Rainfall Parameters

1
Rainfall Parameters for Design Storms
Katrina 23-31 August, 2006
Landscapes Change Climate
Changes Technology Changes
Does (Should) Design ?
Ana P. Barros
Climate Change Workshop, March 29th
2
What need the bridge much broader than the
flood? William Shakespeare , from Much Ado
About Nothing
THE flood? What FLOOD?
the Mississippi flood of 1973 was estimated to
be a 200-year event. the Mississippi flood of
1993 was estimated to be a 500-year event.
At St. Louis
Andrew, 1992
Along the Louisiana Coast
1969 Camille 1992 Andrew (2005 Katrina)
8/25
?
8/24
8/23
3
Depth-Duration-FrequencyIntensity-Duration-
Frequency
(1) X hr T year event (rainfall depth)
http//www.nws.noaa.gov/ohd/hdsc
(2) Hyetograph I(t), 0lttltX (distribution of
rainfall depth over the duration)
Design Manuals (FHWA, State DoTs, NCRS, etc)
foothills
I(t)
24hr Event
plateau
Diurnal Cycle
4
Current NWS Precipitation Frequency Publications
Location 5 min - 60 min 1 hr - 24 hr 2 day - 10 day
Arizona, Nevada, New Mexico, Utah, and Southeast California NOAA Atlas 14 (2003) NOAA Atlas 14 (2003) NOAA Atlas 14 (2003)
Remainder of the Western US Arkell Richards (1986)Frederick Miller (1979) NOAA Atlas 2 (1973) Tech. Paper 49 (1964)
Delaware, Illinois, Indiana, Kentucky, Maryland, New Jersey, North Carolina, Ohio, Pennsylvania, South Carolina, Tennessee, Virginia, West Virginia, and Washington, DC  NOAA Atlas 14, Volume 2 (June 2004)  NOAA Atlas 14, Volume 2 (June 2004)  NOAA Atlas 14, Volume 2 (June 2004) 
Remainder of the Eastern US Tech. Memo 35 (1977) Tech. Paper 40 (1961) Tech. Paper 49 (1964)
Hawaii (ongoing) Tech. Paper 43 (1962) Tech. Paper 43 (1962) Tech. Paper 51(1965)
Alaska Tech. Paper 47 (1963) Tech. Paper 47 (1963) Tech. Paper 52 (1965)
Puerto Rico (ongoing) Tech. Paper 42 (1961) Tech. Paper 42 (1961) Tech. Paper 53 (1965)
5
Uncertainty
6
1day - 100 yr Atlas 2
Actual Risk?
Observations Methodology Climate Change
?
20 cm
1day - 200 yr DB03
35.6 cm
7
(Bonnin et al., 2004 Peer Review )
8
The Joseph Effect
Nilometer 622-1284 A.D.
H0.91
The idea that persistent (0.5ltHlt1.0) movements in
a time series tend to be part of larger trends
and cycles more often than they are completely
random.
White Noise
H0.5
From Koutsoyannis, 2004
(Mandelbrot and Wallis 1977, Hurst 1951)
9
Classical Statistics
Fractal Statistics
H0.88
10
70-80 years
11
The Noah Effect
The idea that persistent (0.5ltHlt1.0) time series
have abrupt an discontinuous changes.
12
Climate Change
13
Precipitation Processes
Clausius-Clapeyron Equation

vapor pressure
saturation vapor pressure
Clouds
Evaporation
???
Heating
(2)
e
(3)
Temperature (T)
T
Td
(1)
Cooling / Lifting
Rainfall
14
Schumacher Johnson (2005)
Regime versus Seasonality
(1999-2003)
Climate Change
Regime Seasonality Antecedent
Conditions
15




warmer climate





Climate Change




LCL

LCL
Rainshadow






moist air
dry air















LCL


colder climate
Rainshadow




LCL

dry air

Orographic Effects and Storm Centering


colder, drier air
16
Recognizing Changing Needs (independent of
climate change)
Design Criteria as a Process
Observations Number, Location, Accuracy
Raingauges Ground-based Radar Satellites GIS,
GPS,..
Methods L-Moments Hurst Statistics
Multifractals
17
(No Transcript)
18
(No Transcript)
19
Landform and the Diurnal Cycle of Rainfall
Day-time
Night-time
20
Multisensor/Spatial Data
Ground Radar Satellite Radar IR Microwave Mo
dels (reanalysis) Data-Assimilation
2.5ox2.5o
Reanalysis
21
Recognizing Changing Needs (independent of
climate change)
Design Criteria as a Process
Observations Number, Location, Accuracy
Raingauges Ground-based Radar Satellites GIS,
GPS,..
Methods L-Moments Hurst Statistics
Multifractals Interpolation Statistics

22
Statistics of Extreme Rainfall from Observations

• (1) Selection of the Probability Distribution -
  Underlying Statistical Model
• (2) Estimation of Distribution Parameters
• (3) Regionalization Generalized Interpolation

23
Current NWS Precipitation Frequency Publications
Location 5 min - 60 min 1 hr - 24 hr 2 day - 10 day
Arizona, Nevada, New Mexico, Utah, and Southeast California NOAA Atlas 14 (2003) NOAA Atlas 14 (2003) NOAA Atlas 14 (2003)
Remainder of the Western US Arkell Richards (1986)Frederick Miller (1979) NOAA Atlas 2 (1973) Tech. Paper 49 (1964)
Delaware, Illinois, Indiana, Kentucky, Maryland, New Jersey, North Carolina, Ohio, Pennsylvania, South Carolina, Tennessee, Virginia, West Virginia, and Washington, DC  NOAA Atlas 14, Volume 2 (June 2004)  NOAA Atlas 14, Volume 2 (June 2004)  NOAA Atlas 14, Volume 2 (June 2004) 
Remainder of the Eastern US Tech. Memo 35 (1977) Tech. Paper 40 (1961) Tech. Paper 49 (1964)
Hawaii (ongoing) Tech. Paper 43 (1962) Tech. Paper 43 (1962) Tech. Paper 51(1965)
Alaska Tech. Paper 47 (1963) Tech. Paper 47 (1963) Tech. Paper 52 (1965)
Puerto Rico (ongoing) Tech. Paper 42 (1961) Tech. Paper 42 (1961) Tech. Paper 53 (1965)
24
L-Moments
(Order Statistics)
X X1,X2,.Xn,.
L-1 E(X) L-21/2 EX1,2-X2,2 L-31/3
EX1,3-2X2,3X3,3 L-41/4EX1,4-3X2,43X3,4X4,4

L-CV L-1/L-2
dispersion
symmetry
L-skew L-3/L-2
L-kurtosis L-4/L-2
peakedness
L-K
Discordant
L-S
Hoskings and Wallis (1997)
L-CV
25
(No Transcript)
26
(No Transcript)
27
Koutsoyannis (2004)
28
Atlas 14, Vol. 1, Version 3.2
29
Atlas 14, Vol. 1, Version 3.2
30
(No Transcript)
31
1000 yr, 24-hr rainfall
Atlas 14, Volume 2
32
Recognizing Changing Needs (independent of
climate change)
Design Criteria as a Process
Observations Number, Location, Accuracy
Raingauges Ground-based Radar Satellites GIS,
GPS,..
Methods L-Moments Hurst Statistics
Multifractals Interpolation Statistics

33
Montana
Atlas 2, 1973 - Classical Parrett, 1988
L-moments
34
2yr-6hr Precipitation
(1) 30.7 mm (2) 23.8 3 mm (R20.30)
(1) 30.0 mm (2) 27.7 2 mm (R20.59)
2
2
2
3
1
(1)
Atlas 2, 1973
(2) Parrett 1998
35
(No Transcript)
36
3day - PMP HMR51
140 CM
37
Pe1 15,000 years
Multifractals as Design Alternative to Probable
Maximum Precipitation Estimation with known Risk
Pe1 1 million years
Douglas and Barros, 2003
38
Multifractals, Scale and And Small Dam Design
Current Practice
Known Risk
Best Estimate
39
One measure does not fit all
Riding on climate variability in the
Mississippi Alluvial Plain (so far.)
500-year event 10s billion 100,000s people Weeks
warning
15,000-year event 100s billion 1000,000s
people Days warning

Dams
Urban Dev.
Katrina
0.2
0.005
1-2
0.0001
10

DouglasBarros 2003
40
Rethinking the Paradigm
Location, Location, Location
Climate cycle
Landscape Dynamics
Human Dimensions
Cascading Effects
Moving toward a comprehensive, nonlinear,
adaptive framework for
risk assessment