Uncertainty in mass-balance calculations of non-point source loads to the Arkansas River

Mueller, Jennifer, author
Gates, Timothy K., author
Colorado State University, publisher
Journal Title
Journal ISSN
Volume Title
Field data on flow and solute concentrations are used in mass balance cal-culations to estimate the non-point source loads of salt to an upstream reach and to a downstream reach along the lower Arkansas River in southeastern Colorado. A similar analysis is conducted to estimate selenium (Se) loads to the downstream reach. The unknown variable calculated to bring closure to the mass balance along each river reach for each sample period is the total unmeasured non-point load in surface water and groundwater that enters along ungauged tributaries and along the main stem of the river. Several sources of uncertainty in the mass balance calculations are considered: instrument error in the data sonde of field electrical conductivity (EC) measurements at multiple locations during a sample period and in laboratory analysis of water samples, uncertainty in assuming that the instantaneous EC measurements at locations represent corresponding EC averaged over the entire sample period and over the cross section, ambiguity in using fitted regression equations to relate EC values to concentrations of total dissolved solids (TDS) and dissolved Se, and error in streamflow measurements at gauging stations. These analyses render models of the input vari-ables to the mass balance equations for each reach as random variables with specified probability distributions. Monte Carlo simulation is used to solve the resulting stochastic mass balance equations to predict distributions of possible values of the non-point source loads for TDS and Se. Results indicate substantial uncertainty in the predicted loads. Annual average coefficients of variation (CV) in predicted TDS loads along the upstream river reach range between 0.37 and 0.91, and annual averages of 90% prediction intervals range from 2,404 to 22,562 (kg/day)/km. Annual average CV in predicted TDS load values for the downstream reach range between 0.09 and 0.20, and annual averages of 90% prediction intervals range from 2,615 to 8,312 (kg/day)/km. For the Se mass balance along the downstream reach, the average CV over nine sample periods within a three-year period is 0.07, and the 90% prediction intervals of Se loads range from 0.008 to 0.016 (kg/day)/km.
2006 annual AGU hydrology days was held at Colorado State University on March 20 - March 22, 2006.
Includes bibliographical references.
Rights Access
Associated Publications