Models for Gas Accumulation in Low-Permeability Reservoirs, Rocky Mountain Region, U.S.A. – An Evolution of Ideas and Their Impact on Exploration and Resource Assessment

Keith W. Shanley¹, Robert M. Cluff¹, and John W. Robinson^2
1 The Discovery Group, Denver, CO
² Consultant, Denver, CO

Introduction

The Rocky Mountain region of the western United States contains several sedimentary basins long known to contain substantial volumes of natural gas. Many of the gas accumulations in these basins are hosted in low-permeability reservoir rock and are widely referred to as ‘basin-center’, ‘deep basin’, or ‘continuous-type’ accumulations. It is the opinion of many that these gas accumulations are distinctly different from gas occurrences in more conventional petroleum provinces and comprise a separate type of petroleum system. The descriptors used in association with these gas accumulations attempt to convey the sense that these fields are thought to have large aerial extent, that they lack clear gas-water contacts, that they occur in a portion of the basin that is gas-saturated, that water has been driven down to near- irreducible levels, and that production is largely determined by the ability to locate ‘sweet spots’ where the appropriate drilling and completion technology can be brought to bear. Because these accumulations are not easily understood in terms of conventional petroleum geology and do not easily appear as discrete gas accumulations, resource appraisal methods tend to be strongly statistical in nature and often take a ‘cell-based’ approach in which the size of cells approximate median drainage areas of individual wells. The consequence of this is the implication that there are enormous resource volumes both in-place as well as technically recoverable and, just as importantly, that many of the risks commonly associated with key elements of the petroleum system such as trap, seal, source, migration, charge, and reservoir are reduced in these gas-saturated regions. Because the magnitude of these projected resources are so large they are thought to comprise a significant portion of the nation’s energy endowment and have become a highly sought-after asset type within the portfolios of many energy companies wishing to balance reserve-to-production ratios and portfolio risk.

The notion of ‘basin-center’ gas accumulations has recently been challenged and it has been suggested that these gas accumulations are essentially conventional in nature. The unusual aspect of these gas accumulations lies not in their occurrence or distribution, but rather in the nature of fluid flow in very low-permeability reservoirs. The challenge to the prevailing paradigm is based on insights gained from the drilling of numerous wells, the collection of numerous field studies, greatly improved methods of core analysis, and an examination of water production within these basins. If gas resources are more discretely distributed and have more affinity to conventional petroleum provinces, then the estimates of available resources have almost certainly been overstated and the perceived risks of finding gas have almost certainly been underestimated. Understanding the resource volumes and the associated uncertainties is of critical importance at the national scale as well as for individual entities evaluating investment decisions.

The remainder of this short paper traces the evolution of ideas concerning the nature of petroleum systems associated with these low-permeability rocks. Rather than searching for a ‘fatal flaw’ in this historical development, we view the collection of key data in light of prevailing concepts at the time the data were collected. We regard the change in perception of these gas-rich, low-permeability systems to reflect the emergence of new data and the introduction of additional concepts with which to interpret those data. We conclude with a brief discussion of where we believe key problems remain.

Development of the ‘basin-center’ gas paradigm

It has long been known that many of the sedimentary basins in the Rocky Mountain region contain prodigious volumes of source rocks, have complex successions of potential reservoir rocks, and have widespread gas shows upon drilling. It has also long been appreciated that evaluating the hydrocarbon potential of these basins is difficult. Prior to approximately the early 1990’s wells drilled in the deeper portions of basins such as the Greater Green River Basin of southwest Wyoming were commonly evaluated through a combination of wireline logs, mud logs, and drill-stem tests. For many operators wireline log interpretation on its own was inconclusive, mud-log shows were common, and drill-stem tests offered an opportunity for the rocks ‘to perform’ and provide critical information concerning the nature of subsurface fluids and, sometimes, subsurface pressure. It was well understood that there were inherent problems with drill-stem tests in these low-permeability regions. Shut-in times were often insufficient for adequate pressure buildup and the interpretation of pressure data required considerable care. Nevertheless, fluid recoveries in these basins were regarded as critical information and pressure data was often used to distinguish ‘tight’ areas from areas with more improved permeability. (Since the early 1990’s industry has, by and large, discounted the usefulness of drill-stem test data in very low-permeability reservoirs as evidenced by the dramatic drop in the number of drill-stem tests run over the last 10-15 years).

In the deeper portions of the Greater Green River Basin, and other similar basins, drill-stem tests of the principle low-permeability Upper Cretaceous reservoir intervals occasionally yielded free-gas recovery, often resulted in mud, or gas-cut mud, sometimes mixed with some free-gas, and rarely yielded any significant free-water or water that was clearly distinguishable from the water cushion used in the drill-stem test itself. This lack of substantial water recovery, coupled with the very low-permeability of the various reservoirs proved critical. General reservoir engineering principles based on a wealth of experience in more permeable reservoir systems suggested that if free-water was not recovered, then the formation of interest was likely at a water-saturation at, or below, critical water saturation (the water saturation at which there is sufficient relative permeability to water for water to flow). This same experience suggested that critical water saturation and irreducible water saturation were very close, with only minimal separation in terms of water saturation. The consequence of this was that the low volumes of water production either in producing wells or in drill-stem tests became associated with the notion that large portions of a basin were at, or near, irreducible water saturation and, therefore, could not produce significant free-water. The corollary of a large area thought to be at irreducible water saturation is a large area fully gas-saturated. From this point forward, geologic models began to focus on methods that might explain how formation waters had been expelled and how the basin had been driven to irreducible water saturation. it was an almost natural outcome that the role of hydrocarbon buoyancy would diminish under this paradigm.

By contrast, drill-stem tests in wells drilled in the more shallow portions of these same basins, where permeability was noticeably better, commonly recovered free-gas, free-water that was clearly in excess of the water cushion, and occasionally mud or gas-cut mud. The interpretation of these test recoveries led to the idea that unlike the deeper basin, the basin margins were clearly not at irreducible water saturation and gas accumulations existed in the presence of free-formation water, similar to conventional petroleum provinces. To explain the apparent juxtaposition of these two distinct types of accumulations two primary models evolved, (1) a ‘water block’ model in which the decrease in relative permeability from gas saturated to water saturated rock forms an effective updip barrier, or (2) downdip hydrodynamic flow into the basin center that prevented updip migration of gas.

The emergence of a potentially unique type of gas accumulation required innovative approaches to resource evaluations. If large portions of sedimentary basins were gas-saturated and discrete accumulations were not present, and if the role of buoyancy were reduced, then traditional approaches of examining field-size distributions would not adequately address these targets. In an attempt to address the resource potential of these areas a cell-based approach evolved in which play areas, thought to be gas-saturated were divided into cells that approximated the median drainage area of individual wells and resource volumes were tied to the performance of existing wells that had already tested some of the cells. Minimum-size recovery volumes associated with individual cells were a fraction of the minimum field-sizes considered in more conventional analyses. Additional statistical treatments of the data allowed for scattered areas to develop (i.e. sweet spots) with better than average recovery. The result, however, was the prediction of very large in-place, as well as technically-recoverable resources. In the resource assessment community, these areas became known as ‘continuous-type’ or ‘basin-centered’ gas accumulations. With the acceptance of this new resource type, the technical challenge began to move from one of finding a resource, long recognized as a risky endeavor in conventional provinces, to asking whether a known resource could be technically extracted.

An alternative hypothesis

The ‘basin-center’ gas model was developed and generally accepted in the petroleum geology community from about the mid- 1980’s onwards. Additional data were incorporated to account for basin-to-basin differences, however, the interpretation of fundamental observations relating to water production, irreducible water saturation and therefore gas-saturation, and the diminished role of buoyancy remained essentially unchallenged until recently. Several broad sets of data have recently come together that have caused a re-examination of all observations within these low-permeability regions and have led to the emergence of an alternative model in which these low-permeability gas accumulations are viewed as essentially conventional in nature.

Prior to the mid-1990’s the collection of effective permeability data in low-permeability reservoirs was generally limited to a few samples in a few key study areas. Much of the data that was collected remained in the engineering community and was neither widely appreciated by the geoscience community nor incorporated into discussions of ‘basin-center’ gas. Beginning in the mid-1990’s, however, a large dataset of effective permeability has emerged with a concentration of data from low-permeability reservoirs in the Greater Green River Basin. For many of the samples in this dataset capillary pressure data were also collected allowing for a comparison of critical gas and water saturation to irreducible water saturation. These data combined with permeability modeling suggest that in low-permeability reservoirs critical water saturations are substantially less than irreducible water saturations and that critical gas saturations occur in the vicinity of 50% water saturation. These observations are captured in a model widely referred to as ‘permeability jail’. The implications are significant. The lack of water production is no longer directly linked to the concept of irreducible water saturation. The lack of water production simply means that the water saturation is less than critical water saturation, but could be substantially higher than irreducible water saturation. If the rocks were no longer at irreducible water saturation, then they need not be fully gas-saturated.

A key dataset has emerged from a re-examination of water production. Most operators describe their fields as producing ‘essentially water free’ with the vast majority of produced water assigned to ‘water of condensation’ in the gas stream. While it is true that most wells in these basins produce very small volumes of water, it is also true that most of the wells produce at relatively low gas rates as well. Although absolute volumes of water are indeed quite small, water-gas ratios (bbls water/MMscf gas) are much higher than can be explained through phase considerations of the gas stream. In the Greater Green River Basin, for example, waters of condensation should be less than approximately 1.0 bbl water/MMscf gas. Of more than 7500 producing gas wells in the Greater Green River Basin, 70% of the wells have water-gas ratios in excess of 1.0 bbl water/MMscf gas. Those wells account for almost 50% of the basin’s gas production. Similar results have been found in the Piceance Basin (4,739 wells) and Uinta Basin (3,725 wells) where approximately 70% and 80% respectively, of the wells have water-gas ratios in excess the volume that can be accounted for as ‘water of condensation’. Clearly, these basins are not water-free and formation water is common and widespread. In many fields, water-gas ratio data show a clear increase in water-gas ratio towards the downdip margins of gas accumulations, something that is commonly observed in conventional petroleum provinces where buoyancy is a dominant process.

A compilation of field studies on major fields throughout the Greater Green River Basin indicated that 100% of the fields examined occurred in conventional trapping geometries. Approximately 38% of the gas fields involve structural traps accounting for 50% of the gas production, 41% of the gas fields involve stratigraphic traps accounting for 30% of the gas production, and 21% of the fields occur in combination traps contributing 20% of the gas production. In no case, was a significant gas field found to occur as a ‘sweet spot’ within a background matrix of poor rock. Nor was there a significant gas field that could be explained as simply the preferential occurrence of natural fractures. In fact, the data suggest that the presence of natural fractures were more likely contribute to increased water production.

None of the findings of these broad datasets were predicted, nor are they adequately addressed, by the basin-center model. The conclusions we reached when all the data were amassed are that (1) gas accumulations are conventional in nature and occur in structural, stratigraphic, or combination traps and are buoyancy-driven, (2) multiphase flow in very low-permeability rocks is such that our traditional ideas of the relationship of irreducible water saturation to critical water saturation must be re-considered, (3) the occurrence of permeability jail is such that there is a broad range of water-saturations in which there is little-to-no fluid flow despite the presence of both gas and water, (4) water production in these basins is common and widespread and cannot be dismissed as water of condensation, (5) all the risks commonly associated with the various elements petroleum systems analysis occur in these basins and these are not inherently low-risk provinces, and (6) it is likely resource assessment have significantly overstated the resource potential and understated the potential risks associated with the elements of the petroleum-system.

So What?

The perception of reduced exploration and development risk in these regions is dependent on whether these are actually continuous-type accumulations or not and whether the risk elements of the petroleum system are significantly less than in conventional provinces. At the regional and national scale, considerations of energy supply are presently predicated on the existence of a very large resource base that is accessed through improvements in cost-of-supply functions (i.e. improvements in extraction technology). Economic models presently have a very large ‘reserve bank’ that can be drawn on depending on changes in cost-of-supply functions. Virtually none of these models considers a dramatic variation in either the absolute size of the available resource, or the probability that the resource size is present. The notion that resources might be conventionally distributed has not been rigorously addressed in most assessments or the subsequent economic scenarios.

Prudent decisions require a sound scientific basis and understanding of uncertainty irregardless of whether those decisions concern national energy supply and national energy policy, or a corporate decision to enter a gas-play in the Rocky Mountain region, or an individual or corporate decision to drill a particular well or invest in a particular deal.

The Future

Petroleum exploration forces us to constantly try to explain a multi-dimensional system with insufficient data and to make predictions at times when we lack sufficient knowledge. However, we need to focus on understanding the fundamental controls on gas accumulations in these difficult settings and the uncertainties that surround our understanding. The following is a partial list of areas where future research might shed light on our understanding of these important gas provinces.

1. Source-rock, gas-generation, and basin-modeling: Despite the drilling of thousands of wells, there remain very few quantitative, calibrated studies (and even fewer publicly available studies) that describe the relationships between source-rock potential, burial and uplift history, expulsion timing, generative volumes, reconstructed migration pathways, and trap timing. If these basins are indeed dominated by conventional traps then an understanding of these elements should improve our assessments of less-densely drilled areas and help us better understand the risks associated with certain trap styles.
2. Reservoir-description: Attempts to predict reservoir performance are, fundamentally, based on ideas of rock-typing. The majority of efforts to describe reservoir performance attempt to do so on the basis of routine porosity and permeability data, and occasionally on the basis of stressed data. It is abundantly clear that multiphase fluid flow in very low-permeability reservoirs is not only very stress-sensitive, but is different from our more traditional reservoir types. Reservoir description should begin to deal with this observation and attempt to develop theoretical links between laboratory and wireline measurements. Progress along this front will allow for much more intelligent drilling and completion decisions.
3. Pressure evolution: The original ‘basin-center’ model suggested that overpressure in these basins was due to source-rock maturity and gas expulsion. Well data, however, suggests that there is a much more complicated process at work. Maps that attempt to describe the onset of overpressure have significant topography that sometimes changes on a high-frequency basis. Some initial work suggests to us that pressure-history and burial history are closely linked in these basins. Because overpressured areas have lower effective stress and are able to contain greater volumes of gas, they are also likely to have improved production characteristics. Advances in our understanding of subsurface pressure in these basins may allow for more effective exploration strategies.
4. Assessment methodologies: The simple fact is that large financial decisions at the corporate and governmental level are based, in part, on an understanding of resource volumes and the associated uncertainties. If the gas accumulations in some of these provinces are indeed conventional, then a comparison of assessment methodologies (continuous-type and conventional) in the same basin using similar datasets might prove illuminating and might indicate the magnitude of the problem of having multiple paradigms. The use of cell-based methodologies should not necessarily be restricted to continuous-type accumulations. Some reservoir-bearing intervals are so heterogeneous that cell-based approaches may prove useful in the prediction of resource volumes in ‘conventional-type’ accumulations.

The available dataset with which to examine these important gas provinces is impressive. We firmly believe that integration of data on multi-disciplinary levels will result in dramatic improvements in our understanding of these low-permeability gas provinces. The emergence of new interpretations and model is healthy. It means that there is new data coming to light and that we, as a community, are asking difficult questions and striving for better solutions.