux-x86-64.so.2 (0x000055c633768000)

Apparently, this R binary was build with gcc-4.8.2. Now, load the correct gcc version with

module purge
module load gcc/4.8.2 r

and rerun the package installation given above.

In order to not to run into this problem again and again, I defined the following alias in my .bashrc:

alias R='echo "module purge ..."; module purge; echo "module load gcc/4.8.2 r ..."; module load gcc/4.8.2 r; echo "module list ..."; module list; R --quiet'

Then, running R includes the following:

$ R
module purge ...
module load gcc/4.8.2 r ...
module list ...
Currently Loaded Modulefiles:
  1) gcc/4.8.2              2) r/3.5.3

Fesom help

cdo -P $(nproc) -remapycon,global_0.25 -setgrid,/pool/data/AWICM/FESOM1/MESHES/core/griddes.nc $fin $fout

Contribute

Run $ lib/./set_links.sh to set git hook links for automatic table of content updates in the README.md on git commit using the nobackup branch of the github-markdown-toc repo https://github.com/jordantrizz/github-markdown-toc forked from https://github.com/ekalinin/github-markdown-toc:

$ cat lib/set_links.sh
#!/bin/bash

echo
echo "******** run set_links.sh **********"
home=$(git rev-parse --show-toplevel)
echo "ln -s $home/lib/.pre-commit $home/.git/hooks/pre-commit"
ln -s $home/lib/.pre-commit $home/.git/hooks/pre-commit
echo "ln -s $home/lib/.post-commit $home/.git/hooks/post-commit"
ln -s $home/lib/.post-commit $home/.git/hooks/post-commit
echo "******** finish set_links.sh **********"
echo

References

Danilov, S., G. Kivman, and J. Schröter, 2004: A finite-element ocean model: principles and evaluation. Ocean Modelling, 6 (2), 125–150, doi:10.1016/S1463-5003(02)00063-X.

Wang, Q., S. Danilov, D. Sidorenko, R. Timmermann, C. Wekerle, X. Wang, T. Jung, and J. Schröter, 2014b: The Finite Element Sea Ice-Ocean Model (FESOM) v.1.4: formulation of an ocean general circulation model. Geoscientific Model Development, 7 (2), 663–693, doi:10.5194/gmd-7-663-2014.

Appendix: Available variables

This is an automated chapter generated by rfesom_varlist2readme.r based on the longname definitions in namelist.var.r.

Variable name
Sea Surface Temperature
Potential Temperature
Salinity
In Situ Density
Potential Density
In situ Buoyancy
Potential Buoyancy
Zonal Velocity
Meridional Velocity
Horizontal Velocity
Zonal Velocity Squared
Mean Zonal Velocity Squared
Meridional Velocity Squared
Mean Meridional Velocity Squared
Zonal Geostrophic Velocity
Meridional Geostrophic Velocity
Horizontal Geostrophic Velocity
Horizontal Eddy Momentum Flux
SGS Zonal Velocity
SGS Meridional Velocity
Horizontal SGS Velocity
Mean Zonal Advective Temperature Flux
Mean Meridional Advective Flux of Temperature
Mean Horizontal Advective Flux Temperature Flux
Total Horizontal Advective Temperature Flux
Eddy Zonal Temperature Flux
Eddy Meridional Temperature Flux
Eddy Horizontal Temperature Flux
Total SGS Zonal Temperature Flux
Total SGS Meridional Temperature Flux
Total Horizontal SGS Temperature Flux
Zonal Advective Flux of Salinity
Meridional Advective Flux of Salinity
Horizontal Advective Flux of Salinity
Zonal Eddy Salinity Flux
Meridional Eddy Salinity Flux
Norm of Horizontal Eddy Salinity Flux
SGS Zonal Salinity Flux
SGS Meridional Salinity Flux
Total Horizontal SGS Salinity Flux
Zonal in situ Density Flux
Meridional in situ Density Flux
Norm of Horizontal in situ Density Flux
Zonal Eddy in situ Density Flux
Meridional Eddy in situ Density Flux
Norm of Horizontal Eddy in situ Density Flux
Relative Vorticity
Relative Vorticity / f
Relative Vorticity / `\|`f`\|`
Horizontal Strain (normal part)
Horizontal Strain (shear part)
Horizontal Strain
Squared Relative Vorticity
Okubo-Weiss Parameter
Ertel Potential Vorticity
PV_bc
PV_vert
Gradient Richardson Number
Mean Kinetic Energy
Total Kinetic Energy
Eddy Kinetic Energy
Horizontal Reynolds Stress
Vertical Reynolds Stress
Kinetic Mean -> Kinetic Eddy Conversion
wb (Potential Mean -> Kinetic Mean Conversion)
Norm of horizontal mean buoyancy flux
Potential Mean -> Potential Eddy Conversion
wb (Potential Eddy -> Kinetic Eddy Conversion)
Norm of Horizontal Eddy Buoyancy Flux
Vertical Velocity
grad_h T
grad_h B
grad_h MLD
Sea Surface Height
Mixed Layer Depth
Buoyancy Frequency Squared
Barotropic wavespeed
Mode-m baroclinic gravity-wave speed
Mode-m baroclinic long rossby-wave speed
Horizontal velocity baroclinic m-mode
Vertical velocity baroclinic m-mode
Vertical Diffusivity
Horizontal Diffusivity
GM Thickness Diffusivity
Passive Tracer
Mean Wind Stress Energy
Eddy Wind Stress Energy
div_h(u_h rho)
div_h(u_hrho)
div_h(u_h b)
div_h(u_hb)
div_h(u_sgs_h b)
Divergence of mean horizontal temperature flux
Divergence of eddy horizontal temperature flux
div_h(u_h t)
Divergence of total horizontal SGS temperature flux
Divergence of mean horizontal SGS temperature flux
Divergence of eddy horizontal SGS temperature flux
grad_laplace_inv_div_h(u_h t)
Divergent part of depth-integrated eddy temperature flux (eq 4 JM02)
Zonal divergent part of depth-integrated eddy temperature flux (eq 5 JM02)
Meridional divergent part of depth-integrated eddy temperature flux (eq 5 JM02)
Divergent part of depth-integrated eddy temperature flux (eq 5 JM02)
Divergence of mean horizontal salt flux
Divergence of eddy horizontal salt flux
Divergence of total horizontal SGS salt flux
Divergence of mean horizontal SGS salt flux
Divergence of eddy horizontal SGS salt flux
dz(wrho)
dz(wb)
dz(wT)
dz(wS)
dz(wb)
Horizontal Buoyancy Diffusion
Vertical Density Diffusion
Vertical Buoyancy Diffusion
Vertical Temperature Diffusion
Vertical Salinity Diffusion
Isoneutral Slope x
Isoneutral Slope y
Isoneutral Slope
Isoneutral Slope Squared
Mean Bottom Stress Energy
Eddy Bottom Stress Energy
MOCw
MOCv
Air Temperature 2m
Runoff
Freshwater Flux
Air Specific Humidity
Atmosphere Shortwave Radiation
Atmosphere Longwave Radiation
Latent Heat Flux To Ocean
Sensible Heat Flux To Ocean
Net heat flux to ocean
Net freshwater flux to ocean
Wind Speed
Curl of Wind Speed
Meridional Wind Stress
Zonal Wind Stress
Norm of Wind Stress
Wind Stress Curl
Ekman Pumping
Total Wind Energy
Mean Wind Energy
Eddy Wind Energy
Virtual Salt
Saltinity Relaxation
Drag Coefficient
Temperature flux to ocean
Salt flux to ocean
Salt flux to ocean
Thermal density flux to ocean
Thermal density flux to ocean
Haline density flux to ocean
Haline density flux to ocean
Density flux to ocean
Density flux to ocean
Density flux to ocean
Thermal buoyancy flux to ocean
Thermal buoyancy flux to ocean
Haline buoyancy flux to ocean
Haline buoyancy flux to ocean
Buoyancy flux to ocean
Buoyancy flux to ocean
Buoyancy flux to ocean
Sea Ice Zonal Velocity
Sea Ice Meridional Velocity
Horizontal Ice Velocity
Sea Ice Concentration
Sea Ice Thickness
Sea Ice Extent
Sea Ice Volume
Snow Thickness
Growth Rate of eff. Ice Thickness
Transport
Bathymetry
Norm of horizontal bathymetry gradient
Scalar product of horizontal velocity and horizontal bathymetry gradient times -1
coriolis parameter over depth
Resolution
Resolution
Mesh Area
First Barolinic Rossby Radius of Deformation

chrisdane / rfesom