Streamflow hydrology and simulation of the Salt River Basin in central Arizona.

Abstract

A continuous simulation streamflow model (i.e., SSARR - an acronym for Streamflow Synthesis and Reservoir Regulation) was evaluated and used to study winter streamflow from the Salt River Basin. This 4,306 square mile basin, which ranges in elevation from 2,200 to 11,500 feet, is associated with a diversity of watershed, vegetation, climatic and hydrologic characteristics. Program modifications allowed computation of potential evapotranspiration within the model. This provided improved flexibility in delineating simulation units and reduced the necessary time-dependent data inputs to daily values of precipitation and temperature. Refinement of initial parameter estimates and relationships was accomplished by trial and error methods. Four years of hydrometeorological data were utilized for model calibration and an additional four years used to test the validity of parameter estimates. Simulated hydrographs generally underestimated peak flows and overestimated recession flows following major rainfall events. The standard error of the estimate for simulated winter flows (November through May) was only 30,000 acre-feet for the calibration period but increased to approximately 120,000 acre-feet for the validation period. Average winter flows during the calibration and validation periods were 840,000 and 690,000 acre-feet, respectively. Approximately 25 to 45 percent of the winter runoff occurring after February 1, March 1 and April 1 could be predicted from simulated snowpack water equivalent and soil moisture conditions on the prediction date. High elevation portions of the basin were more efficient at producing streamflow from a given precipitation input than were the lower elevations. Maximum daily watershed efficiencies (ratio of generated runoff to rain and snowmelt inputs) usually occurred on the date of snowpack disappearance for the relatively high elevation simulation units. Simulated runoff volumes increased exponentially with increased basin precipitation but decreased linearly with increased basin air temperatures. For a selected winter period, simulations indicated streamflow response to a one-inch change in winter precipitation would be approximately 100,000 acre-feet, or 3.4 times greater than expected from a one-degree change in winter temperature.