Research-Operations Interaction in Severe Storms Meteorology


Chuck Doswell

Posted: 09 May 2009 Updated: whenever

This is yet another expression of opinion. Should you wish to discuss it, feel free to send me an email at (use the email link or cut and paste the address after replacing # with @). I may choose to append any exchange between us to this webpage, so if you don't care to have your views presented here, don't waste my time responding. If you're afraid to express yourself for fear of retribution by someone (not me!), you can ask me not to be identified, but I won't post anything from anyone who refuses to identify themselves to me.

Some background

Time spent in Kansas City

After completion of my doctorate, my very first real job was as a research meteorologist working for the Techniques Development Unit (TDU) at the National Severe Storms Forecast Center (NSSFC) in Kansas City, MO. This particular position, doing research for the Severe Local Storms Unit (SLS), had been a dream of mine ever since I'd first been employed during the summers as a student trainee coming out of undergraduate school. During my summers in Kansas City as a new graduate student, I'd learned from the veteran forecasters that the research and operational forecasting sides of meteorology didn't communicate very well. It was as a student trainee that the disconnect between research and operations first became apparent - what I was learning in my classes didn't seem to be consistent with what I saw most forecasters doing on the job. It was at that point that I became determined to devote my career to trying to bridge that gap.

As a short summary of how I chose to go about that, it was clear that to be a fully accepted member of the research community, I had to go on to complete my doctorate. From that point, I would have to conduct research and publish the results in refereed journals. On the other side of the coin, to be an accepted member of the operational community, I had to sit in a forecaster's seat at some point in my life. My first post-doctoral position at TDU was the best opportunity I could imagine to accomplish my career goals because I was tasked both to do research and to work shifts at the SELS forecast desk (primarily fill-in shifts so that the full-time forecasters could do other things), first as a SELS Assistant Forecaster and then as a SELS Lead Forecaster. To be at the research-operations interface, I had to have a solid basis in both worlds.

For six years, I worked at my dream job and I began to realize some things about how best to accomplish the interaction between the research I was doing and the task I was asked with on the forecast desk.

Time spent in Boulder

After leaving NSSFC to return to a full-time research position, I maintained my contacts with the SELS forecasters and did my best to help my research colleagues recognize that the SELS forecasters (at least some of them) were worthy of consideration in research projects related to severe convective storms. I also helped one of my research colleagues in Boulder, CO (where I moved after leaving NSSFC, to work with the Weather Research Program [WRP] within the Environmental Research Laboratories [WRP]) to implement some of his ideas in an operational forecast office. I believe my experience helped him overcome resistance to his new ideas. My boss at WRP was Bob Maddox, whose career path had taken him into research from his beginnings in the Air Force as an operational severe weather forecaster working with the pioneering severe storms forecaster, Robert Miller. Bob Maddox understood what I was trying to do and was supportive because his goals and mine as researchers were quite similar, despite the differences in the paths we took to get to WRP. One important thing Bob did for me was to encourage me to participate in the Flash Flood Forecasting Course (FFFC)at the NWS Training Center (in Kansas City, MO). During this time, I was able to see how effective Bob was as a teacher because of his operational experience. He was able to establish rapport with the forecasters in the course by showing that he understood what challenges they confront every day on their jobs. I tried to emulate his methods as best I could for getting meteorological content across to operational forecasters. We realized that many operational forecasters simply don't have enough education and training to be as effective at weather forecasting as they could be. I recognized that I had a lot to learn about forecaster training, but that NWS training is far from being even marginally adequate for their forecasters. It's very difficult for researchers and forecasters to communicate when they don't have a substantial base of shared information. That base has to be the science of meteorology.

For four years, I worked with some of the best meteorologists in ERL. My operational experience could be applied to the research in many ways, including the development of field programs designed to advance our understanding in ways that could be applied to severe weather forecasting. This lead to some new thoughts about research-operations interaction.

Time spent in Norman

When you have a boss you admire, you want to stay with that person wherever s/he goes. Bob Maddox accepted the position as Director of the National Severe Storms Laboratory (NSSL), and since I had some other reasons to want to return to Norman anyway, I transferred without hesitation. From the outset, I wanted to establish my intentions to work on the research-operations interface, so Bob allowed me to have my office in the NWS Weather Forecast Office (WFO) that had recently moved from the airport at Oklahoma City to Norman, just across the street from NSSL. The meteorologist-in-charge (MIC) at the time was Ken Crawford, whom I had known for several years. I admired Ken a great deal as a manager and truly enjoyed my first several years at NSSL because I was "embedded" in an operational office that was managed by a person who understood and supported the sorts of things I wanted to do. We not only ran several research field programs out of the Norman WFO, but Ken even developed an underfunded version of the EFC that was the best he could do under the constraints imposed on him by NWS management. Clearly, this was just what I wanted to see happen, so for a time I was like a kid in a candy store. But all good things come to an end, it seems. The space in the WFO was needed for an expansion of the AFOS computers, so I reluctantly had to surrender my WFO office - I literally was replaced by a computer!

That ended my unique relationship with the WFO. For some time after moving to Norman, I continued to contribute to the FFFC in Kansas City, but that course was eventually abolished, to be replaced by something that didn't require our participation. I began working with colleagues at NSSL and it became a period when I was more productive in research than ever. The rest of my story is irrelevant here. It can be seen from the preceding that I've developed certain concepts of how the research-operations interaction should proceed, based virtually entirely on my own experience. However, it's become evident to me of late that I've been blind to some fascinating new developments here in Norman that don't fit my notions, but clearly have been quite successful at fostering my goal of research-operations interaction.

The Hazardous Weather Testbed

As discussed in my paper about the history of research in severe convective storms (Doswell 2007 - available here or here), the formation of SELS (which was the predecessor to the present-day Storm Prediction Center [SPC]) included a research component. Unfortunately, for several reasons, the researchers and forecasters began to clash and this was eventually reflected in a physical separation: the research team - the National Severe Storms Project (NSSP) - moved to Norman and became NSSL, whereas SELS stayed in Kansas City. This drastically reduced the research-operations collaboration regarding severe convective storms for decades, but the breach began to heal with the formation of TDU and eventually was transformed when SLS became the SPC and rejoined the researchers at NSSL. With the move to the so-called "National" Weather Center (NWC), the SPC and the Norman WFO came to be located side-by-side, while NSSL researchers were distributed in various locations within the NWC. Between the WFO and the SPC, a collaboration facility (physical space with access to data) - dubbed the Hazardous Weather Testbed (HWT) - was created wherein various experiments could be conducted in direct proximity to both NOAA operational forecasting agencies here in Norman.

Thanks to the creative visions of Steve Weiss (SPC) and Jack Kain (NSSL), among others, the HWT has been used in an annual "spring program" (Kain et al. 2003 - available here or here) aimed at exploring the utility of numerical models in the context of SPC forecasting operations. With time, this annual exercise has become "the place" for researchers to come and either test out their particular models and or participate in a meaningful interaction with other scientists and SPC forecasters. Friendships and collaborations have grown out of this and it's become so popular among external scientists and forecasters that participation has to be limited - get your foot in the door early or you won't be able to participate.

I suppose I could argue that the HWT is just a particular realization of the EFC concept we developed years ago, but my general sense of this successful exercise it that it's rather far outside of the narrow bounds that I envisioned (see above). Since my research history has only recently included much work with numerical models, it never occurred to me how important it would be for forecasters and modelers to get together and share their impressions of model-based forecasts in direct comparison with observations. Recent conversations with some of the participants this spring has made me realize what this has accomplished in a way I never imagined. Perhaps it's too late to teach this old dog many new tricks, but I think I still recognize a good thing when I see it.

The HWT is also being used to give a number of other experimental ideas coming out of NSSL (and elsewhere within the Norman weather community) some operationally-oriented tests. Although the Norman WFO has something of "closed-door policy" about non-NWS people coming into the WFO during operations (This policy has a long history, whatever I might feel about it.), even they feel free to come in and talk about things going on within the HWT. The whole program is being done with little or no funding resources, other than what can be contributed by the SPC staff and the researchers themselves. People are paying their own travel just to come to Norman for the spring program - that speaks volumes to me. I'm coming to realize that if you can create any opportunity, however modest, for people of like mind to work together, they'll figure out a way to make it happen even without much in the way of support. It's rather akin to the "Build it and they will come!" concept from the movie Field of Dreams.

As with any successful program, it's impossible to predict what the future holds. As the personalities change and the external world moves on, circumstances can change and the whole enterprise might fold up and die, or it may prosper even beyond its current level of success. Nevertheless, the principle endures that when individuals want something to happen and they're willing to do whatever it takes to make it happen, then good things can be the result. This is what being a professional is all about. You're not putting out this level of effort because you expect to paid in cash for your time. You're not going the extra mile because it's in your performance plan and your boss is pushing you that way. You're doing it because you care about it and want desperately for it to succeed and bear fruit, not just for you but for the benefit of everyone. You're doing it because you have to and the chance exists, however small, that you can succeed. I'm proud to know such people and I'm very pleased even to be only cheering from the sidelines.