Sunday, March 04, 2012

Back From The Severe Weather Workshop


As many of you know, I spent the past few days in Norman, Oklahoma, at the National Severe Weather Workshop.  Several topics were discussed including the April 27th, 2011 outbreak, the May Oklahoma outbreak that followed, and the EF-5 Joplin tornado.  They even threw a little dual pol radar training which I enjoyed.

The thing that struck me the most was the National Weather Service assessment report after the Joplin tornado.  I have posted a very interesting portion of that below.  If you have the time, I highly recommend reading it!

When I read through it, the numerous mentions of tornado sirens brought up a point I wanted to share with you.  Did you know the purpose for tornado sirens is to warn those outdoors NOT indoors and are certainly not intended to be your primary source for severe weather warnings?  

I talked with John Robinson at our local National Weather Service office in North Little Rock.  He told me the only role his office has in sounding the sirens is in an advisory role to local officials in Pulaski county.  All other counties throughout the state rely on local and county government officials to activate the sirens.

Think about it.  How many of you are heavy sleepers and don't even hear thunderstorms roll in at night?  How in the world will a siren wake you up?  What if you live far enough away from a siren not to hear it inside your house?

In light of what happened in Joplin and the recent outbreak across the Ohio and Tennessee River Valleys,  I think this is a very important point I want to make to all of you.  YOU MUST HAVE A RELIABLE SOURCE FOR WARNINGS AND YOU MUST TAKE IMMEDIATE AND APPROPRIATE ACTION WHEN THEY ARE ISSUED.

I highly recommend buying a programmable weather radio and/or signing up for WeatherCall 7.  Did you know WeatherCall 7 works throughout the entire country?   If you have friends or family out of state, you can sign them up as well.  This service is designed to your specific address whether it's in Arkansas, Indiana, Kansas, Iowa, etc.  They will receive a phone call just like you when severe weather approaches.  The only thing that might be strange is that they will hear the voice of Meteorologist Ned Perme.  But who across this county wouldn't want to hear from Ned whether they know him or not? LOL!  Regardless, they are still getting the warning.

Again, below is a portion of the assessment from the Joplin tornado.  You should read it!


Finding #2b: The majority of surveyed Joplin residents did not immediately go to shelter upon hearing the initial warning, whether from local warning sirens, television, NWR, or other sources. Instead, most chose to further clarify and assess their risk by waiting for, actively seeking, and filtering additional information.
In order to gain a sense for how social models of understanding influenced perceptions of risk and warning response, particular attention was paid to the ―worldviews‖ held by residents. Worldview6 is defined as an overall perspective of how people interpret their environment or the world around them – and is generally informed by things like one‘s experience, education, and cultural values.
Interviews showed aspects of worldview that influenced risk perception and warning response included: previous experiences with tornadoes, apathy, familiarity with seasonal weather patterns in southwest Missouri, optimism bias, perceived frequency of siren activation in Joplin, social networks as mechanisms for warning dissemination, avid fear of tornadoes, and the number of deadly tornadoes earlier in the year.
Previous experiences with tornadoes were shown to have an influence in the way residents perceived their risk and responded to the warnings. As one resident indicated, the tornado he experienced prepared him mentally for appropriate response action during this event. Another resident commented that time spent in Oklahoma City made him complacent to the possibility of a tornado in Joplin.
Similarly, familiarity with seasonal weather in southwest Missouri played a major role in risk perception and warning response. Most individuals commented that severe weather in southwest Missouri during spring is common; however, tornadoes never affect Joplin or themselves
6 Roncoli, Carla, Keith Ingram, Christine Jost and Paul Kirshen (2003) ―Meteorological Meanings: Farmer‘s interpretation of seasonal rainfall forecasts in Burkina, Faso.‖ In Weather, climate, culture. S. Strauss and B. Orlove (eds). Oxford; New York: Berghahn Books. Pp. 181-200.
personally. It was common in the interviews to hear residents refer to ―storms always blowing over and missing Joplin,‖ or that there seemed like there was a ―protective bubble‖ around Joplin, or ―there is rotation all the time, but never in Joplin.‖. One city employee stated, ―... don‘t think it can‘t happen in your community, because that‘s what I thought.‖ This sense in which people believe their personal risk from a hazard is less than the risk faced by others is referred to as optimism bias and can lead to diminished perceptions of threat and influence response.
Although not as common, social networks as mechanisms for warning dissemination were found to generally amplify perceptions of risk and lead to warning response. For example, one woman reported eating dinner with family, receiving a text message about the tornado, and then receiving a phone call shortly after informing her of a storm travelling through Joplin. This heightened the woman‘s belief that a threat existed and prompted her and her family members to take shelter at the restaurant.
Similar to seasonal weather patterns, the perceived frequency of siren activation (false alarms) led a large number of participants to become desensitized or complacent to this method of warning. Many noted that they ―hear sirens all the time[sirens] go off for dark clouds,‖ they are ―bombarded with [sirens] so often that we don‘t pay attention,‖ ―the ―sirens are sounded even for thunderstorms,‖ and ―all sirens mean is there is a little more water in the gutter.‖
The diminished severity or absence of a threat (complacency) held by Joplin residents can be understood as resulting from their normalization of the threat. Normalized responses toward severe impacts are likely to occur in groups frequently exposed to hazardous weather7. The same could be said for residents in Joplin who, based on their perceived frequent exposure to local warning systems (and NWS warnings) during spring, normalized their reactions to the activation of the first siren and characterized it as just another aspect of springtime in Joplin.
Finding #2c: Familiarity with severe weather and the perceived frequency of siren activation not only reflect normalization of threat and/or desensitization to sirens and warnings, but they also establish that initial siren activation has lost a degree of credibility for many residents. Credibility is considered to be one of the most valued characteristics for effective risk communication8.
7 Bankoff, Greg (2007) ̳Living with risk; coping with disasters: hazard as frequent life experience in the Philippines‘. Education about Asia. 12(2): 26-29
8 Trumbo, Craig W., and Katherine A. McComas (2003) ―The functionality of credibility in information processing for risk perception.‖ Risk Analysis 23(2): 343-353.
It should be noted that stakeholders in the warning process, such as media or Emergency Managers, were less likely to think over-warning or desensitization to NWS warnings was an issue. During interviews, Emergency Managers in particular felt the frequency of warnings was appropriate, while media staff were split with some saying ―most warnings were ̳cry wolf,‘‖, while others emphasized the importance of advance warning for all tornadoes regardless of false alarms. In general, these groups have a sense of obligation for the safety of their constituents which influences their worldview of the warning process and risk perception.
Though risk signals tended to elevate awareness, there were certain signals that stood out, added important credibility to the warning, and acted as triggers in prompting a belief in the threat and taking protective action.
Finding #2d: The majority of surveyed Joplin residents did not take protective action until receiving and processing credible confirmation of the threat and its magnitude from a non- routine trigger.
While searching for additional information concerning the severe weather threat constitutes ―taking an action,‖ the actions many residents described taking were not the immediate life- saving measures desired with the issuance of a tornado warning. In most cases, these life-saving actions, or the decision to find shelter, were associated with additional extraordinary risk signals. This was generally achieved in different ways, including:
  1. Physical observation of the environment (seeing the tornado approach).
    While significant numbers of people actually did this, the approach was complicated by having a ―rain-wrapped‖ tornado that made the tornado more difficult to recognize until it was very close. There were numerous accounts of people running to shelter in their homes just as the tornado struck, despite significant advance warning of the risk.
  2. Seeing or hearing confirmation of the threat on radio or television, seeing the large tornado on the air, or hearing on-air instructions to ―take cover now.‖
    When the tornado began moving into Joplin, most local electronic media switched to ―wall-to-wall‖ coverage of the event, which included live video from tower-cams. As coverage quickly evolved, and the magnitude of the event became clear, on-air commentators implored those in the path to take cover immediately. This kind of media coverage helped convey the seriousness and urgency of the situation, and prompted many listeners and viewers to find shelter.
  3. Hearing a second, non-routine, siren alert at approximately 538 pm CDT. 8
It is the Joplin emergency management policy to sound the sirens onetime for a severe weather alert. Because of the length of time that had elapsed since the initial siren alert, and as reports came into central dispatch of a tornado moving into Joplin, the Emergency Manager made the decision to activate the local warning sirens a second time. This second siren activation came about 20 minutes after WFO Springfield issued the Tornado Warning for southwestern Jasper County, including Joplin (issued at 517 pm CDT). A large number of those interviewed noted that this non-routine second siren alert raised their level of awareness, confirmed the alert, indicated the seriousness of the warning, and prompted them to get to the best available shelter.
It is unclear how many of those killed in the tornado failed to take shelter, or if a change in response time and behavior would have impacted survival rates. In the case of the Joplin tornado, it was somewhat fortunate that the tornado was moving at a relatively slow forward speed (~20 mph), and the initial siren alert occurred more than 20 minutes before the tornado struck the city.
Lastly, several of the people interviewed indicated a desire for different levels of warning (applied to local siren policies) as a means to clarify the seriousness/magnitude of the threat. Specifically, these comments spoke to some desired differentiation in warnings and siren tones between life-threatening emergencies and threats to property. These persons noted, ―maybe there should be two levels of warning... a regular warning and a panic button warning for when it will be really bad,‖ ―I wonder if there shouldn‘t be different types of sirens for different types of warnings,‖ and another noted that there is a difference between a warned big event and a warned small tornado or funnel cloud.
After processing a variable number of risk signals and reaching a decision to act, the majority of surveyed Joplin residents took shelter in the most appropriate location available to them. This included basements, interior rooms or hallways, or crawl spaces. This suggests campaigns to promote severe weather safety practices are effective.
Even if this action was taken in the last available secondsken in the last available seconds, in many cases it was a life-saving measure. Unfortunately, due to a number of factors, below ground shelters (basements) are not common in the Joplin area, and some people likely still found themselves in situations that were not survivable. It is unclear to what degree this contributed to the tornado mortality in Joplin. Preliminary analysis done by the Joplin Globe newspaper revealed that most fatalities occurred in residences (54%), followed by non-residential buildings (32%), and in vehicles or outdoors (14%). This type of analysis is beyond the scope of the Service Assessment and is being addressed in separate studies by both the National Institutes of Standards and Technology and Federal Emergency Management Agency.
The majority of businesses interviewed had a plan for receiving warnings and sheltering patrons. While lives were lost in these non-residential buildings, the toll certainly would have been much higher if not for the action plans and employees in these businesses.
In addition, a significant number of fatalities in Joplin occurred in vulnerable populations such as the elderly, infirm, or disabled. These populations typically require additional time and/or assistance to get to the best available shelter.
Best Practice #1: NWS outreach and severe weather safety education programs should continue to emphasize and assist area businesses with severe weather safety action plans via the StormReady program or other similar mechanisms. This kind of outreach and planning assistance should also be extended to vulnerable populations in nursing homes, group homes, hospitals, etc.
Responding to warnings is not a simple act of stimulus-response, rather it is a non-linear, multi- step, complex process. Relationships between false alarms, public complacency, and warning credibility are highly complex as well. While residents of Joplin addressed these in terms of local warning siren systems, they also relate directly to the content and skill of NWS warnings and the weather enterprise as a whole. As a rudimentary evaluation of NWS warning skill, severe weather verification statistics were compiled for the period from 10/1/2007 to 4/1/2011. These are listed in Table 2.
Table 2: NWS Severe Weather Warning verification statistics from 10/1/2007–4/1/2011
Finding #2e: Nationally, 76% of all NWS Tornado Warnings, in their totality, are false alarms. This means 24% of all tornado warnings are eventually associated with an observed tornado indicating limited skill in differentiating between tornadic and non-tornadic events; however, 68% of EF0-1 tornadoes receive advance warning of near 12 minutes, while 94% of EF3-5 tornadoes receive advance warning of near 18 minutes, indicating an ability to better detect strong/violent tornadoes.9 Just over half (54%) of all severe weather warnings coincide with a
9 Because NWS warnings do not differentiate between weak and strong/violent tornado warnings, a calculation of false alarm rate for strong/violent tornadoes is not possible.
Probability of Detection
False Alarm Rate
Initial Lead Time
All Tornado
12.5 minutes
All Severe
18.6 minutes
EF0-1 Tornado
11.9 minutes
EF2-5 Tornado
16.3 minutes
EF3-5 Tornado
17.8 minutes
severe weather event, indicating moderate skill in distinguishing between severe and non-severe thunderstorms.
While there are no guarantees that simply decreasing false alarms will significantly impact warning response behavior, the results of the Joplin residents interviews appear to indicate a relationship between perceived false alarms, degree of warning credibility, and complacency in warning response. Nonetheless, as indicated by the report findings, there are a number of ways NWS can explore to improve effective decision making within the warning response process.
Recommendation #2: To improve severe weather warning response and mitigate user complacency, the NWS should explore evolving the warning system to better support effective decision making. This evolution should utilize a simple, impact-based, tiered information structure that promotes warning credibility and empowers individuals to quickly make appropriate decisions in the face of adverse conditions. This structure should:
  1. a)  lessen the number of risk signals processed before protective action is taken (finding 2b)
  2. b)  provide a non-routine warning mechanism that prompts people to take immediate life-saving
    action in extreme events like strong and violent tornadoes (finding 2d).
  3. c)  be impact-based more than phenomenon-based for clarity on risk assessment (finding 2a)
  4. d)  be compatible with NWS technological, scientific, and operational capabilities (finding 2e)
  5. e)  be compatible with external local warning systems and emerging mobile communications
    technology (finding 2a)
  6. f)  be easily understood and calibrated by the public to facilitate decision making (finding 2a)
  7. g)  maintain existing “probability of detection” for severe weather events (finding 2e)
  8. h)  diminish the perception of false alarms and their impacts on warning credibility and
    response (finding 2c)


Anonymous said...

My mother has a weather radio that during severe weather announces non-stop weather in various parts of the state. I find it irritating and useless. I signed her up for Weather Call. I have it at home also. I've set both up for alerts in case of tornado only, which specifically targets our address, so when we get "the call" we know it's time to take cover. The call will come in on our home and cell phones. I also have a weather radio in our safe area, but that is a last resort or for just hearing what's going on while we're hunkered down. Mom also has a cheap small battery powered radio, which one of tv stations will also broadcast on radio during severe weather. That cheap little radio which fits in the palm of your hand, broadcasts loud and clear to let you know what's happening. Thankfully, our town only sounds the sirens when there is actually a tornado in our area. Thank you for your great work in trying to keep us safe!

Greg Reddin said...

I'll preface my comment by saying I'm a huge supporter of the NWS and I take their work very seriously. But I see something in that table that has really concerned me for a long time. It basically indicates that we don't have a reliable way to look inside a thunderstorm and detect a tornado. They've managed to warn 70% of tornadoes by warning anything that might be a tornado. So 76% of the tornado warnings are false alarms. I wonder if there should be three levels of warning: severe thunderstorm, high probability of tornado, and tornado confirmed.

I understand that, once the third level is reached, there very likely may not be ample time for people to act accordingly, so maybe that's not a good enough answer and maybe there's not a good answer. But a 76% false alarm rate? And we wonder why people don't take warnings seriously?