CI north of Cape Kennedy

Forecaster comments from EWP blog...

A storm cell over Volusia County, Florida apparently ran over a sea-breeze front and developed a lot of rotation.  A hook echo on the reflectivity and strong gate to gate shear was quite evident on Melbourne’s radar.  We had been monitoring MRMS data including reflectivities at -10C and -20C.  Looking back at Convective Initiation, there was a strong signal at 2154z.

Twenty minutes later the satellite picture and NLDN lightning data looked like this:

MRMS data at 2224z showed 57 dbz echoes at the -10C level.

By 2225z, velocity data from Melbourne’s radar compelled us to issue an experimental tornado warning.  Time from the first Convective Initiation signal to issuing the warning was 30 minutes.

After issuing the experimental warning, we got the 3DVAR domains set up properly and here is a screen capture of the 2km vorticity and winds for the storm of interest.
Here is a four panel of various 3DVAR data for the storm at 2230z.

At the top of the hour, the storm was still quite organized, and a second experimental tornado warning was issued.

Both CI products showing development over MLB

Forecaster comments on the EWP blog...

Here is a nice example of the two CI products and a healthy storm that formed in Florida.  We had just changed CWAs so it was more a matter of going back and seeing what they said rather than using in real time but they do show indications of a rapidly developing storm. See the red 95% blob (UAH CI product) on first image, the UW cloud top cooling on 2nd and storm getting big on third.

Simulated ABI imagery in the EFP

Today's focus area for convective initiation extended from eastern Iowa to central lower Michigan.  Although the forcing was adequate for deep convection, the low-level moisture was marginal with surface dew points only in the upper 40s so that the instability was weak.  Based on the ensemble model output (SSEO, AFWA, CAPS), the area of greatest likelihood for convection was located across northern Illinois, southeastern Wisconsin, and central Michigan.  The location of these thunderstorms if they were to occur would have a large impact on air travel through Chicago.  Most of the ensemble members predicted that convective initiation would occur between 21 and 23 UTC across this area.  The NSSL-WRF model, however, had the main convective area further to the west and was the most aggressive with the convective development across eastern Iowa, northern Illinois and southwestern Wisconsin.  The initiation and subsequent evolution of the simulated thunderstorms was clearly depicted by the CIMSS synthetic ABI imagery.  The simulated ABI visible imagery was also useful for depicting these thunderstorms.

When preparing the convection forecasts in the morning, the NSSL-WRF model solution was discounted given the tendency for the ensemble members to prefer convective development further to the east; however, as of 22 UTC, it appears as though the NSSL-WRF solution had a more accurate depiction of the convective development.  The images below show a comparison of the simulated ABI and real GOES infrared channels across the central U.S from 21 UTC.  Notice that the linear band of deeper convection across eastern Iowa and southwestern Wisconsin in the synthetic ABI imagery agrees well with the real GOES imagery.

- Jason Otkin

Monitoring & Forecasting DVN Convection

Forecaster comments from EWP blog...

Despite strong indications from the NSSL-WRF SimuSat imagery, convection in the DVN CWA has stayed rather benign.  The NSSL-WRF appears to have overestimated the low-level prefrontal moisture.

The GOES-R NearCast Theta-E Low vs. Mid-Level Theta-E image indicates very little destabilization ahead of the front (in fact, it actually indicates increased stability).  This is consistent with surface dewpoints in the 30s and 40s reported ahead of the front.

Quiet day

Forecaster comments from EWP blog...

Being LWX today for another rather quiet day. Nearcast products and simulated satellite (and HRRR and 3dVAR) both showed marginal instability and a few air mass sort of thunderstorms and that is what has been happening.  It has let us see a few flashes on the pGLM products but not really enough to evaluate usefulness in warning situation. Image below shows nearcast theta e product in lower right (showing not much) and simulated IR in upper right showing spotty but unimpressive convection.  Left two panels are real IR & Vis sat showing that the simulated stuff not bad.  The CI products seem to be hitting areas that do end up as the small thunderstorms but not sure they are showing more than can be diagnosed by looking at the satellite imagery so not sure if I would warn on this product any more often than the satellite and radar alone but good quick check to make sure I am not missing anywhere I should be watching.

NSSL-WRF SimuSat Prompts Move to DVN

Forecaster comments from EWP blog...

The simulated satellite imagery based off the NSSL WRF is handling a cold frontal passage in the Midwest quite well this afternoon.  The 2000 UTC comparison between the WRF “SimuSat” and observed IR image:

 The NSSL-WRF simulated satellite imagery suggests that storms will erupt over the next hour:

This model output, combined with hits from the UAH Convective Initiation product in the Cu field along the front, has prompted the RNK group to shift domains to DVN for the rest of the afternoon.

Weak Convection in the RNK CWA

Forecaster comments from EWP blog...

We are monitoring convection in the RNK (Blacksburg, VA) CWA this afternoon, but so far the development has been rather weak.

The 3D-VAR analysis has been useful for monitoring updraft intensity and anticipating possible stronger updrafts (top-right indicates instantaneous updrafts, bottom-right indicates 30-minute updraft history).  However, none of the strong updrafts have maintained themselves beyond one “scan”.

The convective initiation algorithms have struggled due to all of the cirrus blanketing the northwestern half of the RNK CWA, but some clearing in the southeast has produced some CI from the UAH algorithm.  Using the new Strength of Signal output, nothing has gotten above ~50 (the included sample is just 46 in the south-central portion of the CWA), and indeed there has been little significant development in those areas.  Nothing has triggered the UW Cloud Top Cooling algorithm yet in our area as of 1902 UTC.

EWP daily debrief 5/15

Today we had our first opportunity to discuss with the week 2 participants their feelings on some of the products from the previous day's forecasting activities.  They didn't have much chance to look at the products in any detail yesterday because they were getting used to the system.  But, there was one in depth discussion we had regarding experiment logistics that I thought was an interesting point to raise for future demonstrations.

One of the forecasters mentioned, "There was a bit of challenge with the number of products we had and the number of procedures... it took a long time to get through it all and then we moved to another CWA so we had to start all over again.  Even the procedures we were using opened up over TX so we had to reset everything... I ended up just loading everything in the volume browser which took a lot of time."

It may seem like a very small part of these demonstrations for the local coordinators to set up procedures for the forecasters for the experimental products when you also have to worry about data flow, display, IT issues, organizing visiting scientists, etc.  Making things as easy to display for the forecasters (or whatever users you have) is very important for an effective demonstration of experimental products.

Dusty CI

Forecaster comments from EWP blog...

Not much going on in EWX as accurately portrayed by nearcast (theta e difference product) showing max instability just across the border in Mexico.  Very nice looking convection over there and an outflow boundary over there kicking up quite a bit of dust.

The boundary/dust storm did trigger a tracked object on the UAH CI product.

CI and Cloud-top cooling over FL

Forecaster comments from EWP blog...

As we were getting started looking at weather in Florida, a good case for the Convective Initiation and Cloud Top Cooling algorithms took place over the southern county of MLB’s CWA.

 

Convective Initiation gave first indication of developing convection at 1925z.

Cloud Top Cooling gave a strong signal at 1940z and 1945z.

Max Estimated Hail Size (MESH) indicated Hail at 1.1 inch at 2026z.

A combination of the Convective Initiation and Cloud Top Cooling used together could have raised awareness of a possibly severe storm up to 45 minutes ahead of time.