Convection firing in the Ohio River valley

A supercell formed early on in Indiana on May 21, 2014. This storm had strong satellite growth rates (derived from GOES cloud products) by 16:37 UTC and a probability of severe of 38% (MESH was only 0.3 in.) As the MESH increased, ProbSevere values increased into the 70-80% range. With some weakening of the storm, the probability of severe decreased to 45%, but then increased again to 70-90% when it was warned on by the NWS at 17:32 UTC. From when ProbSevere first crossed 60%, there was 48 minutes of lead time. This storm would later produce 1.75" hail. It should be noted that forecasters shouldn't necessarily issue a warning when ProbSevere crosses some probability threshold, but rather monitor the ProbSevere and radar trends, using the satellite growth rate information to hopefully increase confidence and lead time to severe hazards (since stronger satellite growth rates tend to be exhibited by more severe storms).

-Cintineo/Sieglaff

Thought Expirament

While in Glasgow, our office about 4 to 5 years ago experienced a point where the lightning had taken out the main radar during severe weather…. about 5 minutes later another strike made all land line phones go dead.

What to do? Well at that time everything we had the backup office had. So, it was up to us.

At that time Satellite, lightning, and the one office emergency cell phone went into action.

It just so happens that a large portion of the products now being tested including nearcast, LAPs, rapid scan satellite with cradle to grave diagnostic algorithms for it, and total lightning and each have nothing to do with radar. I would like to see what happens when the radar goes down and how much difference there will be between me and the neighbouring forecaster for warning boxes. Finding things that work the best and seeing how much a box can be physically shrunk down as compared to an operator with radar is key in this scenario.

Grant H.

May 21 Simulated Satellite Vs. IR Sat

When comparing the NSSL-WRF simulated satellite to the IR satellite the timing of convection as you can see is right on.  However, the location of convection is farther west in central Indiana so it does make me hesitate for where convection will actually move too. ~Vollmar

3..2..1..INITIATION – EWP Operations Update – Wednesday 5/21 – 2:10pm

We were greeted post-EFP briefing with an explosion of development across across the Ohio Valley and out west in Colorado. Forecasters were quickly allocated to work out of the Indianapolis, IN (IND) to diagnose the rapidly developing cumulus field from Montgomery Country westward.

0.5 deg reflectivity from KIND from 1813z – 1929z on May 21, 2014

Our second team is working out of the Boulder/Denver, CO (BOU) CWA diagnosing the rapidly developing convection from N/S near the Front Range Mountains.


Super-Rapid Scan visible imagery from 1917z-1929z

With a Severe Thunderstorm Watch issued for the Ohio Valley and a Tornado Watch issued in CO/WY region until 9pm, it looks like a busy afternoon ahead.


WW #164 from the Storm Prediction Center.


WW #165 from the Storm Prediction Center.


-Darrel Kingfield
EWP Week 3 Coordinator

Daily Summary: Week 3, Day 2 (May 20th, 2014)

What some people (including me) called a marginal day last night were pleasantly surprised with the amount and extent of convection observed in all CWAs utilized today. For a majority of the day we operated in the Cheyenne, WY and Quad Cities, IA/IL CWAs (CYS and DVN). As a supercell fired east of Denver, CO (BOU) forecasters leveraged this opportunity to look at the lightning jump detection algorithm and track total lightning on the tornado warned storm. In the late afternoon as the storms in the CYS area weakened, we shifted our product evaluation to Chicago, IL (LOT) for a DVN/LOT side-by-side warning operations session.

By splitting our operations into West vs. East, we were able to take advantage of (and blog about) almost all products brought into the experiment…with the exception of OUNWRF (sorry Gabe :( ). The tracking meteogram was put to the test again today with  forecasters wanting to track multiple product types simultaneously (radar + MR/MS, etc.) with mixed results. All forecasters issued their fair share of warnings (especially Danielle!) this afternoon. The training storms in the DVN region provided quite the challenge and convection is firing like crazy in LOT right now. I think it will be hard to pull the forecasters away from their computers tonight.

Feedback (BL)

Simulated Satellite Imagery:
- It matched pretty well in the morning, and continued throyughout the forecast
- I really liked it (WV) it captured SW and speed max in KANSAS. This gave me confidence that later in the day the speed max would be good.
- Boudnaries were setting up correctly, off by about a county, gave me a few that this was an area to look at

NearCast

- Wonderful. For all three systems. It will be helpful for inititiation

- In boulder, theta-e stream feeding storm

- In Chicago, convection developed along the northern gradient. As soon as cu hit the boundary, storms went up

- Looked at pw and theat-e products. I had to do some interpolating where data was unavailable which was fine. 

- I liked around 18z, there was a nice cu field in Kansas to Wyoming, that is where highest instability was, and where deepest moisture was.

- Color bar took sonme getting getting used to, but I was able to easily calibrate. 

- All really liked the wide spectrum of colors so it would easily show gradients. 

GOES-R CI

- Mixed results in Chicago area. Highlighted boundary where we had towering cu. Wide variety of values, from 10 to 80, but whole boundary went.

- Mixed feelings. Initially you can see thin line of cu firing across northern iowa. I don’t think the ci product captured that development. Highest value was maybe 40%> there was a 70 further south, nothing developed

Prob Severe

- I liked prob severe, it was pretty on point with targeting big storms that formed, looking at trends was helpful. They started low then amped up so I warned.
- Initially a little hesitant about the product. But by end of shift was really confident in it because it was doing really well.
- It drew the eye to the storm (good color scale).
- In chicago area,, to the south, storms were growing and splitting. Typically you pay attention to right split. Both storms actually had high values. You wouldn’t normally think this one would have as high a prob, this product made you want to look at it. The left split did have a severe hail report
- It didn’t work as well out in Wyoming. Storms were marginally severe (occasional quarter) probs were always low (below 50). On front range
- Im colorblind, so I changed colors. The gradients weren’t good for me. I changed it to – anything below 50 was black, then blocks of color every 10 to 100.
- For convective maintenance it was helpful. When the probs went down, the hail sizes reports went down.

OTD
- Chicago storm that had big hail had an OT. Several others had OTs but did not meet criteria of algorithm, the one that met criteria was longest lasting with biggest hail. Consistant detection. Certainly highlighted biggest storm

Tracking Tool
- I had mixed results with this. i tracked base velocity. The meteograms started to multiply so I stopped
- With boulder storm, I looked at composite reflectivity, all the panes locked up.
- I had major issues. It was great initially. But trying to load with dual pole (many panes) too m,any graphs. Switched to reflectibity and velocity, tab went blank. Reloaded, time with velocity did not match reflectibity. Velocity would update, reflecitivity did not. Red banner kept po[ping up, loading error. But I think it whas potential

-Darrel Kingfield
EWP Week 3 Coordinator

A NearCast Update for DVN

After a long afternoon/evening of analyzing convection over the northern half of DVN’s CWA, I took a few minutes to take a look back at the NearCast theta-e and PW difference products. Starting with the PW difference product, it was good to see the increased moisture advecting into the region and it lined up exactly with where storms initiated and persisted throughout the day. It did take some interpolating since there were some blacked out spots, similar to the theta-e difference product, but both are still helpful tools to see local storm environment as well as what’s coming upstream. The same trend could be seen in the theta-e product and combining these two, it raises my confidence level in being able to use these products when performing a mesoscale analysis.

~Linda

Changing Colors for Prob Severe

I found it nice to change the default color scheme on the prob severe procedure. The default colors sceme doens’t provide high contract between different levels of probability of severe. I changed the color scheme to highlight higher values of prob severe.

-JB

NearCast, Prob of Severe, and MESH Verify

I issued a severe t’storm warning for Linn County thanks to the Prob of Severe/MRMS data (bottom left) product and also looking at the NearCast (top right) which showed me that the environment was unstable to keep the storm severe.  Shortly after, I looked at the MESH (top left) and it shows hail around an 1″ or so and then a spotter reported 1″ hail in Alburnett which is right near the 1″ area.  All and all these products were good and my warning was verified. ~ Vollmar

My First Go at the Tracking Meteogram Tool

Well… unfortunately, my first crack at this tool has not gone well. When I initially pulled up the tool, it looked fine but then at some point, I had closed the Tracking Meteogram tab and could not figure out a way to pull it back up without reloading it all over again. I consulted with folks and we could not figure out how to bring a closed tab back up. It is unfortunate because I had already created a track and my points so I had to redo it.

Another problem I ran into was when the tab was opened, the meteogram would sporadically go blank/missing. Additionally, I received constant red banners that said An internal error occurred during: “Loading Tracking Meteogram.,” whether the meteogram was visible or missing. Lastly, I was testing the meteogram with KDVN 0.5 Z/V products and the Z meteogram did not line up temporally with V. Every time a new radar scan would come in, the V meteogram would update with the latest scan but the Z meteogram increased in the amount of points on the x-axis without accurately or correctly updating the time (x-axis)/max reflectivity value (y-axis).

I do like the ability to move individual points (instead of the entire line like with similar tracking tools in AWIPS-I) and it was interesting to be able to increase/decrease the circle size to capture certain storm features. I also like the capability that it has to offer to be able to see trends laid out in graph form for a particular product or set of products.

~Linda

Three Meteogram Boxes One Product

I loaded the the base velocity product and then the meteogram program and it seemed to work well initially, but as the radar product updated the meteorgram split into two boxes, one for 8 bit and one that just indicated velocity. It later added a 3rd box that included 4-bit after the screen updated again.

-JB