Terry Kelly, visionary of computerized weather, elected Fellow of the American Meteorological Society

July 9, 2018 | Jean Phillips

Before the era of computer-generated images, television weather forecasts required hand-drawn and colored graphics painstakingly aligned on a wall in the studio so the camera could pan from picture-to-picture, or they used “pull-up boards” and magic marker.

Terry Kelly, UW-Madison alumnus and entrepreneur, charted a new course for real-time television weather broadcasting. Credit: Bryce Richter UW-Madison

Terry Kelly, University of Wisconsin-Madison alumnus and entrepreneur, turned consulting and television weathercasting on its end through his revolutionary work in computerized weather. Last year, the American Meteorological Society (AMS) honored Kelly with its highest distinction — Fellow of the AMS — for his leading role in charting a new course for real-time television weather broadcasting.

Like another UW visionary, Verner Suomi, Kelly brought timely weather information in visual formats out of the research laboratory and into the homes of the public.  As SSEC worked to provide satellite and radar imagery in real time and in higher resolution, Kelly and his team invented the real-time weather displays required to make such data useful.

According to Kelly, though, finding his path to meteorology was anything but straightforward. His journey took him from learning to fly airplanes through dangerous weather, to pairing television weather broadcasts with computer advancements.

In fact, he took a leave-of-absence from college because he was undecided about a major. By chance, he saw a tiny ad in the Boston Globe that said, “Come work for us and we’ll teach you to fly!”

“Well, I thought it sounded like fun to learn how to fly a Piper Aircraft but in less than a week on the job, I had caught my hair in the drill press!” says Kelly. Rather than fire him, the union machinist crew took Kelly under their collective wings, situating him at the end of the assembly line where he was involved with quality control — removing bad rivets from planes and replacing them with good ones — rather than operating heavy machinery.

As promised by his employer, Kelly learned how to fly, earning instrument and multi-engine ratings in the process. It was during his flights along the US east coast that Kelly encountered severe weather situations, “at least one or two of which I probably should not have survived,” he notes.

Two graphical representations of weather systems, both created by Kelly and his team, but decades apart. Prior to the team’s work on computerized weather graphics, newscasters would use several hand-drawn pictures and pan the camera from left to right to show the sequence of predicted weather. Side-by-side, they show the evolution of weather graphics and the processes to create and transmit them. Credit: Terry Kelly

From his vantage point above (and in) the clouds and storms, Kelly remembered his childhood fascination with the beauty of the skies and the early days of television weather. According to Kelly, whenever the weather segment aired on television, he would place plastic wrap over the screen, draw the fronts with markers and check his predictive accuracy the next day. Although he recalls being teased about his preoccupation with the weather, it finally dawned on him many years later when flying a Piper Aircraft that he could be a meteorologist, the notion “hitting him on the back of the head,” he says.

So, Kelly arrived at UW-Madison in 1969 to pursue a degree in meteorology. There he met meteorology professor Frank Sechrist, a synoptician who developed early television forecasts for Wisconsin Public Television’s Target program and who became one of Kelly’s mentors. Those 1970s forecasts used SSEC’s Man-computer Interactive Data Access System (McIDAS), a first-of-its-kind system for the display and animation of satellite data. Even with McIDAS, preparation still required hours of graphics development before each broadcast.

“Everyone who goes to college should have this good fortune,” says Kelly, “as I found a mentor in Sechrist.” One summer, Sechrist asked Kelly to do the broadcast program while he was on vacation. It intrigued Kelly, and was the beginning of a sustained interest in television weather.

He would, in fact, later pursue a career in broadcast meteorology while simultaneously developing new computer systems, landing a role at Madison’s WKOW Channel 27 where he became the face of television forecasting for a decade.  As the first on-air meteorologist in the state, he was awarded the AMS Seal of Approval and later the AMS Award for Outstanding Service by a Broadcast Meteorologist.

However, Kelly’s initial attempts to find a meteorology job in 1972, post-graduation, proved challenging. His persistence and networking first led him to a job seeding clouds with silver iodide flares and rockets and running a rain gauge network at seven thousand feet in the Sierra Nevada Mountains. It is one of the few areas, says Kelly, where weather modification has been shown to work. The hydro-electric power generated through cloud seeding created additional snowpack, he says, helping to provide clean hydropower electricity to Southern California. While the pay was low, only $700 per month, Kelly kept the job for two winters. He and his new wife, Mary, lived in a mobile home provided for them in the mountains.

Having endured the hardships of high-altitude living with low pay, Kelly returned to Madison in 1973 to work with Bob Wollersheim at the Space Science and Engineering Center (SSEC) who was the program manager for a new federally funded project at SSEC called Innovative Video Applications in Meteorology (IVAM). The opportunity to work alongside Suomi and Wollersheim allowed Kelly to focus on applied meteorology, an area that fueled his passion for the intersection of technology and meteorology.

The IVAM program was intended to implement the new concept of “nowcasting” by harnessing the latest data processing techniques and information dissemination technologies to provide a forecast to the public while the weather events were still on-going or imminent.

Kelly’s brain was spinning with ideas to put the IVAM concepts into action. He envisioned taking the rapidly evolving tools, displays and output in meteorology to make them useful to ski areas, power utilities, farmers, aviation and food production and insurance companies — any of the weather-dependent industries. And for the public.

During this time, writes Mike Nelson, another UW alumnus, mentee, and colleague who led Kelly’s nomination for the AMS Fellow honor, Kelly began to explore using emerging lower cost computer systems to display weather graphics for television, drawing inspiration from McIDAS that was becoming more sophisticated each year.

Kelly landed funding from the UW’s University-Industry Research (UIR) Program to conduct a nowcasting experiment with Madison Gas and Electric Co., the City of Madison Streets Department, and a local construction company to test the value of better short-range forecasts that could help companies and cities anticipate the weather and craft plans to work safely around it.  UIR was established to encourage connections between university researchers and industry and Kelly’s idea was a perfect fit for UIR’s mission.

These early systems became “a realization of Dr. Suomi’s dream to make geostationary satellite imagery television-compatible, becoming the “eyes” of the weather for scientists and the public,” says Nelson.

Early computers were expensive, says Kelly, “but when the Apple II was released, we knew we could bring this weather technology into the world in a practical way.” With the new computer, Apple introduced a capability of displaying six-color graphics. The image resolution, adds Kelly, was practically nothing in terms of today’s standards, but it was a first step, and systems could be produced and sold at a fraction of McIDAS costs.

Beyond solving the problem of color graphics, the system would need to comply with existing television broadcast standards that dictated everything from frame size and rate to the process for interweaving video and audio. Kelly’s team devised a specialized video conversion box that would convert the nonstandard video into standard video so that it could be used on-air.

Hand-drawn weather forecasts for news broadcasts in the late 60s, complete with notes to the camera operator. Credit: Terry Kelly

Priced at $11,900, Kelly’s new system generated considerable interest at trade shows, but surprisingly few sales. A repeat visitor to their booth suggested to Kelly that in order to be taken seriously, he should increase the “perceived value” to the buyer by increasing the price of his system. He did, tripling the sales price, and in very short order, selling 50 units:  they were well on the way to becoming a multi-million dollar company.

Kelly founded Weather Central in 1974. The company was later purchased by Dynatech Corporation in 1982, though Kelly and his colleagues stayed on to expand the weather graphics industry under the new ownership. Before long, they’d sold graphics systems to hundreds of television stations in the US market, as well as around the world.

During his tenure at Dynatech, Kelly became a group vice president. He credits this era with honing his acumen for managing and developing technology companies and assessing teams, skills that he would rely on again and again.

Under his leadership, Kelly and colleagues invented a suite of tools for modeling and visualizing the weather in 3D. By the late 1990s, Kelly says that he, chief scientist Richard Daly and the Weather Central team demonstrated the first high definition graphics on the air.

The advancements kept coming as they unveiled tools that enabled a localized preview of tomorrow’s weather. Known as futurecast, it used National Weather Service radar and other data to extrapolate storm development or tracks into the future.

“It takes a tremendous amount of computing power to model the atmosphere down to 1km square,” says Kelly.  That early prediction work proved to be of great benefit to insurance companies, he adds, because they were keenly interested in knowing when and where hail storms would form so that adjusters could assess damage with more certainty and rein in claims costs.

MyWeather LLC, a start-up within Weather Central, was established to further develop and commercialize micromodelling and intelligent, instantly available forecasts.

Kelly sold Weather Central to The Weather Company in 2012. The combined companies remain in Madison under IBM Watson, the artificial intelligence wing of IBM.

After years of innovation, Kelly has 13 US patents in his name or as co-inventor, spanning from an array of systems and methods for presenting wind speed and lightning strike information to personalized storm warnings, though he is quick to point out that while many of these advances were based on his ideas, just as many resulted from group efforts.

Currently, Kelly is founder and partner of Venture Management LLC that invests in technology companies, many of which are affiliated with the UW-Madison and are located in Wisconsin. Looking back, says Kelly, his work with Dynatech, and the skills he developed there for assessing and gauging the potential of start-ups, foreshadowed his current endeavors.

Outside of pursuing his continued interest in improving weather forecasts, Kelly spearheaded the establishment of the Climate Science Education Center at the Aldo Leopold Nature Center in Madison to encourage visitors — students, families, teachers — to think and learn about Earth’s climate in an informative, but neutral environment. He served as the Center’s chairman and CEO for more than 20 years.

Kelly’s fingerprints can be seen in other philanthropic gestures, such as his founding investment in Air America Radio, a national progressive network, and Madison’s Rhythm and Booms Independence Day celebration which spanned nearly two decades.

Kelly has spent his career on the leading edge of computerized weather, providing critical information and predictive capabilities for people and industries around the world.

“One thing you’d like to see about the work you do in your life is that it mattered,” he says.

AMS Fellow and past President Bob Ryan adds, “Terry Kelly is an example of what is best about our science and our society … his election as a Fellow [is] as much an honor to the AMS as it [is] to Terry.”

By Jean Phillips