Paper No. 6
Presentation Time: 9:00 AM-6:00 PM

MODELING PHYSICAL AND CHEMICAL HETEROGENEITY OF THE BORDEN AQUIFER USING AN OUTCROP ANALOG


WEISSMANN, Gary S.1, PICKEL, Alexandra2, FRECHETTE, Jedediah D.3, ALLEN-KING, Richelle M.4, JANKOVIC, Igor5, MAGHREBI, Mahdi5, KALINOVICH, Indra6 and MCNAMARA, Kelsey C.7, (1)Earth and Planetary Sciences, University of New Mexico, MSC03-2040, 1 University of New Mexico, Albuquerque, NM 87131-0001, (2)Earth and Planetary Sciences, University of New Mexico, MSC03 2040, 1 University of New Mexico, Albuquerque, NM 87131-0001, (3)Earth and Planetary Sciences, University of New Mexico, Albuquerque, NM 87131, (4)Geology, SUNY Buffalo, 876 Natural Science Complex, Buffalo, NY 14260, (5)Department of Civil, Structural and Environmental Engineering, University at Buffalo, 207 Jarvis Hall, Buffalo, NY 14260-4400, (6)Dillon Consulting Limited, 1558 Wilson Place, Winnipeg, MB R3T 0Y4, Canada, (7)Earth & Planetary Sciences, University of New Mexico, Albuquerque, NM 87131-0001, weissman@unm.edu

In order to evaluate the influence of both physical and chemical heterogeneity on contaminant transport, we excavated a series of outcrop exposure panels in a sand quarry through Borden aquifer sediments, mapped and measured lithofacies distributions in these exposures, evaluated sorption properties of the different lithofacies materials, and modeled the three-dimensional facies distributions using transition probability geostatistics. The sand quarry exposures are located approximately 2-km from the Stanford-Waterloo transport experiments from the 1980s in stratigraphically-equivalent sediments to those of the transport experiments. The exposure panels cover an area that is approximately 20-40m wide and up to 10m high, in total. Individual panels were approximately 15 - 20m wide and 1.5m high. In order to produce an accurate survey of facies locations, each panel location was scanned using terrestrial lidar and photographed at high-resolution. Facies were mapped using a hierarchical approach, where unit types were mapped and classified in the field and textural classes within these mapped units were segmented using edge-detection and textural filtering methods. At the study site, an erosional surface separates two ‘major’ stratigraphic units, thus we evaluated transition probabilities and modeled facies distributions of each unit separately. Markov chain models were fit to transition probabilities measured from the 3D data for each unit. The erosional surface was projected into a 3D model domain, and the Markov chain models were used in sequential indicator simulation to construct realizations of the facies at the site. The resulting realizations of heterogeneity can be used for future transport models at the site.