No abstract

Diurnal fluctuations of hydrological variables (e.g., shallow groundwater level or streamflow rate) are comparatively rarely investigated in the hydrologic literature although these short-term fluctuations may incorporate useful information for the characterization of hydro-ecological systems. The fluctuations can be induced by several factors like (a) alternating processes of freezing and thawing; (b) early afternoon rainfall events in the tropics; (c) changes in streambed hydraulic conductivity triggered by temperature variations, and; (d) diurnal cycle of water uptake by the vegetation. In temperate climates, one of the most important diurnal fluctuation-inducing factors is the water consumption of vegetation, therefore a detailed overview is provided on the history of such research. Beside a systematic categorization of the relevant historical studies, models that calculate groundwater evapotranspiration from diurnal fluctuations of groundwater level and/or streamflow rate have been reviewed. Compared to traditional evapotranspiration estimation methods these approaches may excel in that they generally employ a small number of parameters and/or variables to measure, are typically simple to use, and yet can yield results even on a short timescale (i.e., hours). While, e.g., temperature-based methods of evapotranspiration are simple too, they cannot be applied or become inaccurate over shorter time periods. Similarly, traditional approaches (such as eddy-correlation or Bowen-ratio based) are accurate for shorter time steps but they require a number of measurable atmospheric input variables.

By applying a nonlinear reservoir approach for groundwater drainage, catchment-scale evapotranspiration (ET) during flow recessions can be expressed with the help of the lumped version of the water balance equation for the catchment. The attractiveness of the approach is that ET, in theory, can be obtained by the sole use of observed flow values for which relatively abundant and long records are available. A 2D finite element numerical model of subsurface flow in the unsaturated and saturated zones, capable of simulating moisture removal by vegetation, was first successfully employed to verify the water balance approach under ideal conditions. Subsequent practical applications over four catchments with widely varying climatic conditions however showed large disparities in comparison with monthly ET estimates of Morton's WREVAP model. ª