On Saturday, August 24th, small convective cells developed across parts of northern Alabama and southern Tennessee during the late afternoon and evening hours. The cells did offer the chance to make some evaluation of the UAH GOES-R CI product. In image 1 below, notice that a small, but intense cell was located in SW Lincoln County, TN. Other smaller cells were located in NW Madison County, AL and in NE and central Limestone County, AL. Other cells can also be seen in the far northern part of this particular image, in Middle Tennessee. CI values were 40-50% downstream (to the west, in this case) in the Lincoln/Giles/Limestone/Madison cells. Meanwhile, CI values were around 50-60% near the very small, light cell in south central Limestone County. Unfortunately, a lot of “clutter” also existed in the imagery from this event. So, some of the smaller, lighter cells may be more difficult to detect.
In the next available CI product update, at approx 2132 UTC, CI values had increased downstream of the convection in Lincoln/Giles/Limestone Counties. Notice the 60-70% values there now, indicating a stronger likelihood of convective initiation of clouds that had developed along outflow from the cell in Lincoln County, TN. Farther to the south, CI values of 50-60% and 60-70% had appeared along the border of Madison/Limestone Counties, where no cells were yet readily apparent.
A little later at 2145 UTC, cells can now be seen developing in the area of previous high CI probability, in eastern Limestone County. The cell in SE Giles and SW Lincoln Counties, in the same area that also exhibited high CI values in the previous image has also developed further, with over 50 dBZ reflectivities indicated by KHTX. Also, notice a cell in far NW Jackson County that developed during the previous 15 minutes. However, CI values were not present for this cell, and I was perplexed by this apparent miss initially. Or was there really a “miss” by the algorithm?…we’ll get to that in a minute.
In the latest image I have available from that event, CI values remained high with the cells mainly in Giles and Limestone Counties. Meanwhile, the cell in far western Jackson County was moving into eastern Madison County. However, notice the faint yellow colors that appear in eastern Madison County, to the west of that cell. Since the radar data were loaded last, and overlaid the CI product, the CI values were being covered by the large amount of clutter in the radar. Initially, a cell that I thought had been missed by the CI algorithm, may not have been.
Below is a zoomed image from the area in question above. Notice the few yellow colors from the CI product in far western Jackson County, mostly hidden by the overlaid KHTX radar data and clutter, indicating 60-70% probability for convective initiation.
This next image is a zoom of image 4, which better shows some of the underlying CI values (here, yellow colors indicating 60-70% probabilities of convective initiation) in eastern Madison County at 2202 UTC.
In this case, the UAH GOES-R CI values performed well overall in predicting the likelihood for convective initiation. Cells developed in locations where CI values were typically 50% or greater. In the case of the cell that moved from Lincoln County into Giles County however…I’m not sure this was more a case of cell propagation than it was actual cell development from storm-scale outflow.
Another take-away item from this event though and a lesson-learned, is the appropriate viewing strategy in AWIPS II. AWIPS II allows for the loading of multiple data sets and imagery. Thus, one has to be careful as imagery that are initially loaded may be covered by following imagery. True, I did have another pane that combined the CI with GOES-East Vis imagery, and then IR imagery once low light conditions made that necessary. I also had a pane with CI imagery alone. But, my focus had switched to the CI and radar data at this point, since cells were developing quickly. However, when incorporating radar with CI imagery, it is probably best to overlay the radar imagery with CI values, when trying to make a proper assessment of CI.