The impact of uncertainties in hydrologic measurement on phosphorus budgets and empirical models for two Colorado reservoirs

Abstract

Water budgets and related chemical budgets of aquatic ecosystems commonly are interpreted without reference to uncertainties resulting from errors of measurement. The importance of such uncertainties in the use and interpretation of the phosphorus budgets of two Colorado reservoirs was determined. Water budgets were calculated from data from on‐site measurement of precipitation, evaporation, reservoir stage, outflow, and inflow from the major river of each reservoir. Direction of groundwater flow was determined with existing well information and seepage meters. Contributions of ungauged watersheds and groundwater were estimated from the residual of measured water budget terms. Less than 25% of both water and phosphorus input to Gross Reservoir remained as a residual term: the contribution of errors in budget measurement to empirical model uncertainty was <13%. In Williams Fork Reservoir, 45% of the phosphorus input was calculated as residual: errors in budget measurement contributed almost 50% to empirical model uncertainty. Measured concentrations of total phosphorus in each reservoir in both years were within the confidence limits of empirical model predictions. However, the confidence limits of those models were too large to produce accurate predictions, especially for the Williams Fork Reservoir. Attempts to predict the effect of changes in water or phosphorus budgets on reservoir concentrations of total phosphorus and chlorophyll would be subject to large uncertainties.