batch processing: x and y coordinates instead of i-j indices

[aaschwanden]

For batch processing, I suggest we have to user specifying the x and y coordinates in the projected coordinate system, rather than grid indices. Rationale: this is the more robust approach as it does not depend on the grid dimensions of the data set. I'd prefer not to load a 150m Greenland data set into the GUI but can easily supply the coordinates in the projected coordinate system.

So I suggest changing

parser.add_option("-x", "--x_range", dest="x_range", help="x_min,x_max (in grid indices)")
parser.add_option("-y", "--y_range", dest="y_range", help="y_min,y_max (in grid indices)")

to

parser.add_option("-x", "--x_range", dest="x_range", help="x_min,x_max (in m)")
parser.add_option("-y", "--y_range", dest="y_range", help="y_min,y_max (in m)")

and whatever else is necessary in needed to make this work in pism_regional.py.