Re: PLOT command ambiguous coordinates on T axis

[Ansley Manke <Ansley.B.Manke@xxxxxxxx>]

Hi Steve,
Yes, you could set up a time axis and put each variable on it
(or regrid one variable to the grid of the other) explicitly before
doing the subtraction.

And the reason you get a missing-value result when you do
an @AVE with a zero range is that the computation is to add
up the data in each cell that's included in the range, weighted
by how much of the cell was in the range, and divide by the
length specified by the range.  In your example, the range is zero.
Now, for the 1-dimensional @ave, Ferret detects that you're just
specifying one point and gives you the value at that point.  But
that logic is not in the multi-dimensional @AVE computation,
which requires a positive length in all directions in order to return
a value.

Ansley

> Thanks for your reply. To avoid this informative message should I be setting up the time axis before ferret tries to
> reconcile
> the grids?
> Also, shouldn't the following statements generate the same results?
> yes? list/l=1 tmax[d=1,y=40.25,x=-105.75]
>              VARIABLE : TMAX (degrees C)
>              FILENAME : ncepM1979.cdf
>              LONGITUDE: 105W(-105)
>              LATITUDE : 41N
>              TIME     : 31-JAN-1979 10:29
>          -6.784
> If the lower lat = upper lat I get this:
> yes? list/l=1 tmax[d=1,y=40.25:40.25@ave,x=-105.75:-105.75@ave]
>              VARIABLE : TMAX (degrees C)
>              FILENAME : ncepM1979.cdf
>              LONGITUDE: 105W(-105) (XY ave)
>              LATITUDE : 41N (XY ave)
>              TIME     : 31-JAN-1979 10:29
>         .... (should be -6.784)
> If I leave off the ave transformation on the latitude it computes the ave correctly:
> yes? list/l=1 tmax[d=1,y=40.25,x=-105.75:-105.75@ave]
>              VARIABLE : TMAX (degrees C)
>              FILENAME : ncepM1979.cdf
>              LONGITUDE: 105W(-105) (averaged)
>              LATITUDE : 41N
>              TIME     : 31-JAN-1979 10:29
>          -6.784
>
> Steve

Steve Knox wrote:

> Ansley Manke wrote:
>
> > Hi Steve,
> > It's not an error, but an informational message, generated when
> > Ferret reconciles the grids from the two variables to see if the
> > variables can be compared.  Once this has been done successfully,
> > Ferret keeps track of the information about how the grids were
> > "merged", and so the message isn't issued on the subsequent commands
> > that use the same grids.
> >
> > Ansley Manke
> >
> > Steve Knox wrote:
> >
> > > Why do I get the error on the first plot command but not when I simply
> > > repeat the command?
> > > Any insights would be greatly appreciated.
> > >
> > > ferret -gif
> > >         NOAA/PMEL TMAP
> > >         FERRET v5.50 (64bit CDF-No DODS)
> > >         Solaris 5.8 - 01/23/03
> > >         30-Jan-03 11:38
> > >
> > > yes? use ncepM1979
> > > yes? use ncepM1980
> > > yes? plot/l=1:12 tmax[D=1,y=30:40@ave,x=-120:-60@ave]-tmax[d=2,y=30:40@ave,x=-120:-60@ave]
> > >  *** NOTE: Ambiguous coordinates on T axis: TMAX[D=1,Y=30:40@AVE,X=-120:-60@AVE]-TMAX[D=2Y=30:40@AVE,X=-120:-60@AVE]
> > > yes? plot/l=1:12 tmax[D=1,y=30:40@ave,x=-120:-60@ave]-tmax[d=2,y=30:40@ave,x=-120:-60@ave]
> > > yes? frame/file=d.gif
> > >
> > > The generated gif is ok.
> > >
> > > Steve Knox
> > > NREL
> > > Colorado State University
> >
> > --
> > Ansley Manke  Pacific Marine Environmental Laboratory  Seattle WA  (206)526-6246
>