Picture (above): Green frog on mud. Credit: Jill Heyer (Unsplash).

The main hydrologic observable is streamflow and this is widely observed. Explanations for variations in streamflow begin with the catchment mass balance. In particular, precipitation estimates are routinely available from gridded data, but evaporation and soil moisture remain unobserved over most catchments.

One “trick” used in data analysis is to select a longer period, for example, a year, or perhaps five years, and assume that soil moisture variations over that period are zero. With this “steady-state assumption”, the evaporation is calculated as the difference between precipitation and streamflow. While this assumption is very widely used it has never been formally tested.

In this paper, we use a series of independent streamflow precipitation and evaporation estimates to evaluate how long we need to average the data for the steady-state assumption to hold. We find, to our surprise that the minimum averaging period is around 10 years but in some dry catchments, such as those over much of Australia, period as long as 30 years are needed.

  • Paper: Han J, Yang Y, Roderick ML, McVicar TR, Yang D, Zhang S, Beck HE (2020) Assessing the steady-state assumption in water balance calculation across global catchments, Water Resources Research, 56, e2020WR027392, 10.1029/2020WR027392.