NOAA/CIMSS ProbSevere for an isolated storm in Nebraska

July 9th, 2014 |
NOAA/CIMSS ProbSevere product, and National Weather Service Warning Polygons, 2302-2334 UTC 9 July 2014 (click to enlarge)

NOAA/CIMSS ProbSevere product, and National Weather Service Warning Polygons, 2302-2334 UTC 9 July 2014 (click to enlarge)

The storm in the animation above produced baseball-sized hail in Nebraska (Visible animation is here, courtesy Dan Lindsey from NOAA at CIRA) although MRMS Mesh Hail sizes were “only” in the 1-inch range (that is, nickel to quarter size). How did the ProbSevere product, which product includes MRMS Mesh size as a predictor, perform?

The visible and infrared satellite animation, below, shows quick development in the absence of cirrus obscuration, and the ProbSevere Satellite components from 2230 UTC are both characterized as ‘Strong’. The model components of ProbSevere (MUCAPE around 2000 J/kg, shear exceeding 30 knots) are also strong. Probabilities increased from 40% to >80% before the warnings for the cell were issued.

MRMS values in this case were not extreme; indeed, when the first warning was issued, MESH was still less than 1″ (but ProbSevere was >80%). Satellite growth rates and environmental information in this case compensated for the modest MRMS Mesh values.

GOES-13 Visible (0.63 µm, top) and Infrared (10.7 µm, bottom) from 2200 UTC 9 July through 0200 UTC 10 July (click to animate)

GOES-13 Visible (0.63 µm, top) and Infrared (10.7 µm, bottom) from 2200 UTC 9 July through 0200 UTC 10 July (click to animate)

Severe Weather in Nebraska

June 16th, 2014 |

Unusual twin tornadoes (click here for a summary of photos/videos from the Capitol Weather Gang) formed in northeastern Nebraska (Storm Reports from SPC) late in the afternoon of the 16th of June 2014. How did satellite data anticipate the development and progression of the severe convection? GOES-13 Sounder data painted a picture of ongoing destabilization in the area. For example, the CIMSS NearCast Product, which  arises from a two-layer Lagrangian Transport Model of Equivalent Potential Temperature, shows increasing stability in a forecast for 2100 UTC on 16 June in forecasts from 1800, 1900 and 2000 UTC, below.

CIMSS NearCast forecasts of Theta-e Differences between two layers, all at 2100 UTC, with initial times at 1800, 1900 and 2000 UTC (click to animate)

CIMSS NearCast forecasts of Theta-e Differences between two layers, all at 2100 UTC, with initial times at 1800, 1900 and 2000 UTC (click to animate)

The NearCast output, derived from GOES Sounder data, can predict in advance where axes of instability (and more importantly, where gradients in instability; see also comments on NearCast here and here) will occur. GOES Sounder data can also be used to diagnose the present state of the atmosphere. On this particular day, GOES Sounder estimates of Lifted Index (1400, 2000 and 0000 UTC) and CAPE (1400, 2000 and 0000 UTC) all showed ongoing destabilization over the Plains.

GOES-13 Sounder DPI Analyses of Lifted Index and Convective Available Potential Energy at 1400 and 2000 UTC on 16 June and at 0000 UTC on 17 June

GOES-13 Sounder DPI Analyses of Lifted Index and Convective Available Potential Energy at 1400 and 2000 UTC on 16 June and at 0000 UTC on 17 June

The products above outline the general area where convection might develop. Once the convection has developed, the NOAA/CIMSS ProbSevere product can be used to diagnose/monitor the likelihood of severe weather (large hail, strong winds, or tornadoes) developing — specifically, the likelihood of when severe weather might first occur. The animation below shows the evolution of the tornadic cell as it moved northeastward through Nebraska. Satellite predictors (Normalized Vertical Growth Rate and Maximum Glaciation Rate) for this cell were strong; both were observed at 1925 UTC, nearly an hour before the observed severe weather. ProbSevere first exceeded 50% at 1950 UTC, 13 minutes before the warning at 2003 UTC. 1-inch diameter hail was reported at 2016 UTC. The first tornado report occurred at 2040 UTC.

NOAA/CIMSS ProbSevere model

NOAA/CIMSS ProbSevere model

MUCAPE in the ProbSevere product above is around 4000-5000 J/kg. A special sounding at OAX (1900 UTC) shows similar CAPE values.

The Suomi NPP satellite had a timely overpass over the Great Plains at around 2000 UTC on 16 June 2014. NUCAPS Soundings from Suomi NPP are available, as plotted below, and can be used to estimate instability.

Suomi NPP VIIRS 11.45 µm imagery with NUCAPS sounding positions (Green Dots) Superimposed (Click to enlarge)

Suomi NPP VIIRS 11.45 µm imagery with NUCAPS sounding positions (Green Dots) Superimposed (Click to enlarge)

A sounding at ~42º N, ~97.8º W, below, shows CAPE values around 1000. However, note that the boundary layer temperature and dewpoint are too cool (surface temperature = 21º C) and too dry (surface dewpoint = 12º C). A benefit of the Sounding Software in AWIPS II, however, is that soundings can be easily modified. If the boundary layer is altered such that dewpoints are closer to observed METAR values (20º C), then CAPE values increase to 3000; if the temperatures are modified to be closer to observed values, CAPE increases to more than 4800.

NUCAPS Sounding at 42.07 N, 97.78 W, ~2000 UTC on 16 June 2014 (Click to enlarge)

NUCAPS Sounding at 42.07 N, 97.78 W, ~2000 UTC on 16 June 2014 (Click to enlarge)

Suomi NPP VIIRS data at different wavelengths (0.64 µm visible, 1.61 µm near-IR and 11.45 µm longwave IR), below, give a view of the storm just before severe hail was observed. The 1.61 µm imagery suggests a fully-glaciated anvil, and the 11.45 µm imagery shows evidence of several isolated overshooting tops.

Suomi NPP VIIRS data (0.64 µm, 1.61 µm and 11.45 µm) at 2004 UTC on 16 June 2014 (Click to animate)

Suomi NPP VIIRS data (0.64 µm, 1.61 µm and 11.45 µm) at 2004 UTC on 16 June 2014 (Click to animate)

Click here for a visible image animation from GOES-13; here is an infrared image animation. The famous twin tornadoes in Elkhart, IN, during the Palm Sunday outbreak in 1965 can be seen here.

Severe thunderstorm over the Black Hills of South Dakota

May 27th, 2014 |
GOES-13 10.7 µm IR channel images (click to play animation)

GOES-13 10.7 µm IR channel images (click to play animation)

An isolated severe thunderstorm developed over the northern portion of the Black Hills of South Dakota around 18 UTC (Noon local time) on 27 May 2014, and moved southeastward producing hail as large as 2.75 inches in diameter and a tornado (SPC storm reports), as well as up to 3 inches of heavy rainfall. 4-km resolution GOES-13 10.7 µm IR channel images (above; click image play animation) showed the cold cloud-top IR brightness temperatures associated with the storm (which reached a minimum of -59º C at 21:40 UTC). Convective initiation was aided by convergence of a surface cold frontal boundary with the topography of the Black Hills, as seen here.

A comparison of 375-meter resolution Suomi NPP VIIRS 0.64 µm visible channel and 11.45 µm IR channel images at 19:38 UTC (below) revealed a minimum cloud-top IR brightness temperature of -68º C. Subsequent cumulus cloud development is seen to be suppressed in the stable outflow region in the wake of the storm.

Suomi NPP VIIRS 0.64 µm visible channel and 11.45 µm IR channel images

Suomi NPP VIIRS 0.64 µm visible channel and 11.45 µm IR channel images

A comparison of the VIIRS IR image with the closest available GOES-13 IR image (below) demonstrated the advantage of higher spatial resolution of polar-orbiting satellite imagery, as well as the lack of parallax error associated with the geostationary-orbit GOES imagery. The coldest cloud-top IR brightness temperature on the GOES-13 image was -57º C, compared to -68º C on the VIIRS image.

GOES-13 10.7 µm IR channel and Suomi NPP VIIRS 11.45 µm IR channel images

GOES-13 10.7 µm IR channel and Suomi NPP VIIRS 11.45 µm IR channel images

The NOAA/CIMSS ProbSevere model identified this storm as a potential producer of severe weather. The animation below shows the evolution of the MRMS radar signal with ProbSevere overlain from 1812 UTC through 2000 UTC. The Radar object that is highlighted has strong satellite growth rate (observed at 1725 UTC), MUCAPE of 1750 J/kg and Effective Shear of ~19 knots. At 1810 UTC, when MESH values are 0.53″, ProbSevere is 18%; four minutes later at 1814 UTC MESH increased to 0.94″ and ProbSevere increased to 69%. The National Weather Service issued the first warning at 1905 UTC and severe hail (1.5″ in diameter) occurred at 1910 UTC (when MESH was 1.64″ and ProbSevere was 94% and less than an hour after ProbSevere increased above 50%).

NOAA/CIMSS ProbSevere over South Dakota, times as indicated (Click to enlarge)

NOAA/CIMSS ProbSevere over South Dakota, times as indicated (Click to enlarge)

ProbSevere later in the day, after 2200 UTC, continues to track the hail-producing storm in far southwestern South Dakota. Although ProbSevere is designed to show when the first severe reports might occur, it does continue to provide useful information. In time, as below at 2222 UTC, the satellite growth parameters are replaced by ‘Mature Storm’ as a reminder that this tracked feature is not new.

Annotated NOAA/CIMSS ProbSevere over South Dakota, times as indicated (Click to animate)

Annotated NOAA/CIMSS ProbSevere over South Dakota, times as indicated (Click to animate)

Hail-Producing storm over the Texas Hill Country

May 9th, 2014 |
ProbSevere_09May2014_1400

AWIPS-2 display of a strong thunderstorm over Texas, including ProbSevere Readouts at 1400 UTC

A hail-producing storm (cick here for Storm reports from SPC) moved through Edwards, Real and Bandera counties of Texas after sunrise on May 9th. This storm gives a nice opportunity to compare ProbSevere and GOES-14 SRSO-R depictions of a severe storm. The 1400 UTC image, above, shows ProbSevere > 95%, with a MESH just over 1″. Within the next 10 minutes, MESH increased to 1.48″ at 1406 UTC (when ProbSevere was 96%), and to 1.92″ at 1408 UTC (see image below, when ProbSevere was 99%). A Severe Thunderstorm Warning was issued at 1408 UTC.

ProbSevere_09May2014_1408

AWIPS-2 display of a strong thunderstorm over Texas, including ProbSevere Readouts at 1408 UTC

The storm maintained its strength over the following half hour. Imagery from 1436 shows MESH values near 1.50″ and ProbSevere is at 99%. At 1442 UTC, the radar shapefile vanishes from the display, a result of processing errors, but it is back at 1444 UTC. Note that the Satellite Predictors have flipped at 1444 UTC from values derived at 1015 UTC to ‘Mature Storm‘ (this change has nothing to do with the missed processing at 1442 UTC). Recall that ProbSevere is designed to probabilistically determine whether or not a storm will produce severe weather in the next 60 minutes. ‘Mature Storm’ designations serve as a reminder that Probabilities have been elevated for a long period of time.

At 1458 UTC, below, near the end of the Severe Thunderstorm warning, the MESH value that is incorporated into the ProbSevere computation has decreased to 1.17″, and ProbSevere has dropped to “only” 97%. At 1502 UTC, however, MESH has started to increase again, to 1.21″ (and ProbSevere remains high); a second severe thunderstorm warning is issued at 1504 UTC when MESH is at 1.53″ (and ProbSevere is at 99% again). ProbSevere remains high through 1530 UTC.

ProbSevere_09May2014_1458

AWIPS-2 display of a strong thunderstorm over Texas, including ProbSevere Readouts at 1458 UTC

GOES-14 was in SRSO-R mode during this hail event, allowing an opportunity to see the storm evolution at very high temporal resolution. (The storm initially was right at the edge of the domain). The animation below shows cloud-tops warming around 1440 UTC, before cooling again, consistent with changing updraft speeds that can be inferred by changes in MESH. Cold temperatures occurred at 1501 UTC, 209.1 K. Temperatures were cooler than 210 K only at 1500 and 1501 UTC — that is, for two minutes — demonstrating the importance of 1-minute imagery in resolving without aliasing the coldest features at cirrus level. A second very cold event occurred between 1507 and 1509 UTC (brightness temperatures were cooler than 208 K); it was gone by 1511 UTC (when brightness temperatures were all warmer than 211 K). One-minute imagery is necessary to resolve these very rapid changes at cirrus level.

ProbSevere_09May2014_1458

GOES-14 Visible (0.62 µm) (top) and GOES-14 IR (10.7 µm) (bottom) from 1401 UTC through 1530 UTC on 9 May. Edwards, Real and Bandera Counties are highlighted

The plot below shows the coldest IR Brightness Temperature observed in the GOES-14 10.7 µm channel over the hail-producing storm. Tic-marks along the x-axis are at 5-minute intervals, and there are large differences that occur with time-scales shorter than 5 minutes. This is consistent with the findings of Cintineo et al. in the September 2013 issue of Journal of Applied Meteorology and Climatology (link).

BT_CloudTop_SRSO

Minimum cloud-top IR Brightness Temperature over hail-producing storm, from 1411-1530 UTC. Tic marks on x-axis every 5 minutes.

During the following hours, 1-minute interval GOES-14 0.63 µm visible channel images (below; click image to play animation; also available as an MP4 movie file) showed other areas of convection which produced damaging winds across parts of southeastern Texas.

GOES-14 0.63 µm visible channel images (click to play animation)

GOES-14 0.63 µm visible channel images (click to play animation)