Model specifications for place-specific flood regulation models developed for the Puget Sound case study
Flood source: Precipitation and snowmelt, which can cause floods.
Flood sink: Areas that absorb, detain, or promote infiltration of floodwater, including vegetation and soils that can promote infiltration and evapotranspiration (green infrastructure) and dams and detention basins (gray infrastructure) that can detain flood waters.
Flood beneficiaries: Beneficiaries of flood regulation include residents, farmers, and users of public infrastructure in flood zones.
Flood flow: Flood flows are routed across the landscape through topography and stream networks. Once floodwater is in a stream, it can overtop the streambanks, depending on the amount of floodwater, floodplain width, and the presence of levees. Flood damage can be attributed to upstream flood sources, and mitigated damage can be attributed to upstream flood sinks, which provide the ecosystem service of flood regulation.
Model structure and assumptions
Case studies 1. Western Washington State
Flood source models. We use annual precipitation is the source of floodwater. Flood regulation is For event-based flood modeling, snowmelt is an extremely important variable in seasonally cold-weather climates, such as Western Washington. Since data limitations prevent event-based flood modeling in ARIES, snow presence and snowmelt are not currently included in the flood source model.
Flood sink models. Past ecosystem services studies have essentially mapped flood sinks using spatial data; we drew on these approaches in developing our sink models. Eade and Moran (1996) mapped flood regulation based on soil drainage classifications, while Chan et al. (2006) did so by estimating percent natural land cover, percent natural land cover within riparian zones, distance to the 100-year floodplain, percent agricultural land, and housing units in the 100-year floodplain. Finally, Boyd and Wainger (2003) mapped flood regulation using spatial data including floodplain locations, housing and commercial units and value, percent floodplain as impervious and wetland. Boyd and Wainger also included an environmental justice component to their measures, by mapping median income and percent black or Hispanic populations within their impacted area. We drew on these studies to conceptualize our flood sink models, then extended these approaches by including additional variables considered by others to be important for flood regulation (Eade and Moran 1996, Boyd and Wainger 2003, Chan et al. 2006, Bradshaw et al. 2007). We defined flood sink value, the top-level output of the sink model, as the sum of green infrastructure storage (the sum of infiltration, absorption, detention, or evapotranspiration of potential flood waters by vegetation, soils, and floodplains) and gray infrastructure storage (the sum of storage in detention basins and reservoirs). Both gray and green infrastructure can be “saturated” when their individual components are at full capacity. Because of this, we added the mean days of precipitation per year as an influence to green and gray infrastructure storage in the Western Washington model. This accounts for the fact that green and gray infrastructure are likely to be saturated for more of the year in regions where precipitation is more evenly distributed over the course of a year, allowing soil moisture to remain more temporally uniform. By computing the difference between precipitation and runoff (which accounts for vegetation and soil characteristics), we can estimate the contribution of green infrastructure to flood mitigation. We can thus use the difference between precipitation and runoff as training data for the Bayesian network. Models such as the Curve Number method (CN, SCS 1972), which incorporates data on precipitation, hydrologic soils group, and land use-land cover, can also be used to calculate runoff. We set soil infiltration as a function of impervious surface cover, slope, and hydrologic soils group. These variables have been routinely recognized as predictive variables for potential soil infiltration (i.e., USACE 1998, Tetra Tech, Inc. 2005, Laton et al. 2006, BOR 2007). We considered adding water table depth (available from SSURGO/STATSGO data) as an influence on infiltration but ultimately decided not to include it to maintain tractability in the contingent probability table. Evapotranspiration reduces soil moisture, thereby allowing increased infiltration. In addition, it serves as a proxy for other flood mitigation processes due to the presence of vegetation. We set evapotranspiration as a function of percent tree canopy cover and vegetation type (in both models) and added influences for successional stage and vegetation height for the Western Washington model. Jones and Post (2004) and Moore and Wondzell (2005) note the importance of forest cover and successional stage as drivers of hydrologic processes in Pacific Northwest forests. We discretized mean days of precipitation per year using Jenks natural breaks and estimated its priors on a review of the data for Western Washington. We reviewed GIS data to derive priors for impervious surface cover, slope, and hydrologic soils group. We discretized impervious surface cover to account for ecological thresholds typically present when impervious surface exceeds 10% (Booth and Jackson 1997). We used equal intervals to discretize percent tree canopy cover and Jenks natural breaks to discretize vegetation height for Western Washington. We estimated priors based on spatial data for percent tree canopy cover, successional stage, vegetation height, and vegetation type. For the soil infiltration contingent probability table, we set the highest values of infiltration at low impervious surface cover and slope and hydrologic soils groups A and B. We set the lowest values for infiltration under opposite conditions and interpolated intermediate values. We set the evapotranspiration contingent probability table to its greatest values in cases of greater percent tree canopy cover, later successional stage, tall vegetation (where applicable), and wetlands, and vice versa, and interpolated intermediate values. We set evapotranspiration as slightly lower than wetlands for forests, grassland, and shrubland, and substantially lower for developed and cultivated land use types. We set evapotranspiration and soil infiltration as equivalent influences on the green infrastructure storage contingent probability table. In the Western Washington model, we set mean days of precipitation per year as a strong influence on the contingent probability tables for both gray and green infrastructure storage (i.e., much greater storage when there were very low or low mean days of precipitation per year, and vice versa). We summed values for dam and detention basin storage to quantify gray infrastructure storage, and added this to the value of green infrastructure storage to estimate the total flood sink.
Flood use models. Beneficiaries of flood regulation can be mapped using spatial data and simple GIS overlay operations, eliminating the need for more complex approaches. In these case studies, we identified different beneficiary classes, including farmers, residents, and municipalities with public infrastructure located within the floodplain boundaries. We mapped beneficiaries in both the 100-year and 500-year floodplains in order to differentiate between levels of risk from catastrophic floods of different sizes.
Flood flow models. The source and sink models determine the quantity (in mm/yr) of precipitation falling on the landscape and absorbed or detained by the landscape, while the use model defines the location of potential flood regulation beneficiaries. The flow model routes water from its source locations through the watershed based on the topography of the location. Once the flow of water moving across a landscape intersects a stream, its movement is no longer determined by topography and instead follows the direction of the streambed. Once floodwater is in a stream, it can overtop the streambanks, depending on the amount of floodwater, floodplain width, and the presence of levees. If the downstream flow reaches a dam, floodwater is temporarily detained unless excess water in an already-full reservoir must be released downstream. If floodwater reaches a user, it causes damage. This damage can be attributed to upstream flood sources, and mitigated damage can be attributed to upstream flood sinks, which provide the ecosystem service of flood regulation. Key outputs from the flow models include: 1. Potentially damaging flood flow: The flow route of floodwater across the landscape in the absence of sinks. 2. Potentially damaging runoff: Runoff capable of harming people or damaging property when accounting for flow paths but not sinks. 3. Potential flood damage received: People and property receiving damage when accounting for sources of floodwater and its flow path but not accounting for the action of sinks that reduce potential damage from floodwater. 4. Actual flood flow: The flow route of floodwater across the landscape in the presence of sinks. 5. Flood damaging runoff: Runoff that actually harms people or damages property when accounting for flow paths and sinks. 6. Utilized runoff mitigation: Sinks that actively reduce floodwater, providing the benefit of reduced flood damage for people. 7. Flood damage received: Actual damage received by people and property when accounting for sources of floodwater, flow paths, and sinks encountered. 8. Absorbed flood flow: Flood flows that are absorbed by sinks prior to reaching vulnerable human beneficiaries. 9. Flood mitigated runoff: The portion of the total runoff that is absorbed, detained, or slowed by the action of flood sinks. 10. Flood mitigation benefits accrued: People or economically valuable assets who are spared from flood damage due to the flood regulation activity of sinks.
|Models||Data theme||Source||Spatial extent||Spatial resolution||Year|
|Source – All models||Annual precipitation||PRISM/Oregon State Univ.||United States||800 m||1971-2000|
|Sink – All models||Average annual actual evapotranspiration||SAGE/Univ. of Wisconsin||Global||0.5 degree||1950-1999|
|Average annual runoff||SAGE/Univ. of Wisconsin||Global||0.5 degree||1955-1990|
|Dam storage||National Atlas of the United States||United States||Rasterized point data||2006|
|Hydrologic soils group||SSURGO & STATSGO soil data||United States||Rasterized polygon data at 30 m||n/a|
|Impervious surface cover||NLCD 2001||United States||30 m||2001|
|Slope||Derived from global SRTM data||Global||90 m||n/a|
|Tree canopy cover||NLCD 2001||United States||30 m||2001|
|Sink – Western WA||Vegetation type||NLCD 2001||United States||30 m||2001|
|Detention basins||County GIS offices||King, Pierce, San Juan counties||Rasterized point data||Varies|
|Mean days of precipitation per year||PRISM/Oregon State Univ.||United States||Vector polygon data||1971-2000|
|Successional stage||BLM/Interagency Vegetation Mapping Project||Western Washington & Oregon||25 m||1996|
|Vegetation height||Puget Sound LIDAR Consortium||Parts of Western Washington||30 m||2000-2006|
|Use – All models||Farmland||NLCD 2001||United States||30 m||2001|
|Floodplain extents||FEMA Q3 Flood Data||United States||Vector polygon data||Varies|
|Highways||TIGER/Line files||United States||Vector line data||2000|
|Railways||TIGER/Line files||United States||Vector line data||2000|
|Use – Western Washington||Presence of housing||County assessors’ offices||Clallam, Grays Harbor, Jefferson, King, Kitsap, Mason, Snohomish, Thurston cos., WA||Rasterized point data at 100 m||2004 (Kitsap Co.), 2006 (King Co.); uncertain for others|
|Flow – All models||Dams||National Atlas of the United States||United States||Rasterized point data||2006|
|Floodplain extents||FEMA Q3 Flood Data||United States||Vector polygon data||Varies|
|Flow – Western Washington||Hydrography||Washington DNR||Washington State||Vector line data||n/a|
|Levees||County GIS offices||King, Lewis, Pierce counties||Vector line data||Varies|
Booth, D.B. and C.R. Jackson. 1997. Urbanization of aquatic systems: Degradation thresholds, stormwater detection, and the limits of mitigation. Journal of the American Water Resources Association 33 (5): 1077-1090.
Boyd, J and L. Wainger. 2003. Measuring ecosystem service benefits: The use of landscape analysis to evaluate environmental trades and compensation. Discussion Paper 02-63, Resources for the Future: Washington, DC.
Bradshaw, C.J.A., et al. 2007. Global evidence that deforestation amplifies flood risk and severity in the developing world. Global Change Biology 13: 2379-2395.
Chan, K.M.A., et al. 2006. Conservation planning for ecosystem services. PLOS Biology 4 (11): 2138-2152.
Eade, J.D.O. and D. Moran. 1996. Spatial economic valuation: Benefits transfer using geographical information systems. Journal of Environmental Management 48: 97-110.
Jones, J.A. and D.A. Post. 2004. Seasonal and successional streamflow response to forest cutting and regrowth in the northwest and eastern United States. Water Resources Research 40: W052031-W0520319.
Laton, W., et al. 2006. Estimating runoff quantities for flow and volume- based BMP design. Journal of the American Institute of Hydrology 22 (104): 131-144.
Moore, R.D. and S.M. Wondzell. 2005. Physical hydrology and the effects of forest harvesting in the Pacific Northwest: A review. Journal of the American Water Resources Association 41 (4): 763-784.
Soil Conservation Service (SCS). 1972. National Engineering Handbook, Section 4, Hydrology. SCS: Washington, DC.
U.S. Army Corps of Engineers (USACE). 1998. HEC-1 Flood Hydrograph Package User’s Manual. CPD 1-A (Version 4.1). Hydrologic Engineering Center. Davis, CA.
Acknowledgements and additional contributors
Dave Batker, Jim Pittman, and Paula Swedeen provided data and model review for the Western Washington case study.