Transcript The 100-Year Storm Redefined

A Proposed Method of Categorizing
Extreme Rainfall Events The 100-Year Storm Redefined
I can’t tell these people
anymore that we had yet
ANOTHER 100-year storm!
History of Flooding in Brookfield
Historic flooding events:
March 1897
June 1917
June 1940
March 1960
July 1964
September 1972
April 1973
Recent Flooding in Brookfield
Recent history of flooding in Brookfield
August 6, 1986 – 6.8” in 24 hours
June 20 - 21, 1997 – 6” in 26 hours
August 6, 1998 – 11.35” in 8 hours
June 7 - 8, 2008 – 5.8” in 24 hours
June 19, 2009 – 4.8+” in 3 hours
All 100-year storms or larger!
Wisconsin: 14-Day Precipitation
Valid at 6/16/2008 1200 UTC
Radar image of precipitation in June 2008
Why Call Them 100-Year Storms?
2” rains are different depending where you are
Need something that can be used consistently
across the country
Account for variability in these storms by region
100-Year Storms
Recurrence interval
Based on probability of the storm occurring
How can there be two 100-year storms in a
row?
Let’s Call Them Something Else
1% probability storm
Aren’t we clever?
Public knows 1% = 1/100
They figure out that
it’s the 100-year storm
Now they feel like we tried
to deceive them –
hiding behind probabilities
Probability of the 100-Year Storm
Probability of 100-year event: p = 0.01
100-year event occurring three years in a row?
p3 = (0.01)3 = 0.000001 or 1 in a million
Not always intuitive
100-year event occurring once in 100 years?
1-(1-p)100 = 1-(0.99)100 = 1-0.366 = 63.4%
Public thinks it would occur once every 100
years
Binomial Probability Applied to
Extreme Events
More complex relationships
100-year event occurs 5 times in 25 years?
n!*pY*(1-p)(n-Y) / (Y!)*(n-Y)!
25!*0.015*(0.99)20 / 5!*20! = 1 in 4.4 million
Is this pretty rare?
Consider what people are willing to gamble on
Probability and Statistics
Winning the Powerball Lottery: 1 in 195 million
One Fatal Flaw
People don’t understand statistics
Public’s response to 100-year storm?
“Oh good, we’re safe for another 99 years.”
Two 100-year storms in a row?
“This storm was different than last year’s
storm, so they both can’t be called the 100year storm. Can they?”
Public Confusion
Mark Twain said, “The more you explain it, the
more I don't understand it”
Public more likely to respond this way
Public Response
Disbelief
Disgust
Cynicism
Look for someone to blame
Affects our credibility
Less likely to take our
advice in future
Rating Other Natural Disasters
Earthquakes
Richter scale
Measurement of
Energy released
Strength
Duration of its
seismic waves
Rating Other Natural Disasters
Tornadoes
Fujita scale
(enhanced since
2007)
Measurement of
Intensity
Area affected
Rating Other Natural Disasters
Hurricanes
Saffir- Simpson
scale
Measurement of
Barometric
pressure
Wind speed
Storm surge
Rating Other Natural Disasters
None of these natural disasters are rated
using probabilities or recurrence intervals
Something that implies severity of the
storm, not rarity of the storm
Use a scale instead to rate the event
Easily understood, the higher the number,
the worse the natural disaster
Revise Terminology for
Communication with Public
Keep It Simple
S_ _ _ _ _
Use recurrence
intervals as basis
Don’t call them 100year storms
Use a scale like other
natural disasters
A New Rating System Proposed

RI = 2(G-1)
RI = Recurrence Interval of the rain event
G = the Category of storm
Example:
G-factor of 7 in the formula is 2(7-1) = 26
equals the 64-year event
Categorizing Storms Based on
Recurrence Intervals: RI = 2(G-1)
Category of Storm Recurrence Interval Rain Storm
G-2
2 to 4 year
G-3
4 to 8 year
G-4
8 to 16 year
G-5
16 to 32 year
G-6
32 to 64 year
G-7
64 to 128 year
G-8
128 to 256 year
G-9
> 256 year
Placing the Recurrence Interval
Storm into the New Rating System
Recurrence Interval
Rain Storm: RI = 2(G-1) G - Factor
2-year
2
Category of
Storm
G-2
5-year
3.32
G-3
10-year
4.32
G-4
25-year
5.64
G-5
50-year
6.64
G-6
100-year
7.64
G-7
Simple G-Factor Table
Recurrence Interval
Storm
Duration
2-year
5-year
1 hour
2
3
4
5
6
7
2 hour
2
3
4
5
6
7
3 hour
2
3
4
5
6
7
6 hour
2
3
4
5
6
7
12 hour
2
3
4
5
6
7
24 hour
2
3
4
5
6
7
2 day
2
3
4
5
6
7
3 day
2
3
4
5
6
7
5 day
2
3
4
5
6
7
10 day
2
3
4
5
6
7
10-year 25-year 50-year 100-year
Different Durations of 100-Year
Storms Have Different Effects
2.8 inches of rain in one hour will wash out
roads, culverts, and “flood” streets for a time
7.5 inches of rain in ten days causes rivers to
overtop with widespread flooding
Need to Account for Differences
Develop adjustment factors for G
New formula: GADJ = G x (DAF) x (IAF)
G = the Category of storm
Duration adjustment factor
Intensity adjustment factor
Intensity & Duration Adjusted
Category Storms
Recurrence Interval
Storm
Duration
2-year
5-year
1 hour
0
1
1
2
3
3
2 hour
1
1
2
2
3
4
3 hour
1
1
2
3
3
5
6 hour
1
1
2
3
4
6
12 hour
1
1
2
3
5
6
24 hour
1
2
2
4
5
7
2 day
1
2
3
5
6
7
3 day
1
2
3
5
6
7
5 day
2
3
4
5
6
7
10 day
2
4
5
7
8
9
10-year 25-year 50-year 100-year
Brookfield Flooding Revisited
August 6, 1986: 6.8” in 24 hours = G-8
June 20 - 21, 1997: 6” in 26 hours = G-7
August 6, 1998: 11.35” in 8 hours = G-10+
June 7 - 8, 2008: 5.8” in 24 hours = G-7
June 19, 2009: 4.8+” in 3 hours = G-6
No implication of rarity!!!
Criticism and Scrutiny
No continuous curve for
each category storm
Interpret between
durations to pick the right
category storm
Or develop a look-up
table or formula to
interpolate (any grad
students looking for a
project?)
Criticism and Scrutiny
Does not address:
antecedent moisture,
snowmelt,
soil types,
% of impervious cover,
etc.
Not intended to!
These relate to runoff,
not the rain storm event
Criticism and Scrutiny
Just another name for
the 100-year storm
Precisely. That’s the
point!
Better way to
communicate to public
Severity, not rarity, is
key to public’s
understanding
Advantages of a New System
Preserves the existing data and science
Does not require overhaul of hydrology or
engineering
Generally consistent with rating systems
for other natural disasters
Easy to understand
Bigger the number, the bigger the storm
Advantages of a New System
Provides a more appropriate scaling factor
The 50-year storm is not 50% as big as the 100-year
storm as some might assume
It is almost as big (Typically only 1 unit away)
Uses different terminology than for floodplains
Avoids confusion that 100-year storm always
results in the 100-year flood
Or that the 100-year flood can only occur when
there’s a 100-year storm
Advantages of a New System
Does not imply that big storms are rare
Minimizes criticism, skepticism and
general disregard for hydrologic or
engineering descriptions of the event
Makes it clear to people that these are big
storms that we should take seriously
Incents people to protect themselves
against larger event
Does This Make Sense?
What’s Next?
Does this process have merit?
Academia support?
NOAA, USACE HEC support?
Anyone with me?
Your Turn for Feedback!
For copies of this presentation or white paper:
email me at: [email protected]