[gradsusr] NCEP global SST data and GrADS

[Bill Reilly]

Jennifer:

Thanks for your help.

Degrib doesn't work with these GRIB2 files (rtgssthr_grb_0.083.grib2) 
correctly for some reason... I've tried every possible combination of 
switches and options but I've had no luck.

I ended up using the lower resolution data file (rtgssthr_grb_0.5.grib2) 
which works perfectly, and is good enough resolution for my purposes.

Thanks again for your help.

Regards,

Bill

--

Bill Reilly

(+34) 686-110-748

bill at passageweather.com
www.passageweather.com

bill at brwebworks.com
www.brwebworks.com

On 03/02/2018 14:27, Jennifer M Adams wrote:
> Hi, Bill —
> The problem is that GrADS has determined that your grid does not wrap the globe. The error is in your XDEF declaration. You have
>      XDEF 4321 LINEAR -179.958333 0.083333
> which you say you got from degrib. That is incorrect. The size should be 4320 — the grid as you’ve defined it actually overlaps and therefore fails the globe-wrapping test in GrADS (if longitude at x=1 is roughly the same as longitude at x=xsize+1)
>
> Since your descriptor file is for a flat binary file and not a grib2 file, I suspect you may have stuck an extra column of data onto your grid when converting the format.
>
> I downloaded the grib2 file in question and used g2ctl to get a descriptor and got this for XDEF:
>     XDEF 4320 linear 0.042000 0.083000
>
> Your grid size is clearly 4320 and not 4321. A quick check of the initial lon value confirms that the grib2 grid goes from 0 to 360 and not -180 to 180.
>
> Here is my complete descriptor:
> dset ^rtgssthr_grb_0.083.grib2
> index ^rtgssthr_grb_0.083.idx
> undef 9.999E+20
> title rtgssthr_grb_0.083.grib2
> dtype grib2
> ydef 2160 linear -89.958000 0.083
> xdef 4320 linear 0.042000 0.083000
> tdef 1 linear 00Z01feb2018 1mo
> zdef 1 linear 1 1
> vars 1
> sst   0,1,0   0,0,0  sea surface temperature [K]
> endvars
>
> —Jennifer
>