As part of the GOES-R proving ground, several WFO’s have been evaluating a suite of products from the University of Wisconsin/CIMMS satellite meteorology group. While the host of products includes functionality to access cloud top cooling, overshooting tops, turbulence probability, lightning probability, thermal couplet detection and magnitude and severe weather probability, the main focus has been on the Convective Initiation (CI) product. While the algorithm is fairly complex, the basic idea is to use brightness temperatures to calculate cloud top cooling rates. Considering these cooling rates along with microphysical phase change data, the areas most likely to see convective initiation are highlighted by the algorithm and displayed to the forecaster. A three tiered system is currently in place in which pre-CI cloud growth areas are highlighted in green, areas where CI is likely to occur in yellow and areas where CI is currently ongoing in red.
While the product is designed to recognize CI before significant radar returns develop, it does have some drawbacks in that it will not work well in a cirrus dominated regime. In addition, because of a box-average approach to measuring brightness temperature, fast moving clouds will tend to cause false alarms. Preliminary statistics show that the product detects over half of CI events and false alarm about once every 4 times. At its best, the CI product has been shown to indicate CI as much as 37 minutes before radar returns reach 35 dBz.
Several WFO’s around Eastern Region have been involved in actively evaluating the CI product performance including Greenville-Spartanburg, Raleigh, Charleston SC, Cleveland, Charleston WV, and Binghamton as well as Eastern Region Headquarters. After compiling comments from several of the offices mentioned, the reviews thus far of the CI product are mixed but there is good reason for that. By far, the biggest drawback seems to be that the CI product usually does not provide any significant increase in lead time when used in conjunction with the current GOES-13 infrared satellite data due to the fact that imagery usually only comes into AWIPS every half hour or so. The product was not meant to be used with GOES-13, however, it was meant to be used with the new GOES-R satellite which is slated to first launch in 2015. GOES-R will include an Advanced Baseline Imager (ABI) which will provide 10 infrared channels at a temporal resolution more than five times faster than GOES-13. With this system in place, CI imagery will no doubt prove much more effective in terms of lead time. Feedback from the offices to this point has been fairly consistent.
Here at WFO Raleigh, we have noticed that the product tends to work better in pulse storm situations rather than those with a lot of forcing involved. This probably has to do with the box-average approach and storms propagating outside of the pre-determined area needed to properly evaluate brightness temperatures. Despite this, the potential usefulness of this product should not be overlooked. As operational forecasters, it is easy to discount the usefulness of a tool in its developing stages because there is so much other data to look at. Once fully operational with the GOES-R platform, the CI product could be a useful situational awareness tool. It could be used as a guide for radar operators on where they may need to allocate their resources in the near future, especially for discrete cells that may pop up away from more active areas. In addition, there have been a couple of times this year when we were not sure whether or not radar data would be available for a particular event. The CI product would be extremely useful when using satellite data and adjacent radars to interrogate storm cells. WFO Raleigh has started archiving CI imagery on a daily basis in order to further evaluate the CI product and identify situations where it excels. Below is one example from 1 June 2011 in the southern Appalachians where convective initiation was detected well in advance of the first lightning strikes.
Sieglaff, J.M., L.M. Cronce, W.F. Feltz, K.M. Bedka, M.J. Pavolonis, and A. Heidinger 2010: Nowcasting convective storm initiation using box-averaged cloud-top cooling and microphysical phase trends. Submitted to J. Appl. Meteor. Climatol. January, 2010.