In the spring of 2008 we deployed a Campbell Scientific CS110 Electric-Field Meter, with strobe lights and siren for lightning hazard warning, at the Jimmie Austin Golf Club at the University of Oklahoma. The eighteen-hole course covers approximately 2 square kilometers in an irregular shape. We decided to deploy it first near the clubhouse at one end of the course so that the course staff could keep an eye on it as they gain experience with its operation and so that it would attract the attention of golfers. Indeed, it has evoked significant interest and many questions from golf-course patrons. Upon learning that the system is designed to provide alerts before the first lightning flash in the area, many patrons asked more questions, giving golf-course staff an opportunity to direct them to the computer where the data were displayed in real-time. The opportunity to demonstrate to customers how the system works turned out to be one of the most important aspects of the trial period. Furthermore, while explaining the operation of the system, staff members had an opportunity to stress lightning safety in concrete terms. We present an anecdotal description of our early experience with the system.

At the outset we were not certain whether the arbitrarily chosen thresholds for caution and alert levels would be appropriate for this deployment. We fully expected that it would be necessary to adjust them. The field meter is connected to a siren and three strobe lights, not surprisingly green, yellow, and red. The system calculates a one-minute running average of the electric field. During fair weather, with electric-field magnitude less than 500 V/m, the green strobe light flashes. When the average field magnitude exceeds 500 V/m, the yellow strobe light starts to flash. At the same time, the program running on a computer in the clubhouse announces audibly that the caution threshold has been exceeded. If the average field magnitude exceeds 1000 V/m, the red strobe light begins to flash, the siren sounds, and an audible warning is announced at the computer in the clubhouse. The cessation criteria are based on a ten-minute running average. If the ten-minute running average field magnitude drops below 500 V/m, the red light stops flashing, the yellow light starts flashing, and the siren is turned off. If the ten-minute running average field magnitude drops below 250 V/m, the yellow light stops flashing and the green light resumes flashing. If the CS110 internal self-tests indicate an instrument failure or that maintenance is needed, all strobes and sirens are turned off. These thresholds have worked well so far. We have not had to adjust them as we had expected.

Since deployment, there have been approximately 20 storms in the vicinity that caused the system to go to the caution stage and 14 that caused it to go to the alert stage. It turns out that there is an Electrical Storm Identification Device (ESID) located less than 1 km to the west of the golf course, at a Norman city park. The ESID is a single-station lightning warning sensor formerly manufactured by Vaisala, Inc. which sounds an audible warning in response to the occurrence of CG lightning flashes nearby. In contrast, the electric-field meters measure the atmospheric electric field, so that alarms can be enabled when pre-determined thresholds are exceeded, before the first CG flash in a storm. In several cases in which one of the authors (LDM) was able to compare the response of the ESID to that of the field meter, the latter tended to provide an earlier alert. There was one case in which a CG flash that was visually observed by golf-course personnel to originate several km to the south struck within 1 km of the CS110 while it was only in the caution stage. This was not seen as a failure of the single system but rather as an argument in favor of deployment of multiple field meters over a large enough area to monitor the electric field beneath storms within about 10 km of the course boundaries that might produce a flash that could strike within the boundaries.

Operators of a golf course must walk a fine line between under-warning and over-warning. They must not fail to warn of lightning hazards but also must not be overzealous in situations in which storms would not present a serious threat to patrons. The system performed well in providing hazard-warning decision support for such difficult decisions. For example, on May 1, 2008 a supercell followed a track from near Blanchard, OK, through Moore, OK and Midwest City, OK. The core of the storm was between 15 km and 20 km away at various times as it passed to the west and north. One of the authors (LDM) maintained a close watch on the conditions in the area and on the electric-field display. Even though the storm approached within about 16 km of the course location at one time, it never was a serious threat. Appropriately, the system responded only at the caution level.

In the past the course staff made subjective decisions as to whether to call golfers in during storm situations. The decisions were based mainly upon the occurrence of both audible thunder and visually observed lightning. This method does not allow for pre-strike warning of players out on the course. The signal of cautions and alerts by the field meter gave course staff confidence to make informed decisions to remove players from the course before audible thunder was heard or a lighting flash was observed visually. On the basis of past experience, one of the authors (LDM) estimates that out of the 14 instances in which the system went to the alert stage, there would have been 4 or 5 cases in which the staff would have been reluctant to pull players off the course in the past, but did so on the basis of the electric-field warning, providing a greater margin of safety for golf patrons.