The evapotranspiration (ET) estimation method by Gribovszki et al. (2010b) has so far been validated only at one catchment because good quality discharge time series with the required high enough temporal resolution can probably be found at only a handful of watersheds worldwide. To fill in the gap of measured data, synthetic groundwater discharge values were produced by a 2D finite element model representing a small catchment. Geometrical and soil physical parameters of the numerical model were changed systematically and it was checked how well the model reproduced the prescribed ET time series. The tests corroborated that the ET-estimation method is applicable for catchments underlain by a shallow aquifer. The slope of the riparian zone has a strong impact on the accuracy of the ET results when the slope is steep, however, the method proved to be reliable for gentle or horizontal riparian zone surfaces, which are more typical in reality. Likewise, errors slightly increase with the decrease of riparian zone width, and unless this width is comparable to the width of the stream (the case of a narrow riparian zone), the ET estimates stay fairly accurate. The steepness of the valley slope had no significant effect on the results but the increase of the stream width (over 4m) strongly influences the ET estimation results, so this method can only be used for small headwater catchments. Finally, even a magnitude change in the prescribed ET rates had only a small effect on the estimation accuracy. The soil physical parameters, however, strongly influence the accuracy of the method. The model-prescribed ET values are recovered exactly only for the sandy-loam aquifer, because only in this case was the model groundwater flow system similar to the assumed, theoretical one. For a low hydraulic conductivity aquifer (e.g. clay, silt), root water uptake creates a considerably depressed water table under the riparian zone, therefore the method underestimates the ET. In a sandy, coarser aquifer the flow lines never become vertical even bellow the root zone, so the method overestimates the ET rate, thus the estimated ET values need to be corrected. Luckily the prescribed and estimated ET rates express a very high linear correlation, so the correction can be obtained by the application of a constant, the value of which solely depends on soil type.
|Number of pages||12|
|Journal||Acta Silvatica et Lignaria Hungarica|
|Publication status||Published - Dec 1 2011|
- Numerical model
- Riparian vegetation
ASJC Scopus subject areas