You are here:
HISTORICAL ANALYSIS OF THE RELATIONSHIP OF STREAMFLOW FLASHINESS WITH POPULATION DENSITY, IMPERVIOUSNESS, AND PERCENT URBAN LAND COVER IN THE MID-ATLANTIC REGION
Citation:
JARNAGIN, S. HISTORICAL ANALYSIS OF THE RELATIONSHIP OF STREAMFLOW FLASHINESS WITH POPULATION DENSITY, IMPERVIOUSNESS, AND PERCENT URBAN LAND COVER IN THE MID-ATLANTIC REGION. Presented at World Environmental and Water Resource Congress 2008, Honolulu, HI, May 12 - 16, 2008.
Impact/Purpose:
Presentation
Description:
Methods: This study is an examination of the relationship between stream flashiness and watershed-scale estimates of percent imperviousness, degree of urban development, and population density for 150 watersheds with long-term USGS National Water Information System (NWIS) historical daily mean streamflow datasets in EPA Region 3 (R3, Mid-Atlantic USA). I used decadal Census population data from 1930-2000, proportionally allocated into 2000-era county boundaries to estimate population density. For decades after 1960, higher spatial resolution census data were used along with the LandScan 1998 dasymetric estimation of population density at a 450 m grid cell spatial resolution. Temporal land use/land cover (LULC) data: 1973 North American Landscape Characterization (NALC) data, 1992 National Land Cover Data (NLCD1992), and NLCD2001 were used to estimate a 'percent urban developed' parameter for each watershed. The NLCD2001 Imperviousness data layer and the ArcView ATtILA extension (US EPA, 2004) was used for a 2001 estimator of watershed imperviousness and both the coefficient technique of Jennings et al. (2004) and ATtILA were used to estimate 1992 watershed imperviousness. The Richards-Baker Flashiness Index (R-B Index, Baker et al., 2004) was applied to historical NWIS streamflow to calculate annual flashiness values for the period of record for each station. Results: Historical changes in mean stream flashiness were correlated with county-scale based changes in watershed population density estimates. Streamflow stations that showed significant changes in historical flashiness had a higher mean population density than those that showed no change. The strength of the population-flashiness correlation increased as the spatial scale of the population estimator was reduced, with dasymetric LandScan data giving the best relationship. LULC and imperviousness estimators were equally effective at exploring the relationship between stream flashiness and watershed development. Urban development classes 'None' and 'Rural' were statistically the same while increasing levels of development were associated with statistically significant increases in stream flashiness. Watersheds with less than 20% 'urban' development displayed background levels of stream flashiness and mean flashiness increase with urban development density thereafter. My results support previous research that suggests low intensity development does not substantially alter streamflow. Increasing degrees of development intensity do significantly alter streamflow. One use of this dataset is to search for 'positive outliers' - where predicted stream flashiness is less than anticipated by the level of urban development. Detailed examination of these watersheds may yield examples where BMPs or patterns of development have been successful at mitigating the impact of urban development on stream hydrology.