 |
| Titel |
Numerical study of the evaporation process and parameter estimation analysis of an evaporation experiment |
| VerfasserIn |
K. Schneider-Zapp, O. Ippisch, K. Roth |
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| ISSN |
1027-5606
|
| Digitales Dokument |
URL |
| Erschienen |
In: Hydrology and Earth System Sciences ; 14, no. 5 ; Nr. 14, no. 5 (2010-05-17), S.765-781 |
| Datensatznummer |
250012300
|
| Publikation (Nr.) |
 copernicus.org/hess-14-765-2010.pdf |
|
|
|
|
|
| Zusammenfassung |
| Evaporation is an important process in soil-atmosphere interaction. The determination of
hydraulic properties is one of the crucial parts in the simulation of water transport in
porous media. Schneider et al. (2006) developed a new evaporation method to improve the
estimation of hydraulic properties in the dry range. In this study we used numerical
simulations of the experiment to study the physical dynamics in more detail, to optimise the
boundary conditions and to choose the optimal combination of measurements. The physical
analysis exposed, in accordance to experimental findings in the literature, two different
evaporation regimes: (i) a soil-atmosphere boundary layer dominated regime (regime I) close to
saturation and (ii) a hydraulically dominated regime (regime II). During this second regime
a drying front (interface between unsaturated and dry zone with very steep gradients)
forms which penetrates deeper into the soil as time passes. The sensitivity
analysis showed that the result is especially sensitive at the transition between the two
regimes. By changing the boundary conditions it is possible to force the system to switch
between the two regimes, e.g. from II back to I. Based on this findings a multistep
experiment was developed. The response surfaces for all parameter combinations are flat and
have a unique, localised minimum. Best parameter estimates are obtained if the evaporation
flux and a potential measurement in 2 cm depth are used as target variables. Parameter
estimation from simulated experiments with realistic measurement errors with a two-stage
Monte-Carlo Levenberg-Marquardt procedure and manual rejection of obvious misfits lead to
acceptable results for three different soil textures. |
| |
|
| Teil von |
|
|
|
|
|
|
|
|