Elk Basin Madison experience underscores the need for a good understanding of reservoir heterogeneity and subzone performance early in field life. That understanding was necessary in this case before a fully meaningful engineering analysis of the reservoir could be made. Introduction The Elk Basin anticline is one of the giant oil fields in the U. S. (see Fig. 1). It had more than a billion barrels of oil originally in place and has produced at rates in excess of 70,000 BOPD. Production is from seven horizons ranging in depth from 1,000 to 6,500 ft (Fig. 2). The anticline has dimensions of about 2 X 7 miles. Dip is about 25 degrees on the west flank and about 45 degrees on the east flank. Horizons below the Dakota are included in the Elk Basin Unit, operated by Pan American Petroleum Corp. for a large group of working interest owners. The Tensleep sandstone (Pennsylvanian) and Madison limestone (Mississippian) are the most important reservoirs, each having about a half billion barrels of original oil in place. In 1965, the Madison Unit began to experience waterflooding problems for which there was no readily apparent solution. Peripheral water injection initiated in 1962 did not get response. After part of the water was shifted to the interior, response was obtained, but water breakthrough, wellbore scaling and severe productivity declines soon became problems. The oil producing rate again began to decline. A preliminary review of logs and core analyses indicated highly complex zonation in the 920-ft vertical Madison section. A detailed zonation study was therefore undertaken to determine the degree of continuity in the reservoir, with the hope of finding answers to the waterflooding problems. The study was a coordinated effort by both geologists and engineers. Field Development History The Madison reservoir was discovered in 1946 when a well on the crest of the north high (see Fig. 3) was deepened from the Tensleep to the Madison. A number of cores and DST's established that porous intervals in the entire Madison section were productive and that a major discovery had been made. The discovery well had a flowing potential of 898 BOPD. Development was slow until 1948 when a market was found for the black, asphaltic-base, 28 degrees API crude. Seventeen wells were drilled in 1948 and eight more in 1950. Field producing rate increased to about 7,000 BOPD and remained in the 3,000 to 8,000 BOPD range until 1958 (see Fig. 4). During the late 1950's and early 1960's market demand of about 16,000 BOPD was met by drilling 27 additional producing wells and by installing larger rift equipment. From 1964 to the present, market has been essentially unrestricted, but no significant drilling, other than injection wells, was done until 1965 when the first wells were drilled as a result of the zonation study. Early Reservoir Concepts Early ideas about the Madison reservoir were influenced by the fact that most wells were completed open hole through the entire 920-ft section. First production performance was markedly water driveonly a slight pressure decline occurred in the first 10 years of field life. JPT P. 153ˆ

A study of 20 Denver Basin peripheral waterfloods showed that the rate of waterflood oil recovery was influenced by the rate of water injection and by the water cycle volume; but principally it was influenced by waterflood timing. Discussed here are the magnitude and relative influence of these factors on waterflood cash flow and present worth. present worth. Introduction Peripheral injection is frequently chosen as the best Peripheral injection is frequently chosen as the best means to waterflood a reservoir because the peripheral wells contribute less to current income. Also, the peripheral wells are often advantageously located downdip from the oil column. Because of their structural position they may have experienced some water influx or may have had water originally present in part of the sand section. part of the sand section. On the other hand, a peripheral waterflood is especially sensitive to operating decisions made before and during the life of the waterflood. If there are several rows of producers to be stimulated by one row of injectors, a slow waterflood start or a low injection rate can result in severely delayed waterflood response at interior producers. If large volumes of injected water are removed at close offsets to injectors, further delay or even a stalled waterflood may result. The many completed or essentially completed peripheral waterfloods in the Denver basin provide an peripheral waterfloods in the Denver basin provide an excellent opportunity to study waterflood experience and to see how decisions concerning the major operating alternatives - that is, waterflood timing, the injection rate, and the water-cycle volume - have influenced peripheral waterflood cash flow. Waterflood Recovery An earlier paper reported the results of an analysis of waterflood recovery data from 23 Denver Basin waterfloods (20 were peripheral-type floods). The analysis indicated that ultimate recovery was influenced by the timing of a waterflood - the earlier a waterflood is conducted after reservoir pressure reaches the bubble point, the higher is the ultimate recovery. The reduction in recovery was shown to be due to shrinkage effects as expressed in Eq. 1. (1) where Erwf = waterflood recovery efficiency, percent, Cr = relative conformance, percent, E = (This is water-drive recovery efficiency from Arps et al., assuming no pressure drawdown.) Bob = reservoir volume factor at bubble point, Bowf = reservoir volume factor at waterflood start. Eq. 1 indicates that waterflood recovery will be reduced significantly if the reservoir volume factor is high and if a waterflood is delayed until most of the primary oil has been produced. primary oil has been produced. Table 1 shows operating data from 23 waterfloods. P. 1320