EWP Blog

Our morning real time case has taken us to Iowa with an ongoing broken line of thunderstorms. While there have been a few signs of mesocyclonic development, I noticed a known issue with TORP object detection as a QLCS approached KDMX along the radial from the south southwest. While Az Shear was generally unimpressive with this line, the QLCS’s orientation with the radar caused a couple TORP objects to generate. While their probability values were low (near 30%), it still caught my eye as something to check out and then immediately write off after checking radar data. This hiccup wasn’t necessarily disruptive to my forecasting process, but was useful to know about before starting off today so I wouldn’t be confused when I see it.

-Wx Warlock

Initially, prob severe detected a small area for it’s algorithm and it tagged it with low severe probabilities. As the storm progressed, prob severe picked up a larger area for the object, and when it did that, the probabilities increased. This could indicate that the storm is gaining strength, but radar presentation has remained almost steady state. This could be a detection issue as when it picks up a larger area, the probabilities increase due to size.

-Stormy Surge

Along a line of storms riding a surface-based instability gradient and good environmental SRH, TORP had a large jump from nearly nothing to a max of 70% in a matter of a few scans. AzShear didn’t look all that impressive as you can see the very meager jump in max values on the graph. Comparing to KDMX data, there was a general signature in the velocity data for one scan with some convergence noted, but it never really tightened up enough that I would have expected to see a that high of a TORP value. The data was real, but it looks like it was sampling the meso which was relatively weak as well. The range from the radar was around 100 km so it was approaching the known distance where it’s less trustworthy, but the takeaway here from me is that I think the high TORP value didn’t match what I had in my head when I looked at velocity data. -newt

I still don’t know if this will turn into anything, but AZ shear signal that is being picked up along with the trend is concerning. This storm is just starting to show some inbound velocity and this may be a very early warning signal. Definitely worth watching at this point.

Quick example of a very small false signal (low prob around 30%) that appears in AZ shear data. That said, I think this is good. It forced a quick look at V/SRM data to see if it was real, and it could be quickly eliminated – and it quickly disappeared from AZshear in the next couple of minutes. Showed up on multiple radars – also good. I also think setting filter lower to say 15% would be better. Having the extra circles there is not problamatic and does not prevent you from seeing the data behind it.

Radar scans came into AWIPS and with TORP at 1922, 24, 27, and 28z. However, an extra update came in to  TORP at 1925z. This is strange given that there wasn’t really an extra radar scan, but the data looked real with legitimate updates. It was helpful too as the TORP increased from 51% at 1922z to 82% at 1925z with an evident increase in the AzShear, which is a big confidence booster in getting a proactive tornado warning out which isn’t easy to do in a messy QLCS setup. I guess glitch or not, can’t argue having an extra mysterious update. -newt

Today’s archived training case starts us off with an already fairly active (and wet) broken line of storms dragging through Georgia.  As I’m getting acclimated with looking at Az Shear and TORP, I found an interesting drop out of data. While using KJGX radar, TORP was tracking a mesocyclone for a few frames before dropping out as said mesocyclone continued to approach the radar site. The mesocyclone did not seem to lose intensity or be cycling, so I was confused by TORP’s sudden disinterest.

I was then reminded that other radars may be reporting “closer” TORP objects. So I flicked on the ability to see TORP objects from multiple radars and low and behold, several objects were still tracking the meso.

The probabilities of these TORP objects were rather high with values of 64%, 68%, and 80%. I’m still a little confused on why one of these is read as closer than the actual closest radar of KJGX, but it did give me an idea. It would be neat to allow a sort of threshold limit to TORP objects from other radars to pop in when at a high enough probability so you have less but more potent objects on screen.

-Wx Warlock