HR: 0830h
AN: H31B-0464    [PDF]
TI: Non-Linear, Bayesian Hydrogeophyiscal Inversion in the Vadose Zone
AU: * Hou, Z
EM: hou@uclink.berkeley.edu
AF: CEE, UC Berkeley, 2100 Shattuck Ave., Berkeley, CA 94720 United States
AU: Rubin, Y
EM: rubin@ce.berkeley.edu
AF: CEE, UC Berkeley, 2100 Shattuck Ave., Berkeley, CA 94720 United States
AU: Hubbard, S
EM: sshubbard@lbl.gov
AF: LBNL, 1 Cyclotron Road, Berkeley, CA 94720 United States
AB: A comprehensive approach for modeling the dynamics of soil moisture profiles in the vadose zone under conditions of parameter uncertainty, using an information-based Bayesian-stochastic framework, is proposed and applied to a field site in Napa Valley in California. The Bayesian framework is combined with the one-dimensional Richards' equation and van-Genuchten-Mualem soil models, with initial and boundary conditions provided by geophysical and meteorological measurements. To identify the model parameters such as the van-Genuchten-Mualem soil water retention parameters $\alpha$ and n, the porosity $\theta$$_{s}$, and the saturated hydraulic conductivity K$_{s}$, soil texture analyses are performed and soil moisture content obtained from TDR, neutron probe and crosshole GPR are collected as a function of time. Soil texture information, combined with a hydrologic database -- Rosetta, is used to derive prior information for the model parameters, which are considered as random variables. The prior probability density functions (pdfs) for the parameters are developed from minimum relative entropy considerations. Soil moisture content measurements using geophysical methods are then used to update the model parameters and to improve the prediction of the soil moisture profiles. Results show that as additional measurements are incorporated into the updating procedure, the uncertainty level associated with the model parameters is reduced. The approach also provides estimates of the statistical moments of the output soil moisture content. For both dry and wet seasons at the study site, the method predicts the soil moisture profiles very well.
DE: 1829 Groundwater hydrology
DE: 1866 Soil moisture
DE: 1869 Stochastic processes
DE: 1875 Unsaturated zone
DE: 1894 Instruments and techniques
SC: Hydrology [H]
MN: 2003 Fall Meeting