Global lightning detection with Prof. Robert Holzworth

The WWLLN – World Wide Lightning Location Network - is an experimental, collaborative project headquartered at the University of Washington in Seattle. They gather data from sensors located around the globe, then process the data so it can be plotted and regularly updated on maps for scientific use. The project is continually being updated and improved, and recently they installed a new main computer for the central processing work. Professor Robert Holzworth is the director of this project and discussed some of the technical aspects and challenges of this work.

Q: Can you give a brief historical background of your career and the creation of this network project?

RHH: In my PhD experiment at the University of California, Berkeley, I studied space physics problems using electric field measurements on stratospheric balloons. During that project we had to throw away about 10% of our data, because the balloons were too near thunderstorms and therefore those fields had no relationship to the magnetosphere or ionosphere (which have much smaller electric fields than seen near thunderstorms). So, as a result of my PhD research I got interested in lightning and thunderstorms because we had some of the best vector electric field data above thunderstorms. Since then I have launched over 50 scientific balloons, 5 rocket payloads and three satellite instruments, for research projects funded by NSF, NASA and the DOD.

As a result of our good data on thunderstorms and lightning, my research has moved into detailed studies of the upward coupling of lightning and thunderstorm energy into the upper atmosphere and ionosphere. Many of the outstanding problems relate to the global electric circuit which is believed to be driven by the action of global thunderstorms. However, no one had a method for determining the instantaneous level of global lightning activity, so we looked into that problem. In my PhD thesis I developed a crude index of global lightning activity from some radio data that was collected at a few stations round the globe. These stations counted ‘sferics’ or ‘atmospherics’ which are radio noise transients caused by lightning.

In my early research I often worked with a colleague from New Zealand (Richard Dowden, now long retired), who built payloads for some of my balloon experiments. Dowden and his Physics students at the University of Otago studied space physics using radio waves in the VLF (Very Low Frequency) band, which also turns out to be the band with the peak power from lightning. Dowden and his student James Brundell had patented a technique for locating lightning using the VLF radio waves from lightning. Because of my interest in the global physics problems that can be addressed with a global lightning network, I became involved with the fledgling network, originally called the Toga network, operating primarily in the southern Pacific region.. The WWLLN (World Wide Lightning Location Network) grew out of the Toga network, and now covers the globe, locating over 10 million strokes per month globally. I have been directing the WWLLN network since its global expansion and presently operate it with help from colleagues at the University of Washington, as well as two of Dowden’s former students: Drs. James Brundell and Craig Rodger.

24 hr. lightning animation
click for current data

Q: What method does the system use to detect a lightning discharge? Does it use a form of triangulation, similar to the GPS system, to plot location?

RHH: Not triangulation but time of arrival, or TOA. TOA systems use a well timed transient detection from multiple locations to work backwards and figure out where the lightning must have originated in order to reach all those stations at the times they were detected. We have 40 or more receiving stations all around the world, which ‘listen’ to the radio noise from lightning (called sferics). When a station decides a transient pulse is probably lightning, then it time-tags the transient, and sends that exact time, accurate to about a few microseconds, to the WWLLN computers in Seattle, and the backup in Dunedin, New Zealand. These computers then calculate the best fit for lightning strokes to match the times of arrival at the stations. Actually this system works as well as it does because we use the time of Group arrival, or TOGA, not just the time of arrival (TOA). The TOGA is essentially the time of arrival of the peak energy at each station. Since the peak lightning energy is usually generated at about 10 to 15 kiloHertz, the TOGA is not the first energy to arrive, but a time in the middle of the lightning pulse. When we were first using just the TOA, we were only able to achieve location accuracies of within about 100 km. Now, using the TOGA, we can locate lightning anywhere in the world to within about 5 to 10 kilometers, which is about the wavelength of these VLF waves. That means the lightning is located within the storm, since most thunderstorms have dimensions of about 10 km or so.

Q: Is there a threshold where a discharge can't be reliably detected?

RHH: Yes, WWLLN locates lightning which has peak current above about 40 kiloAmps. That means we detect the most energetic tail of the lightning distribution. So we only locate about 10 to 20 percent of all lightning, but our studies have shown that we do locate nearly 100% of lightning-generating storms. This is not an intrinsic limitation on the method. That is, we have to limit the total amount of data sent over the internet so our computers can handle all the calculations in real-time, and so our remote stations with poor or expensive internet connections can keep up. We do that by setting a threshold at each station, so the station does not send too many TOGAs. We therefore concentrate on the largest lightning strokes, which are also the ones that are often of most scientific interest for global studies.

Q: You're now in the process of improving the "plot" accuracy by increasing the number of global detectors. What areas of the globe currently have the best coverage, and what areas would you like to have more?

RHH: We have great coverage in the Pacific, North and South America, Europe and in Southeast Asia. We need to improve our coverage in Africa, and central and northern Asia. Since our technique requires stations to have a continuous broad band internet connection, we have had difficulty getting partners in politically unstable, or developing countries. Since the VLF radiation from lightning propagates very well over oceans, but not so well over land, we have tried to place many of our stations at the margins of the continents, such as our Antarctic stations. No, there is not a lot of lightning in Antarctica, but a station there, along the coast, can detect lightning sferics from great distances, well into the northern hemisphere. We will install over a dozen new stations this year, and hope to have 60 stations by late in 2010.

Q: Does the latency of the Internet have much effect or pose much of a problem with the ideal operation of the system?

RHH: Not really. The longest internet delay for a packet, assuming it gets through at all, is usually less than 1 second. So, we typically accumulate data for several seconds before we start trying to find solutions for the best fit to the TOGAs. We loose a few percent of data because of packet collisions which are never sorted out. This is because we use the UDP packet protocol, which does not have a handshake. It is a send-and-forget system, but it has a very low overhead in internet bandwidth. Each TOGA is sent as a UDP internet packet which has about 50 bytes of information including the time to very high accuracy. Each packet is sequentially numbered, so we know when we loose a packet, and the total lost is very, very small overall. In fact, without the internet , and without GPS for cheap, high timing accuracy we could not do this network.

Q: On your website you're welcoming individuals or organizations to host a new detector location in order to improve the networks detection capability. Can you describe what a detector consists of, what the responsibilities of the "host" would be, and what your criteria is to become a host?

RHH: Each WWLLN receiving station consists of a VLF antenna (about 2-m of wire, inside a 2 inch PVC pipe), along with a preamp, and a GPS antenna. These antennas and the preamp are mounted on a rooftop, and connected to a small electronic box which houses the GPS engine, and electronics for the preamp. This small electronics box sits inside, next to a PC, which is connected to the internet. Both the GPS and the VLF radio signals are fed into the computer from the small electronics box. Except for the VLF antenna (inside a 3-meter 2" diameter PVC pipe), all the rest of the extra electronics fits into a box the size of a small suitcase for shipping. Each host typically provides the PC, which does not have to be the latest and fastest, but it does need to have a Linux operating system to run our code.

Yes, we welcome new hosts especially at locations that will help fill in our holes in north and east Asia, India, and anywhere in Africa. WWLLN provides all the receiving hardware, and the host provides the Internet, as well as keeps the power and Internet connected. In exchange, WWLLN gives ALL the global lightning data to each host, either by monthly CD or weekly download. The host does not need to monitor their equipment, WWLLN operators do that. We inform the host if there is any problem. Many stations have been on and operating without interruption for years at a time. This is possible because we use dedicated Linux systems that are very, very reliable, and originally designed (Unix) for network control (such as the phone system).

We have a lot of information on line at our main website: http://wwlln.net or potential hosts can contact me directly at bobholz@uw.edu

The only limitation we have on hosts is that we need to build and test all these systems before we send them. So, we are concentrating on filling in our weak places before we welcome anyone in any random location. A potential host can look at our maps at our main web page, and decide if your location is, say, about 1,000 km from any other WWLLN station, and then, we would very much welcome your participation immediately.

Q: Our business office last year suffered significant equipment damage as a result of an unexpected lightning strike - even with a multi-layered surge suppression system. Are there practical ways for computer users to receive warning when they should shut-down and disconnect their valuable equipment?

RHH: WWLLN is not really the best system for immediate lightning protection. We do a great job at predicting, or ‘now casting’ the location and movement of severe storms, but I don’t want your readers to think that WWLLN can be effectively used for lightning protection. If you have a golf course, or if you have particularly sensitive equipment that needs to be protected from lightning transients, I recommend one of many commercially available lightning monitors, which don’t really tell you exactly where the lightning is located, but can give you an idea that lightning is occurring nearby. Also, for computers, if you want to isolate yourself from the main power glitches, which could be caused by nearby lightning, then you should power your computer and your network switch with a UPS (uninterruptible power supply, often called a ‘No Break’). Then, when your hair stands on end, or you hear static on your radio, just unplug the UPS from the power mains, without messing up your computer, or your connectivity! A UPS that will power a PC for about 20 minutes costs about $100 in the USA.

Q: How can people access your network data for their own use?

RHH: WWLLN.net has many plots of our data which cover the globe. A casual user can learn a lot from the online plots we provide. If you want more detail, our main web page also shows a link to a Google Earth overlay with all the lightning for the last hour. These are all available free. If the user wants more control over the plotting, or use of the data, then please contact us and we will gladly sell you access to the WWLLN data set. WWLLN is a consortium of universities and government labs, and we operate solely by sale of the data along with the UW contribution of my salary, since this is one of my research project.

Q: What are the future plans and projects that you're working on or excited about?

RHH: We are very actively working on providing the energy per stroke, along with the location of lightning. The energy per stroke is closely related to the Peak Current, so we will then be able to give our users both location and ‘strength’ of the stroke. We already have this working at the individual station level, and hope to have the entire system working in the next 6 to 12 months. It’s tricky because the radio wave propagation is not uniform everywhere. But we have some very smart students, and former students working on solving the problem in an efficient manner.