Atkins Lectures 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
Daniel Medina, senior engineer, Water Resources 21 August 2013 2
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. 4.5 4.0 3.5 3.0 Rainfall depth (in) 2.5 Arapahoe, CO Miami, FL 2.0 Baltimore, MD 1.5 Quillayute, WA 1.0 0.5 0.0 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Percentile
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 Damages avoided = damages without GI – damages 8. 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: d 80 = 80 th 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 ● Physical damage estimating potential losses from ● Economic loss disasters ● Social impacts ● GIS-based – Shelter requirements – Displaced households – Population exposed to scenario floods, earthquakes and hurricanes.
Vulnerability curves Federal Insurance Vulnerability curve Administration (FIA) 30 USACE 25 Flooding depth (ft) 20 15 10 One-story house, no 5 basement 0 -5 -10 0% 20% 40% 60% Percent structural damage
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 Middle James, without GI Upper San Antonio, without GI Average Annualized Losses (AAL) = area under damage curve
Flood losses avoided 1,200 1,000 800 Damages (millions) Damages avoided 600 with GI Without GI 400 200 0 0 20 40 60 80 100 Return period (years)
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: No assets exist in the 2-year floodplain 1. No assets exist in the 5-year floodplain 2. masks No assets exist in the 10-year floodplain 3.
Zero-damage threshold
Zero-damage threshold
Distribution of avoided losses Year 2040 development (2006 dollars) Upper Chattahoochee HUC8 2,500 2,000 Damages (millions) Original Hazus estimate 1,500 No assets in the 2-year floodplain No assets in the 5-year floodplain No assets in the 10-year floodplain 1,000 500 0 0 20 40 60 80 100 Return period (years)
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) $100 01100004, Quinnipiac, New Haven, CT 2-year mask 02030201+02030202, Southern Long Island, Long Island, NY $90 02040205, Brandywine-Christina, Northern DE 03150201, Upper Alabama, Montgomery, AL $80 05120208, Lower East Fork White, Bloomington, IN 12040104, Buffalo-San Jacinto, Houston, TX $70 10190004, Clear, West Denver, CO Losses avoided (millions) $60 12080005, Johnson Draw, West Texas - Odessa, TX 12080007, Beals, West Texas - Big Spring, TX $50 10190003, Middle South Platte-Cherry Creek, East Denver, CO 05140205, Tradewater, West KY $40 16040101, Upper Humboldt, Northeast NV 02050306, Lower Susquehanna, North of Baltimore in PA $30 12090205, Austin-Travis Lakes, Austin, TX 02080201, Upper James, Southern WV $20 03130001, Upper Chattahoochee, Northeast of Atlanta, GA 04080203, Shiawassee, Near Flint, MI $10 07010102, Leech Lake, Northern MN 12100301, Upper San Antonio River, San Antonio, TX $0 0 20 40 60 80 100 02080205, Middle James River, Near Richmond, VA Return period (years)
Distribution of avoided losses Year 2040 development (2006 dollars) $100 01100004, Quinnipiac, New Haven, CT 5-year mask 02030201+02030202, Southern Long Island, Long Island, NY $90 02040205, Brandywine-Christina, Northern DE 03150201, Upper Alabama, Montgomery, AL $80 05120208, Lower East Fork White, Bloomington, IN 12040104, Buffalo-San Jacinto, Houston, TX $70 10190004, Clear, West Denver, CO Losses avoided (millions) $60 12080005, Johnson Draw, West Texas - Odessa, TX 12080007, Beals, West Texas - Big Spring, TX $50 10190003, Middle South Platte-Cherry Creek, East Denver, CO 05140205, Tradewater, West KY $40 16040101, Upper Humboldt, Northeast NV 02050306, Lower Susquehanna, North of Baltimore in PA $30 12090205, Austin-Travis Lakes, Austin, TX 02080201, Upper James, Southern WV $20 03130001, Upper Chattahoochee, Northeast of Atlanta, GA 04080203, Shiawassee, Near Flint, MI $10 07010102, Leech Lake, Northern MN 12100301, Upper San Antonio River, San Antonio, TX $0 0 20 40 60 80 100 02080205, Middle James River, Near Richmond, VA Return period (years)
Distribution of avoided losses Year 2040 development (2006 dollars) $100 01100004, Quinnipiac, New Haven, CT 10-year mask 02030201+02030202, Southern Long Island, Long Island, NY $90 02040205, Brandywine-Christina, Northern DE 03150201, Upper Alabama, Montgomery, AL $80 05120208, Lower East Fork White, Bloomington, IN 12040104, Buffalo-San Jacinto, Houston, TX $70 10190004, Clear, West Denver, CO Losses avoided (millions) $60 12080005, Johnson Draw, West Texas - Odessa, TX 12080007, Beals, West Texas - Big Spring, TX $50 10190003, Middle South Platte-Cherry Creek, East Denver, CO 05140205, Tradewater, West KY $40 16040101, Upper Humboldt, Northeast NV 02050306, Lower Susquehanna, North of Baltimore in PA $30 12090205, Austin-Travis Lakes, Austin, TX 02080201, Upper James, Southern WV $20 03130001, Upper Chattahoochee, Northeast of Atlanta, GA 04080203, Shiawassee, Near Flint, MI $10 07010102, Leech Lake, Northern MN 12100301, Upper San Antonio River, San Antonio, TX $0 0 20 40 60 80 100 02080205, Middle James River, Near Richmond, VA Return period (years)
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 AALA 2040 E 0.45 A N + A R 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
Recommend
More recommend
