Watershed level Risk Analysis with HEC-WAT 01 May 2019 Rockville, MD Will Lehman, Economist Hydrologic Engineering Center Institute for Water Resources US Army Corps of Engineers BUILDING STRONG ® www.hec.usace.army.mil 1
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
BUILDING STRONG®
Risk Management
Center
Watershed level Risk Analysis with HEC-WAT
01 May 2019Rockville, MD
Will Lehman, EconomistHydrologic Engineering CenterInstitute for Water Resources
US Army Corps of EngineersBUILDING STRONG®www.hec.usace.army.mil 1
BUILDING STRONG®
What is HEC-WAT?
■ Provides a plug-in architecture to allow other computational models to be computed in the program sequence
■ Integrates HEC-HMS, HEC-ResSim, HEC-RAS and HEC-FIA models, eliminating manual data exchange.
■ Supports systems and watershed-based studies.
■ Supports risk and uncertainty evaluations.
2
BUILDING STRONG®
What is HEC-WAT?
■ Provides a plug-in architecture to allow other computational models to be computed in the program sequence
■ Integrates HEC-HMS, HEC-ResSim, HEC-RAS and HEC-FIA models, eliminating manual data exchange.
■ Supports systems and watershed-based studies.
■ Supports risk and uncertainty evaluations.
HEC- WAT INTEGRATES
3
BUILDING STRONG®
What is HEC-WAT?
■ Provides a plug-in architecture to allow other computational models to be computed in the program sequence
■ Integrates HEC-HMS, HEC-ResSim, HEC-RAS and HEC-FIA models, eliminating manual data exchange.
■ Supports systems and watershed-based studies.
■ Supports risk and uncertainty evaluations.
HEC- WAT INTEGRATES
HEC- WAT Facilitates Evaluation
4
BUILDING STRONG®
HEC- WAT INTEGRATES
5
BUILDING STRONG®
Plug-in Architecture
HEC-WAT Manages the computes through plug-ins Plug-ins interact
with each other through a centralized database
6
Database
1. Compute
7. Read Input
8. Write Output
Repeat!
BUILDING STRONG®
HEC-WAT Model Integration Models and tools that implement the Plugin Interface can contribute to the
computational process
● Hydrology - HEC-HMS ● Reservoirs - HEC-ResSim● Hydraulics - HEC-RAS ● Economics - HEC-FIA
Communication is defined by the Plugin API and facilitated by HEC-WAT
Data is transferred through a common DSS file.
HEC-WAT Framework
HEC-WATSimulation
With DefaultProgram
Order
HEC-ResSim Plug-In
HEC-RAS Plug-In
HEC-HMS HEC-ResSim HEC-RAS HEC-FIA
HEC-HMS Plug-In
HEC-FIA Plug-In
Model Results (simulation.dss)
7
BUILDING STRONG®
Load Distribution & Consequences
Yakima River
KennewickBurbank
Pasco
Richland
WestPasco
Ice Harbor
Priest Rapids
Li
Cn
Cn+1
Oct Dec Feb Apr Jun Aug Oct1947 1948
10
15
20
25
30
35
40
45
50Plan: BaseCondit:Event1948:RAS-Reach1 River: Columbia Reach: RM 143-Bonn RS: 145.86
Time
Sta
ge
(ft)
Legend
Stage
Qi
Qo
8
BUILDING STRONG®
HEC- WAT Facilitates Evaluation
9
BUILDING STRONG®
HEC-WAT Workflow
■ Import existing models or develop models from within HEC-WAT
■ Develop alternatives
■ Organize & store data
■ Edit models – accessed via plug-ins to view Native model interfaces
■ Run modeling software – via plug-ins
■ View and compare alternative results
10
BUILDING STRONG®
HEC-WAT Interface
Map Window
Content Pane
Message Pane
Study Pane
Display Pane
11
BUILDING STRONG®
12
Development of HEC-WAT Model
BUILDING STRONG®
13
ResSim Models
Development of HEC-WAT Model
BUILDING STRONG®
14
ResSim Models
FIA Models
Development of HEC-WAT Model
BUILDING STRONG®
15
ResSim Models
RAS Models
FIA Models
Development of HEC-WAT Model
BUILDING STRONG®
Model Linking
HEC-ResSim Links
HEC-RAS Links
16
BUILDING STRONG®
Hydrologic Modeling (HEC-HMS)
Reservoir Analysis (HEC-ResSim)
River Hydraulics (HEC-RAS)
Consequence Analysis (HEC-FIA)
Deterministic Compute
■ Single Flood Event● Example: 8 January 1986
to 13 January 1986● Simplest type of compute● Eliminates manual
handoffs between models
■ Period of Record● Example: 1 October 1943
to 30 September 2014● Slightly more complex
compute
17
BUILDING STRONG®
FRA Simulations ■ FRA simulations uses a Monte
Carlo style compute to support risk analyses.
■ Individual applications sample model parameters from a range of values to capture uncertainty.
■ Natural variability and knowledge uncertainty sampled separately.
■ Maintains consistency between alternatives by allowing use of same initial seeds.
18
BUILDING STRONG®
How do we capture a distribution of uncertainty in EAD?
Nested Monte Carlo: A. Sample instances of natural variabilities as flood
events, with enough events to capture the distribution of damage
B. Sample instances of knowledge uncertainties in model parameters to get their impact on the damage distribution
inner loop A varies natural variabilities, computes EAD
A Bouter loop B varies knowledge uncertainty, computes EAD distribution
1 outer loop B = a realization
HEC-WAT/FRA
19
BUILDING STRONG®
Exceedance Probability1 0
Sampling Variability and UncertaintyNested Monte Carlo Simulation sample new
frequency curve (uncertainty)
sample uncertain model parameters
Peak
Flow
(cfs)
CDF
20
BUILDING STRONG®
Exceedance Probability1 0
Sampling Variability and Uncertainty
One Event: sample member i
Nested Monte Carlo Simulation
sample variabilities
sample new frequency curve (uncertainty)
sample uncertain model parameters
Peak
Flow
(cfs)
CDF and then sample events (variability)
21
BUILDING STRONG®
Exceedance Probability1 0
Sampling Variability and Uncertainty
One Event: sample member i
Nested Monte Carlo Simulation
sample variabilities
sample new frequency curve (uncertainty)
sample uncertain model parameters
Peak
Flow
(cfs)
CDF and then sample events (variability)
22
BUILDING STRONG®
Exceedance Probability1 0
Sampling Variability and Uncertainty
One Event: sample member i
Nested Monte Carlo Simulation
sample variabilities
sample new frequency curve (uncertainty)
sample uncertain model parameters
Peak
Flow
(cfs)
CDF and then sample events (variability)
23
BUILDING STRONG®
Exceedance Probability1 0
Sampling Variability and Uncertainty
One Event: sample member i
Nested Monte Carlo Simulation
sample variabilities
sample new frequency curve (uncertainty)
sample uncertain model parameters
For each realization, get an EAD estimate:
…still end with a sample of damages
One Realization
Peak
Flow
(cfs)
CDF and then sample events (variability)
24
BUILDING STRONG®
Exceedance Probability1 0
Sampling Variability and Uncertainty
One Event: sample member i
Nested Monte Carlo Simulation
sample variabilities
sample new frequency curve (uncertainty)
sample uncertain model parameters
For each realization, get an EAD estimate:
…still end with a sample of damages
One Realization
Peak
Flow
(cfs)
CDF
After repeating for many realizations:
and then sample events (variability)
25
BUILDING STRONG®
0
0.1
0.2
0.3
0.4
0.5
0.6
025
0,00
050
0,00
075
0,00
01,
000,
000
1,25
0,00
01,
500,
000
1,75
0,00
02,
000,
000
2,25
0,00
02,
500,
000
2,75
0,00
03,
000,
000
3,25
0,00
03,
500,
000
3,75
0,00
04,
000,
000
4,25
0,00
04,
500,
000
4,75
0,00
05,
000,
000
5,25
0,00
05,
500,
000
5,75
0,00
06,
000,
000
Rela
tive
Freq
uenc
y
Annual Damage ($)
sample of annual damage from one realization(spans natural variability)provides 1 estimate of EAD
mean = average = EAD
Annual Damages
A
26
BUILDING STRONG®
0
0.1
0.2
0.3
0.4
0.5
0.6
025
0,00
050
0,00
075
0,00
01,
000,
000
1,25
0,00
01,
500,
000
1,75
0,00
02,
000,
000
2,25
0,00
02,
500,
000
2,75
0,00
03,
000,
000
3,25
0,00
03,
500,
000
3,75
0,00
04,
000,
000
4,25
0,00
04,
500,
000
4,75
0,00
05,
000,
000
5,25
0,00
05,
500,
000
5,75
0,00
06,
000,
000
Rela
tive
Freq
uenc
y
Annual Damage ($)
0
5
10
15
20
25
30
35
40
025
0,00
050
0,00
075
0,00
01,
000,
000
1,25
0,00
01,
500,
000
1,75
0,00
02,
000,
000
2,25
0,00
02,
500,
000
2,75
0,00
03,
000,
000
3,25
0,00
03,
500,
000
3,75
0,00
04,
000,
000
4,25
0,00
04,
500,
000
4,75
0,00
05,
000,
000
5,25
0,00
05,
500,
000
5,75
0,00
06,
000,
000
Rela
tive
Freq
uenc
y
Average Damage (EAD) $
100 realizations
sample of annual damage from one realization(spans natural variability)provides 1 estimate of EAD
mean = average = EAD
Annual Damages
A B
27
BUILDING STRONG®
0
0.1
0.2
0.3
0.4
0.5
0.6
025
0,00
050
0,00
075
0,00
01,
000,
000
1,25
0,00
01,
500,
000
1,75
0,00
02,
000,
000
2,25
0,00
02,
500,
000
2,75
0,00
03,
000,
000
3,25
0,00
03,
500,
000
3,75
0,00
04,
000,
000
4,25
0,00
04,
500,
000
4,75
0,00
05,
000,
000
5,25
0,00
05,
500,
000
5,75
0,00
06,
000,
000
Rela
tive
Freq
uenc
y
Annual Damage ($)
0
5
10
15
20
25
30
35
40
025
0,00
050
0,00
075
0,00
01,
000,
000
1,25
0,00
01,
500,
000
1,75
0,00
02,
000,
000
2,25
0,00
02,
500,
000
2,75
0,00
03,
000,
000
3,25
0,00
03,
500,
000
3,75
0,00
04,
000,
000
4,25
0,00
04,
500,
000
4,75
0,00
05,
000,
000
5,25
0,00
05,
500,
000
5,75
0,00
06,
000,
000
Rela
tive
Freq
uenc
y
Average Damage (EAD) $
100 realizations
sample of annual damage from one realization(spans natural variability)provides 1 estimate of EAD
mean = average = EAD
sample of mean damage (EAD) from all realizations(spans knowledge uncertainty)provides distribution of EAD
Annual Damages
A B
28
BUILDING STRONG®
0
0.1
0.2
0.3
0.4
0.5
0.6
025
0,00
050
0,00
075
0,00
01,
000,
000
1,25
0,00
01,
500,
000
1,75
0,00
02,
000,
000
2,25
0,00
02,
500,
000
2,75
0,00
03,
000,
000
3,25
0,00
03,
500,
000
3,75
0,00
04,
000,
000
4,25
0,00
04,
500,
000
4,75
0,00
05,
000,
000
5,25
0,00
05,
500,
000
5,75
0,00
06,
000,
000
Rela
tive
Freq
uenc
y
Annual Damage ($)
0
5
10
15
20
25
30
35
40
025
0,00
050
0,00
075
0,00
01,
000,
000
1,25
0,00
01,
500,
000
1,75
0,00
02,
000,
000
2,25
0,00
02,
500,
000
2,75
0,00
03,
000,
000
3,25
0,00
03,
500,
000
3,75
0,00
04,
000,
000
4,25
0,00
04,
500,
000
4,75
0,00
05,
000,
000
5,25
0,00
05,
500,
000
5,75
0,00
06,
000,
000
Rela
tive
Freq
uenc
y
Average Damage (EAD) $
100 realizations
sample of annual damage from one realization(spans natural variability)provides 1 estimate of EAD
mean = average = EAD
sample of mean damage (EAD) from all realizations(spans knowledge uncertainty)provides distribution of EAD
Annual Damages
EAD estimates
A B
29
BUILDING STRONG®
.5.99 .95 .9 .75 .25 0.1 .05 .011000
10000
100000
1000000
est. probability of exceedance
peak
ann
ual f
low
(cfs
)
Curves estimated from 30 data points
Uncertainty in a frequency curve estimated from 30 years
of data
30
BUILDING STRONG®31
.5.99 .95 .9 .75 .25 0.1 .05 .011000
10000
100000
1000000
est. probability of exceedance
peak
ann
ual f
low
(cfs
)
Curves estimated from 60 data points
Uncertainty in a frequency curve estimated from 60 years
of data
BUILDING STRONG®
Concept behind event sampling
1 0Exceedance probability
1 0Exceedance probabilitystage
stageDamage Damage
FlowFlow
32
BUILDING STRONG®
Concept behind event sampling
1 0Exceedance probability
1 0Exceedance probabilitystage
stageDamage Damage
FlowFlow
33
BUILDING STRONG®
Concept behind event sampling
1 0Exceedance probability
1 0Exceedance probabilitystage
stageDamage Damage
FlowFlow
34
BUILDING STRONG®
Concept behind event sampling
1 0Exceedance probability
1 0Exceedance probabilitystage
stageDamage Damage
FlowFlow
35
BUILDING STRONG®
Concept behind event sampling
1 0Exceedance probability
1 0Exceedance probabilitystage
stageDamage Damage
FlowFlow
This results in one event. Rinse and repeat.
36
BUILDING STRONG®
Damages
Number of events
Average of event
damages
EVENT SAMPLING
Annual exceedance probability (for the hydrology)01
Running Average
DAMAGES
COUNT
37
BUILDING STRONG®
Damages
. ..
.Number of events
Average of event
damages
EVENT SAMPLING
Annual exceedance probability (for the hydrology)01
Running Average
DAMAGES
COUNT
38
BUILDING STRONG®
Damages
. ...
.Number of events
Average of event
damages
EVENT SAMPLING
Annual exceedance probability (for the hydrology)01
Running Average
DAMAGES
COUNT
39
BUILDING STRONG®
Damages
. ...
. .. ......
.Number of events
Average of event
damages
EVENT SAMPLING
Annual exceedance probability (for the hydrology)01
Running Average
DAMAGES
COUNT
40
BUILDING STRONG®
Damages
. ...
. .. ....... ..
..
....
.Number of events
Average of event
damages
EVENT SAMPLING
Annual exceedance probability (for the hydrology)01
Running Average
DAMAGES
COUNT
41
BUILDING STRONG®
Damages
. ...
. .. ....... ..
..
..... ... . ...
.Number of events
Average of event
damages
EVENT SAMPLING
Annual exceedance probability (for the hydrology)01
Running Average
DAMAGES
COUNT
42
BUILDING STRONG®
Damages
. ...
. .. ....... ..
..
..... ... . ...... .. ....
.Number of events
Average of event
damages
EVENT SAMPLING
Annual exceedance probability (for the hydrology)01
Running Average
DAMAGES
COUNT
43
BUILDING STRONG®
Damages
. ...
. .. ....... ..
..
..... ... . ...... .. ...... .... .
.Number of events
Average of event
damages
EVENT SAMPLING
Annual exceedance probability (for the hydrology)01
Running Average
DAMAGES
COUNT
44
BUILDING STRONG®
Damages
. ...
. .. ....... ..
..
..... ... . ...... .. ...... .... .
... . ...
.Number of events
Average of event
damages
EVENT SAMPLING
Annual exceedance probability (for the hydrology)01
Running Average
DAMAGES
COUNT
45
BUILDING STRONG®
Damages
. ...
. .. ....... ..
..
..... ... . ...... .. ...... .... .
... . ...
.Number of events
Average of event
damages
EVENT SAMPLING
Annual exceedance probability (for the hydrology)01
Running Average
DAMAGES
COUNT
46
BUILDING STRONG®
Damages
. ...
. .. ....... ..
..
..... ... . ...... .. ...... .... .
... . ...
.Number of events
Average of event
damages
EVENT SAMPLING
EAD
Annual exceedance probability (for the hydrology)01
Running Average
DAMAGES
COUNT
47
BUILDING STRONG®
Damages
. ...
. .. ....... ..
..
..... ... . ...... .. ...... .... .
... . ...
.Number of events
Average of event
damages
EVENT SAMPLING
EAD
Annual exceedance probability (for the hydrology)01
Running Average
^
DAMAGES
COUNT Distribution of Annual
Damages
EAD^
48
BUILDING STRONG®
American River Insurance Study For stability a 2D channel was linked to a 2D overbank area
through a 2D SA connection. Two connections were set to breach Model development took approximately 8 hours
..
49
BUILDING STRONG®
Hydrologic Sampler Events
50
BUILDING STRONG®
Hydrologic Sampler Events
51
BUILDING STRONG®
Hydrologic Sampler Events
52
BUILDING STRONG®
Hydrologic Sampler Events
53
BUILDING STRONG®
54
BUILDING STRONG®
HEC-RAS output by event
55
BUILDING STRONG®
AEP Per Structure
56
BUILDING STRONG®
EAD Per Structure
57
BUILDING STRONG®
58
BUILDING STRONG®
59
BUILDING STRONG®
Conclusion
■ HEC-WAT/FRA is a planning and evaluation tool that conducts risk assessments in a systems context.
■ It includes systems approaches, event sampling, alternative analyses, structural and non-structural analyses, Life Loss, agricultural damage analyses.
■ Is being used nationwide for dam and levee evaluations and assessments, and planning and design studies.
60
BUILDING STRONG®
QUESTIONS?
US Army Corps of EngineersBUILDING STRONG®
www.hec.usace.army.mil
Related Documents
