FERRET FACILITY

26.1.5. FERRET FACILITY#

Ferret : a practical tool for fast visualisation on datarmor

this step is dedicated to basic git usage to manage properly your source code and (potentially) interact with croco’s developers the croco’s repository is so far hosted at Inria (https://gitlab.inria.fr)

  1. Load the appropriate module

    module avail
    module load ferret/7.1__64b
    

    Launch it by typing

    ferret
    
  2. Load your netcdf dataset (yes? is the usual prompt)

    yes? use data.nc
    

    or

    yes? use "/home6/datawork/login/Simulation/data.nc"
    
  3. Visualise the data structure

    yes? show data
    

    you will get a list of all the variables contained in the fill loaded and their dimensions :

    • i : designate the x dimension

    • j : designate the y dimension

    • k : designate the vertical dimension

    • l : designate the temporal dimension

  4. List the numerical values of a section of a given variable :

    From now on let’s consider the variable temp which gets four dimensions (time + x,y,z).

    yes? list /i=10/j=10/k=40 temp
    

    This will list all the numerical data at the level 40, i=10, j=10 of the variable temp.

  5. Plot a one dimensionnal feature by fixing n-1 of the variable dimension number (n).

    yes? plot /i=10/j=10/k=40 temp
    

    this will plot the time series of the variable temp.

    yes? plot /i=10/j=10/l=4 temp
    

    this will plot the vertical profile of the variable temp at time l=4.

    yes? plot /i=10/j=10/l=4 temp
    yes? plot /i=10/j=10/l=40/ove temp
    

    the same as the previous but with superimposition of two profile at two different instants (l=10 and l=40)

  6. Plot a two dimensionnal features by fixing n-2 of the variable dimension number (n).

    yes? plot /i=10/j=10 temp
    

    This will plot a Hovmuller diagram (time vs z) of the variable temp.

    yes? plot /k=40/j=10 temp
    

    In case, k=40 designate the surface layer, this will plot a hovemuller diagram along all longitudes vs time.

    yes? plot /k=40/j=10/lev=(10.,20.,1.) temp
    

    The same as the previous one but setting a color bar that extends from 10 to 20 with bins of 1.

    yes? plot /k=40/j=10/lev=(0)(10.,20.,1.)(30) temp
    

    The same as the previous one but extending the first and last color class respectively down to 0 and up to 30.