NAME =  Kun Yang  
ORGANIZATION =  University of Tokyo  
ADDRESS =  Department of Civil Engineering, University of Tokyo, Hongo 7-3-1, Bunkyo-ku, Tokyo 113-8656, Japan  
COUNTRY =  Japan  
PHONE =  03-5841-6109  
FAX =  03-5841-6130  
E-MAIL =  yangk@hydra.t.u-tokyo.ac.jp  
POSTER_ONLY =  no  
THEME =  T1  
DATE =  13-Jul-04-16:04:07  
ABSID = T1KY13Jul04160407 
TITLE = Inverse analysis of the effect of soil vertical heterogeneity on land surface and subsurface processes 
AUTHOR_1 = Kun Yang 
INSTITUTION_1 = University of Tokyo 
PRESENTER = AUTHOR_1 
AUTHOR_2 = Toshio Koike 
INSTITUTION_2 = University of Tokyo 
AUTHOR_3 = Baisheng Ye 
INSTITUTION_3 = Chinese Academy of Science 
ABSTRACT =   The importance of soil horizontal heterogeneity has been widely addressed in modeling of land surface processes, but relatively few studies explore the effect of soil vertical heterogeneity on land surface and subsurface processes. It is the motivation of this study to investigate the relationship between the soil vertical heterogeneity and the land processes. For this purpose, a single-source land surface model is developed, which approximates a vertically heterogeneous soil by a homogeneous top domain, a homogeneous deep domain, and a transitional domain between the two. The model parameters of the top and bottom domains are inversely estimated from observed soil moisture and temperature profiles, and then the parameters are reasonably interpolated for the transitional layer. The model parameters include soil domain division, soil hydraulic and thermal parameters, and the surface evaporation resistance. 
  Analyses are carried out based on synthetic data and data collected in the Tibetan Plateau. We suggest that the soil vertical heterogeneity plays a key role in land surface and subsurface processes. Simulations with uniform soil parameters cannot reproduce observed soil moisture, temperature, and surface energy budget for vertically heterogeneous soils even though all model parameters are optimized. We also find that inversely estimated soil parameters for each domain could be quite different from laboratory analyzed one, because the former represents the mean properties of each soil domain while the latter represents the properties of a small soil sample. Simulations with the latter may produce erroneous soil moisture profiles. Moreover, we emphasize that a successful simulation of soil moisture and temperature profiles is not an adequate condition for correct simulation of the surface energy budget, so a land surface model should be evaluated based on not only comparisons of soil moisture and temperature but also comparisons of energy partitions.