EWP: UW-CTC captures sea breeze convection

Ahead of the frontally forced convection over the southeastern US, the sea breeze along coastal South Carolina provided a trigger for thunderstorm development late this morning.  Figures 1 and 2 below show the UW-CTC algorithm capturing these newly developing thunderstorms at 1632 and 1645 UTC 09 May 2012, respectively.  These storms went on to produce 50-60 dBZ composite reflectivity (not shown because radar data was not available at those times).   The figures below show GOES IR-window brightness temperatures (top left), GOES visible reflectance and NLDN lightning detection (top right), UW-CTC rates (bottom left), and UW-CTC ice mask (bottom right).

 Figure 1.  Valid 1632 UTC 09 May 2012

Figure 2.  Valid 1645 UTC 09 May 2012

EWP: UW-CTC captures southwest building line of thunderstorms

Another day at the EWP with storms initiating in the late morning so immediately grabbing screen shots and getting forecasters to look at the cloud-top cooling products while they are relevant is a challenge.

Below are some figures of the UW-CTC rate products (bottom left panels) from AWIPS-II  showing the southwestward building convection along a cold front over southern Georgia.  Since the radar environments were just initialized to the southeastern US @ EWP we do not have radar data from this morning, so NLDN CG lightning strikes are plotted atop the GOES visible imagery (top right panels).  The GOES IR-window brightness temperatures are plotted in the top left panels.  The bottom right panel is the UW-CTC ice mask, which identifies where before the inclusion of GOES optical depth retrievals, it was not possible to identify cloud-top cooling rates.

The progression of figures 1-4 below show UW-CTC identifying newly developing cooling cloud tops.  Figure 1 valid at 1615 UTC 09 May 2012 is an example of the UW-CTC algorithm identifying a vertically cooling convective cloud beneath thin cirrus clouds, due to the inclusion of GOES optical depth retrievals (note how the UW-CTC signal is occurring in the area of 'ice mask').  Figures 2-4 are a progression from 1715 UTC - 1745 UTC showing the convective clouds developing along a cold front over southern Georgia.

Figure 1. Valid 1615 UTC 09 May 2012, note the UW-CTC signal over southeastern Georgia (bottom left) occurring in the area of thin cirrus clouds (indicated as blue areas in the bottom right).

Figure 2.  Valid 1715 UTC 09 May 2012.  UW-CTC is identifying developing convective clouds over southern Georgia.

 Figure 3.  Valid 1732 UTC 09 May 2012.  UW-CTC is identifying developing convective clouds over southern Georgia.

Figure 4.  Valid 1745 UTC 09 May 2012.  UW-CTC is identifying developing convective clouds over southern Georgia.

EWP daily debrief 5/9

Once again, we had our debrief with the NWS forecasters participating in the EWP portion of the Spring Experiment.  Below are some comments we got on the products we were demonstrating during yesterday's events...

SATCAST / UW Cloud-top Cooling
- "Convection was ongoing for most of the period, but there were several instances where both CI products would indicate something and we couldn't use radar because there was no coverage... but there were several instances where it definitely showed CI."
- "The CI did really well... the CTC especially did really well where we had no radar coverage... the CTC really showed consistent signatures that were tied to strong updrafts... so there was more than just some value, I think it would be a very helpful tool to keep following."
- "We could see the storms show up in the more sensitive product (SATCAST) first and then it would show up in the CTC and give us an idea of strength... the strength of signal stuff really helped identify the storm progression... we could pick out areas where conditions were improving."
- "One and a half hour lead time on 60 dBZ echo... it was really cool to see that lead time."
- "I really believe that we are going to be able to use the product in diagnosing storm strength and use this for advisory purposes."
- "I can see where we could put out a product before we even see the cells develop... especially for severe weather... we could let people know where we are watching and alert them to the possible threat for severe or tornado."
- "From and aviation perspective, I can see these products being really helpful in forecasting the convective coverage, even if it doesn't become severe."

Nearcast
- "Tried to look at it, but it wasn't too beneficial because there was a lot of missing data due to ongoing convection." (Brownsville)
- "Was very useful in detecting the lower theta-e... saw some successes Monday and saw them again yesterday...  projecting a sharp axis of moisture... came in really handy because we could see that moisture was still present after sunset." (El Paso)

Simulated Satellite
- "I wanted to see before the storms started, I see it as a product you could use beforehand... It was timing the actual initiation of the storm quite nicely... you could see the jet and weak perturbation that was triggering the convection... It really helped out."
- "We made a neat procedure... could show the simulated overlaid with the observed and you could really see how the situation would evolve with time."
- "I was really impressed on how well it picked up on the whole pattern, convective initiation and location... the purpose is to get a whole 3D representation and I liked that aspect."
- "Would be useful to identify low clouds and their trends."

Overall / Training
- Forecaster mentioned that providing one end to end product from growth to severe or tornado threat will be the future... being able to communicate your confidence as the situation develops to the user in a constant fashion rather than having a number of different alerts being put out by the NWS, TV and internet is how they envision communicating hazard information in the future.

MAF Mesoscale Desk @ 01Z

Forecaster experimental discussion from EWP blog...

"23Z GOES Nearcast for 01Z supports strong updraft potential for the Brewster county storms.  Legacy Hail algorithm maintained 2 inch potential for a couple scans, so the potential for severe size hail is likely to continue.  This hail threat should continue through 03Z.  EMs and other decision makers should continue to monitor for potential warnings and statements."

SATCAST UAH CI Strength of Signal Provides Early Detection of Redeveloping Thunderstorms

Provided is a sequence of consecutive AWIPS-II display images highlighting the growth of new convective development on the south side of a thunderstorm complex that has just moved over this region.  The UAH CI "Strength of Signal" (SS) product shows very weak signals (< 30) of near-future CI likelihood early on       (Figure 1) from 1832 UTC, but then steadily increases its index values as new GOES satellite scans come in (Figure 2 through Figure 5).

Once the convective cloud top begins to approach the tropopause by 1932 UTC, its cloud top cooling rate (per the 10.7 um Brightness Temperature) decreases, which corresponds with a decrease in vertical growth.  As this occurs, the SS index values level off and even drop a little (Figure 6).  In the next GOES satellite scan at 1940 UTC, the UAH CI algorithm now begins to consider this convective cloud group as "mature" and ceases to consistently track and process it.  Our attention now turns to the radar, where the first appearance of an associated radar echo is expected in the near future (Figure 7).

Figures 8, 9, and 10 show the sequence of radar images leading to the eventual detection of convective rainfall by radar.  The first detected instance of a 35 dBZ or greater reflectivity associated with this convective cloud development occurs at 2012 UTC, about 1 hour and 40 minutes after the UAH CI algorithm began tracking and processing this convective development.

Figure 1.  UAH CI "SS" valid from 1832 UTC (Left).

Figure 2.  UAH CI "SS" valid from 1845 UTC (Left).

Figure 3.  UAH CI "SS" valid from 1902 UTC (Left).

Figure 4.  UAH CI "SS" valid from 1910 UTC (Left).

Figure 5.  UAH CI "SS" valid from 1915 UTC (Left).

Figure 6.  UAH CI "SS" valid from 1932 UTC... Notice that the index signal value has decreased as the cloud top has slowed its vertical growth and decreased its cooling rate as it has neared the tropopause (Left). At this time, there is still no radar echo associated with this new convective cloud development (Right).  

Figure 7.  At 1942 UTC... Still focusing on the region in the radar where new convective development is expected (Right).

Figure 8.  At 1948 UTC... Still focusing on the region in the radar where new convective development is expected (Right).

Figure 9.  At 2000 UTC... We are beginning to see the first semblance of a radar echo associated with the expected convective development, as highlighted by the UAH CI "SS" product (Right).

Figure 10.  At 2012 UTC... Now we see the convection increasing in intensity, per the radar reflectivity crossing above the 35 dBZ echo threshold (Right).

EPZ Meso Update 2115Z

Forecaster comments from EWP blog...

The synthetic WRF run at 05082012 at 0Z correctly predicted several upper level features and mesoscales features correctly for the afternoon of 05082012. The following 4 images show synthetic WRF images of IR (upper left), WV (upper right), and 10.35u-12.3u IR Band Difference (lower left).  Looking at the WV imagery you can see an upper low over the U.S. Southwest as well as an upper trough over the upper Midwest.  In the IR and WV imagery you can see convection occurring from AZ to NM to TX to CO to Old Mexico.  In the brighter white areas of IR, you can see low level cloud cover.  In the Band Difference image, you can see the area of best moisture and “hot spot” for convective initiation in the southern tip of TX and into Mexico.  The 5th image below is the true observed IR (upper left) and WV (upper right) which verifies the synthetic WRF.  This imagery would be very helpful to forecasters in the field because it seems to do a decent job of predicting upper level features, low level clouds, and convection 12-20 hrs in advance.

17Z Synthetic:

18Z Synthetic:

19Z Synthetic:

20Z Synthetic:

20Z Observed IR and WV:

EPZ Warning Desk update @ 2100Z

Forecaster comments from the EWP blog...

Thunderstorms were strengthening south central NM.  Combination of legacy and GOES-R Nearcast product support continued development.  Another area of storms were located just south of Luna and Dona Ana counties and moving north.  We expect this activity to hold together and are likely to impact these two counties within the 1-2 hours.

EPZ Meso Update 2050Z

Forecaster comments from EWP blog...

CI products were able to correctly depict a small hail producing t-storm about 1.5 hours ahead of time.  In the first image below, the UAH CI showed a 53 index (upper left) indicating significant cloud growth at 1732Z.

The next image below shows CIMSS CI of -7 to -8 K/15 min at 1745Z (lower left) which is typical of weak storms based on recent studies.

 

The next image below at 1901Z depicts the same storm, 1.5 hrs later when it first reaches 60 dbz in KEPZ comp reflectivity (upper right) and 77-80 DVIL (lower right).  At this point the storm is likely producing small hail.

 

Another point to take away from this case is that for this environment, it takes about an index of 50 to 60 UAH CI to produce strong storms.

Convection Initiation over Mex state of Coahulia @ 1732z

Forecaster comments from EWP blog...

At 1732z…UAH-CI showed a signal strength in excess of 90 and the UW-CTC showed cloud top cooling of -9 degC.   Storm did develop in the area where the two values were seen around 1745z and continued thru 1845z.  Feel this is a success for both methods.

UAH CI "Strength of Signal" Probability Forecasts Up and Running at 2012 GOES-R Proving Ground

With the EWP forecasters stationed in southwest Texas today, we were noticing some of the newer instances of convective initiation flaring up just across the Rio Grande River in Mexico.  Provided are a couple of the first views of the new UAH CI  SATCAST Strength of Signal (SS) output, as seen in the AWIPS-II display.

Figure 1.  AWIPS-II screenshot of the UAH CI Strength of Signal output from 1732 UTC (top-left pane) and corresponding satellite IR image (top-right pane), from the same satellite scan.

Figure 2.  AWIPS-II screenshot of the UAH CI Strength of Signal output from 1845 UTC (top-left pane) and corresponding satellite IR image (top-right pane), from the same satellite scan.

This new version of the SATCAST algorithm provides a continuous scale from 1-100 of the strength of the satellite-derived signal for near-future CI.  Per forecaster feedback from last year's 2011 SPC Proving Ground, this new product has taken a step away from the binary "Yes/No" output CI forecasts and now combines all of the available satellite spectral information into a single numerical value.  Higher values correspond to faster rates of cloud top cooling rates, colder/higher cloud tops, and wider updraft widths.

The Strength of Signal values are currently derived from the early beginnings of an ever-growing neural network in which the algorithm is trained from an empirical database with the correct answers of what immature forms of convective clouds go on to produce CI and which ones do not.  As this database continues to grow over the coming weeks, months, and years, the algorithm will "grow smarter".

Furthermore, this new paradigm is paving the way to soon begin training the algorithm with input from Rapid Refresh Numerical Weather Prediction model data. As this transition occurs in the near-future, the output will look the same, but it will become increasingly more meaningful as a true probabilistic product.

This work ties into ongoing research with NOAA's Earch System Research Laboratory (ESRL) in which the SATCAST data is being assimilated into developmental versions of the Rapid Refresh (RAP) and High Resolution Rapid Refresh (HRRR) models in order to "put the convection in the right place" for hot-starting the models, inevitably leading to more accurate short-term model output.  To that end, we envision a truly integrated and data-fused product in which a positive feedback effect occurs as short-term model data characterizing the broad-scale atmospheric environment is ingested by the CI algorithm in order to produce more accurate CI forecasts, and these more accurate CI forecasts are assimilated into the short-term prediction models in order to significantly improve how and where the models are simulating CI.  If the models can get the placement of early convection correct, then their longer term forecasts will take a giant leap forward in terms of accuracy in all aspects of their output.