Model enhancements for urban runoff predictions in the South-West United States 2182

Authors: KENNEDY, J. - Us Geological Survey (USGS); Goodrich, David - Dave; Unkrich, Carl

Submitted to: Book Chapter
Publication Type: Book / Chapter
Publication Acceptance Date: 2/1/2013
Publication Date: 8/1/2013
Citation: Kennedy, J., Goodrich, D.C., Unkrich, C.L. 2013. Model enhancements for urban runoff predictions in the South-West United States. In: Blöschl, G., Sivapalan, M., & Wagener, T. (Eds.). Predictions in Ungauged Basins. Cambridge University Press, pp. 332-337.

Interpretive Summary: Growth and urbanization occurred rapidly in the American Southwest and is projected to exceed the growth of other regions of the United States in the future. Computer models used to predict the effects of urbanization on runoff typically account for the impervious areas (e.g. roads, roofs, driveways) but not the effect of changes in the soil’s ability to absorb rainfall in the constructed area (e.g. yards, common areas). In this study detailed hydrologic measurements collected in a residential development and adjacent natural watershed in southeast Arizona were used with a novel ARS watershed model (KINEROS2) to demonstrate that compaction of soils done for construction resulted in roughly a 50% decrease in the amount of water that can infiltrate into the soils. Because of this change, about 17% of the total runoff due to urbanization was caused by soil compaction with the remainder caused by constructed impervious areas. Overall, the urbanization resulted in a nearly 20-fold increase in runoff over the natural watershed. The study also validated the new KINEROS2 model “urban element” which simplifies how a sub-division can be represented in the watershed computer model. In recent years, the increase in runoff associated with urbanization in the Southwest has begun to be considered as a potentially renewable water source so it is important to accurately estimate the amount of this extra, manageable water.

Technical Abstract: Detailed hydrologic measurements collected in a small residential development and adjacent natural grassland watershed in southeast Arizona, used with a novel rainfall-runoff model, demonstrate that the decrease in infiltrability of constructed pervious areas in the developed watershed accounts for roughly 17 percent of the overall increase in runoff. Directly connected impervious areas, primarily streets and driveways, cause 56 percent of the increase in runoff, and indirectly connected impervious areas, primarily rooftops and sidewalks account for the remaining increase in runoff. The total increase in runoff between the two watersheds is 940 percent. Infiltration measurements indicated that saturated hydraulic conductivity (Ks) is 114 percent greater in the grassland watershed than in the urban watershed, 6.2 mm/hr and 2.9 mm/hr, respectively. Ks in the urban watershed identified using one-dimensional parameter optimization is 5.9 mm/hr. A new component of the KINEROS2 modeling framework, the urban element, forms the basis for the model by simulating a contiguous row of houses and adjoining street. Tests using different levels of discretization illustrate that although Ks and slope remain important parameters to be estimated, the watershed geometry can be greatly simplified when discretizing the urban watershed using the urban element.