Southeastern Section - 63rd Annual Meeting (10–11 April 2014)

Paper No. 7
Presentation Time: 10:20 AM

ESTIMATING RUNOFF WITH A LIDAR BASED GIUH


MASON-DEESE, William, Department of Geology, University of Georgia, Athens, GA 30602 and DOWD, John, Department of Geology, University of Georgia, 210 Field Street, Athens, GA 30602, jdowd@uga.edu

Identifying flow paths and the time at which they contribute is a major goal of watershed research. Flow paths are used to predict source areas where runoff is produced and the transfer of runoff to the basin outlet. In this paper, methods of identifying flow paths in ungauged basins are used to model storm runoff from two forested watersheds, Panola Mountain Research Watershed (PMRW), Georgia, and Coweeta Hydrologic Lab Watershed 18, North Carolina.

In order to compare predicted storm flow to measured discharge, baseflow, defined here as flow that is not directly attributable to current rainfall, must be separated from stream discharge. Previous research at PMRW by Cary (2011) used geochemistry to separate discharge into four source waters. Cary's hydrograph separations were used to calibrate three recursive digital filters, which separate baseflow by attenuating the discharge hydrograph. The digital filters are able to approximate the total quantity of baseflow, although the digital filter sub-hydrographs peak too late and have a smoother shape than the geochemically derived sub-hydrographs. The “baseflow” consisted of both groundwater and soil water sources.

Storm runoff occurs over surface and near-surface pathways. A dense network of surficial pathways is extracted from topography derived from high-resolution LiDAR. The structure of the network, that is, the average length of streams in each Strahler order is used to parameterize the geomorphologic instantaneous unit hydrograph (GIUH), which is a two parameter gamma distribution (Cudennec 2004). The GIUH is used in conjunction with precipitation records to model storm runoff from several storms on both watersheds. The modeled storms are compared to observed runoff by removing baseflow from total discharge. Results are variable, with Nash-Sutcliffe coefficients ranging from 0.95 to less than zero. Nevertheless, this work provides the foundation for a storm runoff model that can be applied in ungauged basins.