Last week, I reported on the National Weather Service’s terrible miss of an obviously strong tornado that did damage in the Whitman, Nebraska area.
In this report, I am going to provide additional information pertaining to that event and tell the story of a fatal NWS tornado warning miss near Detroit last month.
Whitman, Nebraska, Tornado Warning Miss
You can read my initial report here. Since then, I have learned the issue was neither the nationwide outage of some National Weather Service (NWS) data that was in progress nor was there a problem with any system at the North Platte NWS Office. There appears to be an issue regarding the ability of some NWS meteorologists to properly evaluate radar. As we have reported the last few years, that is not a unique issue.
Unfortunately, the National Weather Service, as usual, has gone into defensive mode and has not issued a statement about this fiasco even though some of us expected one. The NWS simply will not provide objective information when a failure occurs. It seems to believe it is not accountable to anyone or any organization, unfortunately.
 |
| Livonia Tornado Path Map |
Livonia, Michigan, Fatal Tornado Warning Miss
A tornado struck the Detroit suburb of Livonia -- without warning -- on June 5 at 3:30pm. Unfortunately, a child was killed by a falling tree. Residents were quoted as saying:
Three members of Congress, including Representative Debbie Dingell, have asked the National Weather Service for a response, which the NWS has not yet provided, even though the tornado was nearly a month ago. Below is what the local NWS has said, so far:
The NWS is wrong that a warning can't have been issued. I forcefully disagree with the explanation pertaining to the radar.... as you will see below.
Here is my analysis of the Doppler wind radar. It was possible for a warning to have been issued.
The radar, from the get-go, was being operated in the wrong mode. During thunderstorms, it should be operated on 80 to 180 second mode -- always! Unfortunately, it was being operated on 240 second mode which is not able to capture the rapid changes that can occur with tornadoes and downbursts (the nearby Detroit Terminal Doppler Radar, which is the primary radar for downbursts, went down ten minutes before the tornado occurred).
Here is the wind data from the NWS Doppler. The tornado began at 3:30pm. At 3:25 (not shown), the first indications that rotation might form were visible. If it wasn't already, the radar should have been set on faster scanning at that time.
At 3:29 (below), there was rotation (circled) but it was not concentrated. However, if a meteorologist wants to have an advance warning, this was the time. When we see rotation increasing, that is the best way to provide advance notice.
At 3:33, it is clear there is a tornado in progress by the red/green touching and increasing in indicated wind speed. Because meteorologists have no way of knowing how long a tornado might last, there is no doubt a tornado warning should have been issued at this time, if one wasn't out already.
The red/green are still touching at 3:37, but it was a less organized presentation than above. I would interpret this as the tornado weakening.
At 3:41, there is still rotation but it is broad. So it is unlikely a tornado is still occurring. The next data interval, 3:45, shows no significant rotation in the area.
If the NWS radar had been operated with more rapid updates or if the Terminal Doppler Weather Radar -- which scans at 1 minute intervals -- had been operative, it may have been possible to get out a warning before the death occurred.
It is unclear to me why National Weather Service offices are having so many issues with radar operations and interpretations during thunderstorm and tornado situations. I have heard widely varying accounts of the quality of the Service's radar training.
Representative Dingell and the other members of Congress are right to be worried about the direction of the weather service's tornado warning program. Even the NWS now admits that tornado warnings are less accurate than a dozen years ago and that lead-time (the interval between the issuance of the warning and the start of the tornado) is down by about 40%!
I call for urgent congressional investigations and hearings into why the National Weather Service's tornado warning program is crumbling. As you know, I believe there is an urgent need for a National Disaster Review Board. This would be the opportunity for Congress and the Administration to study the issue.
It has been over thirty years since the Doppler radars were first implemented across the NWS. Many of the meteorologists who started in the 1990s are like me, retired now for over five years with many retiring every year since. So it makes sense that the experience level has lowered across the NWS. I would think for the two local offices you cite above a good postmortem should at least should be performed.
ReplyDeleteYour suggestion for an independent review board should be seriously considered. I would think the storm surveys led by the local Warning Coordination Meteorologist should suffice for most situations. However, in a situation where deadly tornadic or other severe/flooding situations occur, I agree that an independent review board should be used. Of course, if Congress gets involved, then things may get going into that direction sooner than later.
I would rather have more frequent tornado sirens or sirens in general for a bad thunderstorm and err on the side of caution then not.
ReplyDeleteThe false alarm rating over 80 percent for tornadoes and sirens routinely announcing showers and 45 mph wind gusts in my mid sized Midwest city are a direct result of the NWS reacting to comments like “anonymous” above. This “better safe than sorry” approach will lead to more Joplins. We can do better than this. Keep up the good work Mike.
ReplyDeleteIt is partially due to comments like the above that my mid sized Midwest city has a long term tornado false alarm rate above 80 percent and outdoor sirens that announce rain showers and winds no higher than 45 mph on a somewhat regular basis. The NWS has taken this "better safe than sorry" philosophy to heart, especially over the past several years.
ReplyDeleteWhile training and warning operations procedures may be lacking, the focus on "decision support" and "partner outreach" enacted over the past decade may be a larger factor taking meteorologists' resources away from performing real science and improving warning services.
Another Joplin type event may be unavoidable without some type of intervention. Keep up the good work Mike.
-Retired NWS Meteorologist
I’m just a layman, so please take your dose of salt…. I’ve read that there have been budget cutbacks for NOAA, is that true? I’ve seen corporate situations where cutbacks lead to new or green resources that aren’t backstopped by more experience make bad decisions based partly on fear of reprisals. Setting these folks up to fail virtually guarantees it. Does NWS have resource constraints out in the field?
ReplyDeleteThank you for your comment, Layman. NOAA has not had budget cuts. However, the current management of NOAA has near zero interest in weather (they are interested in climate change). The current management of the NWS are talented people trying to do the best they can but they are hamstrung in a number of ways. I disagree with a number of NWS priorities but the total dollars seem to be adequate.
DeleteReplying to myself. An EF1 went through Central Michigan over the weekend. It touched down at 4:57 EDT but was not warned until 10 minutes later:
ReplyDeletehttps://kamala.cod.edu/mi/latest.nous43.KGRR.html