Considering Hysteresis in LNAPL Lens Geometry, Mobility, and Well Thicknesses

When released to the subsurface, LNAPLs (light nonaqueous phase liquids) will distribute as residual saturation in the form of disconnected blobs and ganglia or as larger accumulations in the form of mobile lenses near the water table. The conceptual model assumed for the distribution of these immiscible contaminants greatly affects important characteristics such as the rate of mass transfer, the mobility of the contaminant, and estimates for the areal extent and volume of contamination. Although the transport and distribution of LNAPLs is strongly affected by subsurface heterogeneities and knowledge of stratigraphic detail is typically limited, a better understanding of NAPL distribution in homogeneous and simple layered soils should result in more accurate conceptualizations of site contamination, better interpretations of observed field data, and more realistic mobility estimates. This paper highlights a new conceptual model for the geometry of LNAPL pools and lenses and illustrates how the phenomenon of saturation hysteresis and the concept of imbibition entry pressures plays a critical role in the description of LNAPL lenses at equilibrium. It is proposed that the geometry of LNAPL lenses cannot be fully understood without considering hysteresis in capillary pressure-saturation relations. In fact, it is proposed that NAPL lenses of any measurable thickness cannot be stable without this hysteresis. Considering this, attempts to model LNAPL distribution must account for this phenomenon, and it can also be important in estimating LNAPL mobility or in interpreting product thickness in monitoring wells. This conceptual model was verified by laboratory flume studies and applied to calculations of maximum LNAPL lens thickness and the location of an LNAPL lens with respect to the capillary fringe. It was found that the model was a good predictor of these values. The proposed model also implies that the mobilization of LNAPL lenses will occur at lower hydraulic gradients than would otherwise be predicted. Additionally, the potential significance of accounting for hysteresis and imbibition entry pressures when interpreting well product thicknesses and other field data is discussed.