The U.S. Geological Survey, in cooperation with the Colorado Water Conservation Board, compared two methods for estimating base flow in three reaches of the South Platte River between Denver and Kersey, Colorado. The two methods compared in this study are the Mass Balance and the Pilot Point methods. Base-flow estimates made with the two methods were based upon a 54-year period of record (1950 to 2003).

The Mass Balance method for estimating base flow is based on a mass balance of all known inflows to and outflows from a given stream reach, with the equation being solved for groundwater flow into or out of the reach. A positive mass balance indicates a gaining reach (base flow) and a negative balance indicates a losing reach. The mass balance was calculated using daily mean streamflow, wastewater treatment plant discharge, and stream diversion data. Monthly mean base flow was calculated as the average of all daily mean mass-balance results for a given month.

The Pilot Point method is based on a daily mean mass balance of all inflows to and outflows from a stream reach. The Pilot Point differs from the Mass Balance method in that extreme daily mass-balance results are constrained utilizing two analytical solutions that represent the maximum possible streamflow gain or loss. Additionally, the Pilot Point method utilizes a smoothing function, based on a moving average of the daily constrained mass-balance results. The moving average for this study utilized a moving-average period, called the bin width, of 61 days. The maximum and minimum base-flow constraints and the smoothing function are utilized to provide base-flow estimates that exhibit reasonable maximum values and temporal variability consistent with the concept of groundwater flow being gradual and slow.

Both the Mass Balance and Pilot Point results provided similar patterns in annual and monthly base flow. All three reaches were indicated to be gaining reaches, particularly after about 1970, with the magnitude of base flow increasing downstream. This degree of similarity between the two methods was expected because both methods are based on a streamflow mass balance. The magnitude of estimates provided by the two methods was measurably different. The stream gains and losses estimated using the Mass Balance method were consistently more variable and of greater magnitude than those estimated using the Pilot Point method. In the Denver to Henderson reach, the median estimated annual mean base flow was 34.0 cubic feet per second (ft³/s) using the Mass Balance method and was 39.1 ft³/s using the Pilot Point method. In the Henderson to Fort Lupton reach, the median estimated annual mean base flow was 50.0 ft³/s using the Mass Balance method and was 40.0 ft³/s using the Pilot Point method. In the Fort Lupton to Kersey reach, the median estimated annual mean base flow was 234 ft³/s using the Mass Balance method and was 214 ft³/s using the Pilot Point method.

The Mass Balance results were quite variable over time such that they appeared suspect with respect to the concept of groundwater flow as being gradual and slow. The large degree of variability in the day-to-day and month-to-month Mass Balance results is likely the result of many factors. These factors could include ungaged stream inflows or outflows, short-term streamflow losses to and gains from temporary bank storage, and any lag in streamflow accounting owing to streamflow lag time of flow within a reach. The Pilot Point time series results were much less variable than the Mass Balance results and extreme values were effectively constrained. Less day-to-day variability, smaller magnitude extreme values, and smoother transitions in base-flow estimates provided by the Pilot Point method are more consistent with a conceptual model of groundwater flow being gradual and slow. The Pilot Point method provided a better fit to the conceptual model of groundwater flow and appeared to provide reasonable estimates of base flow.