Re: advection term

The documentation for all the spherical harmonic functions
state that they require the input grids be global. Based on the
figures you included, your grids are not global.

On 02/29/2012 01:00 PM, Alexander Semenov wrote:
> Hello
>
> I'm trying to calculate the advection term (u*(dT/dx)+v*(dT/dy)). I
> wanted to do this in 2 steps: 1st step is to calculate the gradient, the
> 2nd is to multiply it by u and v
>
> So I calculated gradient as shown in the example:
> http://www.ncl.ucar.edu/Document/Functions/Built-in/igradsF.shtml
>
> T_grad_lon = theta ; create arrays to hold output, same size and type as input
> T_grad_lat = theta
> ; this procedure will overwrite
> ; values in T_grad_lon and T_grad_lat
>
> *gradsf* <http://www.ncl.ucar.edu/Document/Functions/Built-in/gradsf.shtml> (theta, T_grad_lon, T_grad_lat)
> T_grad_lon@long_name ="longitudinal gradient (derivative)"
> T_grad_lat@long_name ="latitudinal gradient (derivative)"
> theta =*igradsF* (T_grad_lon, T_grad_lat)
>
>
> Then I had to multiply gradient by u and v to get(u*(dT/dx)+v*(dT/dy).
>
> I did 2 different ways:
>
> 1st way.
>
> u*T_grad_lat+v*T_grad_lon. You may see on the picture ("1way.png") that the plot looks pretty much the same all over the domain. I don't see the maximum advection where it is supposed to be by the west coast of Greenland.
>
> 2 way
>
> u*theta+v*theta. I got the right pattern, but values are off. There can't be advection 700 K/s.
>
> I think that this way is not correct because theta has 2 dimensions: theta(T_grad_lon,T_grad_lat), so when I do the multiplication (u*theta+v*theta)I end up having:
>
> u*T_grad_lon+u*T_grad_lat+v*T_grad_lon+v*T_grad_lat, whereas I need only 2 highlighted terms
>
>
> Could you help me find out what is wrong and how can I get it fixed?
_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Wed Feb 29 13:29:55 2012