Regression Model for Daily Maximum Stream Temperature

Abstract

An empirical model is developed to predict daily maximum stream temperatures for the summer period. The model is created using a stepwise linear regression procedure to select significant predictors. The predictive model includes a prediction confidence interval to quantify the uncertainty. The methodology is applied to the Truckee River in California and Nevada. The stepwise procedure selects daily maximum air temperature and average daily flow as the variables to predict maximum daily stream temperature at Reno, Nev. The model is shown to work in a predictive mode by validation using three years of historical data. Using the uncertainty quantification, the amount of required additional flow to meet a target stream temperature with a desired level of confidence is determined.