Convective Initiation – Verified!

Forecaster comments from EWP blog...

This series of images shows an ideal case of the CI product detecting a rapid development of storms along a cold front in northeast Arkansas.  The first image shows a strong CI signal (red) at 19:10.

The second image shows the same area about 10 minutes later at 19:25.  The weird tooth-shaped object is the Cloud Top Cooling rate and this indicated -16 deg/15 min.

Finally, at 21:25, about two hours later, the visible imagery showed a developing storm with an overshooting top.

CI Within Nearcasting Instability Max

UW-CTC is depicting strong cloud-top cooling and convective initiation within an area of higher instability that has been forecast by the UW Nearcast model.  The warning forecasters in ALY are watching this storm as it is isolated and ahead of the main line.

VIS and UW-CTC (top left), ENX Base Reflectivity (top right), Nearcast 780-500mb Cape (bottom left), and Nearcast Thta-E Difference (bottom right).

CI products still useful

Forecaster comments from EWP blog...

After the convection developed rapidly, the CI and CTC products became masked due to the TS plumes.  However, the convection at the north end of the CWA may have kicked out an outflow boundary to cause more development in clear air.  This was picked up by the CI product with an increasing signal in subsequent frames.

A stronger (yellow) signal was evident on the next frame.

The resulting composite reflectivity (upper right image), showing development to 45dbz 15 minutes later.  Comp reflectivity was used here due to beam blockage at 0.5 degrees.

Benefits of pre-operational demonstrations

During yesterday's forecasting exercises, a forecaster captured an event that raised a few questions during the case and in our daily debriefing today regarding the cloud-top cooling rate product.  Below is an excerpt from their EWP blog post, including the images they captured...

"Here is the series of visible and infrared satellite pictures for the storm cell that we issued the tornado warnings.  The Convective Initiation product gave a strong signal as this cell began developing, but the Cloud Top Cooling product never identified this storm.  The infrared pictures are sampled at the coolest part of the cell at each time step, about 5 to 10 minutes.  The storm grew explosively and produced one inch hail along with strong rotation on the radar."

Upon some investigation, it was discovered that the growing cloud in question was being affected by the neighboring mature storm a few pixels away within the cloud-top cooling rate products cloud mask algorithm.  This example has been flagged by the developers as a good case for some potential improvements (which they have already quickly identified) to the way the algorithm classifies the scene in which cloud-top cooling rates are calculated.

This is a good example of why bringing products into a operational-type setting and demonstrating them in an experimental sense is very beneficial towards improvement and a successful transfer from research to operations.  In this case, the developers were able to quickly identify the issue that the forecaster expressed confusion with.  Until products are run in real-time over a large domain, some small events may be missed that are still important to an operational user's perspective.  It is essential to identify and address issues within any experimental product prior to operational implementation.

Nearcasting Forecasts Instability in ALY Area

Storms west of Albany, NY are forecast to move into an area of higher instability that is forecast by the UW-CIMSS Nearcasting Model.  New midlevel (780-500mb) CAPE output is depicting an area of 500-700 J kg-1 as the storms approach from the west.  This would support the storms intensifying and growing in coverage as they continue to move eastward.

VIS (top left), ENX Base Reflectivity (top right), Nearcast 780-500mb Cape (bottom left), and Nearcast Thta-E Difference (bottom right).

EWP daily debrief 5/16

Below is some feedback gathered from today's daily debrief of the EWP forecasters...

SATCAST / UW Cloud-top cooling
- For Davenport, IA CWA... "The strength of signal never really got above yellow (~ 60-70%)... the rest were pretty marginal... there were a few interesting CTC signals up to -22 C/15min... the ones that flagged the strongest would shift around, so it was tough to get a feel for the lead time or false alarms.  The CTC signals would appear once and then go away, so it was hard to get some continuity... generally they were -10 to -15 C/15 mins... so that fit our expectation that it wasn't a very favorable environment for severe weather... collectively they both handled the situation well, but if you handle it on a case by case basis it would have been tougher to judge."
- For the Melbourne, FL CWA... "The two CI hits we had in the red (mid 90s %) had about a 24 minute lead time on when it showed up well on radar and turned out to be the stronger storms.  The CTC didn't get the second storm, but the CI did, and that was the one that had the tornado.  The higher CI probabilities really helped me tell that the storms were intensifying versus what I could already tell on visible."
- "One of the interesting challenges with these CI/CTC algorithms... we've had plenty of training on how these products work, but there are still those moments when we're trying to figure out why something didn't happen... that's something that will come with time and more use."

Nearcast
- For Davenport, IA  CWA... "The theta-e difference was a pretty good indicator that it was going to be a pretty marginal day."
- "The theta-e difference picked up nicely on the fact that there were stronger signals in N. MI and that correlated well to the stronger storms"

PGLM
- For the Melbourne, FL CWA... "It showed some signals, but I'm not sure it would have had much of a role in my warning process yesterday."
- "Another application would be where there isn't much radar data... if I was out west forecasting for a fire where there was no radar coverage, it would be very useful in letting people know where the lightning was."
- "From an aviation perspective, especially with a situation where there is a line of storms, this would be very useful in increasing situational awareness."
- "Lightning could have interesting heavy rain/flash flood impacts as well."

Day 2 at HWT

Although there has been a general lack of wide-spread,exciting, edge-of-your-seat, thunderstorms to warn on this week, forecasters have had plenty of opportunity to play with this year's assortment of experimental weather products. UAHuntsville's convective initiation product, or SATCAST, is sporting it's new "strength of signal" look and functionality. Some successes, watching the development of storms along a frontal boundary stretching through Iowa, UAH SATCAST followed the growing clouds from birth to rain, with values ranging from 20-70 on the strength of signal scale. These clouds went on to produce minimum 35dBZ echoes downstream. Overall, the area did not experience any severe weather, but SATCAST was able to differientiate between the weaker storms and the somewhat stronger storms of the day. (See http://goesrhwt.blogspot.com/2012/05/ci-algorithms-over-iowa.htm for more information). Another success, using SATCAST over the Melbourne, FL warning area, the forecasters report good utility of the strength of signal format indicating the growth of a cloud over subsequent satellite images, one of which resulted in a storm that produced a tornado warning. ( See the post at http://goesrhwt.blogspot.com/2012/05/experimental-warnings-north-of-cape.html). Some not as successful points, working in Roanoke, VA in the early afternoon. Cloud cover and cirrus contamination made using the convective initiation products problematic for the forecasters. (See http://goesrhwt.blogspot.com/2012/05/weak-convection-in-rnk-cwa.html). Although surrounding areas were giving indications of CI, the WFOs remained masked out. Overall, the lack of specific severe weather events monopolizing the time of the forecasters really allowed each of the four participants the time to really dig into the products made available to them this week. One hopes that now that the forecasters are more familiar, the weather will cooperate and give them something a bit more interesting to play with. Time will tell.

CI: day vs. night... advantages of higher resolution data

Forecasters operating over the Davenport, IA CWA area brought up some questions regarding what they were seeing in the UAH-CI SATCAST product's output this evening.  There was some confusion why all of the sudden the amount of objects (and shape of objects) being forecast for CI rapidly increased over one satellite scan.  The forecaster spoke with the UAH visiting scientist about this and it was explained very well to the forecaster that the issue is the switch from the products daytime-only cloud mask to the nighttime cloud mask.

The UAH-CI SATCAST product makes use of the visible data within its cloud mask when it is available to aid in identifying small growing cumulus clouds.  During the day, this is extremely beneficial, but since visible data is not available at night, the product is unable to make such high spatial identifications of growing clouds.  This really shows the benefit of having higher resolution satellite data like that which will be available on GOES-R's ABI during all times of the day.... not only for imagery, but for derived products and decision support tools as well.

 UAH-CI SATCAST in daytime mode.

UAH-CI SATCAST in nighttime mode.

Nearcasting CAPE in response to forecaster comments

Early forecaster comments on the nearcasting revealed a general difficulty comparing the vertical theta-e difference and related algebraic representations of the mid-tropospheric instability forecasts from the Lagrangian trajectory model to other stability parameters and other sources (model forecasts and analyses, etc.).

In response to this, product developers Jordan Gerth (University of Wisconsin) and Robert Aune (NESDIS/STAR/ASPB Madison, WI) participating in the testbed this week formulated a simple convective available potential energy (CAPE) calculation using nearcasting model output with the 780 hPa equivalent potential temperature (theta-e) as the parcel and 500 hPa theta-e as the environment.

With CAPE (units of J/kg or m² s^-2), users can easily understand the physical significance of the information provided by the GOES Sounder and watch the evolution.  It is believed that this will be ideal for following the growth and decay of elevated instability.  Even though this is not a full-profile CAPE calculation, the 780-500 hPa layer potential energy is particularly important for storm development and updraft strength.

The CAPE calculation was recently implemented in AWIPS II for the remainder of this year's Hazardous Weather Testbed.  Furthermore, the new layer CAPE output from the nearcasting model is available online at http://cimss.ssec.wisc.edu/model/nrc/ in the "Under Development" section.  It was recommended that forecasters compare to fields in AWIPS II and on the SPC mesoanalysis web page.

CI algorithms over Iowa

Forecaster comments from the EWP blog...

While thunderstorms have not been as impressive over Iowa, Illinois, and Wisconsin, the UAH convective initiation and UW cloud-top cooling algorithms have been increasingly active over the last hour.  To track their effectiveness compared to radar, I overlaid the MRMS merged reflectivity at lowest altitude product with the UAH CI and UW CTC all on the same image (something that wasn’t available with AWIPS 1).  There was an interesting evolution with the UAH CI starting with the 2132 UTC satellite image, when it flagged an area of CI near West Amana, IA (south of Blairstown) with a 28 on the new Strength of Signal scale.

By 2144 UTC, a 20 dBZ shower had developed with that CI object.  By 2200 UTC, CI was up to 44, and the radar reflectivity was up to 48 dBZ.

Shortly thereafter it maxed out at 63 CI SOS; the cell’s reflectivity maxed out between 2226-2234 UTC at 52 dBZ.

A neighboring storm near Lake Iowa triggered both the UAH (74) and UW (-13 C/15 min) algorithms; it too maxed out around 50 dBZ but was shorter-lived.

Viewing the CI and CTC data simultaneously with radar imagery allows forecasters to make direct comparisons to their most frequently-used warning tool.