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

20030528: McIDAS installation at UFRJ (cont.)

>From:  David Garrana Coelho <address@hidden>
>Organization:  UFRJ
>Keywords:  200305280404.h4S44lLd024244 McIDAS IDD-Brazil image sectors

Hi David,

re: When there is no southern sector from GOES, you will end up with either
of two situations:

>       I understand this will happen every now and then, but I believe we
>can live with that (MUCH better than before, when we had none...lol).

OK. Are you familiar with the GOES-12 scanning schedules?

re: RIO sector being cut off may be result of too small a frame; setup
MCGUI for larger frames

>       Exactly like you said, it worked like a charm.

OK, good.

re: After MCGUI is up and running

>       Tried it (again) 5 mins ago, GUI lists no dataset, except by a
>"DATALOC -  - - - - -" which does nothing.

There are two GUIs supplied with the distribution.  The one from
SSEC is 'GUI'.  The one from Unidata (me) is MCGUI.  You can specify
the automatic start of the MCGUI interface from the gui that gets
popped up when you start McIDAS with:

mcidas -config

If you select the radiobutton that says to automatically start the MCGUI
AND select the radiobutton that says to save configuration options,
then your McIDAS session should include the MCGUI interface each
time you start McIDAS with:

mcidas

(not mcidasx!)

If the MCGUI is coming up correctly, then -- if I did things correctly
on brisa -- you should be able to access the RIO dataset from the
Imagery option under the Display dropdown located at the top of
the MCGUI.  This may not have been completely setup early this morning,
but it should be now.

By the way, your ADDE remote server is working nicely.  I pointed my
McIDAS session (running the MCGUI) at brisa when I got into work,
and I have been loading images off of it from the RIO dataset
ever since.

More on image sectors below.

re: chopped image using xview

>       Forget it, like I said above, image is OK after change of
>resolution in MCIDAS conf.

Very good.

re: larger South American sectors

>       The idea of creating full South American images at 4km (including
>VIS) and we choosing the cuts locally is perfect to me.

I was hoping you would say this.  I just added another option to the
script that gets the data, ~ldm/mcidas/bin/get_goese_sectors.tcl.  The
new option -- which is now the one that is running out of cron -- gets
the full South American sector for all IR channels (3.9, 6.5 (aka WV)
10.7 (aka IR), and 13.3 (aka CO2)).  All of these images are at their
primary resolution which is 4 km.  The script also gets the same view
for the VIS band (0.65 um) by reducing the resolution (called
blowing-down in McIDAS) by a factor of 4 in both the Line (top to
bottom) and Element (side to side) dimensions.  I want to let this run
for awhile to see the impact on your network connection.  I don't think
that the impact will be much, but we have to check anyway.  After this
is seen to be working smoothly, I can add in the creation of high
resolution VIS sectors for Belem, Rio, and Buenos Aires.

I just checked and see that the South American sectors are now being
created on brisa.  I will have to adjust the time that these sectors
are grabbed since it seems like the VIS sectors creation on the upstream
ADDE server is delayed by a considerable amount from the IR bands.
Hmm... I just looked at the VIS sectors on the upstream ADDE server
and they have not been updated since 17:15 Z.  I will have to
investigate the cause for this.

>Concerning the 2km
>and 1km VIS images, the areal extent in the links you sent is ok, we just
>need to move the center the closest possible (avoiding empty sectors on
>eastern frontier, like you said) to 23S 43W location (the center of our
>mesoscale domain) for the 1km resolution,

OK, I will do this on the next pass.

>and doing the same at 2km res,
>but if possible, trading some of western extension for southern (15%-20%)
>in this last case, so we can have high res vis scan on cold fronts
>earlier.

Since you can always convert (blow-down) a 1 km image into a 2, 3, 4, etc
(integer) km image locally, I think that it would be best to get as
large a VIS sector at 1 km as possible, one that covers the full
extent you are interested in (within reason, of course, :-)

>       To help you visualize what I mean, below are two links with
>samples of our mesoscale domains, the first is the coverage we would like
>for 1km (no problem if it covers more, the more the merrier =).
>
>http://www.lpm.meteoro.ufrj.br/tempete/chuvatotal.png
>(Desired area for 1km VIS)
>
>       And the 2km cut (In this case, it can cover less due to size of
>image):
>
>http://cardeal.meteoro.ufrj.br/furnas/chuvatotal.png
>(Desired area for 2km VIS)

Excellent, a picture is worth 10000 words!

re: connection speeds

>       In fact I have a DSL 512/256, but for some reason (blame Murphy's
>Law) it was painfully laggy yesterday. And I dont use Linux at home
>(trying to get over workaholic tendencies), I was using X-Win32.:=)

I understand the need to move away from workaholic tendencies.  The
one commodity that I am short of is time!!

re: drawing maps

>       It worked, all I'd like to know now is how you did to place
>country borderlines on it. In the future, we may want to add brazilian
>state borderlines as well...but no rush on it for now.

All invocations run by the MCGUI can be see in the gui Text and Command
window that can be opened by clicking the left mouse button on the
MCGUI button that has the keyboard icon (the one just to the left
of the button with the large Z (for zoom)).  This button has three
functions:

left click -   open and close the MCGUI Text and Command window.  This
               is where one can see the commands that were run, recall
               and edit them, and rerun them.  It is a nice little
               tool for learning how to use the McIDAS command mode.

middle click - open the non-MCGUI McIDAS Text and Command window

left click   - open additional MCGUI Text and Command windows (one per
               click).  These do not contain the output from MCGUI
               commands.  The are useful for producing textual lists
               (the output of a number of commands is strictly textual)
               and especially for development work.

re: need to modify /etc/services

>       Fixed and restarted xinetd.

The remote server on brisa is working nicely.  I have no trouble accessing
it from my machine here in Boulder, and loading of images is not all
that slow.

re: The first step is to get the MCGUI up and running correctly.  The next
step is to explore the options available.

>       On hold.

I am now testing the creation of South American image sectors (as
noted above) on brisa.  I will shortly add the creation of a high
resolution VIS sector that better covers your modeling domain.
McIDAS is setup on brisa to recognize the new images in the
SAMERICA dataset (as opposed to the RIO one).  I will add the VIS
sector to that dataset as well.

re: presentation of IDD and data availability at Brazilian Congress
of Meteorology

>       Glad to see you liked it. Added this idea to my "To do" list. =)

If your "to do" list is anything like mine, I have sympathy for you
;-)

Cheers,

Tom
--
+-----------------------------------------------------------------------------+
* Tom Yoksas                                             UCAR Unidata Program *
* (303) 497-8642 (last resort)                                  P.O. Box 3000 *
* address@hidden                                   Boulder, CO 80307 *
* Unidata WWW Service                             http://www.unidata.ucar.edu/*
+-----------------------------------------------------------------------------+