[Thread Prev][Thread Next][Index]

Re: [ferret_users] arithmetics not working on a variable (broken netCDF file?)

Hi Ryo,
You're right. The modulo length is the length of the repeat, whatever that is. It's most common use is longitudes repeating around the earth.

You can indeed set that to any number that you want. But for a single-point axis, I'm not sure what makes sense. I suppose if you set the modulo repeat to the size of one grid cell, and then ask for a shade plot that is more than that width, Ferret can make the plot.

If you add bounds to the x axis of the netCDF file, then you won't have the problem of the size of the grid cell being larger than the modulo length.

Ryo Furue wrote:

Thanks for your response.

| I was able to download your file after all. I think that the only
| problem here is the *value *of the modulo attribute on your x
| axis. The modulo length is the length of the modulo repeat. For a
| longitude axis the modulo length is always 360 degrees, the length of
| the wraparound of degrees around the earrth.  So this attribute should
| always be either
| | lont:modulo = " " ;
| or
|                lont:modulo = 360 ;

That's news to me!  And I don't think that's right.  Consider the
domain of a numerical model that extends from lon=0 to lon=180.
When modulo = 360, I expect that the model field is missing
for 180 <= lon <= 360, 540 <= lon <= 720, and so on.
But, when modulo = 180, I expect that the model field is defined
for any lon.

I think this is the PURPOSE of the modulo attribute, that is,
to allow for an arbitrary cyclic-domain size.  In addition, Ferret
itself agrees with me! that is, it does exactly what I have
described.  I have an netCDF file in which

float lonv(lonv) ; lonv:units = "degrees_east" ; lonv:axis = "X" ; lonv:missing_value = -9.99e+33f ; lonv:modulo = 100. ;
And Ferret does plot the field for whatever domain:

   fill/k=1/l=1/x=-180:180 u

without any gap.

| Ferret will not draw a shade plot of data on a 1-point by n-point
| region

Hmm, I can see that FILL cannot work if the domain isn't cyclic.
But, I guess that SHADE can work whether the domain is cyclic or
not because it just paints gridcells without interpolation.


[Thread Prev][Thread Next][Index]

Contact Us
Dept of Commerce / NOAA / OAR / PMEL / TMAP

Privacy Policy | Disclaimer | Accessibility Statement