Abstract The standard waterflood prediction method, based on the familiar multilayer reservoir model was modified so that provision was made for resaturating the free gas space in each layer. The method was then applied to an actual reservoir case for various assumed stages of primary depletion prior to waterflooding. Calculations were made with the aid of an IBM 650 computer. Results were significantly different from those obtained by the standard prediction method of Dykstra-Parsons, Stiles, and Suder-Calhoun. Inclusion of the resaturation concept was especially important when trying to evaluate the optimum formation pressure at which to start a flood in order to produce the most oil (primary plus secondary). Other optional modifications allow for layer-to-layer variations of porosity and initial and final saturations. An illustrative example was worked out for a hypothetical reservoir. Results indicated that the effect of the optional modifications is probably not large enough to justify their use in most reservoir applications. Introduction In preparing for a recent hearing before a state regulatory body, the need arose to determine the optimum formation pressure at which to start a waterflood in order to obtain the greatest total yield of primary and secondary oil from a reservoir. In order to evaluate the optimum pressure, it was necessary to modify the standard waterflood prediction methods to allow for liquid resaturation or fill-up of free gas space developed during primary depletion. In the improved method (termed the resaturation method), it Is assumed that either all or most of the free gas space is resaturated with oil in each individual layer of the standard "layer cake" reservoir model. Other optional refinements allow for layer-to-layer variations of porosity and initial and final saturations. DESCRIPTION OF NEW METHOD Callaway pointed out that waterflood predictions must make provision for resaturating the free gas space in the reservoir. The concept when applied to the reservoir as a whole, however, tends to present an over conservative picture for depleted reservoirs. Such a view would be equivalent to saying that all of the depleted portions of the entire reservoir must first be resaturated with oil before waterflood oil is produced. Instead, we prefer to assume that resaturation occurs in each permeability stratum or layer of a reservoir separately. Once a given layer is resaturated, waterflood oil production will start for that layer. Equations were written that allow for resaturation in each layer. The equations also make optional provisions for a residual gas phase behind the water front and for layer-to-layer variations in porosity and initial and final saturations. The reservoir model used was the same as proposed by Dykstra-Pasrsons, i.e., a "layer cake" model with uniform permeability strata arranged in ascending or descending order (Figure 1). This is admittedly a somewhat artificial view of a reservoir. Its justification lies in the fact that such a model is amenable to simple mathematical solutions; and, more importantly, predictions based on such a model have produced reasonably close agreement with actual performance in many cases.