HR: 10:35h
AN: H52D-02    [Abstracts]
TI: On The Propagation of Uncertainty in Upscaling Solute Transport Models
AU: Gomez-Hernandez, J
EM: jaime@dihma.upv.es
AF: Universidad Polit‚cnica de Valencia, Camino de vera, s/n, Valencia, 46071 Spain
AU: * Fernandez-Garcia, D
EM: dafernan@dihma.upv.es
AF: Universidad Polit‚cnica de Valencia, Camino de vera, s/n, Valencia, 46071 Spain
AU: Cassiraga, E F
EM: efc@dihma.upv.es
AF: Universidad Polit‚cnica de Valencia, Camino de vera, s/n, Valencia, 46071 Spain
AB: The scale of measurements is typically orders of magnitude smaller than the computational scale requisite for achieving a functional model. Upscaling is the process by which information at a given small support scale (e.g. core samples) is transferred to a larger support scale given by the numerical model grid-blocks. At the present, upscaling techniques only consider the transfer of small-scale information to ultimately obtain grid-block parameter values. Yet, the impact of smoothing and suppression of the velocity fluctuations within grid-blocks on the uncertainty associated with transport model predictions is important. We present Monte Carlo transport simulations that elucidate how uncertainty in transport models propagates through out the upscaling process. To achieve this, we compare the uncertainty associated with a transport model before and after upscaling the underlying heterogeneity. In order to contemplate a wide range of aquifer realities, two different random function models were evaluated: (i) A Gaussian random function with low spatial connectivity of high and low extreme values; and (ii) A non-Gaussian random function with high spatial connectivity of high and low extreme values.
DE: 1832 Groundwater transport
DE: 1869 Stochastic hydrology
DE: 1873 Uncertainty assessment (3275)
SC: Hydrology [H]
MN: Fall Meeting 2005