Response to Flash Flood Warnings: State of our Knowledge Burrell E. Montz Department of Geography East Carolina University [email protected].

Response to Flash Flood Warnings: State of our Knowledge

Burrell E. MontzDepartment of Geography

East Carolina [email protected]

Topics

• Short fuse events– Flash floods– Tornadoes

• Overview of studies• Summary of findings• So what...?

There remains, then, the need for a mathematical model of human response to warnings, a model that would mimic all essential characteristics of human response in a setting of a local flood warning system and that would enable one to predict the outcomes of decision-event pairs. Krzysztofowicz, R., 1993

The Problem

Reality

Components of Public Response

Hear Understand Believe Personalize Decide to act Respond

What People Say

Gruntfest et al., 2008

What People Do

League, 2008

Actual versus Anticipated Behavior

• Difference between what people say and what they do

• Importance of context and circumstance• Difficult to document impacts of

– Time– Memory– Cognitive dissonance

Tornado Studies: Sources of Information

Schmidlin et al., 2009; Schmidlin and King, 1997; Balluz et al., 2000

Actions and Reasons:35% took shelter

• Positive actions correlated with– Perceived danger– Presence of children– High school education– Hearing warning– Having a basement– Being married

• Negative actions correlated with– Previous damage– Less education– God’s will– Lack of access to shelter– Limited mobility

• No correlation– Age, gender, race– Lead time– Owning NWR– Family size– Previous experience

NWS Service Assessments

• Super Tuesday 2008 Tornadoes– 57 dead– 18 (32%) heard some

warnings– 11 (61%) heeded warnings– 8 (44%) sought shelter– 6 (33%) did not

• Mothers’ Day 2008 Tornadoes– 21 dead– 11 (52%) knew of warning– 10 (47.6%) tried to take

shelter________________________

– 14 groups interviewed– 6 (42.8%) heard official

warning– 6 heard from family or friends– 4 (28.5%) sought shelter– 6 tried but it “came too fast”

Flood Fatalities

Source: League, 2009, http://www.geo.txstate.edu/lovell/IFFL/research.html

Vehicle Deaths

Source: League, 2009, http://www.geo.txstate.edu/lovell/IFFL/research.html

Gender Breakdown*

* Where reported

Source: League, 2009, http://www.geo.txstate.edu/lovell/IFFL/research.html

So what about warnings?

But...

League, 2009

There is a difference between

•Intentional Drivers

•Situational Drivers

False Alarms, Near Misses, and Response

• What we know– Very different definitions

of false alarms• NWS vs public

– Perceptions of accuracy vary

• NWS vs public

– Cry wolf or warning fatigue or neither

– Influence of event type– We don’t know enough

Barnes et al., 2007

And...There is no ONE public

•Different languages

•Different understandings

•Different situations

•Different capabilities

•Different needs

Vulnerability Factors IndicatorsSocio-economic and demographic attributes

Age, gender, income, profession, family situation

Social structures Cohesion of community; social networks

Infrastructure Building quality and types

Attitudinal, psychological, and knowledge factors

Experience, risk perception, views of nature, press coverage

Warning systems Communication channels and relevancy

Public policy/risk management

System of actors; decision-making process

Spatial and temporal aspects of event

Time of day; location; local knowledge

Long way to go...

•NWS mission: Protect life and property•NWS warnings are only the beginning of meeting this mission•Warnings move through various paths to the public•Warnings are received and understood differently•Collaborative effort required to get positive, protective responses•Social science research required to understand why people respond the way they do under what circumstances

Conclusion

Thank you

Any questions you’d like to wade

through?

http://blogs.davenportlibrary.com/sc/wp-content/uploads/2008/04/no-wading.jpg