Defending the G.R. NWSJuly 9th, 2014 at 3:00 am by Bill Steffen under Bill's Blog, Weather
First, I’m sure in hindsight that the G.R. NWS wishes they would have issued the perfect tornado warning well in advance of the Kent Co. tornado Sunday evening and that the sirens had sounded (though I wonder at 10:30 on a Sunday night, when it’s dark and pretty much everyone is inside with the rain and storm outside, that sounding the siren would have prompted (mostly males) to go outside to see what was going on, where they would have been in far more danger than inside their homes).
Second, people have to understand that we’re not going to see every tornado in time to get out the perfect warning (or sometimes any warning). That’s the reality of the technology that we have. Not every surgery is successful today…once in a while your car may be recalled…once in a while your cell call gets dropped. The level of radar technology is certainly better than it was 40 years ago when I started in this business. Back then the station had an old RCA-AVQ-10 radar that showed only white reflectivity showing up on a black background. It was hard to pick up snow (especially fluffy snow) in the winter. Today’s doppler radars are much better, but far from perfect.
If you look at my past blog entries, I started talking about Sunday night storms on July 2. I hit the probably of Sunday evening/night storms pretty hard on July 4 and 5 on TV. Beginning on the Day 3 outlook, the Storm Prediction Center had a Slight Risk Area for S. Wisconsin and N. Illinois and as we approached Sunday, the Risk Area was brought into West Michigan and finally covered pretty much all but the SE corner of Lower Michigan. People looking at weather updates should have known that we had a good chance of thunderstorms Sunday evening/night and that there was the possibility of a severe storm. SPC’s discussions mentioned the possibility of an isolated tornado. SPC didn’t issue a watch here, thinking that severe weather would be too isolated to warrant a watch. Look at the Severe Weather Reports for Sunday, July 6. Thunderstorms formed in Iowa and produced several relatively small and short-lived tornadoes (some at the list are duplicate sightings of the same tornado). As storms developed/moved east into Wisconsin and Northern Illinois, there was only the report of “small branches” down at Harvard, IL., hardly worth a large watch and perhaps not even a county warning, since the wind might very well have been less than 58 mph (warning threshold) to down “small branches”. The GRR National Weather Service issued a severe t-storm warning for S. Ottawa Co. That verified with the one tree down on US 31 at Port Sheldon Rd. However, that’s all that was reported, so the NWS opted not to extend the warning into Kent Co. That was a very reasonable call. The four radar images here are base velocity images (click on each image to enlarge). The first is from 10:18 pm, just a couple minutes before the tornado touched down. The radar is at the airport in the upper right. The green color indicates air moving toward the radar and the red color air that is moving away from the radar. The brighter green color indicates stronger wind. In that first image, you can see the radar picking up some stronger wind gusts along the Kent/Ottawa Co. border. However, you don’t see a tornado and based on just that image, you wouldn’t issue a tornado warning….at most you’d go severe t-storm warning. Considering nothing much was reported in Ottawa Co., it’s a very reasonable decision to let it ride. Now 4-5 minutes later (10:23 pm) you get the next scan and now you see stronger incoming winds and the red on top of the bright green. Here the tornado is on the ground and crossing US 131. This is only one scan, one frame and you’ve at this point had no reports of damage. Do you pull the trigger at this point, or do you wait for another image? A severe t-storm (not tornado) warning was issued at 10:30 pm, about the time the tornado ended. Had it been daylight, someone might have seen the funnel form and alert the National Weather Service (spotters are very important to alerting the public of tornado formation). This tornado lasted about 10 minutes. You can see the tornado in the 3rd image on 52nd St., but it was pretty much done in image #4. By the time damage reports came in, the tornado had dissipated around 28th at Breton. So, even if a tornado warning was initiated when the tornado first touched down, it would be difficult to warn people in the path of the storm before it was already on top of them.
Michigan gets an average of 16 tornadoes per year and most of them are small and short-lived. The last F4 tornado in Michigan was in 1977. The last EF1 in Kent County was in 2001, over 13 years ago. Tornadoes are rare here. Most are relatively small and short-lived. Sunday night we had general showers and storms over much of the area. The National Weather Service has a lot of territory to watch and they had a lot of storms to track at the same time. This isn’t the Plains States where you might have a classic supercell and you can watch the rotation develop before the tornado touches down. Tornado forecasting is much better than 50 years ago and you can see that in the lower death tolls today compared to the past. People are warned more often before tornadoes and with better lead time than in the past. Most people know what to do in a tornado situation. However, these small, brief, isolated spin-ups are extremely difficult to forecast before they are on the ground. We had people on TV saying they ran for the basement. They did the right thing and we had no fatalities. You can’t just throw tornado warnings out for every thunderstorm. People would start ignoring them if the false-alarm rate is too high. The Storm Prediction Center has an excellent and much improved record, with a very high percentage of tornadoes (especially strong tornadoes EF2 and above) in Tornado Watch boxes. The National Weather Service in general is much better than it used to be. Our local NWS in Grand Rapids is staffed by excellent meteorologists. You’ll be convinced of that if you follow their forecast discussions. I’m sure they will study this storm, study each radar image to see if there was some clue that might have indicated that a tornado was developing. But maybe there wasn’t. The radar image on the left is approximately 2 minutes before the first touchdown. There’s certainly nothing there that would confirm a tornado. There will be new improvements made in the future, but you have to realize that the present level of technology is such that there will on rare occasion be that unexpected tornado or that surprise snowfall.