Kevin Bacon game: Flood modeling edition David Ford Consulting Engineers, Inc. Sacramento, CA California Water & Environmental Modeling Forum Oct. 2006
Feb 01, 2016
Kevin Bacon game: Flood modeling edition
David Ford Consulting Engineers, Inc.Sacramento, CA
California Water & Environmental Modeling Forum
Oct. 2006
Kevin Bacon game
According to Wikipedia (so it must be right, huh?)
The trivia game Six Degrees of Kevin Bacon is based on a variation of the concept of the small world phenomenon which states that any actor can be linked, through their film roles, to Kevin Bacon. The game requires a group of players to try to connect any film actor in history to Kevin Bacon as quickly as possible and in as few links as possible.
Kevin Bacon game example
Kevin Bacon game example
• Here is an example, using Elvis Presley:• Elvis Presley was in the movie Change of
Habit (1969) with Edward Asner. • Edward Asner was in the movie JFK (1991)
with Kevin Bacon.
Rules of the Kevin Bacon game: Flood modeling edition
• Rule 1: I made up the game, so I get to make up the rules.
• Rule 2: Forget the movies—no Hollywood here.• Rule 3: Link one flood model/study/modeler
with another flood model/study/modeler with rational linkages.
Round 1: These 2 studies
Comp. Study hydrologic analyses
• Highly regulated stream, so fitting frequency model to gaged data not appropriate.
• Common design storm approach (at right) wouldn’t work. [Where do you position the rainfall in a 26,000 sq mi watershed?]
• Alternative approach, documented by Hickey et al. in ASCE journal, relies on gaged flows, historical patterns, composite floodplain concept. [See John Hickey for reprint.]
Design storm of specified
probability
Calibrated rainfall-runoff-routing
model
Runoff peak of known probability
Comp. Study hydraulic analyses
• Hydraulic analyses use unsteady network model—the mother of all UNET models initially.
• Geometric data collected using digital terrain models and bathymetric surveys (2-ft contour lines.)
• Over 100 routing reaches and approximately 3,000 cross sections.
• Extensive use of advanced modeling features, including hydraulic storage areas, lateral weirs, flow diversions, levees, and bridges.
We love the Comp
Study H&H (mostly)
We love the Comp Study H&H (mostly)
Yuba-Feather basin
Oroville Dam
New Bullards Bar Dam
Yuba-Feather operation challenges
• Contributing watersheds large, with much unregulated flow.
• Rainfall and snowmelt runoff.
• Travel times long.• Dynamic natural +
engineered system.• Operations managed
by several agencies.
Benefit of forecast-coordinated operation
• Reservoirs interconnected, so decisions felt throughout system.
Oroville
New Bullards Bar
300,000 cfs capacity
Benefit of forecast-coordinated operation
• Reservoirs interconnected, so decisions felt throughout system.
• Greatest benefit when operations coordinated. For example, best use of 300 kcfs capacity at confluence considers current and future states of both reservoirs.
Oroville
New Bullards Bar
300,000 cfs capacity
Yuba-Feather F-CO
HEC-ResSimCorps' database
NWSRFS
HEC-ResSim
CDEC database
21
56
3
HEC-ResSim
10
9
8
7
11
4
Reservoir operators CDEC Web server
Corps' Web server
Corps' firewall
Synchronized configuration
database
Relationship of Comp Study and F-CO
• Corps of Engineers did extensive study of hydrology and hydraulics in Central Valley for the Comprehensive (Comp) Study.
• 2005 DWR study used design hydrographs from Comp Study to investigate flood operation options for Oroville and New Bullards Bar reservoirs.
• Findings from study supported concept of F-CO project efforts; models formed foundation for work.
Round 2: These 2 storage facilities
D05 watershed and channels
CAL. EXPO.
EDISON AVE
ET
HA
N W
AY
G ravity ou tlet
S lu ice g ate
D05 p on d
Pu m p
Pu m p d isch arg ep ip elin e
Lower Am erican R iver
Levee
In flow (ru n off)
A
AF lap g ate
How should we determine the 100-yr WSEL @ Howe + Northrup?
Use 100-yr design storm + 100-yr d/s boundary condition (BC)
Use 100-yr design storm + ? d/s BC
Use ? design storm + 100-yr d/s BC
All of the above
“Coincident frequency” analysis
• If statistically independent, use total probability theorem
• Pick a stage• Use eqn to get probability• Pick another stage• Repeat, repeat, repeat• Combine to get
stage-probabilityP ro b a b il i ty th a ts ta g e is w ith in
th is in te rva l= 0 .5 4
27.0 '22 .5 '
31 .0 '
36 .0 '
S tage
P robab ility o f nonexc eedenc e0.00 1.00
= 0 .2 2 = 0 .1 4 = 0 .1 0
))()(()( exterior
conditionsexterior
exteriorinteriorinterior stageFstagestageFstageF
Interior area elevation-frequency function (at pond)
LAR stage (feet) AEP
(1) 22.5 (2)
27 (3)
31 (4)
36 (5)
Total
(6)
0.50 25.37 27.29 27.95 27.95 25.76
0.20 26.00 28.87 32.03 32.26 28.31
0.10 27.27 30.54 32.26 32.41 30.54
0.04 28.34 32.06 32.36 32.65 32.18
0.02 29.51 32.21 32.41 32.83 32.39
0.01 30.68 32.29 32.58 32.98 32.53
0.005 31.66 32.35 32.76 33.12 32.68
0.002 32.17 32.48 32.92 33.62 32.91
Illustration of computation
LAR (exterior) stage (ft)
(1)
F(32.18| LAR stage)
(2)
F(LAR stage)(3)
F(32.18| LAR stage) * F(LAR
stage)(4)
22.5 0.002 0.54 0.001
27 0.02 0.22 0.004
31 0.13 0.14 0.018
36 0.20 0.10 0.020
sum = 0.04
Hydrology for map modernization for San Joaquin basin?
• Although study not yet completed, hydrology tools being developed.
• Can’t use typical simple river approach in lower reaches due to impact of SJ river. So what do we do?
New Melones Don Pedro McClure Millerton
San J oaquin R.
Stanislaus R. Tuolumne R. Merced R.
Round 2: D05 pond and New Melones Dam
• Stages in channel upstream of D05 pond affected by watershed runoff AND by pond elevation, which is affected by LAR stage.
• Need “coincident frequency analysis” to account properly for both upstream and downstream conditions if independent.
• Stages at some locations in channel downstream of New Melones affected by releases and San Joaquin River flow (stage), so need coincident frequency analysis here too.
Round 3: These 2 guys
Martis Dam PMF study
• Joe DeVries managed PMF study (for Corps) of Martis Dam.
• PMF=runoff from severe combo of meteorologic+ hydrologic conditions.
• Caused by PMP.• PMP in this watershed has snowmelt.
FERC snowmelt guidelines
• Compute using HEC-1 energy-budget model.• Requires
• Shortwave radiation• Dewpoint temperature• Temperature sequence• Wind speed sequence• Rainfall sequence• Area covered by snow at start of storm• Snowpack water equivalent• Snowmelt temperature• Change in temperature with elevation
Where do we get data to calibrate/ verify snowmelt model?
Our hero
Round 3: Joe DeVries and David Parker
• Joe DeVries managed PMF study of Martis Dam for Sacramento District of Corps.
• Martis Dam PMF inflow includes snowmelt runoff, so snow data + snowmelt model needed.
• Snow data stored by CDEC.• David Parker manages CDEC data
dissemination system.
What’s my point?
• Kevin Bacon linkages are a fact in flood modeling. Flood studies often related by goals, methods, data, analysts, results.
• We never really finish much. Flood studies may have a long legacy, so:• Do them good enough initially.• Take the time to document them.
• Share your toys. Flood models, carefully assembled, can be re-usable and scalable.