Using uncertainty analysis and groundwater measurements to improve the confidence of river water balance estimates

Abstract

Listen

Translate article

Summary An improved understanding of water balances of rivers is fundamental in water resource management. Effective use of a water balance approach requires thorough identification of sources of uncertainty around all terms in the analysis and can benefit from additional, independent information that can be used to interpret the accuracy of the residual term of a water balance. We use a Monte Carlo approach to estimate a longitudinal river channel water balance and to identify its sources of uncertainty for a regulated river in south-eastern Australia, assuming that the residual term of this water balance represents fluxes between groundwater and the river. Additional information from short term monitoring of ungauged tributaries and groundwater heads is used to further test our confidence in the estimates of error and variance for the major components of this water balance. We identify the following conclusions from the water balance analysis. First, improved identification of the major sources of error in consecutive reaches of a catchment can be used to support monitoring infrastructure design to best reduce the largest sources of error in a water balance. Second, estimation of ungauged inflow using rainfall–runoff modelling is sensitive to the representativeness of available gauged data in characterising the flow regime of sub-catchments along a perennial to intermittent continuum. Lastly, comparison of temporal variability of stream-groundwater head difference data and a residual water balance term provides an independent means of assessing the assumption that the residual term represents net stream-groundwater fluxes.