The following sample session will give you a feeling for how to use
the basic capabilities of GrADS.  You will need the data file
'model.dat' on your system.  This sample session takes
about 30 minutes to run through. 

If you have any problems with running or using GrADs, send me 
email:  doty@cola.umd.edu.

This data file is described by the data descriptor file 'model.ctl'.
You may want to look at this file before continuing.  The data 
descriptor file describes the actual data file, which in the case
contains 5 days of global grids that are 72x46 elements in size.

To start up GrADS, enter:

     grads

If grads is not in your current directory, or if it is not in your 
PATH somewhere, you may need to enter the full pathname, ie:

     /usr/homes/smith/grads/grads

GrADS will prompt you with a landscape vs. portrait question; just
press enter.  At this point a graphics output window should open on your
console.  You may wish to move or resize this window.  Keep in mind that
you will be entering GrADS commands from the window where you first
started GrADS -- this window will need to be made the 'active' window
and you will not want to entirely cover that window with the grahpics
output window.  

In the text window (where you started grads from), you should now see
a prompt:  ga>    You will enter GrADS commands at this prompt and see
the results displayed in the graphics output window.  

The first command you will enter is:                               

    open model.ctl

You may want to see what is in this file, so enter:

    query file

One of the available variable is called ps, for surface pressure.  We
can display this variable by entering:

    d ps

d is short for display.   You will note that by default, GrADS will 
display an X, Y plot at the first time and at the lowest level in the
data set.  

Now you will enter commands to alter the 'dimension environment'.  The
display command (and implicitly, the access, operation, and output
of the data) will do things with respect to the current dimension 
environment.  You control the dimension environment by entering
set commands:

    clear                       clear the display 
    set lon -90                 set longitude fixed
    set lat 40                  set latitude fixed
    set lev 500                 set level fixed
    set t 1                     set time fixed
    d z                         display a variable

In the above sequence of commands, we have set all four GrADS dimensions
to a single value.  When we set a dimension to a single value, we say 
that dimension is fixed.  Since all the dimensions are fixed, when we
display a variable we get a single value, in this case the value at
the location 90W, 40N, 500mb, and the 1st time in the data set. 

If we now enter:

    set lon -180 0                 X is now a varying dimension
    d z

We have set the X dimension, or logitude, to vary.  We have done this
by entering two values on the set command.  We now have one varying
dimension (the other dimensions are still fixed), and when we display 
a variable we get a line grahp, in this case a graph of 500mb Heights
at 40N.  

Now enter:

    clear
    set lat 0 90
    d z

We now have two varying dimensions, so by default we get a contour plot.
If we have 3 varying dimensions:

    c
    set t 1 5
    d z

we get an animation sequence, in this case through time. 

Now enter:

    clear
    set lon -90
    set lat -90 90
    set lev 1000 100
    set t 1
    d t
    d u

In this case we have set the Y (latitude) and Z (level) dimensions to 
vary, so we get a vertical cross section.  We have also displayed two 
variables, which simply overlay each other.  You may display as many 
items as you desire overlaid before you enter the clear command. 

Another example, in this case with X and T varying (Hovmoller plot):

    c
    set lon -180 0 
    set lat 40
    set lev 500
    set t 1 5
    d z

Now that you know how to select the portion of the data set to view, 
we will move on to the topic of operations on the data.  First, set
the dimension environment to an Z, Y varying one:

    clear
    set lon -180 0 
    set lat 0 90
    set lev 500
    set t 1 

Now lets say that we want to see the temperature in Farenheit instead of
Kelvin.  We can do the conversion by entering:

    display (t-273.16)*9/5+32

Any expression may be entered that involves the standard operators of
+, -, *, and /, and which involves operands which may be contants, 
variables, or functions.  An example involving functions:

    clear
    d sqrt(u*u+v*v)

to calculate the magnitude of the wind.  A function is provided to 
do this calculation directly:

    d mag(u,v)

Another built in function is the averaging function:

    clear
    d ave(z,t=1,t=5)

In this case we calculate the 5 day mean.  We can also remove the
mean from the current field:

    d z - ave(z,t=1,t=5)

We can also take means over longitude to remove the zonal mean:

    clear
    d z-ave(z,x=1,x=72)
    d z

We can also perform time differencing:

    clear
    d z(t=2)-z(t=1)

This computes the change between the two fields over 1 day.  We could
have also done this calculation using an offset from the current time:

    d z(t+1) - z

The complete specification of a variable name is:  

name.file(dim +|-|= value, ...)

If we had two files open, perhaps one with model output, the other
with analyses, we could take the difference between the two fields
by entering:  display z.2 - z.1

Another built in function calculates horizontal relative vorticity
via finite differencing:

    clear
    d hcurl(u,v)

Yet another function takes a mass weighted vertical integral:

    clear
    d vint(ps,q,275)

Here we have calculated precipitable water in mm. 

Now we will move on to the topic of controlling the grahpics output.  
So far, we have allowed GrADS to chose a default contour interval.  
We can override this by:

    clear
    set cint 30
    d z 

We can also control the contour color by:

    clear
    set ccolor 3
    d z

We can select alternate ways of displaying the data:

    clear
    set gxout shaded
    d hcurl(u,v)

This is not very smooth; we can apply a cubic smoother by entering:

    clear
    set csmooth on
    d hcurl(u,v)

We can overlay different grahpics types:

    set gxout contour
    set ccolor 0
    set cint 30
    d  z

and we can annotate:

    draw title 500mb Heights and Vorticity

We can view wind vectors:

    clear   
    set gxout vector
    d u;v

Here we are displaying two expressions, the first for the U component of
the vector; the 2nd the V component of the vector.  We can also 
colorize the vectors by specifying a 3rd field:

    d u;v;q

or maybe:

    d u;v;hcurl(u,v)

You may display psuedo vectors by displaying any field you want:

    clear
    d mag(u,v) ; q*10000

Here the U component is the wind speed; the V component is moisture.

We can also view streamlines (and colorize them):

    clear
    set gxout stream
    d u;v;hcurl(u,v)

Or we can display actual grid point values:

    clear
    set gxout grid
    d u

We may wish to alter the map background:

    clear
    set lon -110 -70
    set lat 30 45
    set mpdset nam
    set digsiz 0.2                  Digit size
    set dignum 2                    # of digits after decimal place
    d u

To alter the projection:

    set lon -140 -40
    set lat 15 80
    set mpvals -120 -75 25 65       Map projection constants
    set mproj nps                   North Polar Stereographic
    set gxout contour
    set cint 30
    d z 

In this case, we have told grads to access and operate on data from 
longitude 140W to 40W, and latitude 15N to 80N.  But we have told it
to display a polar stereographic plot that contains the region bounded
by 120W to 75W and 25N to 65N.  The extra plotting area is clipped by
the map projection routine. 

This concludes the sample session.  At this point, you may wish to 
examine the data set further, or you may want to go through the GrADS
documentation and try out the other options described there.  


Brian Doty
doty@cola.umd.edu
July 20, 1992