Making Weather Warnings Work

1
Making Weather Warnings Work Wednesday, July
18, 2007 Sheldon Drobot
2
Outline

• The Warning Process
• Some common theories
• Results from the Warning Project
• Results from a survey of CO residents about the
  December 20-21, 2006, winter storm

3
One view

• Hearing the warning
• Understanding the contents of the warning message
• Believing the warning is credible and accurate
• Personalizing the warning to oneself
• Confirming that the warning is true and others
  are taking heed
• Responding by taking a protective action

Mileti and Sorensen, 1990
4
Another view

• Risk identification Does the threat exist?
• Risk assessment Is protection needed?
• Risk reduction Is protection feasible?
• Protective response What action to take?

Lindell and Perry, 1992
5
A third view
Weather Warning
Medium
Mode
Decision
Environment. cues
Atmosphere
Surface
Trauma
Non-met factors
Risk
Experience
Many others
6
Looking at weather warnings and non-met factors
through case studies

• The Warning Project
• 1000 mail-in surveys collected in Austin and
  Denver
• 90 questions on info sources, trauma,
  demographics, and vignettes
• Dec. 20-21, 2006, winter storm
• 254 internet surveys collected along the Colorado
  Front Range
• Storm was an interesting test of physical and
  societal studies

7
Weather warning flash floods
Cell phone
Austin
Austin
Denver
Weather bug
NOAA weather radio
Other
The Weather Channel
Internet
Local radio stations
Environmental cues
Local TV stations
0
10
20
30
40
50
8
Weather Warnings flash floods
100
90
80
70
60
Percent
50
40
30
20
10
0
Sirens
Phone
Local
TV
Door
NOAA
Cell
E-mail
Radio
Weather
phone
Radio
N 519
9
Weather warning winter weather
Denver Dec. 20, 2006, Winter Storm
NOAA Weather Radio
During
NWS webpages
Leading Up
Other webpages
Newspapers
Cable TV
Public Radio
Friends
Local TV
0
20
40
60
80
100
10
Weather warning winter weather
Stay for some other reason
Left home
18
48
34
Stay based on weather forecast
11

• What percentage of people say they would drive
  through flooded roads stratified by

12
whether they agree or disagree with the
statement Most flash-flood deaths are associated
with vehicles
100
75
43
50
Percent agreeing that they would drive through
flooded roads
31
25
33/106
126/291
0
Agree
Disagree/ Dont know
13
flood experience
100
75
46
50
Percent agreeing that they would drive through
flooded roads
34
25
54/160
102/229
0
Yes
No
14
Post-traumatic Coping Efficacy
100
75
53
50
39
Percent agreeing that they would drive through
flooded roads
25
78/199
32/61
0
High
Low-Moderate
15
Non-met factors Denver Storm
Do you check multiple sources daily?
60
50
40
Percent who stayed home
30
20
10
45/94
38/108
0
Yes
No
16
Non-met factors Denver Storm
Anxiety related to driving in hazardous weather
Percent who stayed home
28/79
20/36
5/24
29/56
None
Low
Moderate
High
Anxiety Level
17
Summary

• Results from previous studies suggest that the
  warning process is not linear and not simple
• Results from my studies suggest that
  decision-making is a complicated processwe need
  to combine physical and social sciences to better
  understand process

18
Resources

• Mileti, D. and J. Sorenson, 1990 Communication
  of emergency public warnings. Available at
  http//emc.ornl.gov/EMCWeb/EMC/PDF/CommunicationFi
  nal.pdf
• Lindell, M. and R. Perry, 1992 Behavioral
  Foundations of Community Emergency Planning.
  Taylor Francis. 320 pp.
• http//ccar.colorado.edu/drobot/publications.html