We will use matplotlib.pyplot for plotting and scipy's netcdf package for reading the model output. The %pylab inline causes figures to appear in the page and conveniently alias pyplot to plt (which is becoming a widely used alias).

This analysis assumes you changed DAYMAX to some multiple of 5 so that there are multiple time records in the model output.

To see this notebook with figures, see https://gist.github.com/adcroft/2a2b91d66625fd534372.

In [ ]:
%pylab inline
import scipy.io.netcdf

We first create a netcdf object, or "handle", to the netcdf file. We'll also list all the objects in the netcdf object.

In [ ]:
prog_file = scipy.io.netcdf_file('prog__0001_006.nc')

Now we will create a variable object for the "e" variable in the file. Again, I'm labelling it as a handle to distinguish it from a numpy array or raw data.

We'll also look at an "attribute" and print the shape of the data.

In [ ]:
e_handle = prog_file.variables['e']
print('Description =', e_handle.long_name)
print('Shape =',e_handle.shape)

"e" is 4-dimensional. netcdf files and objects are index [n,k,j,i] for the time-, vertical-, meridional-, zonal-axes.

Let's take a quick look at the first record [n=0] of the top interface [k=0].

In [ ]:
plt.pcolormesh( e_handle[0,0] )

The data looks OKish. No scale! And see that "<matplotlib...>" line? That's a handle returned by the matplotlib function. Hide it with a semicolon. Let's add a scale and change the colormap.

In [ ]:
plt.pcolormesh( e_handle[0,0], cmap=cm.seismic ); plt.colorbar();

We have 4D data but can only visualize by projecting on a 2D medium (the page). Let's solve that by going interactive!

In [ ]:
import ipywidgets

We'll need to know the range to fix the color scale...

In [ ]:
[e_handle[:,0].min(), e_handle[:,0].max()]

We define a simple function that takes the record number as an argument and then plots the top interface (k=0) for that record. We then use the interact() function to do some magic!

In [ ]:
def plot_ssh(record):
    plt.pcolormesh( e_handle[record,0], cmap=cm.spectral )
    plt.clim(-.5,.8) # Fixed scale here

ipywidgets.interact(plot_ssh, record=(0,e_handle.shape[0]-1,1));

Unable to scroll the slider steadily enough? We'll use a loop to redraw for us...

In [ ]:
from IPython import display

In [ ]:
for n in range( e_handle.shape[0]):