Gembuds,
Just following up here in case there is interest in my attempts to plot
Goes17 images with GEMPAK (gpmap). I'm still trying to determine what are
features and what might be fixable.
There are 3 main issues I'm working on:
1) IR colors are reversed
2) projection
3) map registration
Regarding the IR color reversal:
Here is what the image should look like based on plotting Goes-15 IR with
sabir.tbl:
http://www.met.sjsu.edu/weather/images/test/goes15-IR-sabir.gif
Here is what the file Goes-17 image looks like with sabir.tbl: (colors
reversed)
http://www.met.sjsu.edu/weather/images/test/goes17-IR-sabir.gif
Here is what the file Goes-17 image looks like with reversing the colors by
reversing the order of lines in the file with a perl one liner:
http://www.met.sjsu.edu/weather/images/test/goes17-IR-sabir-reverse.gif
But this puts the emphasis at the wrong end of the scale, so I tweaked a
bit:
http://www.met.sjsu.edu/weather/images/test/goes17-IR-sabir-reverse-2.gif
So, I can get the image to look more reasonable at least, but more work to
do there.
For 2) the projection. If you toggle back and forth between these:
http://www.met.sjsu.edu/weather/images/test/goes15-IR-sabir.gif
http://www.met.sjsu.edu/weather/images/test/goes17-IR-sabir-reverse-2.gif
I'm trying to replicate the first (Goes-15) image. they are both using
PROJ = SAT. Is this fixable?
For 3) the map registration. It seems like what ever map I choose, the
projection of the map onto the image is about 10 kilometers too far south
and east, at least at the lat/lon of California. You can see in this plot
for example that the "corner" of California is not in Lake Tahoe as it
should be:
http://www.met.sjsu.edu/weather/images/test/goes17-vis-map.gif
And here the SF peninsula is clearly off between map and image:
http://www.met.sjsu.edu/weather/images/test/goes17-vis-sf.gif
Any pointers or comments would be much appreciated.
-Mike
On Fri, Jul 26, 2019 at 11:16 AM Mike Voss <mike.voss@xxxxxxxx> wrote:
> Scott,
> Some follow-up on the color reversal. To make sure my system is behaving
> "normal" I plotted a Goes15 IR image with gpmap:
> GAREA= 19.3;-148.3;52.2;-84.4
> lutfil=sabir.tbl
> DEVICE= gif|Goes15-IR-test.gif|1000;700
>
> which produces a normal looking image:
> http://www.met.sjsu.edu/weather/images/test/Goes15-IR-test.gif
>
> Then, using the same gpmap parameters, aside from the positive longitude
> and different name, I plotted a GOES17 image:
> GAREA= 19.3;148.3;52.2;84.4
> DEVICE= gif|Goes17-CH2-sabir.gif|1000;700
>
> which indeed produces an image with the colors reversed.
> http://www.met.sjsu.edu/weather/images/test/Goes17-CH2-sabir.gif
>
> (Also, the positive latitudes work, but there is clearly something
> different there as well)
>
> So, the question now I guess is to understand if this is a feature in
> GEMPAK, and I need to create work around tables with the colors reversed as
> Scott mentioned, or whether there something about the Goes16-17.nc files
> that's "off", or might there be another explanation. I'm a little
> surprised that I'm the only one having this issue.
>
> Thanks for any comments or advice.
> -Mike
>
>
> On Fri, Jul 26, 2019 at 9:25 AM Scott Jacobs - NOAA Federal <
> scott.jacobs@xxxxxxxx> wrote:
>
>> Mike,
>>
>> The reverse colors should be investigated. In the meantime, you can go to
>> the $GEMPAK/tables/luts directory and create a new color look-up table with
>> the reverse order of the lingray.tbl file. Then change the imgtyp.tbl to
>> reference this new table instead of the default "GRAY".
>>
>> Scott
>>
>>
>> Scott Jacobs
>> NOAA/NWS/Office of Central Processing
>> Technology, Engineering and Application Management Team
>> SSMC2 - 12350
>> 301-427-9413
>>
>>
>>
>> On Fri, Jul 26, 2019 at 11:58 AM Mike Voss <mike.voss@xxxxxxxx> wrote:
>>
>>> Scott,
>>> Thanks for your great suggestion. With that technique of GAREA=DSET and
>>> then "cursor garea" I was able to see that this GAREA selection:
>>> GAREA = 30.1;125.5;44;115
>>>
>>> produces this image from the current GOES17 CONUS vis (Channel02):
>>> http://www.met.sjsu.edu/weather/images/test/goes17-CONUS-test.gif
>>>
>>> So, yes the longitude positive is the opposite of standard GEMPAK.
>>>
>>> When I plot the the same GAREA with IR (Channel14) I get this plot,
>>> where the shading looks reversed (white is warm):
>>> http://www.met.sjsu.edu/weather/images/test/goes17-CONUS-IR-test.gif
>>>
>>> ( This is with standard $GEMTBL/sat/imgtyp.tbl values as:)
>>> GOES17 CH14 11.2um IR 0 255 188 2**13 1 GRAY
>>>
>>> I tried tweaking this to
>>> GOES17 CH14 11.2um IR 0 128 188 2**13 1 GRAY
>>>
>>> but I get a white map:
>>> http://www.met.sjsu.edu/weather/images/test/goes17-CONUS-IR-test2.gif
>>>
>>> Where might I look to reverse the color ramp? Or, could there be
>>> something else fairly obvious I'm missing?
>>>
>>> Thanks again for the tips; this is much appreciated!
>>>
>>> -Mike
>>>
>>>
>>>
>>> On Fri, Jul 26, 2019 at 5:32 AM Scott Jacobs - NOAA Federal <
>>> scott.jacobs@xxxxxxxx> wrote:
>>>
>>>> Mike,
>>>>
>>>> Try to use GAREA=DSET with PROJ=SAT. This should give you the entire
>>>> domain of the image file you selected. You can then use "cursor garea" from
>>>> the GEMPAK prompt to zoom in on the area you want. I tried this on the G17
>>>> CONUS image and the longitude values came back as positive - the opposite
>>>> of what GEMPAK normally uses. However, when I used the resultant garea, the
>>>> image looked ok.
>>>>
>>>> Scott
>>>>
>>>>
>>>> Scott Jacobs
>>>> NOAA/NWS/Office of Central Processing
>>>> Technology, Engineering and Application Management Team
>>>> SSMC2 - 12350
>>>> 301-427-9413
>>>>
>>>>
>>>>
>>>> On Thu, Jul 25, 2019 at 3:47 PM Mike Voss <mike.voss@xxxxxxxx> wrote:
>>>>
>>>>> Hi Gembuds,
>>>>> I'm a longtime user of Gempak, and maybe 10 years ago I had a little
>>>>> better grasp of how it all works. Nonetheless, I'm trying to replace some
>>>>> of my outdated satellite images scripts with newer GOES17 images from the
>>>>> NIMAGE feed of GOES17 data. So, not the original nc files, but the
>>>>> adjusted version supplied by Unidata that theoretically can be plotted in
>>>>> Gempak.
>>>>> https://www.unidata.ucar.edu/data/nimage.html
>>>>>
>>>>> My system is running the latest versions of LDM and Gempak at this
>>>>> level:
>>>>> Linux vulcan.XXX 3.10.0-957.21.3.el7.x86_64 #1 SMP Tue Jun 18 16:35:19
>>>>> UTC 2019 x86_64 x86_64 x86_64 GNU/Linux
>>>>>
>>>>> I tried several times to get images to plot via gpmap with various
>>>>> standard GAREA entries, but I kept getting this error:
>>>>> " [GEMPLT -60] NIMCORD - Invalid image display coordinates."
>>>>>
>>>>> Then I tried various geographic identifiers. With GAREA=US I got this
>>>>> to plot for the CONUS image:
>>>>> http://www.met.sjsu.edu/weather/images/test/goes-vis.gif
>>>>> And then interestingly when I plot the Full Disk image and use
>>>>> GAREA=US, I get this:
>>>>> http://www.met.sjsu.edu/weather/images/test/goes-vis2.gif
>>>>>
>>>>> HMM....strange, somethings seems off with the lat lon registration.
>>>>>
>>>>> Then I fiddled with adding lat lon corners and got some to work, but
>>>>> plots are much farther east than expected:
>>>>> E.G.:
>>>>> GAREA = -165;15;-155;25
>>>>> produces: http://www.met.sjsu.edu/weather/images/test/goes-vis3.gif
>>>>> and
>>>>> GAREA=-179;10;-140;30
>>>>> produces http://www.met.sjsu.edu/weather/images/test/goes-vis4.gif
>>>>>
>>>>> So, clearly gempak is not reading something about the lat/lon
>>>>> correctly. Before I go any further in my "pool sweep" level of
>>>>> troubleshooting, might there be a $GEMTBL/sat/imgtyp.tbl adjustment needed
>>>>> here? Or would anyone have another tip to help diagnose my underlying
>>>>> issue.
>>>>>
>>>>> thanks,
>>>>> -Mike
>>>>>
>>>>>
>>>>> --
>>>>> Mike Voss
>>>>> SJSU Meteorology
>>>>> 408.924.5249
>>>>> _______________________________________________
>>>>> NOTE: All exchanges posted to Unidata maintained email lists are
>>>>> recorded in the Unidata inquiry tracking system and made publicly
>>>>> available through the web. Users who post to any of the lists we
>>>>> maintain are reminded to remove any personal information that they
>>>>> do not want to be made public.
>>>>>
>>>>>
>>>>> gembud mailing list
>>>>> gembud@xxxxxxxxxxxxxxxx
>>>>> For list information or to unsubscribe, visit:
>>>>> https://www.unidata.ucar.edu/mailing_lists/
>>>>>
>>>>
>>>
>>> --
>>> Mike Voss
>>> SJSU Meteorology
>>> 408.924.5249
>>>
>>
>
> --
> Mike Voss
> SJSU Meteorology
> 408.924.5249
>
--
Mike Voss
SJSU Meteorology
408.924.5249
