[ncl-talk] error using vinth2p on CCSM4 PI winds
Dennis Shea
shea at ucar.edu
Fri Feb 27 13:55:15 MST 2015
[1] When reading an entire array,
Use
und= f1->U
Not
und= f1->U(:,:,:,:)
The former is more efficient. If you want to explicitly note that U is
4D, I'd suggest adding a commebt
und= f1->U ; 4D or (:,:,:,:)
[2] There is an odd mix of units for the P0, plev and ps variables.
I speculate your 'P0u' is in Pa when it should be in mb (hPa).
P0u = f1->P0
ps = f4->PS ; Pa
P0u = P0u*0.01
P0u at units = "mb"
On Fri, Feb 27, 2015 at 12:18 PM, Lisa Murphy <lmurphy at rsmas.miami.edu>
wrote:
> Hi -
>
> I'm using NCL version 6.1.2. I'm working with the CMIP5 CCSM4 PI
> simulation data. The surface U and V winds are not archived so I tried
> using the the vinth2p function to interpolate to either 850mb or 925 mb.
> However, when I do so the winds are very strong and seem to be offset. I
> checked the lat and lon coordinates and things seem to be fine. I'm not
> extrapolating (extrap = False), but this shouldn't be an issue when I
> interpolate to 850 mb. I also tried both the linear and log interpolation
> (interp = 1 or 2).
>
> The file size is large (13 Gb). Could this be the issue?
>
> Please see the attached image file, and a subset of my script
> below. Thanks for any help.
>
> Lisa
>
> und= f1->U(:,:,:,:)
> lat = f1->lat
> lon = f1->lon
> time = f1->time
> P0u = f1->P0
> ps = f4->PS
>
> ; for comparison with another model interpolate the model data to these
> ; 25 constant pressure levels
> plevs = (/70.,100.,150.,200.,250.,300.,350.,400.,450.,500.,550., \
> 600.,650.,700.,750.,800.,825.,850.,875.,900.,925.,950.,975., \
> 990.,995.,1000./)
> plevs at units = "mb"
> plevs at positive = "down"
>
> interp = 2 ;1 = linear, 2 = log, 3 = loglog
> extrap = False
>
> hyam = f1->hyam
> hybm = f1->hybm
>
> ui = vinth2p(und,hyam,hybm,plevs,ps,interp,P0u,1,extrap)
> printVarSummary(ui)
>
> uil = ui(:,17,:,:)
> copy_VarCoords(ui,uil)
> copy_VarAtts(ui,uil)
> printVarSummary(uil)
>
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20150227/114adede/attachment.html
More information about the ncl-talk
mailing list