Using CSPP QuickLooks and Direct Broadcast data to view gridded NUCAPS fields

August 7th, 2020 |

NOAA-20 852-mb Temperatures, ca. 1750 UTC on 7 August 2020.  Inset:  Global Map of where the data sit (Click to enlarge)

A previous blog post (link) detailed how to access NOAA CLASS to create Gridded NUCAPS (NOAA-Unique Combined Atmospheric Processing System) imagery from those data.  (You can also view some gridded NUCAPS fields here;  click here to see the 850-mb field of Temperature from that site, it is very similar to the imagery above).   This post details how to use the CSPP QuickLooks software package to create imagery at different levels.  These QuickLook fields give good information quickly and at many different levels for Direct Broadcast data.

Download the Sounder QuickLook software for Linux from the CIMSS website here.  Documentation is also available at the download website.  The files to download are shown in this graphic. The package is self-contained and requires only unzipping and un-tarring.

After downloading, define the $CSPP_SOUNDER_QL_HOME variable as the directory where the package sits on your unix platform.  Then, set up the environment with the command:  source $CSPP_SOUNDER_QL_HOME/

This software package works on NUCAPS EDR (Environmental Data Records) files created at Direct Broadcast sites by CSPP (that are also available after some time from CLASS), and those files can be found at websites such as this one: — underneath this are directories for NOAA-20 (‘j01’) and Suomi-NPP (‘npp’). For example, NOAA-20 data from 7 August 2020 from the ~1747 UTC overpass is at (This website is not preserved forever but will go away after about a week. The directory includes an edr subdirectory that contains the files needed; a typical filename looks like this:; it is the EDR for NOAA-20 and it contains data on 7 August 2020 from 1752 through 1753 UTC. The directory will include up to about 18 of these EDRs (the number depends on how long the satellite is within view of the Direct Broadcast antenna at CIMSS).

How do you create the QuickLooks?

  1. Move the EDR files to your machine, and that’s easily done with wget*.  Of course, the yyyy_mm_dd_jdy_hhmm value (2020_08_07_220_1747 above) changes with each satellite overpass!
  2. Create a list of the files in that directory, i.e., files=$CSPP_SOUNDER_QL_HOME/data/NUCAPS-EDR*
  3. Invoke the shell script from $CSPP_SOUNDER_QL_HOME/scripts/ "$files" NUCAPS --dset temp --pressure 850.   This will create an image, shown above, that is a temperature mapping at the closest pressure level to 850 mb in the NUCAPS retrieval. (Pressure levels in the Radiative Transfer Model that is used by NUCAPS are listed here;  in the map label above, note that values are truncated, not rounded).  You can also map dewpoint temperature (dwpt), relative humidity (relh) and mixing ratio (wmix). By default, temperature scaling matches the bounds of the image, but you can specify the bounds if needed, using --plotMin=250.0 --plotMax=300.0, for example.  The time of the image is the time of the first scan line — for this ascending pass, it’s the southernmost line.  In this QuickLook image, the airmass difference between the relatively cool air over Ohio/Indiana and the warmer air to the south is apparent.

You can also create a QuickLook SkewT/logP plots for each scan. This produces one SkewT per ScanLine, at the mid-point along the scanline that contains 30 separate profiles.   The sounding below was produced by this command:

./ ./data/ NUCAPS

NOAA-20 NUCAPS Sounding from 1755 UTC on 7 August 2020 at 41.93º N, 75.81º W (Click to enlarge)

The SkewT has characteristics that suggest the presence of clouds.  What did this particular sounding look like in AWIPS?  That’s shown below.  In AWIPS, the sounding also terminated at about 550 mb, and the temperature and dewpoint lines above that level match the Quick Look sounding shown above.

NSharp AWIPS presentation of NOAA-20 NUCAPS Sounding at 41.93 N, 75.81 W at ~17UTC on 7 August 2020 (Click to enlarge)

NUCAPS Sounding Availability points from this NOAA-20 pass are shown below. The sounding point — in yellow — that is circled in blue is the one shown above. The sounding just to the east of that point — a green point that gives useful information down into the boundary layer — is shown here. Quick Looks choose the mid-point sounding along the line, and sometimes, as in this case, the retrieval that produced the profile did not converge.

NOAA-20 NUCAPS Sounding Availability Points from AWIPS, 1732 UTC on 7 August 2020 (Click to enlarge). The sounding shown above is from the point circled in blue.  This is in the middle of the 30 sounding points along the horizontal line of available profiles.

Gridded NUCAPS fields around nocturnal convection over the southern Plains

April 22nd, 2020 |

GOES-16 Clean Window (10.3 µm) infrared imagery, 0346 – 1021 UTC, 22 April 2020 (click to animate)

GOES-16 Clean Window (10.3 µm) infrared imagery, above (click to animate) shows two regions of convection over the southern Plains, one moving through central/southern Oklahoma, one developing over the Texas Panhandle and moving east). A similar (but slightly later) animation of GOES-16 Low-Level water vapor infrared imagery (7.34 µm) is below.

GOES-16 Low-Level water vapor (7.3 µm) infrared imagery, 0456 – 1441 UTC, 22 April 2020 (click to animate)

At 0821 UTC, two distinct mesoscale convective complexes are apparent, with a clear region between. This time approximated an overpass by NOAA-20; data from the Cross-track Infrared Sounder (CrIS) and the Advanced Technology Microwave Sounder (ATMS) are combined to create NUCAPS soundings.

GOES-16 Low-Level water vapor (7.3 µm) infrared imagery, 0826 UTC, 22 April 2020, along with NUCAPS sounding points (click to enlarge)

During this time, there were three soundings launched at Amarillo — at 0000, 0600 and 1200 UTC.  They are shown below and all three suggest steep mid-level lapse rates.

Radiosonde from Amarillo TX at 0000, 0600 and 1200 UTC on 22 April 2020 (Click to enlarge)

The NUCAPS profile south of Amarillo (in the water vapor image above, the ‘green’ point just south of the ‘red’ point just south of the convective system over Amarillo) is shown below.  It also shows fairly steep mid-level lapse rates.  Click here to see a toggle between the NUCAPS profile below and the 0600 UTC Amarillo Radiosonde.

NUCAPS Profile at 35 N, 101 W, 0821 UTC on 22 April 2020 (Click to enlarge)

Gridded NUCAPS fields allow a forecaster to view thermodynamic information from the entire pass more easily than can be achieved by examination of individual soundings, or by viewing soundings via the pop-up SkewT.  The animation below shows the Total Totals index, the 850-500-mb lapse rate, and the lapse rate from 700-300 mb.  Strong instability (Total Totals values around 50) is indicated downstream of the system over the Texas panhandle; also, lapse rates are steeper between 700 and 300 mb (about 7.5º C/km) compared to those between 850 and 500 mb (about 6.8º C/km).

Gridded NUCAPS data gives timely satellite-derived (and model-independent) estimates of the thermodynamic state of the atmosphere.

Gridded NOAA-20 NUCAPS estimates of Total Totals index, 850-500 mb Lapse Rate and 700-300 mb Lapse Rate (Click to enlarge)

Comparing Gridded NUCAPS data to model fields

February 24th, 2020 |

NUCAPS fields of 850-mb dewpoint Temperature toggled with NAM40 and RAP40 estimates at approximately the same time, ~1800 UTC on 24 February 2020 (Click to enlarge)

Gridded NUCAPS fields include 850-mb dewpoint temperature fields, and this blog post compares the NUCAPS fields to model fields, and this is part of an ongoing series of blog posts on these horizontal fields.  The imagery above compares NUCAPS fields at 850 mb with NAM40 and RAP40 data over the southeastern part of the United States.  Very dry air is indicated over the western Atlantic Ocean north and west of the Gulf Stream.  There is generally good agreement between the NUCAPS and model fields. Model fields appear dryer (or NUCAPS fields are more moist).  Model fields show a pronounced gradient over the upper midwest that, at this scan time, were too far west to be viewed by NUCAPS.  (Click here to view the NUCAPS points — green, yellow and red — for this time, to show something about the data that has been input into the gridded fields).

However, the following pass from NOAA-20 (click here to view NOAA-20 orbit paths) included midwestern data.  Again, the general good agreement is obvious, especially with regard to the placement of the gradient. Is the atmosphere as dry as the model suggests at 850 mb? That’s a hard question to answer given 1200 UTC Soundings at Omaha, Minneapolis/Chanhassen, and Green Bay.  Note that NAM12 and RAP13 data are being shown in this example;  they gave mostly the same answer as NAM40/RAP40 used above.

NUCAPS fields of 850-mb dewpoint Temperature toggled with NAM12 and RAP13 estimates at approximately the same time, ~1900 UTC on 24 February 2020 (Click to enlarge)

Gridded NUCAPS over the southeast United States

February 21st, 2020 |

Analyzed snow depth at 0600 UTC on 21 February 2020 from the NOHRSC (Click to enlarge)

Snow fell over the southeastern United States, principally North and South Carolina, late on 20 February/early on 21 February 2020. This blog post, one in a series, investigates how gridded NUCAPS thermal fields perform in analyzing the rain/snow line. The snow totals are shown above, an image that was taken from this website at the National Operational Hydrologic Remote Sensing Center (NOHRSC).

NOAA-20 overflew the Carolinas shortly after 0700 UTC on 21 February, and gridded values of 950-mb, 900-mb and 850-mb Temperatures are shown below. (Note how the 950-mb field intersects the ground at the western edge of the Piedmont).  The 0º C isotherm at 850 and 900 mb is close to the coast;  it is sub-freezing over most of the land at those levels.  The analysis from 950-mb shows cold air stretching southwestward from southeastern Virginia, and that region is also where the accumulating snow was focused.  This is an argument in favor of the temperature fields in NUCAPS giving useful information about the rain/snow line.

850-mb, 900-mb, and 950-mb analyses of temperature derived from NUCAPS vertical profiles of temperature, 0723 UTC on 21 February 2020. The same color enhancement is used for each level, spanning -40º C to 30º C; 0º C is highlighted by the black line (Click to enlarge)

One of the gridded NUCAPS fields available in AWIPS via the Product Browser (there are many!) is the binary probability of a temperature occurring.  The 850-mb binary probability of 0º C is close to the coast, at 900-mb, just slightly inland.  The 950-mb values also suggest cold air is more likely over the region where snow fell.  There are also some embedded cold pockets at 950 mb over interior North/South Carolina.

Conditional Probability of 0 C at 850, 900 and 950 mb, 0723 UTC on 21 February 2020 (Click to enlarge)

Note that gridded NUCAPS fields include data from infrared retrievals, microwave-only retrievals, and from retrievals that do not converge. The gridding can mask behavior in the vertical profiles that might not necessarily engender confidence in a meteorological analyst. The plot below shows NUCAPS points (Green points are infrared retrievals that successfully converged, yellow points are microwave-only retrievals, and red points occur where the microwave-only and infrared retrievals failed to converge; this is typically where precipitation is falling) plotted on top of the 850-mb temperature analysis.  Note, however, that values do show up everywhere!  Users of the gridded data should keep in mind the quality of the data that goes into the analysis when they use it.  Two vertical soundings from which gridded data are derived are shown at bottom.  Users can decide if they would use those vertical soundings in isolation.

850-mb Temperatures with NUCAPS Sounding points superimposed, 0711 UTC, 21 February 2020 (Click to enlarge)

850-mb Temperatures with NUCAPS Sounding points superimposed, 0711 UTC, 21 February 2020. Two soundings are also shown, from a green point and from a yellow point.  Note that the plot also shows the binary probability of a temperature at 0º C (Click to enlarge)