Month: February 2017

February 2017 Blog Post



In today’s weather world a great deal of historical data is collected and can be accessed in prepared reports.  Lightning data is one example.  The idea of being able to determine where lightning struck the ground was developed in the late 1970’s and initially deployed in the western US to be able to locate where a lightning strike might have started a forest fire.  In the years up to 1989 the technology for what became the National Lightning Detection Network was being developed and coverage spread until it encompassed the entire United States.  So complete lightning data coverage for the US was initiated in 1989 and all of that data is still available.

The National Lightning Detection Network was developed and owned by a company called Global Atmospherics in Tucson, Arizona.  Global Atmospherics was purchased by a Finnish company Vaisala, Inc., in 2002 and has owned and operated the network since that time.  In 2004, a competing data network, the US Precision Lightning Network was established to provide lightning strike data.  In 2009, a third network, Earth Networks Total Lightning Network, was deployed in the continental US. All three of these networks continue to operate and supply data from their own network of ground-based sensors.   Each network uses more than 150 sensors placed throughout the United States and adjacent countries to collect their data.

In discussing lightning, there are several terms to be aware of.  A cloud-to-ground lightning strike is an electrical discharge between the atmosphere and the ground.  The lightning flash is the entire event from beginning to end.    The return stroke is the electrical current flowing through the lightning channel.  Many times there can be more than one stroke, and this is what causes the lightning to appear to flicker.  There can be as many as 20 or more strokes in one lightning flash.  A strike point is the location on the ground where lightning hit. And not all of the return strokes necessarily hit the same point on the ground.

Lightning stroke data collection usually includes the date and time of a strike, the latitude and longitude (location) of the strike, the peak amplitude (in kA or thousands of amperes of current in the stroke), the polarity of the strike (positive or negative), and a determination of whether it was a cloud-to-ground or cloud-to-cloud strike.

A common application for lightning data, is for claims purposes.  An  insurance user can query a lightning network data archive,  for the property where lightning is believed to have struck, and then look at the map with data points that represent strike locations. The search diameter from a specific point/address can be selected between 1 and 15 miles.

In reviewing and using the lightning data reports, it is important to understand the limitations of lightning networks .  In general, US lightning networks have a near 100 percent detection of thunderstorms.  So, if a lightning detection network report shows no lightning in that location on the day in question that information can be believed.

For address-specific applications of lightning data, it gets a little more complicated, because the ability to detect a thunderstorm is not the same thing as detecting every single stroke that contacts the ground.  All networks have a flash detection efficiency of 90-95 percent or greater for cloud-to-ground lightning.  In general, the networks may not record the smallest magnitude lightning strikes very well.  In addition, during intense thunderstorms when there are many strikes, the sensors can be processing data and resetting after a flash when the next flash occurs and they may not see the second flash.  So, all networks can miss some lightning strikes.

All US lightning networks state their median location accuracy is about 1/8 mile.  So, for a property that is a large campus or is located in a lightly populated area that accuracy doesn’t generally affect the interpretation of results.  However, for homes in a heavily populated area the 1/8 mile radius around the reported strike point could encompass a number of separate houses.  So while there is a 95 percent chance that the lightning actually struck within approximately 1/8 mile of the pinpoint shown on the data map that might not be at the house selected as the center of the data area.  This image shows a point data map with a 5 mile search radius.


Image from CoreLogic report for Haag Engr.

In addition to the median accuracy, which can be envisioned as a small circle around the points on the report map, the ultimate accuracy of the location is dependent upon the number of sensors that “see” the flash and their locations.  The greater number of sensors that detect a stroke, the more accurately that stroke will be plotted. The fewer sensors that detect the stroke will have a larger error and this will be visualized with a more elliptical or oval shape.   This is illustrated by comparing the point data on the Lightning Stroke Map above with the same data as presented on the Confidence Ellipses map below.  So rather than the strike point always being within 1/8 mile of the point shown on the map in the data report, the actual strike point could be several miles away in the worst cases.  These ellipses have a 99 percent of encompassing the actual strike point.  This images shows the same point data map with the 5 mile search radius with the 99 percent confidence ellipses shown.  Note the large size of some of the ellipses.

Image from CoreLogic report for Haag Engr.

In summary, a lightning data report can be considered almost foolproof when evaluating whether or not lightning was present in the area of interest on the date and time of interest.  However, the exact location of the strike may not be as obvious as the reports present.  In these cases it is important to use information from the scene, including physical evidence and expert evaluation, to determine if lightning actually struck the property or caused damage at the property.

John D. Stewart P.E.

John D. Stewart graduated from the University of Texas at Arlington with a Bachelor of Science degree in Electrical Engineering.  He is a Principal Engineer with Haag Engineering Co., and a registered professional engineer in the states of Texas and Arizona.  Mr. Stewart is a member of the Institute of Electrical and Electronic Engineers (IEEE), the American Institute of Chemical Engineers (AIChE), the National Fire Protection Association (NFPA), the International Association of Arson Investigators (IAAI), the National Society of Professional Engineers (NSPE), and the Texas Society of Professional Engineers.

 See his profile here.