Re: [ferret_users] Overlaying location of minimum U wind

[Russell Fiedler <russell.fiedler@xxxxxxxx>]

Yes, it looks like the average  of the locations of the minima at each level are being calculated rather than
the minima of the average U.

You can get all the minima at once using the @LOC transformation by subtracting the the value
from the minima at each longitude

 let uwndave = uwnd[z=`zl`:`zu`@ave]
let zero_at_minu=uwndave[y=@MIN]-uwndave   ! negitve every where except at minimum u
let y_at_minu=zero_at_minu[y=@loc:0]

con uwndave
plot/sym/line/ov y_at_minu[x=0:120:2]


On Saturday 05 February 2011 05:38, Ansley Manke wrote:
> Hi-
> I don't know why the locations would be wrong. I wonder if it could be 
> be the order of operations, y=@loc:`minval`, and the averaging in Z?  
> SET MODE DIAGNOSTIC might  help sort that out. You could try defining a 
> variable first,
> 
> let uwndave = uwnd[z=`zl`:`zu`@ave]
> 
> and operate on that to see what difference it makes.
> 
> The reason that /LINE has no effect is that each PLOT command is drawing 
> a separate single-point line, and has no knowledge of the points that 
> came before.  You could save each location after it's computed, and plot 
> at each X a 2-point line using {xpos,xlast}, {ypos,ylast}.
> 
> Ansley
> 
> On 2/4/2011 3:23 AM, Samrat Rao wrote:
> >
> > Hi,
> >
> > I am trying to locate and plot the minimum U wind values on a lat-lon 
> > figure. I am attaching the figure generated.
> >
> > After ploting the U wind contours i use the following:
> >
> >  repeat/range=0:120:2/name=xpos (  \
> >    let minval =  uwnd[x=`xpos`e,y=@min,z=`zl`:`zu`@ave,l=7]; \
> >    let ypos = uwnd[x=`xpos`e,y=@loc:`minval`,z=`zl`:`zu`@ave,l=7]; \
> >    list ypos; \
> >    
> > plot/line=14/color=red/thickness=3/size=0.15/symbol=3/vs/nolabel/overlay 
> > xpos,ypos  )
> >
> > As you can see the symbols are being plotted below the location of 
> > minimum U wind. I have the following questions:
> >
> > 1) The reason for the incorrect location of the symbols. I am not 
> > having this problem with NCEP reanalysis data.
> >
> > 2) I am unable to get the 'line' qualifier connect the different 
> > symbols. How to do this?
> >
> > 3) At some locations there is a discontinuity (for example around 
> > 55E). To me it looks artificial.
> >
> > 4) In case there is a missing value for 'ypos', then how do i ensure 
> > that this does not cause an error and abort in the script execution?
> >
> > I hope you people will have patience to reduce some of my difficulties.
> >
> > Thanks,
> > Samrat Rao.
>