BRO MESO Desk @ 20Z

Forecaster comments from EWP blog...

The following four screen captures depict several favored areas of UAH-CI strength of signal along the outflow boundaries that exist immediately north of CWA and north-south along a line from Los Angeles-Realitos-Encino-Raymondville.   Knowing were the boundaries were located,  UAH strength of signal allowed us to focus on the most favored areas between 20Z-21Z.   (Synthetic Sat showed the most likely CI would take place between 20-21Z, particularly with storms along the Rio Grande.  It didn’t capture any activity over the northern extent of the outflow boundary.)

BRO Afternoon Forecast 1920Z

Forecaster comments from the EWP blog...

Synthetic WRF imagery was useful in making a convective forecast for Brownsville this afternoon.  At 13Z, the synthetic WRF depicted two areas of convection over TX with the south complex near Brownsville being slightly displaced to the south in location.  This convection was occurring in an area of high shear ahead of an upper level low progress slowly across the southwest U.S. (seen in synthetic and observed imagery).

13Z Synthetic:

13Z Observed IR/WV (top 2 panes):

At 19Z synthetic and observed IR/WV imagery matched up well with the southern complex, but over did convection to the south of the northern complex in TX.  Over Brownsville area, visible satellite imagery depicted a low broken cu field at 1914Z which matches low clouds in the synthetic IR imagery.

19Z Synthetic:

19Z Observed IR/WV (top 2 panes):

1914Z Visible Sat:

Conclusion:  The Synthetic WRF model seems to be doing an OK job overall so far today with a few issues.  However, it depicts the upper low driving convection today as well as 2 distinct areas of convection over TX.

For Brownsville area for the rest of the afternoon, synthetic images below predict convective initiation around 20Z with convection strengthening through the afternoon hours (22Z image), and then beginning to decrease in strength at 0Z.

20Z Synthetic:

22Z Synthetic:

0Z Synthetic:

BRO MESO desk @ 1834Z

Forecaster comments from EWP blog...

A few towering cu were present across BRO.   CTC centriods were present in LaSalle county (CRP) an immediately south of Starr county.  In fact, the LaSalle county CTC also showed a high strength of signal on the UAH-CI.

EWP: UW-CTC 45 minute lead-time on severe hail

Another day at HWT we are focusing on southern Texas.  An upper level low over northern Mexico is finally beginning to move eastward, this combined with humid low-level flow off the Gulf of Mexico is setting the stage for severe thunderstorms over southern Texas today.  There were numerous severe thunderstorms ongoing as of late morning.  One of the most recent developing storms is showcased below.  At 1632 UTC 10 May 2012, the UW-CTC product showed cloud-top cooling rates of ~ -6K / 15 min.  The following satellite scan at 1645 UTC the UW-CTC increased substantially to over -20 K / 15 min.  Later images of radar-based maximum expected hail size show estimated severe hail occurring at 1715 UTC and significant hail in excess of 3.00" at 1725 UTC. 

Figure 1.  Valid 1632 UTC 10 May 2012.  The UW-CTC product indicates initial substantial vertical growth of a developing thunderstorm (red circle).

Figure 2.  Valid 1645 UTC 10 May 2012.  The UW-CTC rate on the developing storm explodes into the strong category with cooling rates near -23 K / 15 min (red circle).

Figure 3.  Valid 1715 UTC 10 May 2012.  The bottom right panel shows radar-based maximum expected hail size exceeding severe thresholds (1.00", red circle), 45 minutes after the initial UW-CTC signal.

Figure 4.  Valid 1725 UTC 10 May 2012.  The bottom right panel shows radar-based maximum expected hail size of over 3.00 (red circle)", 55 minutes after the initial UW-CTC signal and 40 minute after the maximum UW-CTC signal.

Another CI success story @ 2032z

Forecaster comments from the EWP blog...

At 2032z…storm was pushing offshore from St Johns County in Fl.  The CTC product indicated strong cooling rates approaching -20 DegC/15 minutes associated with that storm.  There was a short period of time when we did not receive any products.  However…the radar showed a stronger storm with increased lightning activity by 2112z.

EWP: GOES optical depth retrieval enhances UW-CTC

The inability for the UW-CTC algorithm to detect newly developing convection beneath thin cirrus clouds was the primary deficiency identified by forecasters in previous experiments.  In response to forecaster feedback, a recent proposal was awarded to UW/CIMSS to include GOES optical depth retrievals into the UW-CTC algorithm for the purpose of detecting these type of developing thunderstorms.  Thin cirrus clouds are frequently present related to anvil blow off of nearby storms, decayed convection from previous days, and related to jet streaks.  Figure 1 below provides another example of detecting a newly developing convective cloud in areas of thin cirrus.  The UW-CTC value is ~ -9K /15 min and the tower beneath the thin cirrus clouds can be seen in the upper right panel highlighted by the red circle.  Since this storm was out of our focus domain, radar data is not available, but NLDN detected the first lightning strike about 10 minutes later.

Figure 1.  Valid 2115 UTC 09 May 2012.  Note the developing thunderstorm beneath cirrus blow-off from storms to the south (red circle, top left panel).  The corresponding UW-CTC was ~ -9 K /15 min.  The bottom right panel shows the UW-CTC ice mask, which shows areas where detection was not possible prior to the inclusion of GOES optical depth retrievals.

CHS Desk – UAH CI product/Visible-Radar 2014Z

Forecaster comments from EWP blog...

Two images are shown here.  On the left, a line of storms extends along the South Carolina coast and into portions of southern Georgia at 2014Z. Earlier at 1945Z, on the right, the UAH Convective Initiation product showed 3 areas of 60%.  Initiation was predicted along and just ahead of an outflow boundary, and just off the coast oriented west southwest to east northeast.  The product performed very well, with around 30 minutes of lead team for the updrafts that exhibited +35dBZ.

JAX Cloud To Cloud Lightning near MLB 2035Z

Forecaster comment from EWP blog...

Cloud to cloud lightning was observed in a very weak storm in Osceola County.  pGLM flash rates are observed in the upper right and lower left images depicting the cloud to cloud lightning.  Cloud to ground lightning was not observed with this storm as depicted by the absence of strikes in the lower right panel.  This data proves that this weak storm had an updraft strong enough to produce cloud to cloud lightning but not cloud to ground lightning.

I feel that adding cloud to cloud lightning data will really help storm warnings, fire weather applications, aviation forecasts, etc. if added as a baseline product for WFOs because we’re obviously missing a lot of lightning data not only in weak convection as this example shows but in strong/severe storms as well.

EFP simulated satellite evaluation

Part of this year's EFP portion of the Spring Experiment is to examine simulated satellite imagery from the CAPS ensemble.  The goal of this project is to determine how differences in the microphysics schemes in the different model runs produce different cloud objects and features in the simulated imagery.  While this is not directly related to GOES-R, it is exposing the modeling community to satellite techniques and radiative transfer.  The group asked me to help them explain why different clouds seemed "cooler" in the simulated imagery and how features of clouds (such as ice particle concentration, height and optical depth) would affect what an actual satellite would "see".  I was able to alleviate some confusion amongst the participants regarding whether the standard IR window was seeing a layer or more of a surface when it came to different cloud features, such as an optically thick storm updraft core versus a relatively optically thin cirrus cloud associated with an anvil.  It is good to see some of the non-satellite community actively trying to better understand how satellite observations are made.

UAH CI Product Achieves 1.5 hour Lead Time in Tropical Florida Environment

One of the places in which the old version of the UAH SATCAST CI product used to suffer was in the tropical environment of the Gulf Coast with relatively low PODs and shorter lead-times.  Shown below is an example of an output CI forecast that was able to produce a lead time of 1.5 hours over radar (Figures 1 and 2).

Several improvements have been made to the algorithm since its debut as a "cloud object tracking CI product" 2 years ago (Spring 2010). With a newly defined future path as a "doubly data fused" algorithm (using short-term NWP environmental forecasts as input --and-- CI forecast output as NWP assimilated input), the best way to gauge performance is to make sure we are thoroughly bench-marking and improving its accuracy and usefulness as a Decision Support System (DSS).  Otherwise, using it as input into NWP models will prove much less effective. 

Figure 1.  SATCAST Strength of Signal CI product produces forecast output of 70+ index value (on scale from 0-100), valid from 1315 UTC (Left).  The storm motion is to the ENE.  Note the lack of any radar echoes at the time (Right).

Figure 2.  By 1445 UTC (1.5 hours after the above CI forecast in Figure 1), the first 35+ dBZ echo is finally detected by the radar (Right).