The “Above Magenta” Reservoir at Ursa: A Process-Response Model to Explain a Classic Log Signature

Schofield, Kevin, and John Serbeck, Conoco Inc

Extended Abstract

In the Gulf of Mexico, and many other deepwater slope and basin settings, the gamma ray log profile of a blocky, low-gamma unit capped by a “fining upward” trend of increasing gamma-ray response back to the shale baseline is commonly interpreted as the signature of a leveed channel system. The above-magenta sand at Ursa, as exemplified by the MC809-1 well, is a classic example of this profile.

In the “standard” interpretation of this log facies, the blocky sand at the base is representative of the channel fill or lag, and the overlying fining-upward interval the levee. A brief consideration of the process dynamics of a leveed shows that this interpretation is unlikely to be viable, as leveed channels do not migrate laterally over the distances which would be required to deposit a full levee profile over a channel fill. This can be seen empirically in images of modern leveed systems, and has been demonstrated theoretically by Peakall et al. (in press), who show that in their proposed three-stage evolution of a highly sinuous channel, the degree of lateral accretion in “Stage 1” is minimal compared to the width of the overbank system.

We propose an alternative model for explaining the vertical profile observed, at least in the case of the Above Magenta sand. The model provides an explanation for the observed electric log, core and seismic character of the sands, which are not satisfactorily accounted for by the “standard” model. More importantly, it offers a predictive description of the geometry of the reservoir sands which suggests greater connectivity and lateral extent than would be expected from the channel model. The Above Magenta reservoir at Ursa is closely associated with the western edge of the Ursa sub-basin, with reservoir sands deposited adjacent to the basin-bounding “Venus” salt. The seismic signature of the system shows an elongate bright/thick parallel to the salt wall. Interestingly enough, no channel is seen on the seismic either to the north or south of the immediate area of the salt. The core from the interval in the MC809-1 well shows exactly what would be expected from the log signature: an interval of stacked, clean sands, overlain by a series of thinly-interbedded sands and muds, with decreasing net-to-gross upward. Although they have not been cored, subsequent penetrations of the same interval all show a similar log signature, and are presumed to have a similar general facies distribution.

A closer examination of the facies in the cores, however, reveals a rather more complex depositional history than a simple amalgamated channel-fill. There are at least three sandbodies represented in the core, with a hiatus between the first and second during which time there appears to have been sand deposition elsewhere, as shown by a single flow-event bed within the shale-prone interval. The three sands have very different, but internally-consistent sedimentological character:

1. The basal sand, some 17’6” thick, has somewhat variable grain-size between and within flow units, and has abundant sedimentary structures (ripple lamination, climbing ripple sets, planar laminations).
2. The second sand, 21’ thick, has a very consistent grain-size, and is virtually homogeneous, with few amalgamation surfaces of bedding planes. It has abundant scattered mudstone chips, and a characteristic drilling-fluid invasion rim. It is capped by a slurry or debris flow.
3. The third sand, estimated at 18’ thick (the central portion of core was not recovered) is very similar to the second, with fewer mudclasts, and lacking the invasion rim, suggesting an overall less permeable sand.

There is no de facto reason why this facies distribution should not represent the amalgamation of a series of individual bar-forms within a channel. If that were the case, however, we would expect to see evidence of the sandy system to the north and south of the Venus salt. Also, the cored well sits on the periphery of the major amplitude thick, implying that if these are channel-fill facies, the channel must be very much deeper to the west. This does not appear to be geometrically likely, given the thickness of the “levee”.

The model we propose for the Above Magenta sands is that they are not associated with a channel-fill, but rather a mouth-bar, or “HARP” facies associated with a depositional system (possibly initially channelized) which entered the basin from the north or northeast at the start of a sand-prone cycle of deposition. Density currents passing through this system encountered the slope associated with the basin-margin defined by the Venus salt. Assuming that the flows had traveled down the slope to this point in a state of equilibrium (steady/uniform sensu Kneller, 1995), encountering this gradient would probably have changed the flow state to a depletive condition (again, sensu Kneller, op.cit.), encouraging deposition in the region of the basin-edge. The most likely first products of these flows would be from a steady- or waning-depletive flow regime, giving rise to a rapid deposition of massive sands with poorly-defined amalgamation surfaces. In the case of the lower sand in the MC809-1 well, where tractional currents are in evidence, we propose that this is the result of deposition from a flow which had already been substantially depleted upstream, allowing the lower concentration flow to interact with the substrate to produce a much more highly structured bed. The potential flow-path of such a current is illustrated.

Kneller (1995, op.cit.) provides a simple experimental illustration of the shape of the deposit of a single surging flow encountering a lateral 30^º ramp. This shows that deposition adjacent to the ramp is 3 to 4 times greater in thickness than would be expected from an unimpeded flow passing “north to south” through the tank. The deposit is much thicker at the input end of the flow path, waning to zero at the southern end. Thus, we anticipate that in the region immediately adjacent to the “Venus” salt, thickly-amalgamated lobate sands should occur, forming a series of laterally-offsetting sheets, separated by thinner intervals of more distal facies, representing periods of relative inactivity when the thickness of sand “encouraged” deposition to migrate to adjacent lows. Away from the salt-defined edge of the basin, flows would be relatively unimpeded, and flow through the basin with little reason for deposition. As the area adjacent to the basin margin accumulated more sand, then the centre of deposition was able to migrate further east and south. It would not, however, be expected to extend beyond the southern margin of the salt. This model explains the seismically observed (as yet untested in wells) increase in thickness towards the “Venus” salt, and the apparent cessation of thick sand deposition to the south. It also implies better lateral amalgamation of sands than the channel-fill model.

The overlying low net-to-gross package may represent the levee deposits of a channel system which developed in the centre of the basin subsequent to the phase of deposition described above. It may also, however, simply represent a period of unconfined flow and relatively slow deposition as the character of the deposystem evolved from sand-rich to sand-poor through the depositional cycle.

References Cited

Kneller, B., 1995: Beyond the turbidite paradigm: physical models for the deposition of turbidites and their implications for reservoir prediction. In: Hartley, A.J., and Prosser, D.J., (Eds): Characterization of Deep Marine Systems. Geological Society of London Special Publication 94, pp. 31-49.

Peakall, J., McCaffrey, B, and Kneller, B., (in prep.): A process model for the evolution, morphology and architecture of sinuous submarine channels. Jour. Sed Research.