It has been unseasonably quiet in all of the LMA networks the past few days; unfortunately, the upcoming week does not look too promising either.  In the meantime, below is a summary of severe/near severe activity in any of the networks during the project so far including the date/time & corresponding networks…

Note: No systematic verification of jumps has been processed for any of the cases.

April:

4 Apr 2012 – NALMA: approx. 2300 UTC  (possibly out-of-range).

5-6 Apr 2012 – NALMA: 2300-0100 UTC.

9-10 Apr 2012 – WTLMA / OKLMA: 2100-0200 UTC (possibly out-of-range for both networks).

12 Apr 2012 – WTLMA: 0100-0500 UTC.

13-14 Apr 2012 – OKLMA: 1900-0500 UTC.

15 Apr 2012 – WTLMA: 0300-0500 (& 14 Apr OKLMA, 2100+ UTC, likely out-of-range)

17 Apr 2012 – NALMA: 1900-2100 UTC (possibly out-of-range)

20 Apr 2012  – WTLMA/OKLMA: 0000-0300 UTC

20 Apr 2012 – FL-LDAR: 1900-2300 UTC

22 Apr 2012 – FL-LDAR: 0700-0800 UTC.

25 Apr 2012 – WTLMA: 0000-0300 UTC (possibly out-of-range).

26 Apr 2012 – WTLMA: 2200-2300 UTC (possibly out-of-range).

27 Apr 2012 – NALMA: 0000-0100 UTC (possibly out-of-range).

29 Apr 2012 – WTLMA/OKLMA: 0000-0700 UTC.

29-30 Apr 2012 – WTLMA: 2300-0500 UTC.

30 Apr / 1 May 2012 – WTLMA/OKLMA: 2100-0500 UTC.

May (as of 2245 UTC on 14 May 2012):

2 May 2012 – DCLMA: 2000-2230 UTC.

3-4 May 2012 – DCLMA: 2100-0200 UTC.

4 May 2012 – DCLMA: 1600-1700 UTC.

6 May 2012 – NALMA: ~1100 UTC.

6-7 May 2012 – NALMA: 1700-0100 UTC.

7 May 2012 – NALMA: 1900-2100 UTC.

8 May 2012 – NALMA: ~1800 UTC (possibly out-of-range).

14 May 2012 – FL-LDAR: (likely non-severe and under jump threshold; HWT Spring Experiment/EWP operating in domain).

It was an active evening over two of our networks last night, with lightning jumps occurring on multiple storms in OK & TX.  SHAVE was active as well though phone numbers in some of the areas were a bit sparse and calling ended at 9pm CDT, with many storm reports appearing just after this.  Luckily, storm reports in the region were plentiful including one of 3 in hail SE of Plainview, TX.

Oklahoma (Caddo/Grady Co. storm):

Storm trends, jumps and storm reports (SPC, prelim) for this same storm:

There were multiple other storms in Oklahoma that also had lightning jumps.  These storms were along the line from central Canadian Co through NW Oklahoma Co into Logan, Lincoln and Payne Counties.  Watching the realtime feed it seems that many of the ‘jumps’ were due to storm mergers along the line.  It will be interesting to compare storm cluster size with timing of the jumps in post-analysis.  SHAVE did call along this line, but found only dime-to-nickel-size hail.

Plainview, TX storm:

(details to be completed)

—–

SHAVE closeout at Fri 20 Apr 2012 015711 GMT:  SHAVE made ~103 calls today.  We had 30 hail reports with 2 being severe (1.0″+), 0 being significant (2″+).

SPC storm reports:

http://www.spc.noaa.gov/climo/reports/120419_rpts.html

Earlier entries (to be updated) have gone over a few of the technical issues we’ve found during these first couple of events.  Managed to find a new one on Friday.

Our flash sorting algorithm (i.e., w2lmaflash) seems to be having difficulty dealing with timestamps across multiple minutes from multiple domains coming in at the same time. For example, lets say the current time is 2012 UTC… the OK-LMA is sending 2009 UTC data (the more activity the slower the data transfer gets, up to about 4 min latency), but meanwhile there is no activity in the NA-LMA domain so the timestamp on the LDM feed the most recent data is 2012 UTC. This seems to be leading the w2lmaflash algorithm to be ‘prune’ (or drop) flashes that are more than a minute older (our current processing interval) than the most recent time from an LMA network…

At least that’s what I believe the problem is right now since it’s not reproducible in post-processing when it pulls from all the networks at the same time.

—-

I pulled all the data from Friday off the realtime feed and re-ran it over the weekend over the Oklahoma domain.  The Norman tornadic storm is labeled as Cell 1 in all the figures and tables below using scale 1 from the w2segmotionll (kmeans) storm tracking.

The total flash rate (TFR) rate for this storm peaked at 282 flashes per min at 2244 UTC, with jumps occurring at 1835, 1844, 1853, 1940, 1952, 2012, and 2027 UTC (Fig. 1).  Hail times from Local storm reports associated with this storm  were recorded at 2041 (1 in), 2055 (1 in), and 2142 (1.5 in).  The EF1 tornado moved through Norman, OK between 2102 and 2113 UTC (estimates from TDWR).

I’m discounting the lightning jump at 2012 UTC since there was no real-time radar merger between 2000 and 2008 UTC, hence nothing for the algorithm to track.  Oddly, during this same time period [2002-2008 UTC] four stations in the OKLMA were also not communicating, leading to too few stations to do any quality control for flash sorting.  So it is unknown whether the storm would have produced a lightning jump at this time had the merger/tracking been working correctly.

What is really interesting with this storm is how the character of the lightning changed prior to 2000 UTC and after 2010 UTC as the storm began it’s deviant motion right (towards Norman).  The animation below (5 min intervals) catches this character as lightning initiations grow not only in number but also across the storm. The flash rate continues to increase with initiations spreading into the anvil region by the time the core of the storm is over Norman.  This increase in flash rate and expanded coverage of lightning is reflective of change in the strength and size of  the main updraft core over the same time period (linked via storm electrification processes in around the updraft region).

20120413_animation (quicktime movie, looks a bit better than gif animation below. Either way, you’ll need to click on the link above or image below to see the animation)

There was also a tornado reported in Mustang at 12:45 am CDT, but I haven’t had time to evaluate that storm yet.

Feel free to leave any questions or comments below or send them directly to me at kristin.kuhlman -at- noaa.gov

Since the LMA networks depend on a line-of-sight for the time-of-arrival detections, distance from the network center is a key aspect to quality control.  While accuracy of a any VHF detection in the center of the network is around ~10 m, this drops off dramatically with range.  Flash sorting helps to remove some of this dependency, however, we still want to make sure jumps are not occurring simply because a storm is moving into to the LMA domain.

Originally, this network range issue was planned to be part of the flash sorting algorithm, but after looking at the data we decided to change it to a strict 150 km range (Fig. 1).  But I was still seeing an effect of storms entering the network, so I decided to reduce the range again to 125 km (Fig. 2).

We’ve now switched to 125 km for the majority of the networks (100 km for FL).  This means that if any part of the identified k-means cluster overlaps the region within 125 km of the network center it is eligible for a lightning jump. (And, yes, this does mean that the majority of a cluster can be outside the range, hence why we went a little stricter and confined it to 125 km instead of 150 km).

The new network range map can be seen below:

I don’t have the exact location of the OKLMA southwest extension yet, nor is the data feed working in realtime so that is being left out for the moment.

Moving an algorithm into real-time operations is full of pitfalls and challenges.  This first week of operations for the lightning jump algorithm was no exception.  No matter how much archive data you test on, there are certain aspects that change once you’re at the mercy of a real-time feed. Add to that, working with weather data where the outcomes are unknown and you get a great recipe for failure.

This first week we had three events to test data on, with the SHAVE project active on 2 of the days.  The below images (screenshots from the event) will give you a chance to see our visualization of the lightning jump for the SHAVE participants and what you can see real time during future events at:  http://ewp.nssl.noaa.gov/projects/shave/data/

Basically, the jump is currently being visualized by colored circles along the track of the centroid of the identified storm cluster. When a storm has a lightning flash rate > 10 flashes per min, but does not have an active ‘jump’ a gray circle is placed along the track.  When the storm meets the jump criteria, the circle changes to red.  The activity this first week was our first chance to see the visualization and get the data over to kml format (Kiel Ortega was instrumental in getting all of this up and going).

–

5 Apr 2012: Northern Alabama

(to be filled in w/ info on any jumps)

SHAVE closeout on Fri 06 Apr 2012 at 011011 GMT:  SHAVE made ~471 calls today. Today we had 98 hail reports with 11 being severe (1.0″+), 0 being significant (2″+).  We had 1 wind reports today.

SPC storm reports for this event: http://www.spc.noaa.gov/climo/reports/120405_rpts.html

–

7 Apr 2012:  Oklahoma

SHAVE was not active on this day and there were no warnings or storm reports over any of the LMA regions.

–

9 Apr 2012:  Oklahoma

Activity on 9 Apr 2012 was primarily outside the central range of the OKLMA network, though there was a chance that storms would move into range. I’ve included it here mostly for illustration of storm cluster identification and storm tracking.

The kmeans storm identification works on 3 separate scales, simply thought of small, medium and large (or 1, 2 and 3).  Each of these is using Reflectivity at -10 C, with thresholds set such that all reflectivity less than 20 dBZ becomes 0 and all greater than 50 dBZ appears at 50 dBZ to the algorithm.  We are also using a percent filter such that pixels nearby that are partially filled or surrounded by pixels that are filled match the neighbors.  This acts to expand the cluster slightly, with the goal to pick up more of the lightning flashes.  Looking at the example in the screenshot above, all the CG flashes within the colored area would be counted as occurring with that storm, anything outside the area, even closely nearby will not be associated.  Sometimes (at scales 1 & 2 especially) this may combine storms that the human forecaster would deem to be separate entities.  This is typically a problem with smaller storms occurring in close proximity to each other as we had from this example on 9 Apr.

The corresponding visualization in Google Earth (or similarly on Google Maps) can be seen below.

SHAVE closeout at Mon 09 Apr 2012 234803 GMT. SHAVE made ~109 calls today. We had 27 hail reports with 13 being severe (1.0″+), 10 being significant (2″+)

This is the initial “welcome” post for the Lightning Jump Algorithm test.  Please come back using the direct link to the “Lightning Jump Algorithm” category for additional posts.

Greg Stumpf, CIMMS/MDL