Daily Summary: Week 4, Day 4 (June 4, 2015)

The final day of forecast activities for week 4 didn’t see much change geographically, with groups operating in the Boulder, Goodland, and Hastings CWA’s. The Goodland group would later move to Dodge City as activity to the south looked a little more promising. Once again, the Boulder group found the PGLM total lightning Flash Extent Density and Lightning Jump information to be quite valuable in monitoring fluctuations in storm/updraft intensity. One particularly strong storm near Simla, CO produced a tornado from which we were able to view live streaming footage (see image below). Activity in the other CWA’s, was a little slow to develop, giving participants ample time to evaluate the pre-convective products.

– Bill Line, SPC/HWT Satellite Liaison and Week 4 EWP Coordinator

Forecaster comments from weekly debrief:

GOES-R LAP

- Would be nice to see this generated with other nwp – rap, nam, euro

- Used in conjunction with other tools, reaffirm what we might be thinking

NUCAPS

- Number of gridpoints is nice, if clouds around, probably something nearby that might be semi representative

- Very useful out in west, don’t usually have previous nights cloud debris, data sparse, so useful out there

- Midday time frame is good, gives an idea, comparing to 12z, see how atmosphere is evolving

- I think  its useful for sites like my office, we don’t do upper air, not any close, except Charleston, but that’s more of a coastal environment, 

- Probably not gonna be as useful to me operationally in Florida, pretty good coverage of other obs networks. 

GOES-R CI

- Southeast didn’t like it, but found it really helpful yesterday, might depend on type of convection, and how widespread. When lots of tiny discrete cells, lots of colors, hard to see what s going on

- Filtered out 0-40, because I never saw anything from that

- Real-time time verification would be helpful. So I know how well its doing in the current situation, what the thresholds are

- I think it might do well out west with the type of convection we get in Reno

- Might be useful writing tafs,  giving you a heads up on convective development

ProbSevere

- I think it did a good job overall, did well as far as trends, categorization.

- Favorite product for week for me, in SC, we have a lot of pulse convention, so a lot going on, this is an easy way to show which ones to look at. I know other people in my office would use it

- From a  purely convective standpoint, using mucape, yesterday, made it seem a lot higher than the prob should be. Maybe have an elevated versus SB convection probs. Mesh never got above an inch, but probsevere 70s and 80s, didn’t drop much

- Since its mainly for initial severe, if a a certain amount of time has passed, would it make sense to drop the probabilities, but still track 

- Outline change to different color when its “mature”

- In pulse severe, nice to see trends in ind probs

- Incorporating lightning data into prob severe would be nice

- Someway to incorporate  mesocyclone  information would be nice

- DCAPE gradient good for winds

- Incorporate climatology, pwat climatology

SRSOR

- Must have it

- Big tool for radar sparse areas.

- Maybe have canned locations/ predefined regions. Then a WFO can say, I want region 3 today because..., for example.

- Perhaps have a webpage where the relevant person may select the center point for the sector, which will then show the sector space. They then submit this.

- Q: You can get 5min over the full disk. A: I don’t want 5min data, I want 1 min data

Lightning Jump

- I like having the different colors, as opposed to increasing brightness

PGLM

- 6min summation- I like the short time res better than 6 min.

- Different color curve for the summation is necessary (larger values)

- Yesterday case with storm approaching Denver, we had 16 min between when I first saw pglm and when I first got cg

- I think this will be great. 4th july concert a few years ago, saw storm, had to evacuate everyone. Made call when we saw cg lightning. Pglm would have increased lead time.

- Would use IC lightning data for DSS 

- Need training on lightning!

ProbSevere near Copeland, KS

Dime sized hail reported at ~0011z near Copeland, KS. While it’s subsevere, the ProbSevere product did jump from 12% at 0000z to 72% at 0006z.

Jason Williams

Vis sat loop of northern DDC CWA updrafts

Click for loop. You can see a couple of updrafts come shooting up. Reflectivity also increased, although ProbSevere has remained quite low (6%, if it even has a value show up at all). Would hesitate to go SVR given the very low ProbSevere numbers. I’m not seeing a TBSS, and ProbSevere MESH has been around 1/4″.

Jason Williams

Lightning Jump Miss

This is a case where I was surprised that there was not a lightning jump. The most that the lightning jump algorithm went up to was 1 sigma. However, we can see in the bottom right panel that the prob severe was gradually increasing (up to 78%). The PGLM data was showing what you could refer to as a ‘jump’ as flash density went from 14 to 29 to 38 to 53.

-Helen Hunt

7 sigma lightning jump precedes tornado development by 16 minutes

We have been tracking a persistent supercell near Simla, CO for the past couple hours.   The lower left panel shows a seven sigma lighting increase, followed by a three sigma increase preceded the eventual tornado development by around 15 to 17 minutes.  It is thought the vertical stretching of the updraft into in the mid levels as indicated by a significant increase in lightning eventually may have manifest in the lower levels by eventual development of a tornado.

John Pendergrast

Parallax on storm SW of DDC

Something for forecasters to keep in mind is that parallax is a factor. Below is the 1min vis and NEXRAD scan at 2306z.

Jason Williams

PGLM Utility

I am finding the PGLM lightning data to be very helpful in detecting lightning before there are any cloud to ground flashes. This data would be helpful in the Columbia, SC FA when storms are approaching outdoor events. It would allow us to give people a heads up sooner than CG data can. The animation below shows the PGLM Surface Flash Extent image overlaid with the 1-min NLDN lightning plot. Near Denver, CO, you can see how the PGLM data picks up on lightning much earlier than the NLDN first CG strike appears (16 minutes earlier!). 16 minutes is critical when it comes to trying to move a large number of people out of an event.

(Click image to see animation)

As far as other uses for the PGLM data, I’m not sure that it’s something that I would actually use after CG lightning appears and when I’m trying to decide whether or not to issue a severe thunderstorm warning. At that point, I think it would be useful to have the PGLM data incorporated into the ProbSevere data rather than looking at it on it’s own.

-Helen Hunt

CI and a storm’s anvil

50% CI at 2208z; however, that particular cloud element ended up fading away. There was one just to its west that ended up initiating; however, it was underneath the cirrus shield from storms to the west and thus wasn’t caught by the CI product.

Jason Williams & CoonieCatEye

Thoughts on 1-min vis imagery

Even though we haven’t had anything manage to sustain itself in the GLD CWA, there has been a lot of utility in the 1 minute visible imagery. Using imagery with coarser time resolution would almost certainly have us missing much of the (albeit brief) life cycle of those little pulses that tried to get an updraft going before getting sheared off.

Having such fine time resolution really helps to pick out which clouds might have the most promise of going up, and being able to see them begin to get sheared almost as it’s happening is valuable in assessing not just the individual cell’s situation, but the environmental conditions as a whole. Clearly, even though CAPE values were being shown as quite high, updrafts were not able to be sustained. Having real-time information on actual conditions was very helpful in having an idea of how the short-term would likely evolve, rather than simply relying on short-term model data and/or satellite imagery that only would come in every 15+ minutes.

Jason Williams

CI false alarm, yet another sheared-off cloud

Click for animation. CI was 75% on another cloud, which began to show signs of getting sheared off not too long into vertical growth. The 1 min imagery showed that it was still trying to have growth going below that level, but it never managed to get its act together.

Jason Williams & CoonieCatEye

Lightning Jump and ProbSvr Used in Warning Decision

I had been noticing a gradual increase in ProbSvr and at 2134Z, it reached 79%. At the exact same time, I got a 2 sigma lightning jump. This increased my confidence that the storm was intensifying and that a severe thunderstorm warning was needed. I went ahead and issued the warning based on ProbSvr and Lightning Jump alone. After the warning was out, the prob severe continued to increase to 94%. The WFO has not issued a warning on the storm. It will be interesting to see whether or not it verifies.

-Helen Hunt

Low CI verifies in GLD

First two images are from 2135z showing a little cellular blob (middle of the image) trying to go up. CI was fairly low at 38%, and given how things have gone thus far, we figured the chances of it going up were pretty small. The third image is from 2154z, showing that indeed that cloud had pretty much completely dissipated.

Jason Williams & CoonieCatEye

Supercell west of Limon, CO matures

Sustained total lightning as indicated by surface flash density and CIMSS prob severe model 99% indication on the cell west of Limon, CO. MESH is topping out around 1.25 inches.  Due to the sustained high lightning activity lightning jump sigma levels remain  around 1 or less.  I min visible imagery shows a developed inflow region marked by low level cu field streaming into the southeast flank of the cell.

John Pendergrast

Thunderstorm tried to get going…but sheared (GLD)

GOES-14 1-minute visible satellite data showed a towering cumulus that tried to grow across southern portions of Greenley county at 2050 UTC…

which started to build nicely over the next 10 minutes…

and grew further, showing what looked like a nice anvil top at 2110 UTC.

However, the low level southeast winds clashed with the stronger mid level southwest winds, which tore the top of the thunderstorm away from the low level convection just 10 minutes later.

A light shower was noted on KGLD radar at 2118 UTC near the southern Greenley county border that quickly dissipated.

CoonieCatEye/Jason Williams

Situational display for GLM lightning and cell growth

Upper Left: is the Boulder Regional Flash Exent, Upper Right: 1 Min Vis imagery along with CIMSS Prob Severe Model, Lower left: Lightning Jump, Lower Right:  0.5 deg Reflectivity.



John Pendergrast

False Hopes, CI 90% (Hastings, NE)

Just an FYI, probably saw the highest CI I’ve seen since I’ve been here (90% eastern Osborne CO).  Went on to just be run of the mill cu with no radar reflectivity seen. I continue to monitor CI but it has not instilled much confidence this week. –Cattywampus

Fig 1: 90% CI in eastern Osborne Co.

Far southern GLD development?

It hasn’t amounted to anything yet, but we’re watching a cloud that’s gone up in the far southern end of our CWA. The two images are from 2045z and 2100z. CI hasn’t flagged it at this time, but having the 1-min imagery really helped us see how it was growing – I picked these two images because they were 15 minutes apart and indicate how much that spot has grown in that time.

Jason Williams & CoonieCatEye

Hastings, NE (SVR #2): Losing ProbSevere guidance due to cell merger

Was monitoring a developing cell mainly with base/dual-pol data. Was on the edge of a warning decision and looked to ProbSevere for added confidence. During the 20:08z scan had a ProbSevere of 54% (fig 1) with no growth of glaciation rate. During the subsequent 20:10z scan the cell merged with the northern cell and quickly jumped to 84% (fig 2). Lost confidence in this prob due to the merger. Ended up warning on base data info along and just the trend of the increasing ProbSevere from the previous scans. –Cattywampus

Fig 1: 20:08z ProbSevere 54%

Fig 2: 20:10z ProbSevere 84%, Cell Merger

CI Success Story

This is an example where the CI product performed very well. At 1915Z, the CI probability increased from nothing to 61% (yellow) and 70% (yellow-orange) over two spots in the Boulder FA. In the next satellite scan (15 min later), you can see the significant cloud growth that occurred in the exact area where the 70% CI product had been. This storm continues to grow in the proceeding images and eventually becomes severe. The cell that showed 61% CI also intensifies significantly, but it takes a little longer (about an hr).

I was pretty pessimistic about the CI product after using it yesterday in the southeastern part of the country where there were widespread discrete and multicells. There were so many areas that the CI picked up on that it just become too busy to be of much use to me. However, my faith has been restored in the CI product after seeing how well it is performing out west today. I think that the utility of the product will depend on the type of the convection expected and the coverage of the expected convection.

(Click image to see animation)

-Helen Hunt

Ultimate 4-Panel Display

I call the image below the ‘Ultimate 4-Panel Display’. I probably would have incorporated the ENI data somewhere into my display, but I wasn’t allowed to use it. This 4-panel might be a good recommendation for forecasters.

Upper Left: PGLM Surface Flash Density w/ range rings

Upper Right: GOES 1-min Vis w/ CI

Lower Left: Lightning Jump

Lower Right: 0.5o Reflectivity w/ Prob Severe

-Helen Hunt

Clouds inhibiting convective initiation…for now (GLD)

GOES-14 1-minute satellite imagery showing bands of cumulus clouds covering most of the CWA early this afternoon. This appears to be inhibiting convection, at least for now.  However, some breaks in those clouds may be moving into eastern areas from KMCK to KHLC.  These areas will need to be monitored for possible towering cumulus development over the next 1-2 hours.

The 1-minute satellite loop shows this nicely.

(Click to animate above image.)

CoonieCatEye/Jason Williams

Comparing PGLM Color Tables

When comparing the default, LMA1 and LMA2 color tables for viewing the PGLM lightning data, I found that I don’t really like any of them. All three images below show the PGLM data at 1942Z using the various color table options. The highest pixels (58 & 53) in the northern storm show up as the the gray/green colored pixels which are in the center of the default color curve. I would have assumed that the bolder blue hues would have been the higher values in this case, which is why I don’t care for the default color table. For the same reason, I don’t like LMA1 or LMA2 either. Improving the color table will be important moving forward with this product.

Default

LMA1

LMA2

-Helen Hunt

First SVR for GID

Probsevere jumped to near 70% with this storm.  Large core of 50+ dBz at 30kft.  The -30 is around 24kft based on the LBF sounding.  There was also a jump in lightning according to the ENI timeseries and data.  Decided to issue a warning for this storm.

MacGyver

Seeing between sfc obs with LAP CAPE

We don’t have many obs in our CWA, but we do see that temps are around 90F just outside our CWA. Looking at LAP CAPE, we do see a CAPE gradient not just to our SE but also within our SE CWA. This does make one wonder if warmer temps have indeed made it into our CWA that we aren’t able to see with the lack of obs, but can infer from the higher CAPE values.

Jason Williams

Severe indicated cell produces first HWT warning near Cheyenne

A prodigious lightning producer early developing north of Cheyenne has a CIMSS indicated prob severe nearly 90 percent.  The cell is exhibiting strong inflow along with sustained high lightning levels as shown by the GLM Mapper, PGCO Surface Flash Density regional mapper and MRMS Verically Integrated Ice (click second image for loop).

John Pendergrast

Boulder FA Mesoscale Discussion

CIMSS products were helpful in determining which portions of the FA are most favorable for convection. The CIMSS PWAT, CAPE, LI, and SI values are all most favorable in the northeastern quadrant. CAPE is up to 2200 J/kg, LI values up to -7 and PWAT up to 1”. My initial analysis of only the CIMSS products is consistent with the SPC Day 1 Categorical Outlook which shows a slight risk over the eastern part of the FA and only a marginal and general area further south and west. 0-6 km bulk shear is also most favorable in the east/northeast quadrant with values up to 40 knots. This shear is sufficient to promote organized convection including supercells. The main threats today will be wind, hail and even tornadoes.

-Helen Hunt

GLD AFD

Surface low moving into SW KS; progged to move over CWA through the afternoon and evening. Dew points currently in the 65-70F range in most of the CWA with temps of 75-80F. Cu field ongoing in NW CWA; sfc boundary bisecting from SW to NE both at surface and aloft.

Southwest flow at 500mb with weak shortwave troughs embedded in the flow. 12z DDC sounding has a large amount of CAPE above about 675mb (2700 J/kg) but has a large cap below. Convective temp is 95F.

HRRR shows 850mb winds rapidly increase this evening to 30kt by 00z in the eastern CWA; also indicates this is where convection would take place. NAM also shows this.

Based on surface obs and satellite trends, would expect the eastern CWA to indeed have the highest chance of convection.

Jason Williams & CoonieCatEye

Gravity Waves on SSRO

These gravity waves should lead to an interesting afternoon across the GID CWA.  The one min ssro data is EXTREMELY useful and helpful with trying to figure out what is going on.  I think these waves may lead to some CI or storm intensification later today.

MacGyver

Daily Summary: Week 4, Day 3 (June 3, 2015)

Today’s activities took us to Cheyenne, Boulder, and Jacksonville, with the Cheyenne group moving to Topeka before ending in Dodge City. The PGLM total lightning Flash Extent Density within the Colorado LMA in NE Colorado saw heavy use with storms that developed in the Boulder CWA. It was nice to have a group operate in the Jacksonville today as it allows us to demonstrate the tools in a different region of the US. This is important considering most of the experiment has been spent west of the Mississippi River thus far. This group, however, did not have access to the 1-min imagery, something they missed dearly! The Boulder group found it helpful to monitor the LAP PWAT products as  moisture quickly approached the front range from the east, tightening the moisture gradient leading up to convective initiation. The approaching feature was also apparent in the 1-min imagery. The NUCAPS, CI, ProbSevere and ENI Lightning products all saw use throughout the day as well.

– Bill Line, SPC/HWT Satellite Liaison and Week 4 EWP Coordinator

Forecast comments from daily debrief:

GOES-R LAP
- Higher pwats bulging into foothills was helpful, you could see moisture coming in
- Cape vaues, as clouds thinned, could see cu increasing there
- Comparing to everything else I have, compare to rap meso, etc. get a good quick check. Meso seems to be following it
- Broad environmental analysis
- Would be cool to use a first guess form other models too
- In southeast was useful yesterday, northern part of cwa had higher moisture, cape, etc.
- Later in the day it was confusing
- Initial convection that died went up before gradient tightened. Later stuff was when gradient was tightened

NUCAPS
- Waiting for cap to break. Had 18z special sounding, looked at nucaps, you could see that it wasn’t far from breaking. Could see a sign of a capping in uncaps. Location in nucaps of feature was close to that on the sounding.

- In Cheyenne area, showed high cape,

GOES-R CI
- In southeast, it showed development pretty well
- Ci not my favorite product, yesterday, there was so many areas, with many different colors, a lot going on, I couldn’t focus on one area over another.
- There was an area that was 94, couldn’t tell if that was the one that developed.
- Ci was great for 5-15 , but when you have 1-min imagery, you can really see stuff happening.
- Might be nice to see values like probsevere.

- Set below 40 to black, didn’t single with 3 main cells. Cloud cover that was there might have squashed it. Would use it with main convection, after that, losing utility

ProbSevere
- It showed that the storms we had weren’t going to do much. Low end stuff. Anticipating when is core going to go up. But probs stayed low. So it nailed that.
- I liked using it with the eni time-series, because both of them increase my confidence whether I will issue a warning., both will show me fav signals, my confidence would be higher
- Case yesterday, all products were going up simultaneously, gives you confidence in the storm.
- Not too much information with lightning
- Make a easier delineation at 50%
- Whatever color-tables are decided on, you would get used to , or change it to preference
- During our busy time with 3 cells, it was good for SA, with other little cells developing around big cells.

- Southeast, over 70 was the threshold for the day

SRSOR
- Was neat to see increase in eni lightning, then overshooting top thereafter
- Had puffy cu field that was generally going west across cwa, when we got big storm on southern part, you could see cu field being injected into it. Upper clouds going east. Seeing that, gives you knowledge that it will be a long lived cell
- Seeing updrafts reorganize themselves within storms. Repoisitioning itself into more favoriable part of storm, gives you confidence that that storm may thrive. In near real time that is big
- In pre convective environment, looking at ll cu field, denoted were higher levels of moisture was present, you could basically see the moisture setting up. No more guessing where it is.
- No 1min data in the southeast today, but had a case where it would have been helpful. Convection developed on outflow boundary during FD (30 min) scan. This development was missed by routine imagery.
- In summer time, if there was 1min ir, could use it to identify thermal belts.
- 1min winds with fires, doing dss, if you have a plume and you think it is going to change based on winds in sat iamgery, that gives you lead time to make the call and say the plume will be changing direction... from imet perspective. Would want to warn of impending wind shift. Any heads up would be great.
- Overall we just need it all the time, across the country.

- If I had it, I would keep it up all the time, and keep ot detection up all the time, because its rare, it wouldn’t impeded your obs.

Lightning jump

- Easy to read, makes jumps obvious

- Saw sigs on base data sometimes before jump, during or after

- When storm was so pulsey, id look at this and pglm

- Storms were pulsey at times, and jumps in this and increase in pglm brought me back to the storm, couldn’t see it in other data.

PGLM

- We had 3 main cells we were monitoring, went through cycles of pulsing and dying. Looked liked it died on radar, then pglm came back, kept me aware of storm evolution, good SA tool, kept an eye on it

- Interesting to see extent of how far lightning got away from core of storm into anvil, anvil crawlers. Benefit of having extent

- I liked default colorbar, had familiar cold to hot.

- Green yellow red purple might be better. 

- Idea of defining colors with respect to flash rate. Blues 30, purples 40.

Large increase in pGLM lightning and strong supercell growth.

Saw a rapid increase in pGLM Flashes(106/1min) and then witnessed a very strong increase in reflectivity and supercell with a hook appendage. This definitely got my attention and ended issuing a SVR for it. –Cattywampusg

Fig 1: (Gif loop) Rapid pGLM increase in NE Douglas.

Fig 2:  (Gif loop) Subsequent supercell development. Loop to long (sorry) its the last warning in the series.

ENI Time Series and Prob Severe Influence Warning Decision

I noticed a very dramatic increase in lightning activity and prob severe as a storm approached the Jacksonville FA (from Lanier County), so I decided to go ahead and pull the trigger on a severe thunderstorm warning. The prob severe didn’t quite make it to the 80% threshold (only 78%), but since it accompanied the ‘lightning jump’, my confidence was higher that it was intensifying and was capable of producing severe weather. There was also an overshooting top associated with this storm as seen by the overshooting top detection.

The warning has not verified yet.

(Click image to see animation)

-Helen Hunt

Vigorous updraft well sampled on 1 min imagery

The persistent cell over eastern Kansas shows good evidence of strong updrafts on the shadowed 1 minute imagery.  Deep convection with good evidence of rotation on radar imagery likely is producing severe hail.  CIMSS algorithm currently shows 99 percent likelihood of severe with hail size hear 2 inches.

John Pendergrast

Huge increase on PGLM prior to DRC?

Might have had a DRC from 2346z to 2354z on our middle cell. PGLM values of around 30-40 were present in the storm’s core at 2345z, jumping to 57 two minutes later, 70 at 2350z, and 99 a minute later. Possible hook echo showed up at 2359z and persisted through this post at 0007z.

Jason Williams

Value in GLM data

The storm south of DIA has a very obvious core, but of note is how far away lightning is being observed from the parent updraft. NLDN shows a CG in the upper-right part of the lightning field, which is a very long distance from the main storm. In fact, there isn’t even any rain falling at that distance.

Jason Williams

pGLM and Lightning Jump continued benefit for pulsing cell

Have been monitoring a long-lived supercell which appeared to finally look like it was falling apart. Had a subsequent increase in lightning on pGLM 2325z (fig 1) over west Elbert Co. coupled with a 2 sigma jump .  Reflectivity then showed an increase with 60+dbz about 10 minutes later (fig 2 – gif loop 2325-2345z). The lightning was the only piece of information which showed the potential for this to occur. Indispensable in this case.  –Cattywampus

Fig 1: Increase in lightning near cell core over western Elbert Co. (2325z)

Fig 2: Gif loop of increase in reflectivity. (2325-2345z)

Multiple Overshooting Top Examples

This loop shows a few examples of the overshooting top detection algorithm in action. Even though the overshooting tops were already pretty obvious in visible satellite, the detection algorithm really helped grab my attention.

(Click image to see animation)

-Helen Hunt