Re: strange happenings!

[a.larkin@xxxxxxxx]

Dear Ferrets

The problem was not actually solved by the help below, but I did manage to get rid of the ambigous time axis error message. I think the major problem was wantin to plot values that are too small for ferret to want to deal with. This was because I was testing  the routine using unrealistic data. I have since got it working now though with better data.
Thanks Steve
Alice
> From hankin@pmel.noaa.gov Thu Feb 18 01:24:58 1999
> Envelope-to: alarkin@mailhost.sp.ph.ic.ac.uk
> Delivery-date: Thu, 18 Feb 1999 01:24:58 +0000
> Date: Wed, 17 Feb 1999 17:34:46 -0800
> From: Steve Hankin <hankin@pmel.noaa.gov>
> Subject: Re: strange happenings!
> To: a.larkin@ic.ac.uk
> MIME-version: 1.0
> Content-transfer-encoding: 7BIT
> X-Accept-Language: en
> 
> Hi Alice,
> 
> A hurried answer (fitting an answer into a space between meetings)
> 
> Assuming that the grid from xaafba.pjh9jan.nc is potentially applicable
> to both data sets (not enough information in your message to say) I
> suggest that you try:
> 
>  let uvals=U[d=xaafba.pjh9jan.nc,X=@ave]
>  let
> uave=U[d=xaafaa.pjhi10all.nc,X=@ave,T=@ave,g=U[d=xaafba.pjh9jan.nc]]
>  let usd=uave^0.5
> 
> Now the grid ambiguities should be resolved. (Or use
> g=U[xaafaa.pjhi10all.nc] if preferable).
> 
> If this fixes your problem, then please post a brief message.
> 
> Sorry for giving such a hurried answer -- i hope I havn't gone off into
> left field.
> 
> 	- steve
> 
> =====================================
> 
> a.larkin@ic.ac.uk wrote:
> > 
> > Hello,
> > please could someone help me on this, I  am very confused...(and probably begin daft)
> > 
> > I write the following script...
> > 
> > set data "xaafaa.pjhi10all.nc" - a file with 10 years of 12 months each
> > set data "xaafba.pjh9jan.nc" - a file with 1 month of data
> > 
> > I am basically practising commands to do a t-test. So, I want to use the simple formula, (uvals-uave)/usd  where uvals are my values to be tested, uave are my average values and usd is my standard deviation. (I am looking at zonal wind)
> > 
> > let uvals=U[d=xaafba.pjh9jan.nc,X=@ave]
> > let uave=U[d=xaafaa.pjhi10all.nc,X=@ave,T=@ave]
> > let usd=(U[d=xaafaa.pjhi10all.nc,X=@ave,T=@var])^0.5
> > ;then to calculate my t-test values I do...
> > let utest=(uvals-uave)/usd
> > 
> > ;then I contour utest
> > contour/x=1.88E:1.88W/y=88.75S:88.75N/z=100:997/t="15-jan-1979":"24-oct-1988" utest
> > ;this gives me 2 error messages, one I understand and the other I do not.
> > 1. ***NOTE: Ambiguous coordinates on T axis: (uvals-uave)/usd
> > 2. ***ERROR: all data have the same value: data are all flagged as bad
> > 
> > I understand that by using both uvals and uave or uvar in the same formula, the time axis will be ambiguous but this is not usually a problem.
> > The strangest thing is that if I contour just any of (unorm-uvar), (unorm+uvar), (unorm*uvar) or (unorm/uvar) I get the same errors, but if I contour (uvar+unorm), (uvar-unorm), (uvar*unorm) or (uvar/unorm) I get no error messages.
> > Could someone please enlighten me as to what is happening.
> > Thanks
> > Alice Larkin
> 
> -- 
> 
>                 |  NOAA/PMEL               |  ph. (206) 526-6080  
> Steve Hankin    |  7600 Sand Point Way NE  |  FAX (206) 526-6744
>                 |  Seattle, WA 98115-0070  |  hankin@pmel.noaa.gov
>