JUNE 2014
BLAIR AND LEIGHTON
591
Assessing Real-Time Tornado Information Disseminated through NWS Products
SCOTT F. BLAIR AND JARED W. LEIGHTON
NOAA/NWS/Weather Forecast Office, Pleasant Hill, Missouri
(Manuscript received 9 October 2013, in final form 24 January 2014)
ABSTRACT
Real-time confirmation of a tornado specified in National Weather Service (NWS) warnings and statements
is believed to increase the credibility and urgency of these critical warning messages for the end user, because
it represents the greatest degree of certainty that the hazard exists. This timely tornado information disseminated in official NWS products and relayed through multiple sources by private and public partners may
help the public believe, personalize, confirm, and respond to the warning message. This is the first study to
explicitly assess the frequency of real-time confirmation of ongoing tornadoes within NWS products and
explore what unique conditions may facilitate or hinder this process. Tornado reports and their respective
NWS warnings and statements during a 5-yr period from 2007 to 2011 across the central contiguous United
States were compiled and examined. Overall, 40% of tornadoes were confirmed in NWS products in real time.
Increasing tornado pathlength, duration, and intensity subsequently resulted in an increasing likelihood of
real-time confirmation prior to the tornado dissipating. The time of day was a factor; nighttime tornadoes
were 20% less likely to receive real-time confirmation than daytime events. Additionally, increasing tornado
forecast risk in products issued by the Storm Prediction Center corresponded to an increasing likelihood of
real-time confirmation. Analysis of these data reveals specific scenarios when tornadoes are more or less likely
to be reported in real time, providing some guidance for when timely ground-truth information may or may
not be available.
1. Introduction
National Weather Service (NWS) tornado warnings
are one of the most essential services provided to the
general public, private sector, and the integrated warning team (IWT), which is composed of broadcast media,
emergency managers, and other decision makers in the
realm of public safety (Doswell 2005; Dix and Fieux
2011; NWS/Warning Decision Training Branch 2013).
These tornado warnings, communicated through multiple sources, in addition to improvements in technology
and public education, can be partially credited for the
overall decreasing trend in tornado-related deaths over
the past century (Brooks and Doswell 2001; Coleman
et al. 2011).
The actual human response to tornado warnings often varies due to complex demographical, behavioral,
situational, cultural, and economic factors (Mileti and
Sorensen 1990; Perry and Lindell 1991; Legates and
Biddle 1999; Sorensen 2000; Sherman-Morris 2010;
Corresponding author address: Scott F. Blair, NOAA/NWS/
Weather Forecast Office, 1803 N. 7 Hwy., Pleasant Hill, MO 64080.
E-mail: [email protected]
DOI: 10.1175/WAF-D-13-00126.1
Brotzge and Donner 2013; Sherman-Morris 2013). Mileti
and Sorensen (1990) proposed a model in which a user’s
response to a warning progressed through several steps:
hearing the warning, understanding the contents of the
message, believing the warning is credible, personalizing
the warning, and confirming the warning is true, before
a physical sheltering response occurs. When a tornado
warning is issued, the public frequently seeks out additional information beyond the initial warning before
taking protective action in order to believe, personalize,
and confirm the message (Mileti and Sorensen 1990;
Hayden et al. 2007; NWS 2009, 2011a,b; Lindell and
Perry 2012; Sherman-Morris and Brown 2012). Several
studies have shown that a critical component to help
elicit a response is looking for or hearing confirmation of
the threat within the warnings themselves, perhaps because it increases the perceived credibility and urgency
of the warning message (Hodler 1982; Tiefenbacher
et al. 2001; Andra et al. 2002; Hammer and Schmidlin
2002; NWS 2009; Schumacher et al. 2010; NWS 2011a;
Nagele and Trainor 2012).
Ground-truth reports of tornadoes may be received
from storm spotters, the public, or other sources that in
turn can be disseminated in the initial tornado warning,
592
WEATHER AND FORECASTING
subsequent warning statements, and local storm reports.
Real-time information included in these NWS products
serves as an integral component of the warning and
sheltering process. In turn, these messages can be used
by broadcast media, decision makers, and other users to
further clarify the track and location of the hazard as
well as to convey the seriousness of the situation to
a mass audience. The public frequently turns to television broadcasts as their primary source of tornado information (Legates and Biddle 1999; Brown et al. 2002;
Hammer and Schmidlin 2002; Mitchem 2003; Paul et al.
2003; Sherman-Morris 2005, 2010; Hayden et al. 2007;
Demuth et al. 2009; Coleman et al. 2011), with real-time
tornado information aiding in the public’s willingness to
believe the warning message, encouraging a stronger
public response.
This is the first study to assess the frequency of realtime confirmation of ongoing tornadoes within NWS
products and explore what unique conditions may facilitate or hinder this process, by quantifying these realtime occurrences over a 5-yr period across some of the
most tornado-prone regions of the United States. An
overview of the data and methods used is provided in
section 2, followed by the results and their applications
in section 3. Discussion and concluding remarks follow
in section 4.
2. Data and methodology
The domain for this study was confined to the central
contiguous United States, generally the states east of the
Continental Divide and west of the Mississippi River.
This region was chosen to encompass a geographic area
with the most frequent supercell storm mode and tornado activity (Brooks et al. 2003; Smith et al. 2012). The
time interval for this study was from 1 January 2007
through 31 December 2011, to closely coincide with the
current era of NWS storm-based warnings and the enhanced Fujita (EF) damage scale rating system that
began during 2007 (McDonald and Mehta 2006; NWS
2007).
a. Tornadoes
The Storm Prediction Center (SPC) tornado database
compiled from the National Climatic Data Center
(NCDC) publication Storm Data (NCDC 2007–11) was
used to examine tornado events that occurred within the
study domain. Doswell and Burgess (1988) detailed the
myriad of challenges within the climatological tornado
database inherent with F/EF-scale ratings (hereafter
EF), including those classified as ‘‘weak’’ tornadoes
(EF0 and EF1). Much uncertainty exists with EF0-rated
events, as many tornadoes occur in rural areas and
VOLUME 29
produce little or no damage, ultimately defaulting to the
lowest damage rating. Tornadoes rated EF1 and greater
are traditionally confirmed by more formal damage assessments, lending slightly better confidence in climatological studies and databases (Rose et al. 2004;
Verbout et al. 2006). Therefore, all tornadoes rated EF1
or greater were incorporated into the study.
Information about specific location, time, duration,
pathlength, and EF-scale rating were collected for every
tornado included in the study. Tornadoes were also
classified based on pathlength as short (,24 km; ,15 mi),
moderate (24–48 km; 15–30 mi), or long (.48 km; .30 mi).
In addition, tornadoes were further classified as ‘‘day’’ or
‘‘night’’ occurrences to examine whether visibility differences affect real-time confirmation. All sunrise and sunset
times were obtained for each event, and nighttime tornadoes were classified as the time period from 1 h after sunset
to 1 h prior to sunrise. Civil twilight ranges from approximately 25 to 40 min across the domain annually, yet uncertainty exists to the reduction of available brightness in
a storm environment. Therefore, the daytime–nighttime
criteria were chosen to ensure that all of the ambient light
during civil twilight was eliminated for nighttime tornado
events.
b. Real-time confirmation
A ‘‘real-time confirmation’’ of a tornado was defined
as an explicit mention of a confirmed tornado within an
NWS text product that was issued prior to the dissipation of the tornado. In cases when an NWS product
confirmed that a tornado was sighted but was issued
after the tornado dissipated, or when no mention of
a reported tornado was included in the products, the
tornado event was classified as ‘‘no real-time confirmation.’’ Figure 1 graphically illustrates this methodology,
using a hypothetical case of real-time confirmation for
a tornado-warned (and tornado producing) supercell
and its associated tornado path.
To obtain these data, all available NWS tornado
warnings, follow-up warning statements called ‘‘severe
weather statements,’’ and preliminary local storm reports associated with each tornado were acquired from
the Iowa Environmental Mesonet application (http://
mesonet.agron.iastate.edu/cow/). The beginning and ending
times of a tornado as recorded in Storm Data were
compared to the issuance time of the NWS product that
first confirmed the presence of an ongoing tornado. It
should be noted there are some limitations to the precision of specific genesis and dissipation times of tornadoes as documented in Storm Data, leading to some
uncertainty. In rare situations when multiple tornadoes
were ongoing or occurred during the duration of
a warning, and NWS products confirmed the sighting of
JUNE 2014
BLAIR AND LEIGHTON
593
FIG. 1. A hypothetical NWS tornado warning (white polygon) with a tornado path (black
heavy line) is shown for a supercell storm (cartoon radar image). Times for warning and tornado duration are listed. The shading (gray) within the warning polygon represents the period
prior to the tornado confirmed through NWS products and the period of real-time confirmation
(gray hatched).
at least one tornado, all tornadoes within the warning
were considered confirmed in real time after initial
confirmation in the product.
c. SPC products
SPC ‘‘day-1 convective outlooks’’ and their respective
‘‘tornado probabilities’’ as well as any ‘‘convective
watches’’ in effect for each tornado were recorded and
assimilated with the other dataset obtained from Storm
Data to examine whether an increasing forecast risk of
severe weather may correlate to an increasing number of
‘‘eyes’’ in the field to report severe weather. The outlook
and its tornado probability issued at least 8 h prior to
a tornado report were used in the study. This 8-h timeframe was subjectively determined to provide sufficient
lead time for emergency managers and other decision
makers to organize, contact, and eventually deploy
storm spotter networks that might have led to a potential
increase of observers in the field.
and tornado locations, are more frequent in the Great
Plains, with more complex storm modes found in the
eastern portion of the study domain (Brotzge and
Erickson 2009; Smith et al. 2012). These natural influences likely play some role in supporting or hindering
the ability of storm spotters and the public to observe
a tornado.
3. Results and applications
A total of 1362 tornadoes rated EF1 or greater were
included in the study, and approximately 40% (543) of
these tornadoes were confirmed in real time through
NWS products prior to dissipation. Geographically,
there appears to be a slight tendency for tornadoes to be
confirmed in real time more frequently across the Great
Plains states, specifically across the high plains (Fig. 2).
Possible reasons for this higher frequency may stem
from better visibility across the plains than the more
varied terrain found across portions of Missouri, Arkansas, and Louisiana, where hills, dense vegetation,
and higher low-level relative humidity impede observations (Ashley 2007; Brotzge and Erickson 2010).
Additionally, storm mode may also play a role where
supercells, with more easily identifiable storm structure
FIG. 2. Study domain with the beginning tornado locations for
real-time confirmation (543; red triangles) and without real-time
confirmation (819; black squares) from 2007 to 2011.
594
WEATHER AND FORECASTING
VOLUME 29
FIG. 3. (left) Day (1013 events) and (right) night (349 events) distributions of beginning
tornado locations for real-time confirmation cases (red triangles) and events without real-time
confirmation (black squares) from 2007 to 2011.
a. Daytime versus nighttime tornadoes
Tornadoes occurring during the nighttime hours, defined as the period from 1 h after sunset to 1 h prior to
sunrise, were confirmed in real time at a much lower rate
when compared to daytime events. Approximately 45%
of tornadoes during the day were confirmed in real time
versus only 25% at night. Using a Student’s t test, the
rate of real-time confirmation of daytime tornadoes is
statistically greater than that of nighttime tornadoes, to
the 99% confidence level. These findings suggest that
the visual and timely confirmation of tornadoes at night
is potentially hampered by several factors. One of the
biggest hindrances to nighttime confirmation is likely
the difficulty of observing tornadoes in the dark (Ashley
et al. 2008; Brotzge et al. 2011). The study reveals that
a large number of these nighttime events occurred in
areas of complex terrain and ample tree cover (Fig. 3),
resulting in additional detection challenges beyond
the lack of light [Doswell et al. 1999; Fig. 8c in Ashley
(2007)].
Nocturnal tornadoes enhance the vulnerability of the
public due to a variety of factors, not only limited to
a population asleep at night, as demonstrated by previous studies (Simmons and Sutter 2005; Ashley et al.
2008). It is possible that the lower rate of real-time
confirmation of nighttime tornadoes may also delay or
decrease the warning response and increase the vulnerability for those that receive notification at night.
Operationally, ground-truth reports of tornadoes during
nocturnal events generally may become less frequent,
and in some cases the quality less reliable due to the
observing obstacles. This is not surprising because the
lowest levels of a storm become increasingly difficult to
observe at night. These nocturnal challenges may be
indirectly inferred from the 20% reduction of real-time
confirmation between daytime and nighttime tornado
events. It is important for the operational meteorology
community to be aware of the potential diurnal changes
of timely severe weather reports, and does not interpret
a lack of reports at night to mean that the hazard is no
longer occurring. Furthermore, the other components of
the IWT should be aware that real-time confirmation of
tornadoes through NWS products may become less
frequent or absent at night, but should not construe the
shortage of reports as a lower risk when a tornado
warning is in effect.
b. Tornado characteristics
The total tornado duration was a notable discriminating factor between events that were confirmed in real
time and those that did not receive validation in NWS
products prior to tornado dissipation. Minor overlap was
found between real-time events versus no real-time
confirmation, and median values of tornado duration
indicate a large discrepancy between the two datasets,
with 14 min for the real-time occurrences as opposed to
only 5 min for the events with no real-time information
(Fig. 4). The duration of tornadoes that are confirmed in
real time is statistically longer than the duration of tornadoes that are not, to the 99% confidence level.
JUNE 2014
BLAIR AND LEIGHTON
FIG. 4. Box-and-whiskers plot of tornado duration (min) from
543 real-time confirmation cases (gray) compared to 819 cases of
no real-time confirmation (black) from 2007 to 2011. The shaded
box covers the 25th–75th percentiles, the whiskers extend to the
10th and 90th percentiles, and the median values are marked by the
heavy horizontal line within each shaded box.
Typically, there is an inherent amount of time that
lapses from when a tornado is observed in the field to
the issuance of an NWS product confirming the threat.
This process normally can be described in five steps:
1) a tornado is observed by the public, law enforcement,
or storm spotters; 2) an individual reports their observation to local law enforcement, a spotter network, or
directly to an NWS office through some form of communication; 3) the NWS receives the information; 4) the
NWS gauges the credibility of information through environmental and radar-based clues; and 5) the NWS
disseminates the ground-truth information through
a variety of products and services. In an ideal scenario,
this process may only take a few minutes from start to
finish, especially if a tornado is anticipated by the NWS.
However, there are several potential challenges that can
delay the entire communication and dissemination process. While not intended to be a comprehensive list of
pitfalls, these challenges that result in a delay of information can stem from 1) confusion of the individual’s
location and/or the location of the observed tornado,
2) a communication delay or breakdown between the
individual reporting and the receiving party, 3) a lack of
knowledge of the local NWS office and/or their phone
number, and 4) conflicting or unexpected information
within the report. Any one or any combination of these
challenges can greatly impede the process of disseminating this critical information; therefore, logic would suggest
that a tornado persisting for longer periods of time would
invite a greater probability for the ground-truth reporting
and NWS dissemination process to confirm the threat
prior to tornado dissipation.
Increasing tornado pathlength corresponded to
a higher frequency of confirmation of events in real
595
FIG. 5. As in Fig. 4, but for tornado pathlength (km).
time (Fig. 5). Median pathlengths of 10.5 km (7 mi) were
associated with real-time confirmation events, compared
to 4.8 km (3 mi) for events with no confirmation in NWS
products. Pathlength is statistically greater for tornadoes
that are confirmed in real time versus those that are not,
to the 99% confidence level. Furthermore, long-track
tornadoes (.48 km; .30 mi) were confirmed in 84% of
the cases, whereas only 36% of short-track events
(,24 km; ,15 mi) were verified in real time (Fig. 6). It is
believed that increasing tornado pathlength, along with
a longer duration, results in an increasing probability that
the vortex will move across sufficient geographical area
for the public or storm spotters to observe the event.
The amount of time a tornado persisted following the
initial real-time confirmation disseminated through
NWS products showed a dependence on tornado pathlength and duration (Fig. 7). In cases with positive
identification issued in real time, NWS products were
able to confirm the threat for short-track tornadoes with
FIG. 6. Tornado pathlength distribution (%) of real-time confirmation for short- (1204), moderate- (127), and long-track (31)
track from 2007 to 2011.
596
WEATHER AND FORECASTING
VOLUME 29
FIG. 7. As in Fig. 4, but for the tornado duration (min) following
real-time confirmation through NWS products, distributed by total
tornado pathlength (km).
FIG. 8. Tornado distribution (%) by EF-scale rating (EF1, 314;
EF2, 130; EF3, 71; EF4, 24; and EF5, 4) for real-time confirmation
events from 2007 to 2011.
a median of 8 min remaining of the tornado’s lifespan.
This greatly improved for moderate- to long-track tornadoes, with medians of 23 and 49 min, respectively.
Intense tornadoes resulting in EF3–EF5 damage were
confirmed in real time at a much higher frequency than
weaker tornadoes rated EF1 and EF2 (Fig. 8). The most
violent tornadoes of EF4 and EF5 intensity were confirmed in 92% and 100% of the cases, respectively,
which is in stark contrast to EF1 tornadoes, where only
33% of these events were confirmed in real time. While
EF0 tornadoes were not investigated in this study, it is
presumed that a similar rate of confirmation to EF1
events, if not lower, would exist for the weakest and
traditionally shortest-lived events. The rate of real-time
confirmation is statistically greater for weak (EF1) tornadoes versus strong (EF2 and EF3), and for strong
versus violent (EF4 and EF5) tornadoes, both to the
99% confidence level. This tendency for the rate of realtime confirmation to increase as intensity increases also
mirrors the trends seen with pathlength and duration,
which reflects the finding of Brooks (2004) that tornadoes with increasing pathlengths generally result in increasing F-scale ratings. Additionally, Brotzge and
Erickson (2010) showed that increasing F-scale ratings
and pathlengths resulted in a higher probability of
a tornado warning being issued, which was speculated to
spawn from improved visual or radar-based detection;
these findings not only relate tornado intensity and
pathlength as in Brooks (2004) and the current study,
but also hint that these scenarios may more often result
in real-time confirmation. The results in this study show
that longer pathlength and duration and also higher EFscale rating may result in a higher frequency of these
tornadoes being confirmed in real time through NWS
products.
The results are encouraging when considering that the
most destructive, deadly, and long-track tornadoes are
likely to be confirmed in real time through NWS products. The ability to confirm the presence of these extreme events in a timely manner benefits and enhances
the decision makers’ and broadcasters’ roles in conveying the warning message. However, it is important to
clarify the limitations of what these data suggest. While
it is probable that a tornado rated as violent (EF4/EF5)
will be observed in the field and disseminated through
NWS products in real time, there is no guarantee that
the actual strength, magnitude, or potential impact of
the event can be accurately gauged during the tornado’s
lifespan on any regular basis (NWS 2011b). Simply receiving the confirmation of an ongoing tornado does not
indicate that a tornado is a specific strength. Only limited research has been conducted to correlate rotational
signatures from the Weather Surveillance Radar-1988
Doppler (WSR-88D) super-resolution velocity data with
tornado intensity at the surface, with varied success
(Kingfield et al. 2012; LaDue et al. 2012; Toth et al. 2013).
To date, no robust scientific guidance exists to accurately
correlate tornado intensity and its potential ground-level
impacts with WSR-88D data available in real time to
operational forecasters.
c. SPC outlooks and watches
The authors hypothesize an increasing forecast tornadic
risk should be proportional to an increase in preparation
and involvement by the severe weather community, ultimately resulting in more eyes in the field to observe severe
weather. Previous studies have shown that SPC convective outlooks and watches are heavily used in the preevent planning activities of emergency managers and
storm spotter groups, among others (Doswell et al. 1999;
JUNE 2014
BLAIR AND LEIGHTON
FIG. 9. Distribution (%) of SPC forecast risk of general thunder
(86 cases), slight (688 cases), moderate (460 cases), and high (126
cases) from day-1 convective outlooks for tornado events with
(gray) and without (black) real-time confirmation.
Baumgart et al. 2008). Generally, a forecast containing
an increased risk of tornadoes traditionally motivates
members of the IWT to use a more aggressive approach
in preparation. For instance, local media may dispatch
a fleet of television storm chasers and journalists into
the field. Emergency managers may alert local law enforcement and deploy local storm spotters in strategic
viewing locations across the community (Baumgart et al.
2008). Local and national experienced storm chasers
may flock to the highest risk area. The public may also
experience a heightened awareness for the potential of
hazardous weather and keep a keener eye to the sky.
Additionally, NWS staffing is likely to increase during
severe weather to balance the workload, potentially allowing for more frequent dissemination of products
and services immediately following reception of severe
weather reports. In an ideal scenario where an elevated
threat of tornadoes exists, a weather-savvy society is in
place prior to convective development, producing the
highest number of amateur and professional weather
observers available in the field. This in turn should result
in an increased probability of (visible) tornado observation. In contrast, a low risk or a poorly anticipated
severe weather event may result in a muted awareness of
a tornado threat and, thus, fewer resources in the field.
The results revealed that SPC convective outlooks,
issued at least 8 h prior to each tornado, showed an increased number of tornadoes confirmed in real time with
an increased forecast risk of severe weather (Fig. 9). In
cases where ‘‘general thunderstorms’’ were forecast,
tornado events were infrequently confirmed in real time
(24%) whereas ‘‘high risk’’ forecasts were associated
with the majority of tornadoes being confirmed through
NWS products prior to dissipation (54%). Because the
categorical convective outlooks include all threat types
597
FIG. 10. Distribution (%) of SPC forecast tornado probabilities
of 2% (215 cases), 5% (374 cases), 10% (297 cases), 15% (266
cases), and 30% (112 cases) from day-1 convective outlooks for tornado events with (gray) and without (black) real-time confirmation.
of severe weather, forecast tornado probabilities contained in the outlook were also investigated. A similar
trend is found with SPC-forecasted tornado probabilities within the outlooks, where the majority of tornadoes
were confirmed in real time for the highest probabilities
(Fig. 10). In addition, tornadoes occurring in a ‘‘tornado
watch’’ were much more likely to be confirmed via
NWS products in real time compared to ‘‘severe thunderstorm watches’’ (Fig. 11). This was especially true
when ‘‘particularly dangerous situation’’ (PDS) tornado
watches were issued; in these situations, 53% of tornadoes were confirmed in real time.
SPC short-term forecast products reflected the trends
revealed in section 3b, with increasing forecast risk also
relating to a higher probability of real-time confirmation. This relationship is likely a consequence of the
FIG. 11. Distribution (%) of SPC convective watches of severe
(169 cases), tornado (752 cases), and PDS tornado (324 cases)
for tornado events with (gray) and without (black) real-time
confirmation.
598
WEATHER AND FORECASTING
design of SPC products to forecast the areal probability
of severe weather and its magnitude based on anticipated environmental conditions and convective evolution. Tornado events that are long tracked and intense
empirically have a tendency to be well forecasted
through the suite of SPC short-term products. Additionally, the results at least partially support the notion
that an increasing forecast risk of tornadoes may lead to
more observers in the field, ultimately providing increasing potential of timely storm reports and subsequent confirmation in NWS products. Therefore, the
forecasted tornado environment in the IWT planning
stage may increase anticipation that real-time tornado
information may or may not be available.
4. Discussion and conclusions
Continued advances in technology over the past decade have resulted in several new avenues for spotters
and chasers to provide severe weather reports in real
time to the NWS (Pietrycha et al. 2009; Brotzge and
Donner 2013). Social media platforms have also enabled
the public to quickly share photos or text reports of
hazardous weather (Hyv€
arinen and Saltikoff 2010; Blair
and Leighton 2012; Mass 2012). Spotter network reports, streaming video from storm chasers and television
stations, Facebook, and Twitter reports have all been
utilized during warning operations, helping to accelerate
the dissemination of ground-truth information to the
multiple users of NWS products. However, as technology advances, it remains critical that adequate investments are made to both technology infrastructures
and proficiency training for emerging, diverse tools in
the operational setting.
New technology should allow for real-time confirmation to apply beyond human observations with the nationwide completion of the dual-polarization upgrade to
the WSR-88D network. A ‘‘tornado debris signature’’
is distinguished when debris is lofted within the parent
circulation, revealed on radar as a nonmeteorological
echo by a sharp reduction in the correlation coefficient product collocated within a strong velocity couplet
(Ryzhkov et al. 2005; Kumjian and Ryzhkov 2008).
While this signature will not improve lead time, it has
great potential as another tool to identify ongoing
damaging tornadoes relatively close to a radar site. This
is especially true in limited-visibility, high-precipitation
situations or at night in the absence of human groundtruth reports, potentially improving some of the diurnal
disparity of timely tornado information. Future research
will help better determine how this remotely sensed information can be communicated in an effective and understandable channel to the public. Today’s technologies
VOLUME 29
and those of the future will provide more opportunities
for the IWT to monitor ongoing tornadoes, contributing
to a quicker transmission of more detailed information
to the end user.
Real-time confirmation of a tornado through official
and broadcast sources is thought to increase the credibility and urgency of the warning message, which may
help play a role in encouraging the public to seek shelter.
This study provides a climatology of tornadoes confirmed and disseminated through NWS products prior to
dissipation from 2007 to 2011 over the central contiguous United States. Data were also investigated to determine what available signals existed that may facilitate
or hinder receiving or transmitting timely tornado reports. The results illustrate the following:
d
d
d
d
d
d
Approximately 40% (543) of the tornadoes in the
study sample were confirmed in real time through
NWS products.
The high plains region had a higher occurrence of realtime confirmation, possibly attributed to less vegetative and terrain obstructions, more frequent supercell
storm mode, and a high concentration of storm chasers
during the months of climatologically favored tornado
activity.
A strong discriminating factor for a tornado to be
confirmed through NWS products in real time was
tornado duration (median of 14 min compared to
5 min for nonconfirmed events).
The large majority of the most intense and long-track
tornadoes were confirmed in real time [approximately
94% of EF4 and EF5 events; 84% of tornadoes
pathlength .48 km (.30 mi)].
Nighttime tornadoes were 20% less likely to receive
real-time confirmation in NWS products compared
to daytime events (only 25% of nighttime events
confirmed).
SPC short-term forecast products (convective outlooks, tornado probabilities, and watches) showed an
increasing probability for real-time confirmation as
the forecasted risk increased.
This study provides a set of general guidelines for
anticipating situations when real-time confirmation of
a tornado is more (or less) probable, which may be used
to help IWT members and other partners adjust their
expectations for the availability of ground-truth information within NWS products. It is important to remember that each tornadic scenario is unique and some
deviations from the results found herein are possible.
Communication and reporting challenges in the field,
variations in population density, seasonal anomalies of
severe weather events, and NWS workload or technology issues may potentially lead to further abnormalities
JUNE 2014
BLAIR AND LEIGHTON
in the dissemination of timely tornado information. It is
important for broadcasters, decision makers, and other
NWS partners to value the information contained within
the initial science-based tornado warning, regardless of
whether or not validation of the ongoing threat is
available, especially because, as this study demonstrates,
certain conditions may delay or prohibit confirmation.
NWS offices should continue to strive to proactively
issue timely and reliable tornado reports through the
suite of available tools, products, and services. An aggressive approach to confirm the threat should be utilized, including the monitoring of social media platforms
and electronic field reports and videos from spotters.
Additionally, other IWT members should immediately
share tornado reports with each other through NWSChat
or other means to attain a wider dissemination of critical
information. Such team efforts improve the speed of essential tornado updates in official NWS products—in
some cases by several minutes—ultimately enhancing the
warning message to the end users and potentially
achieving a stronger public response to take appropriate
action in the face of hazardous weather.
Acknowledgments. The authors gratefully acknowledge Julie Demuth, Jennifer Laflin, Amos Magliocco,
Albert Pietrycha, and Jeff Manion for their thorough
and beneficial reviews. We also thank two anonymous
reviewers for helping to improve the manuscript. The
views and opinions expressed in this paper are those of the
authors and do not necessarily represent an official position, policy, or decision of the National Weather Service.
REFERENCES
Andra, D. L., E. M. Quoetone, and W. F. Bunting, 2002: Warning
decision making: The relative roles of conceptual models, technology, strategy, and forecaster expertise on 3 May 1999. Wea.
Forecasting, 17, 559–566, doi:10.1175/1520-0434(2002)017,0559:
WDMTRR.2.0.CO;2.
Ashley, W. S., 2007: Spatial and temporal analysis of tornado fatalities in the United States: 1880–2005. Wea. Forecasting, 22,
1214–1228, doi:10.1175/2007WAF2007004.1.
——, A. J. Krmenec, and R. Schwantes, 2008: Vulnerability due to
nocturnal tornadoes. Wea. Forecasting, 23, 795–807, doi:10.1175/
2008WAF2222132.1.
Baumgart, L. A., E. J. Bass, B. Philips, and K. Kloesel, 2008: Emergency management decision making during severe weather. Wea.
Forecasting, 23, 1268–1279, doi:10.1175/2008WAF2007092.1.
Blair, S. F., and J. W. Leighton, 2012: Creating high-resolution hail
datasets using social media and post-storm ground surveys.
Electron. J. Oper. Meteor., 13 (3), 32–45. [Available online at
http://www.nwas.org/ej/2012/2012.php.]
Brooks, H. E., 2004: On the relationship of tornado path length and
width to intensity. Wea. Forecasting, 19, 310–319, doi:10.1175/
1520-0434(2004)019,0310:OTROTP.2.0.CO;2.
——, and C. A. Doswell III, 2001: Normalized damage from major tornadoes in the United States: 1890–1999. Wea. Fore-
599
casting, 16, 168–176, doi:10.1175/1520-0434(2001)016,0168:
NDFMTI.2.0.CO;2.
——, ——, and M. P. Kay, 2003: Climatological estimates of local
daily tornado probability for the United States. Wea. Forecasting, 18, 626–640, doi:10.1175/1520-0434(2003)018,0626:
CEOLDT.2.0.CO;2.
Brotzge, J., and S. Erickson, 2009: NWS tornado warnings with
zero or negative lead times. Wea. Forecasting, 24, 140–154,
doi:10.1175/2008WAF2007076.1.
——, and ——, 2010: Tornadoes without NWS warning. Wea.
Forecasting, 25, 159–172, doi:10.1175/2009WAF2222270.1.
——, and J. W. Donner, 2013: The tornado warning process:
A review of current research, challenges, and opportunities.
Bull. Amer. Meteor. Soc., 94, 1715–1733, doi:10.1175/
BAMS-D-12-00147.1.
——, S. Erickson, and H. E. Brooks, 2011: A 5-yr climatology of
tornado false alarms. Wea. Forecasting, 26, 534–544, doi:10.1175/
WAF-D-10-05004.1.
Brown, S., P. Archer, E. Kruger, and S. Mallonee, 2002: Tornadorelated deaths and injuries in Oklahoma due to the 3 May
1999 tornadoes. Wea. Forecasting, 17, 343–353, doi:10.1175/
1520-0434(2002)017,0343:TRDAII.2.0.CO;2.
Coleman, T. A., K. R. Knupp, J. Spann, J. B. Elliott, and B. E.
Peters, 2011: The history (and future) of tornado warning
dissemination in the United States. Bull. Amer. Meteor. Soc.,
92, 567–582, doi:10.1175/2010BAMS3062.1.
Demuth, J. L., B. H. Morrow, and J. K. Lazo, 2009: Weather
forecast uncertainty information: An exploratory study with
broadcast meteorologists. Bull. Amer. Meteor. Soc., 90, 1614–
1618, doi:10.1175/2009BAMS2787.1.
Dix, D., and J. Fieux, 2011: Atlanta Integrated Warning Team
workshop: A WAS*IS success. Weather and Society Watch,
Vol. 5, No. 4, NCAR Societal Impacts Program, Boulder,
CO. [Available online at www.sip.ucar.edu/news/volume5/
number4/highlights1.php.]
Doswell, C. A., III, 2005: Progress toward developing a practical
societal response to severe convection (2005 EGU Sergei
Soloviev Medal Lecture). Nat. Hazards Earth Syst. Sci., 5, 691–
702, doi:10.5194/nhess-5-691-2005.
——, and D. W. Burgess, 1988: On some issues of United States
tornado climatology. Mon. Wea. Rev., 116, 495–501, doi:10.1175/
1520-0493(1988)116,0495:OSIOUS.2.0.CO;2.
——, A. R. Moller, and H. E. Brooks, 1999: Storm spotting and
public awareness since the first tornado forecasts of 1948. Wea.
Forecasting, 14, 544–557, doi:10.1175/1520-0434(1999)014,0544:
SSAPAS.2.0.CO;2.
Hammer, B., and T. W. Schmidlin, 2002: Response to warnings
during the 3 May 1999 Oklahoma City tornado: Reasons and
relative injury rates. Wea. Forecasting, 17, 577–581, doi:10.1175/
1520-0434(2002)017,0577:RTWDTM.2.0.CO;2.
Hayden, M. H., S. Drobot, S. Radil, C. Benight, E. C. Gruntfest,
and L. R. Barnes, 2007: Information sources for flash flood
warnings in Denver, CO, and Austin, TX. Environ. Hazards,
7, 211–219, doi:10.1016/j.envhaz.2007.07.001.
Hodler, T. W., 1982: Residents’ preparedness and response to
the Kalamazoo tornado. Disasters, 6, 44–49, doi:10.1111/
j.1467-7717.1982.tb00743.x.
Hyv€arinen, O., and E. Saltikoff, 2010: Social media as a source of
meteorological observations. Mon. Wea. Rev., 138, 3175–3184,
doi:10.1175/2010MWR3270.1.
Kingfield, D. M., J. G. LaDue, and K. L. Ortega, 2012: An evaluation of tornado intensity using velocity and strength attributes from the WSR-88D mesocyclone detection algorithm.
600
WEATHER AND FORECASTING
Proc. 26th Conf. on Severe Local Storms, Nashville, TN,
Amer. Meteor. Soc., 3.2. [Available online at https://ams.
confex.com/ams/26SLS/webprogram/Paper211488.html.]
Kumjian, M. R., and A. V. Ryzhkov, 2008: Polarimetric signatures
in supercell thunderstorms. J. Appl. Meteor. Climatol., 47,
1940–1961, doi:10.1175/2007JAMC1874.1.
LaDue, J. G., K. L. Ortega, B. R. Smith, G. J. Stumpf, and D. M.
Kingfield, 2012: A comparison of high resolution tornado
surveys to Doppler radar observed mesocyclone parameters:
2011-2012 case studies. Proc. 26th Conf. on Severe Local
Storms, Nashville, TN, Amer. Meteor. Soc., 6.3. [Available
online at https://ams.confex.com/ams/26SLS/webprogram/
Paper212627.html.]
Legates, D. R., and M. D. Biddle, 1999: Warning response and risk
behavior in the Oak Grove—Birmingham, Alabama tornado
of 08 April 1998. Natural Hazards Research Applications and
Information Center Quick Response Rep. 116, Boulder,
CO. [Available online at http://www.colorado.edu/hazards/
research/qr/qr116/qr116.html.]
Lindell, M. K., and R. W. Perry, 2012: The Protective Action Decision
Model: Theoretical modifications and additional evidence. Risk
Anal., 32, 616–632, doi:10.1111/j.1539-6924.2011.01647.x.
Mass, C., 2012: Nowcasting: The promise of new technologies of
communication, modeling, and observation. Bull. Amer. Meteor. Soc., 93, 797–809, doi:10.1175/BAMS-D-11-00153.1.
McDonald, J. R., and K. C. Mehta, 2006: A recommendation for an
enhanced Fujita scale (EF-Scale). Wind Science and Engineering Center, Texas Tech University, Lubbock, TX, 111 pp.
[Available online at http://www.depts.ttu.edu/weweb/EFScale.
pdf.]
Mileti, D. S., and J. H. Sorensen, 1990: Communication of emergency public warnings: A social science perspective and stateof-the-art assessment. Oak Ridge National Laboratory Rep.
ORNL-6609, 145 pp.
Mitchem, J. D., 2003: An analysis of the September 20, 2002 Indianapolis tornado: Public response to a tornado warning and
damage assessment difficulties. Natural Hazards Research and
Applications Information Center Quick Response Rep. 161.
[Available online at http://www.colorado.edu/hazards/research/
qr/qr161/qr161.html.]
Nagele, D., and J. Trainor, 2012: Geographic specificity, tornadoes,
and protective action. Wea. Climate Soc., 4, 145–155, doi:10.1175/
WCAS-D-11-00047.1.
NCDC, 2007–11: Storm Data. Vols. 47–51. [Available from National Climatic Data Center, 151 Patton Ave., Asheville, NC
28801-5001.]
NWS, 2007: Storm-based warnings team report. National Weather
Service Rep., 45 pp. [Available online http://www.nws.noaa.
gov/sbwarnings/docs/Polygon_Report_Final.pdf.]
——, 2009: Super Tuesday tornado outbreak of February 5–6, 2008.
NWS Service Assessment, 48 pp. [Available online at http://
www.weather.gov/os/assessments/pdfs/super_tuesday.pdf.]
——, 2011a: Joplin, Missouri, tornado—May 22, 2011. NWS Central Region Service Assessment, 35 pp. [Available online at
http://www.nws.noaa.gov/om/assessments/pdfs/Joplin_tornado.
pdf.]
——, 2011b: The historic tornadoes of April 2011. NWS Service
Assessment, 76 pp. [Available online at http://www.nws.noaa.
gov/os/assessments/pdfs/historic_tornadoes.pdf.]
NWS/Warning Decision Training Branch, cited 2013: Integrated
warning team training. [Available online at http://www.wdtb.
noaa.gov/Training/course/index.asp?courseList25IWT.]
VOLUME 29
Paul, B. K., V. T. Brock, S. Csiki, and L. Emerson, 2003: Public
response to tornado warnings: A comparative study of the
May 4, 2003, tornadoes in Kansas, Missouri, and Tennessee.
Natural Hazards Research and Applications Information
Center Quick Response Rep. 165. [Available online at http://
www.colorado.edu/hazards/research/qr/qr165/qr165.html.]
Perry, R. W., and M. K. Lindell, 1991: The effects of ethnicity on evacuation decision-making. Int. J. Mass Emerg. Disasters, 9, 47–68.
Pietrycha, A. E., S. F. Blair, T. J. Allison, D. R. Deroche, and R. V.
Fritchie, 2009: Emerging technologies in the field to improve
information in support of operations and research. Electron.
J. Oper. Meteor., 10 (2). [Available online at http://www.nwas.
org/ej/2009-EJ2/.]
Rose, S. F., P. V. Hobbs, J. D. Locatelli, and M. T. Stoelinga, 2004:
A 10-yr climatology relating the locations of reported tornadoes to the quadrants of upper-level jet streaks. Wea. Forecasting, 19, 301–309, doi:10.1175/1520-0434(2004)019,0301:
AYCRTL.2.0.CO;2.
Ryzhkov, A. V., T. J. Schuur, D. W. Burgess, and D. S. Zrnic, 2005:
Polarimetric tornado detection. J. Appl. Meteor., 44, 557–570,
doi:10.1175/JAM2235.1.
Schumacher, R. S., D. T. Lindsey, A. B. Schumacher, J. Braun,
S. D. Miller, and J. L. Demuth, 2010: Multidisciplinary analysis of an unusual tornado: Meteorology, climatology, and the
communication and interpretation of warnings. Wea. Forecasting, 25, 1412–1429, doi:10.1175/2010WAF2222396.1.
Sherman-Morris, K., 2005: Tornadoes, television and trust—A
closer look at the influence of the local weathercaster during
severe weather. Environ. Hazards, 6, 201–210, doi:10.1016/
j.hazards.2006.10.002.
——, 2010: Tornado warning dissemination and response at a
university campus. Nat. Hazards, 52, 623–638, doi:10.1007/
s11069-009-9405-0.
——, 2013: The public response to hazardous weather events:
25 years of research. Geogr. Compass, 7, 669–685.
——, and M. E. Brown, 2012: Experiences of Smithville, Mississippi
residents with the 27 April 2011 tornado. Natl. Wea. Dig., 36, (2),
93–101.
Simmons, K. M., and D. Sutter, 2005: WSR-88D radar, tornado
warnings, and tornado casualties. Wea. Forecasting, 20, 301–
310, doi:10.1175/WAF857.1.
Smith, B. T., R. L. Thompson, J. S. Grams, C. Broyles, and H. E.
Brooks, 2012: Convective modes for significant severe thunderstorms in the contiguous United States. Part I: Storm
classification and climatology. Wea. Forecasting, 27, 1114–
1135, doi:10.1175/WAF-D-11-00115.1.
Sorensen, J. H., 2000: Hazard warning systems: Review of 20 years
of progress. Nat. Hazards Rev., 1, 119–125, doi:10.1061/
(ASCE)1527-6988(2000)1:2(119).
Tiefenbacher, J. P., W. Monfredo, M. Shuey, and R. J. Cecora,
2001: Examining a ‘‘near-miss’’ experience: Awareness, behavior, and post-disaster response among residents on the
periphery of a tornado-damage path. Natural Hazards Center
Quick Response Rep. 137. [Available online at http://www.
colorado.edu/hazards/research/qr/qr137/qr137.html.]
Toth, M., R. J. Trapp, J. Wurman, and K. A. Kosiba, 2013: Comparison of mobile-radar measurements of tornado intensity
with corresponding WSR-88D measurements. Wea. Forecasting, 28, 418–426, doi:10.1175/WAF-D-12-00019.1.
Verbout, S. M., H. E. Brooks, L. M. Leslie, and D. M. Schultz, 2006:
Evolution of the U.S. tornado database: 1954–2003. Wea.
Forecasting, 21, 86–93, doi:10.1175/WAF910.1.