John Holbrook
University of Texas at Arlington
Fluvial sandstone layers can be very complex internally, making them particularly frustrating reservoirs to develop. What appears as a uniform sheet of high-permeability sand may turn out to be riddled with numerous and seemingly random clay-draped discontinuity surfaces that may complicate flow models and/or compartmentalize otherwise continuous reservoirs.
Because sea level cycles tend to juxtapose porous fluvial sandstone over shelfal marine source rock during regression, then conveniently cover them with sealing marine shale during subsequent transgression, such sands are highly sought as offshore targets.
The concept of "base-level buffers and buttresses" was developed to explain internal complexity in fluvial strata encased by marine shale in the now emergent Dakota Group of the US Western Interior Cretaceous, and builds upon existing sequence stratigraphic models. These existing models predict that channels will build upward (aggrade) slowly during marine lowstand, forming sand-rich intervals with internal architecture typified by amalgamated and generally progressively stacked channel belts. As transgression ensues, the more rapidly rising sea level causes rivers to rapidly aggrade, resulting in disseminated and progressively stacked sandy channel belts separated by muddy floodbasin deposits.
This pattern is typical of Dakota fluvial strata proximal to the marine lowstand limit where the models were intended to apply. Further up depositional dip, however, fluvial intervals equivalent to the sand-rich lowstand phase reveal multiple generations of valley-scale incision and filling cycles that do not appear to be sea-level driven. A complex hierarchy of potentially flow-impeding scour surfaces within otherwise well-developed sandstone sheets and lenses preserves these cycles. Current models do not address this dip-oriented change in architectural complexity.
Buffers and buttresses expands upon existing models to take these complex factors into account and serves as a predictive model generally applicable to fluvial deposits now submerged on broad shelves, where the sort of fine-scale architectural detail inherent to Dakota sandstone layers is impractical to observe directly.
null
null
General model
Rivers tend to develop a characteristic longitudinal elevation profile that has a shape dictated by the contemporary balance between the forming river's ability to incise versus aggrade valleys at any location. The ability of a river to locally incise or aggrade depends on the balance between stream power (local erosive/transporting energy) and sediment supply. Every profile will be anchored to some base-level "buttress," which defines the downstream point to which the stream must adjust, with sea level being the most common example.
Sea level may control stream power by controlling the elevation and location at which the river must finally debauch. Thus, it controls the average gradient of the river between source area and the strand line sink.
Several other factors also control stream power and/or sediment supply, including water discharge, substrate erodability, tectonic uplift, source-area erosion mechanisms, etc. These factors are more often dependent on upstream climatic and tectonic controls than on downstream control by sea level.
Extensive fluvial sedimentary units tend to record deposition on scales of 105-106 years, during which time all profile-controlling variables may see considerable change. Charac-terization of river deposition by movement of a single longitudinal profile in trail of the sea is thus unrealistic for enduring fluvial deposition.
An alternative approach is to consider downstream control by sea level as a "buttress," which can be held stable for some period of time. Then consider that the fluctuation in the other upstream variables over this stable period will result in a range of instantaneous profiles that will fall between a maximum and a minimum possible river profile.
These maximum and minimum profiles constitute "buffers" between which the instantaneous river profile must remain. The reasonable preservation space for fluvial sediments, excluding that from tectonic subsidence locally beneath the profile, is some part of the space between the buffer profiles.
Since both the upper and lower buffers must meet at the anchoring buttress and they record maximum and minimum possible profile conditions by definition, they diverge updip for some unresolved distance. Rivers in the downstream areas have very little vertical preservation space between the buffers, and are not prone to aggradation or incision unless driven by movement of the buttress.
Further updip, the river has more vertical preservation space and may aggrade or incise freely within these confines without the need for a shift in the buttress position. This can result in repeated incision/aggradation cycles to the same elevations between the buffers at valley-fill scales. This can result in a complex hierarchy of crosscutting bounding surfaces like that typical of the architecture for Dakota lowstand equivalent sandstone in updip areas.
Buttresses such as sea level can move, of course, and the buffer profiles must move with them. Because the buffers are narrow near the buttress, rivers just updip are highly sensitive to any buttress shift. Rivers near the strand are prone to aggrade or incise linearly with change in sea level elevation, and their architecture can be predicted well by current fluvial sequence stratigraphic models. Effects of sea level rise and fall on profile elevation, however, will dissipate updip. As a result, the buffer elevations are not necessarily forced to move because of sea-level rise.
Likewise, most sea-level rise and fall across low-angle shelves does not substantially alter coeval profile slope even at the strand, but only shortens or lengthens this profile, producing minimal effect on aggradation in even the downdip reaches.
Elevation of the buffers may not be substantially changed in updip areas by sea level change further downdip, permitting rivers to aggrade and incise to similar levels throughout long portions of the sea-level rise and fall cycle. This buffer stability appears to have produced the complex architecture observed in more updip areas of the Dakota lowstand deposits.
null
Production implications
Modern exploration tools and concepts can readily yield a good understanding of the spatial relationship between a fluvial reservoir and its enveloping strata, the contemporary shelf edge, and/or its equivalent lowstand marine deposits. Identifying at least roughly where an offshore fluvial target lies within the overall scope of its contemporary hinterland source to depositional sink system is quite feasible. Constraining the fine-scale internal architecture of the reservoir in a specific region targeted for reservoir development can prove more difficult.
Buffers and buttresses offer a conceptual model that allow first-cut predictions of the expected architecture of these deposits and permeability structure. It does this from general knowledge of a fluvial reservoir locally and a sense of where the reservoir is located within the broader contemporary fluvial system.
The model predicts that extensive fluvial layers, usually ascribed to lowstand, will have a tendency to alter their architecture somewhat systematically in the updip direction away from the equivalent marine lowstand depocenter. Near the depocenter, the buffers are narrow, and rivers may not aggrade or incise appreciably from the elevation of concurrent sea level. Sea-level rise during the final stages of lowstand and the initial stages of transgression forces the channels to aggrade with this rising buttress, resulting in an architecture of progressively stacked channel belts.
The rate of aggradation compared to the sediment supply determines the degree of amalgamation of channel belts, and how sand-rich the reservoir. Since this phase is typified by slow aggradation, vertical amalgamation and sand content tends to be high. If the sediment supply is very high with respect to aggradation, vertical space is insufficient to store these sediments. As a result, rivers carve laterally in response and form very broad exploration-friendly sheets of laterally continuous and highly amalgamated channel-belt sands.
Further updip (a few tens of kilometers from the lowstand strand), these same rivers deposit sediment within a thicker buffer zone with much more vertical freedom. Rivers here aggrade and degrade within the buffer zone in response to changes in climate and tectonics without notable control by sea-level shifts. Because the rivers cannot aggrade or incise beyond the buffers, however, the thickness of this deposit is dictated by the thickness of the buffer zone, which in turn is dictated somewhat by the position of the anchoring sea-level buttress.
If sediment supply is strong and the pro-cesses driving incision/aggradation cycles are vigorous, rivers aggrade and incise multiple times within some constrained preservation space between the buffers. This results in intensive reincision of prior fluvial deposits at the valley scale, creating an architecture of laterally and vertically cross-cutting valley fills and preserving of a complex hierarchy of internal scour surfaces ranging from channel to valley scale. These valleys commonly avulse where they fill to the top of the preservation zone.
Similarly to above, very high sediment supply compared to thickness of the preservation space between the buffers can cause valleys to amalgamate laterally to form laterally extensive "multi-valley" sheets.
Understanding and predicting the stacking architecture and the makeup of internal scour surfaces is the key to predicting permeability distribution in fluvial sandstone. Each scour records a discontinuity surface across which flow conditions have changed. Thus, the primary location where contrasting lithofacies should be expected to juxtapose is where fluvial units stack across these scour surfaces.
The juxtaposition of high and low permeability units can also occur across channel scours, channel-belt boundaries, or any surface in the overall hierarchy of fluvial scours. Likewise, each scour records a moment of erosion followed by a period of deposition. It is common that after such scouring, the river contains a surplus of suspended sediment, and it is also common that flow wanes dramatically after intensive scour events.
This means that scour surfaces, especially the smaller ones, have a tendency to be draped by thin but impermeable slackwater muds. These mud drapes can follow parts or all the lengths of scour surfaces and serve as powerful flow barriers. The distribution of these flow-impeding drapes is dictated by the architecture of the internal bounding scours.
