I noticed that when the “Aggressively-Filtered” TORP objects (likely sampling mountain peaks) were displaying the 5,10,15-min tornado probabilities, they would decrease or jump around or increase. In all of the other cases, the probability seemed to only increase with the greater amount of time. Once we finally got a TORP object sampling a real storm off KBBX, the probabilities only increased as you got closer to the 30-minute time frame, matching what I would expect and what I was seeing in all of the other cases.

Kilometers

Noted a DivShear increase along that thin line of precipitation as it increased in intensity and started to produce several lightning strikes.

Kilometers

It was interesting to see a Torp object identified due west of the BBX radar around 21 UTC with really weak reflectivity and what looks like suspicious velocity data with AZshear in the high category for almost 20 min. This is interesting that this one persists even with the “display aggressive filtered objects” off.

Flash

In the morning supercell case set in OUN CWA, I looked at the pretornado probabilities on the second warning I issued and found it interesting and encouraging that the 5-30 min pretornado probabilities showed increasing values at all 5 min time steps and were higher than the observed Torp probability which added some confidence in the detection and confirmed what I was seeing in the base radar data.

Flash

In this case which involves significant terrain on each side of the valley, Torp objects were identified for many areas where poor velocity data in the higher terrain was observed when turning on the “Display aggressive filtered auto objects”. However, when this is turned off almost all of the Torp objects are removed so it appears the filtering technique being employed is working as intended at least early on in the case.

Flash 

Comparison of TORP object markers in Central California.

First image is the PHI tool with “Display Aggressive Filtered” turned off.
Second image is the PHI tool with “Display Aggressive Filtered turned on.

Most of the object markers align with the edges of the mountains and it’s highly likely terrain blockage clutter is resulting in these erroneous markers.

This example is from a shallow supercell case. The cell exhibited weak but grdually increasing Vrot, but Vrot was generally less than 30 knots. AzShear (pictured) and TORP were useful in capturing the uptick in rotational intensity. The decision to issue a tornado warning was made around or just prior to 2215z (but AWIPS hiccups prevented it from going out until closer to 2225z). The decision to warn was aided by a rotational signature in AvShear (the image below is from 2228z, not the time that aided warning decision) and TORP increasing to between 40-50%. A tornado was reported around 2230z.

-Orange Lightning

Looking back through the data after receiving the tornado report at 2232, there was a peak in Torp probability of 50% around 2217 meaning if that threshold was used as a decision point of going with a warning, given other radar data supported a warning, it would have provided 15 min of lead time. VR Shear never got to 20kts and while there was a couplet the values were not that impressive and no strong gate to gate signature. Also of note the DivShear continued to show a stronger signal with higher values than AzShear in this weaker storm case. I think this case also shows the importance of mesoanalysis and knowing that there was a boundary (warm front) this storm was attached to.

Flash

A new plugin was made available to us for this afternoon case that adds pretornado probabilities from 5-30 min at 5 min intervals as seen in the sampling on this image. This is interesting information that I am curious to see how it does through this event.  This would be great to have available to warning forecasters if it shows skill.

Flash

A storm that has been meager to this point along what we believe to be the warm front is picking up strength and signals in the DivShear and to a lesser degree the AzShear caught my eye around 2210 UTC. There is a developing weak velocity signature as well but this might be the first time I have seen the DivShear pick up on a signature more aggressively than AzShear.

Flash