HR: 1330h
AN: H22A-0910    [PDF]
TI: Runtime and Inversion Impacts on Estimation of Moisture Retention Relations by Centrifuge
AU: * Sigda, J M
EM: jsigda@gis.nmt.edu
AF: New Mexico Bureau of Geology and Mineral Resources, 2808 Central SE, Albuquerque, NM 87106 United States
AU: Wilson, J L
EM: jwilson@nmt.edu
AF: Department of Earth and Environmental Science, New Mexico Institute of Mining and Technology, Campus Station, Socorro, NM 87801 United States
AB: Standard laboratory methods in soil physics for measuring the moisture retention relation (drainage matric potential-volumetric moisture content relation) are each limited to only part of the moisture content range. Centrifuge systems allow intensive accurate measurements across much of the saturation range, and typically require much less time than traditional laboratory methods. An initially liquid-saturated sample is subjected to a stepwise-increasing series of angular velocities while carefully monitoring changes in liquid content. Angular velocity is held constant until the capillary and centrifugal forces equilibrate, forcing liquid flux to zero, and then a final average liquid content is noted. The procedure is repeated after increasing the angular velocity. Centrifuge measurement time is greatly reduced because the centrifugal body force gradient can far exceed the driving forces utilized in standard lab methods. Widely-used in the petroleum industry for decades, centrifuge measurement of moisture retention relations is seldom encountered in the soil physics or vadose hydrology literatures. Yet there is a need to better understand and improve the experimental methodology given the increasing number of centrifuges employed in these fields. Errors in centrifuge measurement of moisture retention relations originate from both experimental protocol and from data inversion. Like standard methods, centrifuge methods assume equilibrium conditions, and so are sensitive to errors introduced by insufficient runtimes. Unlike standard methods, centrifuge experiments require inversion of the angular velocity and average sample moisture content data to a location-specific pair of matric potential and moisture content values, The force balance causes matric potential and moisture content to vary with sample length while the sample is spinning. Numerous data inversion techniques exist, each yielding different moisture retention relations. We present analyses demonstrating the impact of different runtimes and different inversion techniques on estimated moisture retention parameters. Moisture retention data were collected for a number of poorly lithified sands and indurated deformed sands using the UFA centrifuge system (Conca and Wright, 1990). Parameters for the van Genuchten model were estimated for short and long runtimes with one inversion technique. Model parameters were re-estimated for one other inversion technique and a simple averaging approach which does not involve inversion. Our results demonstrate that the averaging approach greatly underestimates the van Genuchten n parameter relative to the inversion techniques. Insufficient runtimes also have a significant impact on estimated parameters. Our analysis indicates a need, barring method standardization, for practitioners to include information about inversion technique and runtime criteria when presenting centrifuge moisture retention results.
DE: 1829 Groundwater hydrology
DE: 1866 Soil moisture
DE: 1875 Unsaturated zone
DE: 1894 Instruments and techniques
SC: Hydrology [H]
MN: 2003 Fall Meeting