Nationwide study of the benefits of green infrastructure .../media/Files/A/...Region of Influence...
Transcript of Nationwide study of the benefits of green infrastructure .../media/Files/A/...Region of Influence...
Nationwide study of the benefits of green
infrastructure (GI) for flood loss avoidance
in the United States
2013 International Low Impact Development Symposium, Minnesota, USA
Atkins Lectures
2
Daniel Medina, senior engineer,
Water Resources 21 August 2013
Background
Objective: Estimate flood losses avoided by the implementation of green infrastructure (GI) for new development and redevelopment.
Upcoming stormwater rule proposal in the USA:
● Based on GI
● Capture and retain on site a high percentile storm
● Example capture standard:
– 90th percentile for new development
– 85th percentile for redevelopment
● Assumed to start in 2020; snapshot at 2040
● Environmental Protection Agency is not proposing GI for flood control; these are side-benefits to water quality benefits.
Retention standard definition
Xth percentile storm: The event in which precipitation depth is greater than or equal to X% of all storm events over a given period of record.
The retained volume must be infiltrated, evapotranspired or harvested for beneficial use.
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Ra
infa
ll d
ep
th (
in)
Percentile
Arapahoe, CO
Miami, FL
Baltimore, MD
Quillayute, WA
Study plan
Rationale: smaller run-off volume leads to smaller floodplains and thus fewer flood damages
Evaluate 20 Hydrologic Unit Code (HUC) 8 watersheds with and without GI-based retention
Estimate monetary flood losses for each scenario
Benefits = losses without GI – losses with GI
Scale results nationwide.
Methodology overview
Sample of twenty HUC8 watersheds
Publicly available datasets
Stream gauge hydrology
Hydraulic modelling
Hazus damage estimation
Nationwide scale-up.
Sample watersheds
Datasets
US Geological Society (USGS) streamflow records
USGS National Elevation Dataset (NED) (10-metres)
National Hydrography Dataset (NHD) Plus hydrography
National Land Cover Dataset (NLCD)
STATSGO2 soils
Census 2000 economic activity
Integrated climate and land use scenarios economic growth projections.
Procedure
1. Peak flow distribution from USGS gauge data
2. Adjust peak flows to 2040 hydrology without GI
3. Hydraulic modelling to estimate flood depths without GI
4. Estimate monetary damages without GI
5. Adjust peak flows to 2040 hydrology with GI
6. Hydraulic modelling to estimate flood depths with GI
7. Estimate monetary damages with GI
8. Damages avoided = damages without GI – damages with GI.
Hydrology
Flood frequency analysis with USGS’s PeakFQ software
Region of Influence (RoI) technique for spatial interpolation of peak flows (Eng et al., 2005)
Peak flows at any location in HUC8, existing conditions
Estimate future run-off volumes to approximate future peak flows, with and without GI.
Estimation of future hydrology
Use run-off volume ratios to adjust peak flows (MMSD, 2005)
Run-off volume from TR-55 methodology (curve number)
Future conditions (2040), no GI
Future conditions, with GI
Example: d80 = 80th percentile depth
Hydraulic modelling
Rapid Flood Delineation (RFD) model
As accurate as HEC-RAS
High speed hydraulic profile calculation (6,000 miles per CPU hour)
Automatic cross sections
Depth grids
Flood damage estimation
● FEMA’s methodology for estimating potential losses from disasters
● GIS-based
● Physical damage
● Economic loss
● Social impacts
– Shelter requirements
– Displaced households
– Population exposed to scenario floods, earthquakes and hurricanes.
Vulnerability curves
Federal Insurance Administration (FIA)
USACE
-10
-5
0
5
10
15
20
25
30
0% 20% 40% 60%
Flo
od
ing
dep
th (
ft)
Percent structural damage
Vulnerability curve
One-story house, no basement
Flood damage computation
Hazus uses General Building Stock (GBS)
Assumes uniformly distributed assets on Census blocks
Results:
Sample damage distribution
Middle James, without GI
Upper San Antonio, without GI
Sample damage distribution
Average
Annualized
Losses (AAL) =
area under
damage curve
Middle James, without GI
Upper San Antonio, without GI
Flood losses avoided
0
200
400
600
800
1,000
1,200
0 20 40 60 80 100
Dam
age
s (m
illi
on
s)
Return period (years)
with GI
Without GI
Damages avoided
Zero-damage threshold
Damages begin to occur when:
● Flood waters enter the floodplain, and
● Water reaches exposed assets.
Zero-damage threshold
GBS uniform distribution of assets on Census blocks:
● Some assets appear at risk when they are not
● Damages can be overestimated.
Zero-damage threshold
Flood event at which damages begin to occur:
1. No assets exist in the 2-year floodplain
2. No assets exist in the 5-year floodplain
3. No assets exist in the 10-year floodplain
masks
Zero-damage threshold
Zero-damage threshold
Distribution of avoided losses Year 2040 development (2006 dollars)
0
500
1,000
1,500
2,000
2,500
0 20 40 60 80 100
Dam
age
s (m
illio
ns)
Return period (years)
Upper Chattahoochee HUC8
Original Hazus estimate
No assets in the 2-year floodplain
No assets in the 5-year floodplain
No assets in the 10-year floodplain
100-year event no GI
100-year event with GI
Two-year event no GI
Two-year event with GI
Distribution of avoided losses Year 2040 development (2006 dollars)
$0
$10
$20
$30
$40
$50
$60
$70
$80
$90
$100
0 20 40 60 80 100
Loss
es
avo
ide
d (
mill
ion
s)
Return period (years)
2-year mask01100004, Quinnipiac, New Haven, CT
02030201+02030202, Southern Long Island, Long Island, NY
02040205, Brandywine-Christina, Northern DE
03150201, Upper Alabama, Montgomery, AL
05120208, Lower East Fork White, Bloomington, IN
12040104, Buffalo-San Jacinto, Houston, TX
10190004, Clear, West Denver, CO
12080005, Johnson Draw, West Texas - Odessa, TX
12080007, Beals, West Texas - Big Spring, TX
10190003, Middle South Platte-Cherry Creek, East Denver, CO
05140205, Tradewater, West KY
16040101, Upper Humboldt, Northeast NV
02050306, Lower Susquehanna, North of Baltimore in PA
12090205, Austin-Travis Lakes, Austin, TX
02080201, Upper James, Southern WV
03130001, Upper Chattahoochee, Northeast of Atlanta, GA
04080203, Shiawassee, Near Flint, MI
07010102, Leech Lake, Northern MN
12100301, Upper San Antonio River, San Antonio, TX
02080205, Middle James River, Near Richmond, VA
$0
$10
$20
$30
$40
$50
$60
$70
$80
$90
$100
0 20 40 60 80 100
Loss
es
avo
ide
d (
mill
ion
s)
Return period (years)
5-year mask01100004, Quinnipiac, New Haven, CT
02030201+02030202, Southern Long Island, Long Island, NY
02040205, Brandywine-Christina, Northern DE
03150201, Upper Alabama, Montgomery, AL
05120208, Lower East Fork White, Bloomington, IN
12040104, Buffalo-San Jacinto, Houston, TX
10190004, Clear, West Denver, CO
12080005, Johnson Draw, West Texas - Odessa, TX
12080007, Beals, West Texas - Big Spring, TX
10190003, Middle South Platte-Cherry Creek, East Denver, CO
05140205, Tradewater, West KY
16040101, Upper Humboldt, Northeast NV
02050306, Lower Susquehanna, North of Baltimore in PA
12090205, Austin-Travis Lakes, Austin, TX
02080201, Upper James, Southern WV
03130001, Upper Chattahoochee, Northeast of Atlanta, GA
04080203, Shiawassee, Near Flint, MI
07010102, Leech Lake, Northern MN
12100301, Upper San Antonio River, San Antonio, TX
02080205, Middle James River, Near Richmond, VA
Distribution of avoided losses Year 2040 development (2006 dollars)
$0
$10
$20
$30
$40
$50
$60
$70
$80
$90
$100
0 20 40 60 80 100
Loss
es
avo
ide
d (
mill
ion
s)
Return period (years)
10-year mask01100004, Quinnipiac, New Haven, CT
02030201+02030202, Southern Long Island, Long Island, NY
02040205, Brandywine-Christina, Northern DE
03150201, Upper Alabama, Montgomery, AL
05120208, Lower East Fork White, Bloomington, IN
12040104, Buffalo-San Jacinto, Houston, TX
10190004, Clear, West Denver, CO
12080005, Johnson Draw, West Texas - Odessa, TX
12080007, Beals, West Texas - Big Spring, TX
10190003, Middle South Platte-Cherry Creek, East Denver, CO
05140205, Tradewater, West KY
16040101, Upper Humboldt, Northeast NV
02050306, Lower Susquehanna, North of Baltimore in PA
12090205, Austin-Travis Lakes, Austin, TX
02080201, Upper James, Southern WV
03130001, Upper Chattahoochee, Northeast of Atlanta, GA
04080203, Shiawassee, Near Flint, MI
07010102, Leech Lake, Northern MN
12100301, Upper San Antonio River, San Antonio, TX
02080205, Middle James River, Near Richmond, VA
Distribution of avoided losses Year 2040 development (2006 dollars)
Average annualized losses avoided
(AALA) AALA = AAL without GI – AAL with GI
Year 2040 development
Nationwide scale-up
Regression of AALA vs. watershed properties
● Exposure
● Climate
● Development forecast
Extrapolation to HUC8s not modelled
Relationship with watershed
properties
AALA2040
E
AN + AR 0.45
A
Losses avoided (two-year mask)
Avoided losses in 2040 = $730 million
Present value (2020-2040) = $5 billion
Losses avoided (five-year mask)
Avoided losses in 2040 = $330 million
Present value (2020-2040) = $2.3 billion
Losses avoided (10-year mask)
Avoided losses in 2040 = $110 million
Present value (2020-2040) = $0.8 billion
Validation tests
Diagnostic case studies, not calibration
Stream gauge approach vs. hydrologic modelling
Zero-damage threshold
NED terrain vs. LiDAR terrain
GBS vs. user-defined facilities (UDF)
Conclusions
When applied watershed wide, GI is effective at reducing
● Peak flows for large events
● Flood elevations
● Flood losses
Benefits can be quantified by the average annualized losses avoided (AALA)
GI is necessary for water quality/stream health
GI adds community resiliency and environmental protection
GI reduces future federal expenditure and protects existing federal investments in flood control
Need to improve messaging about the link between run-off volume reduction and flood loss avoidance.
For more information contact: Dan Medina
Celebrating 75 years of
design, engineering and
project management
excellence.
www.atkinsglobal.com