Modeling recession and low flow characteristics of Iowa streams

Available mean daily streamflow data from 143 gaging stations were initially analyzed to characterize recessions and low flows of Iowa streams. The period of record varies between 10 and 115 years for different stations. A computerized mathematical model was used to derive two master recession curves (MRC\u27s)--one for winter and the other for summer--and to calculate the median storage delay factor (K med) for each individual station. These factors were used to estimate the diffusivity of the aquifer(s) contributing base flow to each stream;The MRC\u27s can be used to forecast the general behavior of streams during a period of low flow. The calculated K med factors were also used as indicators of the overall effect of basin geology to develop predictive statistical-based models for estimation of low flow indices at ungaged sites on Iowa streams;Two calculated low flow indices which were selected as dependent variables for establishment of multiple regression models are (1) the Iowa regulated protected low flow defined as the flow rate on the duration curve which is equaled or exceeded 84% of the time (Q84%) and (2) the minimum average of seven consecutive daily discharges that is expected once a decade (Q7,10);Four explanatory variables were initially chosen as potential independent variables. They are (1) drainage area, (2) datum elevation at the gage, (3) mean annual streamflow as a substitute for annual precipitation, and (4) average storage delay factor, which was calculated in the first phase of this research. Multivariate regression techniques with different screening procedures were applied to formulate the models. Advanced diagnostic methods were utilized to improve performance of the models, and a complete residual analysis was conducted to check their adequacy;About 10% of the data were deliberately left out to be used later as a new set to evaluate the models in terms of validity and predictive ability