April 26 provided us with our busiest weather day of the experiment thus far, as widespread severe convection developed throughout the southern plains. One of our groups operated in the Norman CWA, where they were busy issuing warnings all day. Our other group began in Wichita, which started out busy but quieted down towards the end of the day as storms advanced eastward. They would move to Omaha briefly, before ending the day in San Angelo.
LAP
- An area of convection in Texas developed west of the CAPE gradient in the stable air, due to errors in the GFS (which is used as a first guess for LAP).
- Looking at the CAPE field, the locations where GFS was used were underdone. The retrievals (clear and cloudy sky) where available bumped up the CAPE to more reasonable levels.
- Low level PW depicted the dry line really well. The location compared well with surface observations.
- These fields could be especially helpful in data sparse regions.
- We would like theta-e derived from LAP.
CI
- Severe CI had false alarms and misses yesterday
- With one of the false alarms ahead of the line of convection, you could see initial development, then it just dissipated quickly. Severe CI went up to 60.
- Regular CI generally seemed to hit on locations of future convection.
- In Wichita, we had relatively high severe CI values where severe convection did develop later.
- In Midland, there was a very high severe CI near 90 that developed into strong elevated convection.
- What I take from this product is a general situational awareness to target areas.
ProbSevere
- A handful of times, it didn't give much lead-time over when I would have normally warned. Radar signatures were there, but probsevere was below 50.
- When probsevere increased at the same time as radar, it was a confidence booster in out warnings.
- There were some instances when radar was on the edge for warning criteria, but ProbSevere jumped, to we went with a warning.
- It is tough to calibrate to the cwa and the weather situation at hand quickly.
- I saw 3 different occasions that once it got above 60, the storm intensified quickly.
- For reissuing, it helped with confidence, in addition to the radar.
- I felt like I was still ahead of ProbSevere for initial warnings when using basic radar methods.
- One situation with severe hail, radar signatures came down a bit to the point where I wouldn't reissue typically. But probsevere stayed up, so I kept warning, and later on the storm did amp back up and was clearly severe. In this case, I did reissue based on what probsevere was showing in a situation where I probably would not have otherwise.
- Yesterday storms were growing so fast that it didn't provide that much elad-time.
- It would be nice if it could focus more on individual storm cores within a line. Perhaps it could track higher up, like x dbz above -20Z
- In the far south, a pretty significant line of storms developed. The bulk of the line was severe, but the part that was in the southern part of our CWA was not severe. However, probsevere still had it at 90 since it was highlighting the whole line.
- I would like to see the degree of LJ in the probsevere readout.
1-min SRSOR imagery
- It was great having the ability to see all of the overshooting tops.
- It was great to see the progression of the dry lines in this manner.
- I could see the anvil forming and the feeder clouds before the anvil got too big. I could also infer shear from these features.
- It was easy to identify and track boundaries and gravity waves.
- It is very nice to have this for situational awareness. Much better than having to wait 15 minutes for an image.
- At the top of the storm, you could see rotational patterns in cells.
- I could clearly see the outflow boundary, which I then compared the SPC mesoanalysis with. This combination would give me high confidence to which counties should be included in a convective watch.
- For IR, I compared the tops with a sounding. From this, I could prove whether a feature was at or above the tropopause and an OT.
- I would like to see some algorithms take advantage of the 1-min data, such as CI
- With the firehose of data and busy warning operations, we can't sit and watch the imagery constantly, so it is nice to have products that highlight significant details in the imagery such as cooling, OT. This will be helpful to have from 1-min data with GOES-R.
Lightning Jump algorithm
- High lightning jumps often correlated with increases in probsevere
- LJ was occasionally useful, but wasn't consistent from storm to storm.
- The main cells stood out, but these were already obvious to me.
- We would like to correlate LJ with intense rainfall rates. There is usually a ton of lightning with heavy rainfall in Texas.
- I don't feel calibrated to it, so I don't know what the threshold should be to warn. 7+ were obvious, but 2-4 range I was unsure if it was significant.
- I would like a mouseover that shows you the details. This would also help one learn what exactly is a LJ.
- LJ's do help to make the significant storms stand out more than others, attracting your attention.
- I would prefer an outline for this product, but then overlaying with probsevere would get messy.
- Yesterday I had probsevere, lj, and warnings on one display. This seemed like a good SA display.
PGLM
- We would like to correlate lightning rate with intense rainfall rates. There is usually a lot of lightning activity with heavy rainfall in Texas.
- I want to make sure we can utilize it on top of other products; clean to look at.
- Contours would be nice as opposed to the blocks.
- I think ultimate use will be to get a quick glance for situational awareness.
NUCAPS
- There was a moist layer around 400 mb that was apparent in the 12z and 18z RAOBS that was captured by NUCAPS.
- I sampled on both sides of the dryline, and saw the expected difference in NUCAPS profiles.
No comments:
Post a Comment