EWP Week 1 Day 1 Summary

During the first day of the 2017 GOES-R/JPSS Spring Experiment, our two pairs of forecasters operated in the Boston and Raleigh CWA's.

-Bill Line, NWS

• GOES-16 ABI Imagery and products
• Imagery
• Seeing cooling tops in the IR combined with structure of cu development in the VIS (sandwich product) with 1-min resolution keyed my eyes to development along a line before anything else.
• Used imagery mostly for identifying cloud top warming of storms. We did not warn on storms that had warming tops. The storms we did warn on had OTs
• We identified weak/weakening storms from warming tops, or uniform on vis
• Initial convection could be seen going up much sooner in satellite than on radar
• Nice to have 1-min 2-panel IR and VIS on the side next to my radar/warning display. It provides a  very clean, situational awareness display
• Would be nice to time synch 2 panel with single panel on the side
• Derived products (winds, dsi, tpw)
• Winds did not look well. Color coding by wind speed was not helpful.  Color coding by height or pressure level would be better. Barb tells you the wind speed already, no color coding needed. Also many of the winds were completely wrong in speed and direction. Compared with obs (surface winds). Few vectors, very limited.
• Dsi and tpw - There were many holes in the baseline TPW and CAPE, making it difficult to interpret. If too many holes, I will look at other tools.
• Using GFS as a first guess is not ideal. It would be better to have a separate CONUS sector using RAP or HRRR as the first guess, something that updates with new sfc obs every hour.
• Channel differences
• RGB composites
• GLM
• This is the first time we’ve seen GLM data. The lightning placement is a bit off geographically compared to radar. Flashes seem most intuitive, similar to what we are used to seeing from NLDN. I want to know when I would use GLM instead of NLDN? As for gridded events, not sure how I’d use it for warning ops. Color tables worked fine for me.
• I liked the gridded color scale you put together. I like just a simple scale. Brighter colors focus my eyes.
• I didn't notice anything that was too off. With pulse convection, you could see convection pop up quick, and then flatten out quick. Confidence that stuff wasn't going to keep strengthening early on because not really any rapid lightning increases
• ProbSevere
• In our case it wasn't working  too good today.  For a storm that had high velocity in radar, probs were too low because not much azimuthal shear. Spurious data in azimuthal shear product can also cause bad ProbSevere probs.
• Pulse nature of what we were seeing was highlighted in probsevere. Seeing probsevere probs pulsing verified the environment we were in. So it was useful in this sense.
• Used it more for confidence building and not as much warning decision.
• NUCAPS
• Experimental modified nucaps came in very late. But comparisons showed that modification did improve thermodynamic values.
• Operational pass was a little late to be helpful. 1.5 hr to 2 hour after timestamp.
• Experimental data with modified sounding was better than the one that wasn’t. CAPE was closer to that on mesoanlaysis and helped to verify environment we were in.

KRAH-SVR 2300z

Issued an SVR at 2300z as shown below due to sustained Prob Wind hovering over 50% for several scans, and cooling cloud tops on IR and indications of growing updrafts on visible imagery

Also continued  to see an increase in lighting on GLM data

My thoughts were the significant increase in lightning, plus a sustained probability over 50% on Prob Severe for Wind and cooling cloud tops led me to issue an SVR with the main threat damaging winds. 

An LSR was issued by KRAH at 2215 for damaging winds in Richmond county, before we issued a warning.

-Ironman

NUCAPS Usage for storms moving into MA

HWT: June 19, 2017

My partner and I were monitoring a line of storms moving into KBOX CWA.  We got lucky.  NUCAPS soundings became available ahead of the storms.  Attached are 3 images of a sounding near the MA/RI state line:

1) NUCAPS2MARI  (original sounding near the MA/RI border)

2) NUCAPS2MARI EXPERIMENTAL  (already modified)

3) NUCAPS2MARIMODIFIED  (We modified)

1) The un-modified sounding NUCAPS2MARI only had SB CAPE 125, MU CAPE 125.   However, it had a dew point of around 18 deg C.  METARS in the area were showing dew points near 22 deg. C (pretty high for MA!).

2) The Modified Experimental sounding used about 22 deg. C. and gave SB CAPE 2397, MU CAPE 2397.  Quite a change.

3) When we did our own modification  we got SB CAPE 1914, MU CAPE 1914.

We did end up issuing SVR's for most of the area for 60 mph winds, penny sized hail.

Conclusion:  It did help to modify the sounding to the surface METARS in the area.

One point of note:  The temperature from the un-modified sounding was pretty close in this case, just the dew point was low.

Davis Nolan
WKRN News 2 Meteorologist



 

NUCAPS Soundings over KRAH area

The NUCAPS sounding data was useful to assessing the environment in the KRAH area. The data did arrive a little later and it was not available to use at beginning of warning operation. However, I did look at a sounding within the region where storms were developing. The SNPP NUCAPS sounding indicated CAPE values 1500-2000 J/KG.

The experimental SNPP NUCAPS sounding (below right) had much higher CAPE, around 4,500 J/KG at 18z. 

 SPC mesoanalysis at the same time had CAPE around 3,500 J/KG which is closer to the experimental SNPP NUCAPS at 18z.

While the experimental NUCAPS may be a bit overdone, it did lead to a potential reason why the storms in RAH area did not sustain themselves. Shear values were much lower, meaning storms likely intensified and then quickly rained themselves out as shear was too weak to sustain the updrafts.

-Ironman

Prob Severe and More Confidence in Storm Strength

Much of the storm environment across the Raleigh CWA was pulse.  I noticed that the Prob Severe was also "pulsing" up and down throughout the loop below. Updrafts would intensify and then quickly weaken and Prob Severe was also indicating this trend. This gave confidence that any storm would quickly strengthen before it quickly weakened and warnings were not necessary. When Prob Severe would show an increase in probability, mainly for wind, the next update would show a quick decrease.

The "pulsing" of the Prob Severe gave me confidence a warning is not necessary and that the storms would quickly weaken. 

-Ironman

ProbWind Under Forecasting Probabilities

Here is an example where it appears ProbWind under forecasts the probability of severe wind:

In the example above, there is widespread >=50 knot velocities associated with a line of thunderstorms; however, ProbWind only had a 1% chance of severe winds. This was likely because of relatively low azimuthal shear due to weak azimuthal gradient.

It would be beneficial to include some variation of maximum radial wind (such as  to ProbWind to account for the cases where the azimuthal shear is relatively weak.

-Lost Met 

AWIPS SA Procedure

Tested a recommended 3panel SA layout (below) with vis, IR, and radar data. Overall, simple and easy procedure use. If it were possible, it would be nice to be able to time lock all three panels together. In this case, just the left 2 panels were synced while the radar was separate. It would be nice to evaluated all 3 simultaneously in time and also with a cursor readout to establish trends in reflectivity and also IR temps. Think this is just a limitation with AWIPS2. --Tahoe

GLM Color Scale

I utilized the 5 minute GLM Event (1 minute update) with the 5 minute flash (1 minute update to monitor trends in lightning along with other GOES-16 data VIS/IR and radar data.

Here is an animation of the GLM data over the KRAX domain.

I do prefer the color scale above over the more typical gridded color scale. However, I think it may be useful to have the grey colors more white so the higher values stand out a little better. This could be an issue if being overlayed with Visible data. Increasing GLM events and flash data coincided very well with when cells were strengthening.  When GLM events and flash data decreased, cells were weakening. The faster updating GLM data also helped show trends in the storms quicker than just viewing radar data alone.

-Ironman

ProbWind potential enhancement?

One great part of the EWP at the HWT is the ability to discuss ideas with forecasters in response to something they are viewing in realtime.  One forecaster covering the Boston WFO and I were discussing the ProbWind values with the squall line in western MA around 2030 UTC.  The forecaster noted the ProbSevere (ProbWind) values diminished (from 60-70% to 20-30%)--largely in part due to decreased MRMS AzShear values.  The screenshot below shows the single radar 0.5 degree base velocity imagery.  The imagery shows two important features 1) the lack of a gradient in the velocity leading to somewhat low AzShear values and 2) at approx 3300 feet above ground level winds near 55 knots.  The forecaster looked at base velocity values from earlier in the day when severe reports occurred and noted the current velocity was more intense--which lead to him issuing an experimental severe thunderstorm warning.

Figure 1.  KENX 0.5 degree base velocity imagery valid at 2034 UTC 19 June 2017.

While such a product does not exist, we discussed the potential utility of an MRMS Maximum Velocity Magnitude product (for a series of levels or layers like 500m, 1000m, 1500m, 2000m, etc. AGL) and how such a product could potentially further improve the ProbSevere ProbWind predictions.  These types of interactions are extremely valuable and perhaps such a MRMS product is worth considering.

-Sieglaff

Use of GOES-16 Vis/IR 2 Panel and Radar Situational Awareness Screen

One way I was able to monitor trends in convection was using a 2 panel display of 1-minute visible and 1-minue IR next two a radar display which included Prob Severe. Warming cloud tops as well as a more smooth look on the visible imagery gave me confidence the storms of concern were weakening. This coincided with Prob Severe values decreasing fairly rapidly.

Other cells are developing on an outflow boundary well to the NW of the first band of convection, and this display has been useful in monitoring the trends over the last 30 minutes.

-Ironman