[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

20010322: hoping for wisdom...mixing ratio advection



Paul,

I never heard back as to whether the RELH values in the region
in question were negative.

This is a possibility in the grids from NCEP. 
Can you list out the values of MIXR (or RELH) in the areas
(either with gdlist or gdmap) to see what values are being used
in the computation.

For example, in todays ETA 211 grids, 500mb F000 grid,
use GFUNC=slt(RELH,0) in GDMAP, this shows a negative RELH area
SW of California. If you try to contour GFUNC=adv(mixr,wnd) in this
region, you will have missing values (-9999.) so no data values to contour.

Steve Chiswell
Unidata User Support

On Thu, 22 Mar 2001, Paul Ruscher wrote:

> Hi, in doing some routine calculations in preparation for some possible
> automation, we find strange behavior in (I think) gempak 5.6 and I was
> wondering if anyone had a suggested workaround.
> 
> Chiz has suggested some things we've tried to no improvement.
> 
> We are calculating moisture advection (basically) using adv(mixr,wnd)
> using eta and avn model grids, and whenever there is a pronounced minima
> in the mixing ratio, gempak is unable to calculate (or plot) mixing ratio
> advection in the vicinity of that minimum.  Not every minimum, but at
> least one minimum mixr value on nearly every chart we have looked at has
> an associated area of no contour plotting/grid values, so any further
> calculation is moot.  And of course this happens always in a region of
> meteorological interest.
> 
> We have tried working with SUBBOX and CONTUR and CINT, and I've tried
> 
>       GFUNC=SM9S(ADV(MIXR,WND)) (with two iterations of sm9s even)
> 
> and
> 
>       GFUNC= ADV(SM9S(MIXR),WND)
> 
> and along with parameters above (SUBBOX, CONTUR) we can get a change in
> behavior, but not a fix for the problem.
> 
> Let's call this a preliminary inquiry in case this rings a bell with
> anyone; it does not happen with other scalar advections we've tried,
> including temp, avor(), etc...
> 
> Thanks in advance for any wisdom.
> 
> Paul Ruscher / FSU
> 
>