Wednesday, April 2, 2008 : 1:20 p.m.

Use of Pattern Recognition to Evaluate Alternative Recharge Models

Ming Ye, Ph.D. and Dan Lu, Florida State University

This study develops a pattern recognition method to evaluate five alternative recharge models developed for the Death Valley regional flow system including Nevada and the Death Valley area of California. The five models are (1) the Maxey-Eakin model, (2 & 3) a distributed parameter watershed model with and without a runon-runoff component, and (4 & 5) a chloride mass balance model with two zero-recharge masks, one for alluvium and one for both alluvium and elevation. The five independently developed models have been used to simulate groundwater flow and contaminant transport at the Death Valley regional flow system. Since recharge is a major driving force, it is critical to evaluate the recharge models and assess their uncertainty. A pattern recognition method based on belief propagation is developed for this purpose. The method does not intend to evaluate theoretical basis of the five models, but relative correctness of the recharge estimates by examining patterns of recharge estimates of the five models, First, recharge estimates of the five models are transformed into categories from high to low recharge based on a single cumulative distribution function obtained from the recharge estimates. At each point, majority of recharge categories of the five models is identified and constitutes a reference map of the recharge estimates, which is then back transformed into recharge estimates based on the cumulative distribution function used for the transform. Difference between the reference recharge and recharge estimate of each model is estimated, and the difference used to estimate model probability using a constrained maximum entropy method. While the model probability is by definition a prior probability, it is not absolutely subjective, since it is based on quantitative comparison of the recharge estimates. Using a Bayesian method through model calibration, posterior model probability can be estimated and used to assess conceptual model uncertainty.

Ming Ye, Ph.D., Florida State University Dr. Ming Ye is an Assistant Professor in the School of Computational Science and Department of Geological Sciences of the Florida State University. Before joining the Florida State University, he was an Assistant Research Professor of the Desert Research Institute, and post-doc of the Pacific Northwest National Laboratory. He received his Ph.D. in hydrology from the University of Arizona in 2002, and a B.S. in geology from Nanjing University, China, in 1997. His research interests include groundwater modeling in saturated and unsaturated porous and fracture media, parameter estimation, applied geostatistics, and uncertainty analysis of groundwater modeling.

Dan Lu, Florida State University I am a doctoral student in the School of Computational Science of the Florida State University. I received my Master degree in Hydrogeology from China University of Geosciences and Bachelor degree in Environmental Engineering from Shijiazhuang University of Economics. My major is Computational Science in Geology, and my adviser is Dr. Ming Ye. My research area is groundwater modeling.


2008 Ground Water Summit