TORP seems to struggle a bit when there are gaps in between lines on reflectivity. Noticed that several areas of low probabilities existed in gaps throughout the event. Can be a bit distracting as it is eye catching and takes time away from other areas that may need more attention.

Nice clover signature on DivShear. This appeared about 3 minutes after the original tornado report, so did not provide a lead time. However, it is good to see this signature verify.

-FujitaReno

TORP had an issue handling the tracking of objects along a linear AzShear maximum where a weak tornadic circulation was also developing. Initially, an object with ID 340 was identified with low probabilities (27%) along a linear AzShear feature associated with a region of convergence northwest of LWX.

Note: Ignore the trend on the left throughout this post – it was slow to update on objects being clicked in the PHI Tool and doesn’t necessarily represent the actual trend in objects being discussed.

On the next scan, ID340 continues to track the same part of the linear AzShear feature, still with low probabilities, while a new object (ID171) appears with a 95% probability just to the south where a circulation is rapidly strengthening and a tornado is ongoing or forms shortly thereafter.

In the next scan, ID340  jumps south to the tornadic circulation, where ID171 is still tracking the circulation for its second scan in a row.

ID171  then drops off on the next scan and ID340 becomes the primary object tracking the tornadic circulation for the next several scans.

– Mr. Peanut

Issued a tornado warning on the cell just east of Philmont. There was a long broken line (hard to see here as I zoomed in quite a bit) and one cell can be seen broken off,  looks like an embedded weak supercell.  Issued the warning based off of the line break, weak but tight couplet and a very slight CC drop aligned with the couplet. Was watching this thing for a few scans and the rotation slowly wrapped up and the CC drop even had some vertical continuation. At this time, there was no TORP object over the rotation. I have noticed TORP struggles in QLCS type events. When TORP did eventually put an object over this rotation, its trend line came down from the north, which makes no sense in this case. I noticed this a few other times with this simulation and another simulation. It will make huge jumps along a line and not realize that they are two different circulations. This storm did eventually have damage reports,  5 mins after the warning and 3 mins after the TORP object.

First pic shows the storm I warned on just east of Philmont (no warning showed here).

The second picture shows no TORP object over the storm I warned.

The third picture shows the TORP object a few scans later, with the “line history” coming down from the north.

-Gary

Noted a difference in TORP and PHI regarding the reported tornado. Forgot to turn on the setting in PHI that allows for near radar imagery to be seen. However, after I turned this on I noticed that PHI had a 95% ring, while TORP at the same time showed nothing.

I noticed that as the northern end of the line of convection became more parallel to the radar beam, several low-probability TORP objects occurred (see below):

However, when I go back to the 30% threshold the one that is left is correlated to a deeper area of rotation using a traditional All-Tilts analysis (see below the 1.8 deg elevation angle of SRM). This was encouraging that TORP was able to identify the most legit circulation with highest probabilities.

-dryadiabat

Just re-worked every procedure to simplify my view in order to take in the TORP, Az/Div-Shear products more easily. There were time match issues, and missing data and this seems to have helped some. I’m running a simple 0.5 Z/SRM w/ TORP and 4-panel Z/SRM/Az/Div now and it seems to have helped.

Had several issues with thresholds and various settings that resulted in only seeing the TORP for the tornadic circulation after it developed with 95% prob, and no history using the PHI tool to display.

-dryadiabat

Even with the reset of procedures and time matching to TORP,  still having trouble displaying TORP in AWIPS.

Unsure what all elements in the readout for TORP are. 

-dryadiabat

The AzShear product is useful for UDCZ identification. Curious to see its use in the more difficult cases where the UDCZ location can be ambiguous. Will be watching for such examples in this experiment.

-dryadiabat

Really difficult with time match issues. Concerned I’m not getting timely velocity data if time matched to TORP.

Noticed a 25% TORP well west within the stratiform area to the west of the deeper convection, where I would’ve discounted from my manual analysis of radar data.

The Az and Div stood out with the mesovort/possible brief tornado in eastern Prince Williams County. 

Noticed a very distant TORP object from KLWX in southern Pennsylvania. Obviously there are beam height issues and the velocity is relatively low compared to earlier circulations closer to the radar. So, this is encouraging as it seems to account for the beam broadening issue with greater range.

More brainstorming, exploring, and unfiltered blogging. Noticed that once the convection moved east and the beam is more perpendicular that for the UDCZ the DivShear product really shows a strong signal, as would be expected given beam orientation. See earlier post above about AzShear showing a stronger signal when the beam was more parallel to the UDCZ.

Good example of the contrast of radar perspective, KILX vs KDOX.

Similar beam height but a stronger rotation on KDOX. Shows the value of having multiple radar TORP overlaid onto a single radar display, for situational awareness.

Monitoring another area where the line orientation should be favorable for QLCS mesovort/tornado concerns. There are areas within this segment of rotation and a slight upward trend in the latest scan as I type this. TORP has identified this area with 33% prob.

Was curious to see if anticyclonic rotation would have a TORP object. It did not in this example east of D.C.

A TORP detection along the cold front, from KDOX which is much further than KIWX. Reflectivity thresholds must have been just enough. 33% is higher than I would’ve considered from my manual assessment. 

Follow up to the last post. Convection has re-oriented to (1) be more favorable for mesovort generation given the shear vector orientation to the UDCZ, and (2) to minimize TORP objects along the leading edge since the beam orientation is more perpendicular. -dryadiabat

Despite a messy linear structure oriented down-radial, AzShear initially proved valuable in highlighting a weak circulation and TORP even highlighted it with ~30% probability. I was pleasantly surprised to see this after witnessing some tracking and object identification issues in quasi-linear structures during the real time event Monday afternoon.

Things went a bit downhill after that. It became apparent that TORP was over-identifying in an area that actually showed convergence in velocity data, but oriented to the radar in a way that it also lit up AzShear (which I suspect is the leading cause). Granted,

I should note that the above is with no probability threshold set. With a 30% threshold, most of these disappeared, so while it appears TORP may be over-identifying, the probabilities associated with these objects are low and fairly easy to successfully filter out.

– Mr. Peanut

Along this line southwest of the radar, AzShear values increased but SRM still looked benign. TORP did pick up on probs above 20% and defined it as an object

TORP increased in area of the line break along with three ingredients method for meso vortex generation, decided to follow CR CWIP guidance and issue SVR w/TOR Possible TAG.

The portion of the line where the updraft/downdraft convergence zone (UDCZ) caused AzShear to jump and looks to be prompting higher TORP probabilities. It will be very critical for warning mets to know this limitation and be able to identify it, similar to vertical sidelobe contamination. These kinds properties may nearly require a RAC like training, even for experienced forecasters.

See photo below, noticing we are getting a lot of TORPs above 20 % when we get a line break with reflectivity, which is one of the CWIP nudgers

After turning on the range filtering in the Phi Tool, noticed an area closed to LWX had a 95% probability within the SVR w/TOR Possible tag. This was also near a line a break. Went ahead and issued a warning. Had I seen this object earlier, I would have considered a tornado warning earlier, but I only noticed it after the Vrot on SRM drastically increased. But these probs seem to matching some of the 3 IMs as stated in Schaumann and Pryzblylinzky (2012). (See two photos blelow)

Az_Shear appears to be catching on to increasing V_rot about 1-2 scans prior to when I as a forecaster catch onto it. This would certainly be useful to notice a trend is happening before it becomes apparently visible to the eye via SRM. (see below)

Object 88 near Myersville around 1244z, not sure why it would be there (below)

“Don’t time match on reflectivity”.

As of 1252z, coming up with better ideas for D2D procedure bundles and perspective to integrate AzShear/DivShear into a smoother work flow. I will have to compare MRMS rotational track, as well as MRMS 0-2km Azmithal Shear. These products may be able to replace a lot of MRMS product used for velocity fields. However, MRMS will still be needed for hail.

The workload of looking at AzShear and DivShear has decreased the amount of time am looking at Dual Polarization products. However, this could be in part because of only having one AWIPS screen, where at the office I would have 3. And perhaps having GR2 opened in on a Windows PC could also change this for easier viewing of Dual Pol products. Will need to make a consciously effort to incorporate Dual Pol with these products into my workflow.

Another example of TOPR going up with line break. Also, it appears that we get a “couplet”, or a “dipole” like feature in DivShear where enhanced convergence is next to divergence (see 2 pictures below)

TORP probabilities of 30-50% (especially is showing an increasing trend) get my attention on radar and has me exploring base products more (and I need to be better about dual pol). TOPR above 70% has pushed me issue a tornado warning. Even in times of false alarm, I still get a problematic Vrot that I would likely issue a warning on ~1-2 scans later. TORP trends appears to be helping to add some lead time, especially if probs start crossing above 50%. Here on Day 2 of the experiment, still trying to mentally calibrate AzShear/DivShear values with this TORP output.

Noticing low probabilities from where the radial is parallel to the leading edge of the convection. -dryadiabat

Some low probability false alarms associated with ground clutter near radar.

No TORP showing up for fairly strong rotation.

Finding myself issuing tornado warnings around 40% TORP compared to 50% for practice case earlier this afternoon. Possibly due to a better feel for the environment in mind. However, this was not true for all storms, especially those further away from radar where personal confidence was lower.

Example of side lobe contamination possibly impacting TORP.

It would be nice if the future trend portion of the graph was shaded differently to make it stand out better at quick glance.

Showing similar DivShear values/highlighting, though only one area is highlighted by the TORP. Therefore, divshear may not be very useful for low TORP values.