Diploma Thesis
Application of the evapotranspiration-model AMBETI to the calculation of the evapotranspiration in the Weißenstädter area (Fichtelgebirge)
Christian Neuner (07/2000-12/2000)
Support: Thomas Foken
n this study a modified version of the Soil Vegetation Atmosphere Transfer (SVAT) model AMBETI (Agrometeorological Model for the Calculation of Evaporation, Transpiration, and Interception) of the German Weather Service is used, in order to calculate the evapotranspiration on field scale in the Weißenstädter basin, in the Northeast of Bavaria. The evapotranspiration is determined for the years 1998 and 1999, dependent on land-use (summer barley and meadow), amount of precipitation and soil type. The modelled evapotranspiration is compared to high-quality measurements of the VOITEX-99-experiment, which was carried out in the Weißenstädter basin in July 1999, and in which the latent heat was measured by employing the eddy-covariance-method. In a further step the parameterisation of the turbulence part in AMBETI is improved using the parameterisation approaches according to Businger in the standardisation of HÖGSTRÖM (1988) and according to SKEIB AND RICHTER (1984). A comparison of the results of the modified models and both the results of the original model and the measurements of the VOITEX-99-experiment takes place. Apart from these two main parts of the study a sensitivity analysis is realised for a few model parameters in AMBETI by employing the GLUE method (Generalised Likelihood Uncertainty Estimation) of BEVEN AND BINLEY (1992).
The resulting evapotranspiration values of the years 1998 and 1999 show, that on the extremely wet conditions in the Weißenstädter basin, the different soil types almost do not influence the yearly values of the evapotranspiration. Increased respectively decreased amounts of precipitation lead to only small changes in the yearly evapotranspiration rates, as well. The soils in the Weißenstädter basin are saturated and as a consequence the actual evapotranspiration approaches potential values. The yearly evapotranspiration values simulated over summer barley are higher than those simulated over meadow. In a comparison of the simulated latent heat and measured values of the VOITEX-99-experiment, it is remarkable, that the yearly and daily sums of the latent heat can be reproduced by AMBETI in a relatively good way. In the hourly values, though, significant deviations can be observed. The modelled course of the latent heat cannot follow the dynamics of the turbulent exchange under extremely unstable conditions on days with a high amount of incoming radiation. The improvements in the parameterisation of the turbulence according to HÖGSTRÖM (1988) and SKEIB AND RICHTER (1984) do not have a significant influence on the calculation of the evapotranspiration in AMBETI. With regard to the original version of AMBETI only minor changes of the evapotranspiration occur. Compared to other programme parts of AMBETI, like for example the parameterisation of the total stomata resistance, the parameterisation of the turbulence influences the simulated values of the evapotranspiration in the considered period of the VOITEX-99-experiment only slightly. The results of the sensitivity analyses using the GLUE-method show on the one hand, that the parameters minimal stomatal resistance and soil water potential, from which on water stress occurs, have, compared to the likewise considered parameters leaf width and soil water potential, from which on the stomata are totally closed, a noticeable meaning in fitting the model to the data. Many parameter sets exist, which reproduce the measured data equally well. This suggests that the model is possibly overparameterised. On the other hand, the results demonstrate, that with an appropriate choice of the values of the calibration parameters of AMBETI the determination of the evapotranspiration can be improved.