Laserfiche WebLink
<br />USE OF RAINFALL-SIMULATOR DATA IN <br />PRECIPITATION-RUNOFF MODELING STUDIES <br /> <br />By Gregg C. Lusby and Robert W. Lichty <br /> <br />ABSTRACT <br /> <br />Results of a study using a rainfall simulator to define infiltration <br />parameters for use in watershed modeling are presented. During 1981-82, a <br />total of 23 rainfall-simulation runs were made on 5 small plots (about 2,500 <br />square feet) located on 4 representative soil-vegetation types of the Willow <br />Gulch watershed, located about 50 miles east of Denver, Colorado. During the <br />summer of 1982, data for 3 observed rainfall-runoff events were recorded by <br />gages on 4 of the plots. Runoff data from both simulator runs and observed <br />rainstorms were used to develop best-fit parameters of the Green-Ampt <br />infiltration equation. <br /> <br />In all fitting attempts, the hydraulic conductivity term, KSAT, grossly <br />ontro11ed the goodness of fit. High variability in soil-water uptake found <br />from soil samples taken before and after simulator runs confirms the empir- <br />ical nature of KSAT. Best-fit values are plot-average values that reflect <br />both the limitations of the Green-Ampt equation, and the inherent (natural) <br />variability of soil-water properties of field soils. Results of fitting KSAT <br />to reproduce runoff from rainfall-simulator runs, and results of fitting KSAT <br />to reproduce runoff from observed rainfall-runoff events are inconsistent. <br />Summer runs on plots located in the upland area of ponderosa pine give little <br />indication of runoff potential from observed rainstorms. In contrast, results <br />for plots located in the lowland prairie area are in reasonable agreement with <br />results from observed rainstorms. Fall runs on upland plots indicate that <br />cooler soil temperatures may influence the infiltration process. In contrast, <br />fall runs on lowland plots show no consistent effect of cooler soil temper- <br />atures; reasons for these anomalous results are unknown. <br /> <br />The drainage area of the North Fork Willow Gulch watershed was partitioned <br />into homogeneous hydrologic-response units (HRUs), and a conceptual flow-routing <br />network of plane and channel segments was developed to characterize required <br />input to a precipitation-runoff modeling system, PRMS. The application of <br />PRMS to three'storms in 1982, using estimates of KSAT based on simulator runs, <br />produced predicted runoff volumes that were 70 percent less than those observed <br />in the first two cases and 40 percent more than that observed in the third <br />case. Using estimates of KSAT based on observed rainfall events on plots 1-3 <br />improved the prediction for two events, and degraded the result for the third <br />event. Adjustments in KSAT specfications, and adjustments to the storm rainfall <br />confirmed that accuracy of predicting peak flow rates is controlled by the <br />amount of water that falls on the ground, and the amount of water that infil- <br />trates. Runoff routing is adequately represented by the conceptual network <br />of plane and channel segments. <br /> <br />1 <br />