20030827: Re: GOES Sndr product imagery - color enhancements

------- Forwarded Message

>From: "Gary S. Wade" <address@hidden>
>Organization: NOAA/NESDIS/ORA/ARAD/ASPT at UW-CIMSS
>Keywords: 200308272121.h7RLLtLd010767

Hello all,

        An inquiry from Professor Jim Koermer (about color enhancement tables 
and other pertinent information about GOES Sounder derived product 
imagery) precipitated the following response below for some general 
background on the GOES DPI.  Tom Yoksas has said that this type of 
information has been previously made available for the Unidata 
community.  Nonetheless, if a question is asked, it seems that it may be 
worthwhile to re-present the material again now (with some brief 
rationale and explanations).  To re-consider some issues raised here may 
be enlightening and beneficial for the community.

        Not having been a subscribed Unidata member, my initial e-mail was 
refused to the e-mail list.  Now, I'm subscribed; here's the response 
below to Professor Koermer.  Thus, for whatever general or specific 
interest this may raise, read on.  Thanks much.

                                                Gary




Gary S. Wade wrote:

>Greetings Professor Koermer and other interested Unidata users,
>
>    In response to your inquiry about displaying and understanding GOES 
>Sounder DPI (Derived Product Images), I'll start with a brief overview 
>of the DPI concept and then provide a few sources for more detail and 
>examples.  First, the DPI are really just a presentation of the GOES 
>Sounder derived atmospheric profiles (or their derivatives, such as 
>total precipitable water (TPW) or lifted index (LI) stability) in clear 
>air in an image format, blended with traditional 11 um window band 
>imagery where clouds exist.  The profiles (or TPW or etc.) could also be 
>plotted or contoured for viewing; however, the DPI lend themselves well 
>to animation and provide a complete scene, with both retrieved GOES 
>product as well as cloud depiction (thus, permitting the display of 
>moisture evolution as well as convective development, for example). 
>Within McIDAS, a DPI "area" is typically just separated with values 
>0-200 for the product and 201-255 for the "cloud".  This breakdown holds 
>true for the "profile" DPI, including TPW, LI, CAPE, skin temperature, 
>etc.  A different separation is used for the complementary "cloud" DPI, 
>which includes cloud top pressure (CTP) and effective cloud amount 
>(ECA); values 0-75 are for clear (window) temperatures, while values 
>76-255 are for the cloud product.
>
>    Hourly examples of many of these GOES Sounder DPI are available for 
>viewing in near-realtime on the CIMSS Realtime GOES Page (at 
>http://cimss.ssec.wisc.edu/goes/realtime/realtime.html).  I do not think 
>that there is any one magical enhancement table (ET) that displays the 
>essence and content of a DPI.  Many factors come into play, not the 
>least of which is consideration for its aesthetic appeal.  Nonetheless, 
>at CIMSS, we have used some ETs, or very close cousins, consistently for 
>many years (even back in the mid 80's).  These basically continue to be 
>shown on the CIMSS page as well as the NESDIS/FPDT pages in Washington 
>(http://orbit-net.nesdis.noaa.gov/goes/sdpi/).  However, others have 
>chosen to use different palettes.
>
>    Some past discussion on selecting effective ETs for the DPI may be 
>found at http://cimss.ssec.wisc.edu/goes/dpiet/dpiet.html.  I'll just 
>mention here two general points of interest, to me, on color 
>enhancements.  First, I find distinct color bands to be useful in 
>allowing me to quickly associate a semi-quanitative value to the DPI 
>(for example, I know that "red" is modestly high moisture in the TPW - 
>light red 40mm+, darker red 45mm+ with finer intensity still available; 
>yellow is definitely less (30-39mm)).  Secondly, being one of those ~10% 
>of the population who suffer from color perception difficulties ("color 
>blindness"), I need to emphasize intensities, rather than hues, and 
>limit the number of colors (as only so many remain "independent").  Just 
>imagine that bright yellow and bright green look the same, or cyan and 
>white, or purple or blue, and so forth.  (I'm glad that I don't need to 
>use the radar ETs that often are used with NWS radar displays - it does 
>not work for me, with plenty of repeating colors.)  Anyway, effectively 
>coloring images is not necessarily a simple process.
>
>    Finally, for the nuts and bolts of the ETs that we show on the web 
>(whose displays are created on McIDAS), check the available files at
>http://cimss.ssec.wisc.edu/seegsw/dpiets/readme.  Especially note the 
>simple batch files that can be run to build (EU) the ETs (such as 
>MAKPW1.BAT to make "GDPW1.ET").  The command PRDUTIL (via APPW.MCB, for 
>example) will enable appropriate use of the "D" key in McIDAS to 
>interrogate a pixel in a DPI area.  The BAR command is an alternate to 
>using virtual graphic scales to label the gray scale bar at the bottom 
>(but, I'm not sure if the units will be as desired).
>
>    So, enough of an introduction for now.  Please make comments, ask 
>questions, and/or provide suggestions.  Thanks much.
>
>
>                    Gary
>
>
>-------- Original Message --------
>Subject: Color Palette/Values for Derived GINI Products
>Date: Mon, 25 Aug 2003 21:12:46 -0400
>From: Jim Koermer <address@hidden>
>Organization: UCAR/Unidata
>To: address@hidden, address@hidden
>
>
>Hello,
>
>Now that GINI lifted index, precip. water, skin temperature, and cloud
>top pressure pressure are coming over the GOES NOAAPORT channel, I was
>trying to come up with a color schemes that I could use with either WXP
>or McIDAS displays of these data and also trying to understand the
>units/ranges of these data.
>
>We have started looking at the imagery and tweaking various color values
>and they look somewhat similar to regular imagery (i.e. 0 to 255), but
>tend to be clustered in different regions within this range. However, I
>don't have any concept of how these values relate to the normal
>non-satellite derived values that I normally follow.
>
>Right now, I have a student trying different color combinations to find
>some that more clearly define the values and gradients. However, the
>question is how to interpret the resulting quantities. Has any one
>looked into these products.
>
>Jim


-- 

  * --- -- -      * --- -- -      * --- -- -      * --- -- -

           Gary S. Wade         (... watch the sky ...)
      NOAA/NESDIS/ORA/ARAD/Advanced Satellite Products Team
  (1225 W. Dayton St.  Madison, WI 53706  Rm 435  608-263-4743)
  (e-mail:  address@hidden  or  address@hidden)
        [ http://cimss.ssec.wisc.edu/aspt/aspthome.html ]
   [ http://cimss.ssec.wisc.edu/goes/realtime/realtime.html ]

  - -- --- *      - -- --- *      - -- --- *      - -- --- *

------- End of Forwarded Message