Attending the EWP was an excellent experience, and I appreciate all the time and effort that people put into it. The EWP kept the attendees very busy with a series of relatively intense exercises. One’s skill at issuing warnings is much enhanced by issuing so many warnings in such a short period of time. I wish my forecasters had the opportunity to go through these exercises as well. In fact, one of the things I carried away from the EWP was the power of hands-on training of this kind. We try to use WES cases in the field for training purposes to get something like this effect, but the experience in Norman is superior to what we can do with the WES.
Both the CASA radars and the PAR were found to be valuable for their rapid update capabilities. We were able to get routine 1-minute volume scans from both. In several instances, warnings were issued several minutes earlier than possible with 88D radars. Also, from some cases, the high-resolution of the CASA radars helped identify severe features that might have gone unnoticed in 88D data. If anything like a national network of CASA radars is ever developed, we will need to decide if we want to warn for every little vortex these radars are able to detect. CASA radars do suffer from beam attenuation problems, though. The original concept was for CASA to use phased-array antennas, but this has not been achieved as yet. Also, the evaluation of a larger CASA array is probably needed and, I’m told, is planned.
On the CASA radars, there is a sense in the field that CASA radars are primarily valuable as potential gap-filling radars. However, this is not the original intent of the CASA program, and, in fact, gap filling radars have been available for decades from a number of vendors. The ground-breaking collaborative properties of a CASA network are not widely appreciated. Still, gap-filling radars are much needed in the west, probably more so than collaborative or phased array radars. If new money is available for more radars, solving the gap problem may be more of a priority than an innovative new technology.
The LMA I found to be pretty interesting. The thousands of VHF sources detected from lightning channels are mapped into a vertically integrated lightning density product. When color-contoured, this product looks similar to a radar composite reflectivity map, with comparable resolution. Any electrically active storm can be imaged. It was also possible to look at the 3D images of the lightning channels for storms, but this was just too much information. LMA also provides 1-minute updates. What is still being learned is how to associate the severity of a storm with its lightning density history. With some experience, we were able to expedite warnings on storms based on a lightning pulse. The ability to image storms from lightning, I found to be a fascinating concept, and I found the LMA to be a valuable companion to radar when deciding whether to issue a warning. As I work in a CWA with large radar gaps, having LMA data would be particularly valuable. Even with good radar coverage, the detailed LMA data helps to fill-in the picture we get from radar, and at a small cost. The cost of a nationwide LMA capability would seem to be a small fraction that of a radar network, making the LMA attractive from a cost-benefit stand point.
The MRMS algorithms are run off of existing operational data sources. For one real-time case we warned for, the storm went right over the top of the radar, so the multi-radar approach was shown to be valuable in this case. One of the new MRMS algorithms is for hail size. We found this to be pretty good, and tended to agree well with reports. It was at least as good as the “VIL of the day” concept. The “rotation track” product was also useful. As MRMS is derived from available data, some of the products duplicate what might be easier to see another way. Storm heights, for example, might be better found by consulting a cross-section in FSI
We considered the problem of information overload in integrating new data sources into operations. For any of these technologies to succeed, they need to make it easier for a forecaster to issue a product. Having to evaluate a flood of new data every minute may be paralyzing to some. Even though dense new data sources have more information for producing more precise warnings, they need to be integrated in some user-friendly way into operations. This leads to a need for better algorithms for analyzing some of these data streams in the background.
There are no set plans that I am aware of currently to expand CASA, PAR, LMA, or MRMS nationally. Of these, MRMS would be the easiest to implement as it only requires the fielding of algorithms. The LMA may also be cost-effective. CASA and PAR are both expensive technologies, each one at least as expensive as the current 88D network.
Bill Martin (NWS Glasgow MT – 2009 Week 5 Evaluator)
