Groundwater Flow, Sandstone, Denver Basin_AAPG1994

8/3/2019 Groundwater Flow, Sandstone, Denver Basin_AAPG1994

1/21

ABSTRACT

The gray diagenetic facies of the Permian LyonsSandstone, which is associated with all knownpetroleum accumulations in the formation, formedlate in the history of the Denver basin as an alter-ation product of the formations red facies. The redfacies that makes up most of the sandstone containsiron oxide coatings, quartz overgrowths and calcitecements. The gray facies, which occurs locally in thedeep basin, is distinguished by pore-filling dolomiteand anhydrite cements and by a lack of iron oxideand calcite. The dolomite and anhydrite cementsoverlie bitumen that was deposited by migrating oil,and hence formed after oil was first generated in thebasin, late in the Cretaceous or early in the Tertiary.The isotopic composition of oxygen in the dolomiteranges to such light values that the cement musthave formed deep in the basin in the presence of

meteoric water.The gray facies likely formed in a regime ofgroundwater flow resulting from Laramide uplift ofthe Front Range during the Tertiary. In our model,

saline groundwater flowed eastward through Pennsylvanian Fountain Formation and thupwelled along the basin axis, where it discharginto the Lyons Sandstone. The saline water mixwith more dilute groundwater in the Lyons, drivinreaction that dissolved calcite and, by a common-effect, precipitated dolomite and anhydrite. Tfacies gray color resulted from reduction of feoxide in the presence of migrating oil or the Fotain brine. Underlying source beds by this time hbegun to generate petroleum, which migratedbuoyancy into the Lyons. The association of the gfacies with petroleum accumulations can explained if the Fountain brines discharged acraquitards along the same fractures that transmitoil. As petroleum accumulated in the Lyons, newly formed cements prevented continued migtion, as is observed in shallower strata, by sealinginto the reservoirs from which it is produced toda

INTRODUCTION

Geologists have long puzzled over the naturediagenetic alteration in the Permian Lyons Sandstof the Denver basin because the occurrence of dgenetic cements in the formation mirrors the disbution of petroleum reservoirs. Most of the Lyoconsists of a red diagenetic facies. The red Lyonfamiliar to many geologists because it crops outhe basins western margin on the Flatirons, a seof hogbacks along the Front Range of the RocMountains, and because of its use as a distinct

building stone in the area. The facies, which is bren of petroleum, is characterized by hematite staclay minerals, quartz overgrowths, and calccements. A gray diagenetic facies occurs locally dein the basin. The gray Lyons, distinguished by andrite and dolomite cements that overlie bitumstains and by a lack of calcite and hematite, hostsknown oil fields in the formation.

There has been no convincing explanation of

2

Copyright 1994. The American Association of Petroleum Geologists. Allrights reserved.

1Manuscript received, August 5, 1992; revised manuscript received,September 13, 1993; final acceptance, September 24, 1993.

2Department of Geology, University of Illinois, Urbana, Illinois 61801.Present address: Department of Geology, Temple University, Philadelphia,Pennsylvania 19122.

3Department of Geology, University of Illinois, Urbana, Illinois 61801.We thank Roger Burtner of the Chevron Oil Field Research Company for

supplying Lyons core samples, Keith Hackley and Jack Liu of the IllinoisState Geological Survey for helping make isotopic measurements, and

Richard Hay, Randy Cygan, Wendy Harrison, and Steve Johansen for adviceand helpful discussions. We appreciate assistance in the laboratory fromTom Anderson, Steve Altaner, Jay Matthews, and Eric Daniels, and help incomputer programming from the Hydrogeology staff. Jessie Knox and JoanApperson drafted the figures. This study was supported by National ScienceFoundation grants EAR 85-52649 and EAR 8601178, and the generosity ofAmoco Production Research, Arco Oil and Gas, British Petroleum Research,Chevron Oil Field Research, Conoco, Du Pont, Exxon Production Research,Illinois State Geological Survey, Japan National Oil Company, LawrenceLivermore National Laboratory, Marathon Oil Company, Mobil Research andDevelopment, Sandia National Laboratory, Texaco, and Union Oil ofCalifornia.

Groundwater Flow, Late Cementation, and PetroleumAccumulation in the Permian Lyons Sandstone,Denver Basin1

Ming-Kuo Lee2 and Craig M. Bethke3

AAPG Bulletin, V. 78, No. 2 (February 1994), P. 221-241.


2/21

origin of the gray facies nor the reason for its associa-tion with petroleum accumulations. The anhydrite anddolomite cements of the gray facies might have formedsoon after the Lyons was deposited, as suggested bythe high intergranular volume preserved in the sand-stone. Anhydrite occupies up to 25% and dolomitecements occupies up to 15% of the rocks pre-cementpore volume. For this reason, Dimelow (1972) suggest-

ed that these cements might have precipitated from for-mation water deposited with the sediment.

Dolomite and anhydrite might also have precipi-tated in an environment open to flow of salinegroundwater. For example, they could have formedwhere evaporated seawater infiltrated carbonateplatforms in sabkha environments (Butler, 1969; Pat-terson and Kinsman, 1977; Achauer, 1982); somegeologists working in the Denver basin have sug-gested such an origin to us. Levandowski et al.(1973), on the basis of the cement distribution in thesandstone, argued that the cements precipitatedwhen underlying strata compacted during burial,driving saline groundwater upwards through theLyons.

Anhydrite and dolomite cements of the grayLyons Sandstone might also have formed in an envi-ronment open to regional groundwater flow, longafter the formation was buried. Such an origin is notunusual: dolomite cements can form where freshand saline groundwaters mix (Badiozamani, 1973;Land, 1973; Harrison, 1991), where migratinggroundwater reacts with host sediments (Banner etal., 1988; Gregg and Shelton, 1989; Bethke and Mar-shak, 1990), and where carbon dioxide escapes frommigrating brine (Leach et al., 1991).

In this study, we use hydrologic and geochemicalmodels to develop an explanation of the origin ofgray Lyons facies that is consistent with petrographicobservations as well as the stable isotopic composi-tions of the Lyons cements. The results of our model-ing suggest a strong relationship between diageneticalteration of the sandstone, the basins paleohydrolo-gy, and the accumulation of petroleum in the Lyons.

GEOLOGIC SETTING

The present-day Denver basin is an elongate,asymmetric structure whose axis runs north-south inColorado and Wyoming, just to the east of the Front

Range uplift. The Laramide and Front Range upliftsbound the basin to the west. Basin strata extendeastward across parts of eastern Colorado, southeast-ern Wyoming, southwestern Nebraska, and north-western Kansas. The deepest part of the basin liesnear Denver, Colorado, where more than 4 km ofsediments are preserved. Because of the basinshydrocarbon potential, its structure, stratigraphy, andtectonic history have attracted considerable study

(e.g., Hoyt, 1962; Martin, 1965; Izett, 1975; Trimble,1980; Sonnenberg and Weimer, 1981) .

The Denver basin was a marine shelf that subsid-ed slowly through most of the Paleozoic and Meso-zoic. Figure 1 shows the stratigraphic column nearthe Front Range. Pennsylvanian sediments consistmainly of carbonate and shale, although sandstonespredominate in the west along the uplift of theancestral Rocky Mountains. During the Permian, abroad sea intermittently covered a surface of low

relief where sediments, including the Lyons Sand-stone, were deposited in environments ranging fromfluvial to normal marine to hypersaline. During theMesozoic, Triassic and Jurassic sediments consistingmainly of sandy shale buried the Paleozoic section.The basin downwarped rapidly in the Cretaceous,when an interior seaway covered much of westernNorth America. During this period, more than 3 kmof shale and shaly sandstone were deposited at the

218 Lyons Sandstone, Denver Basin

Figure 1Stratigraphic column for the Denver basinnear the Front Range (after Weimer, 1973).


3/21

basins depocenter.With the onset of the Laramide orogeny near the

end of the Cretaceous, the area west of the present-day basin began to be uplifted. Orogeny continuedinto the Tertiary and reached its peak during theEocene. The basin tilted eastward as the Front Rangebegan to emerge. Uplift resumed during the middleOligocene (Trimble, 1980; Morse, 1981). The most

recent uplift, about 56 million years ago, causedabout 450600 m of Tertiary sediments to be erodedfrom the basin (Trimble, 1980). With this final adjust-ment, the basin assumed its present configuration.

For purposes of hydrologic analysis, strata in ourstudy area can be divided into five hydrostratigraphicunits (Belitz and Bredehoeft, 1988): (1) Pennsylvani-an-Permian, (2) Permian-Triassic, (3) Triassic-Jurassic,(4) basal Cretaceous sandstones, and (5) Cretaceousshales. Mississippian carbonates form a sixth unitthat, because of extensive secondary porosity, consti-tutes an important aquifer system in the northernGreat Plains (Downey, 1982; Back et al., 1983). Thislowermost unit, however, is absent in our study area.

The Pennsylvanian-Permian unit consists primari-ly of interbedded carbonates and shales, but sand-stones predominate toward the basins western mar-gin. In the western basin, sandstones of the FountainFormation lie at the base of the unit. These sand-stones grade eastward to dolomite and shale. Sec-ondary porosity and permeability are less welldeveloped in carbonates of this unit than they are inMississippian limestones, which constitute aquifersin the southern basin.

Interbedded red beds and evaporites dominatethe strata of the Permian-Triassic unit. In eastern Col-orado, the Permian Lyons Sandstone lies betweenanhydritic siltstones. The Lyons, a regional aquifer,serves as a supply of potable water where it is shal-lowly buried and as a petroleum resource in thedeep basin. The overlying Triassic-Jurassic unit con-tains shale interbedded with local sandstone layers.The unit acts as a regional aquitard.

The basal Cretaceous sandstones, which includethe Dakota Sandstone, make up the most importantaquifer system within the basin, providing the mainwater supply for consumption and irrigation in thecentral Great Plains area. The Cretaceous shale unitconfines the aquifer system. This unit, the thickest inthe basin, is predominantly shale, but contains minoramounts of limestone, sandstone, and chalk.

DEPOSITION AND DIAGENESIS OF THE LYONSSANDSTONE

The Denver basin in the Permian contained twomajor subbasins where evaporites accumulated: theAlliance basin to the north and the Sterling basin tothe south (Figure 2). The Lyons Sandstone formed as

a nearshore deposit along a band between slightly emergent ancestral Rocky Mountains to west and the evaporite basins to the east. Fluvmarine, and eolian processes transported sand to coast (e.g., Blood, 1970; Walker and Harms, 19Adams and Patton, 1979), where the Lyons wdeposited eastward about 150200 km from the psent-day position of the Front Range. Figure 3 shothe relationship of the Lyons Sandstone to aquitaand other aquifers within the basin. Anhydrite-rsiltstones bounded the Lyons above and below. TPennsylvanian Fountain Sandstone underlies Lyons to the west, but pinches out into carbon

and shale facies as it grades eastward.Lyons sediments have been extensively altesince deposition in a pattern suggesting that the pcess of diagenesis was closely related to oil accumlation (Levandowski et al., 1973). Figure 4 shows the zone of greatest cementation lies close toknown Lyons oil fields, including Pierce, Black Hlow, and New Windsor fields. Two diagenetic facof the Lyons (one red and the other creamy gray)

Lee and Bethke

Figure 2Isopach and lithofacies map of Permian sments in the Denver basin (after Martin, 1965). Line shows the cross section considered in this study, whruns from Big Thompson Canyon to the Keota oil fie


4/21

be distinguished by mineralogy and color. The redLyons is nonpetroliferous. Hematite and clay rim thegrains of quartz sand and much of the pore space isfilled by quartz overgrowths and calcite cement (Fig-ure 5A). Textural relations in the facies suggest a gen-eral paragenetic sequence of hematite and clay over-lain by secondary quartz and then calcite. Detailedpetrographic study of samples from outcrops alongthe Front Range (Hubert, 1960), however, shows thatthe Lyons contains at least two generations of calcitecements that precipitated at different times. Some cal-

cite formed even earlier than the hematite, which iscommonly cited as the first phase to form in the dia-genetic history of the sandstone.

The gray Lyons facies, which is associated with oilfields, is distinguished (Levandowski et al., 1973) byanhydrite and dolomite cements that overlie organicmatter (Figure 5B). The organic matter, which isbitumen deposited by migrating petroleum, coversquartz grains and their overgrowths. For this reason,

the anhydrite and dolomite cements must haveformed after oil first migrated into the Lyons. Togeth-er, the anhydrite and dolomite occupy as much as


Figure 3Cross section AA' (see Figure 2 for location) through the Denver basin from Big Thompson Canyon to theKeota oil field (after Levandowski et al., 1973). The section shows the relationship of the Lyons Sandstone to otherhydrostratigraphic units within the Denver basin.

Figure 4Relationship of areas of the most intensive

anhydrite cementation (ruled pattern) to Lyons oilfields (solid pattern). Open circles show locations ofwells where samples were taken for isotopic analysis byLevandowski et al. (1973) and in this study.


5/21

40% of precement porosity, or about 8% of the for-mations total volume.

Calcite cements are ubiquitous in the red Lyonsfacies, but have yet to be observed in the gray facies.Calcite cementation almost certainly preceded deposi-

tion of anhydrite and dolomite because some calcitein the red facies underlies the early hematite cement(i.e., Hubert, 1960), which in turn underlies the quartzovergrowths, whereas the anhydrite and dolomiteoverlie the quartz overgrowths and later bitumen. Ifthe gray facies is an alteration product of the redfacies, the alteration process must have consumed thecalcite. Alteration, in contrast, seems to have con-sumed little quartz. The photomicrograph of the detri-tal quartz grains (Figure 5C), made by scanning elec-tron microscope after carbonate and sulfate in thesamples had been chemically digested, shows shal-low solution pits indicating that quartz surfacesbeneath the late cements were only slightly etched.

ISOTOPIC COMPOSITIONS OF LATE CEMENTS

Oxygen in Carbonate Cements

Levandowski et al. (1973) analyzed the 18O and13C compositions of carbonate cements; their resultsare shown in Table 1. The analyses did not distin-guish among carbonate minerals, but the dominantcarbonate mineral from gray facies samples isdolomite. The cements range in 18O from +8.8 to+21.2 relative to SMOW. These values are signifi-cantly lower than would be expected in marine car-bonates, which normally fall in the range +30 to+32 (Figure 6).

The low oxygen ratios implicate meteoric water inthe emplacement of the late dolomites, especiallythose samples strongly depleted in 18O, and suggestthat the cements formed at depth in the basin. Usingthe equation of Northrop and Clayton (1966), wecan estimate at any temperature the fluid composi-

tion that would be required to explain dolomite ogiven composition. At 25C, groundwaters with values in the range 26 to 13 would be requirSuch fluids are much lighter than any seawaespecially seawater concentrated by evaporatiand are difficult to reconcile with the likely comptions of meteoric water in the latitude of about 40

Lee and Bethke

Figure 5(A) Photomicrograph showing (undercrossed nicols) the red facies of the Lyons Sandstone.Hematite (arrows) coats the original surfaces of quartzgrains (Q) beneath overgrowths of secondary quartz(O). Calcite cement (C) fills some pore spaces. Scale bar= 100 m. Sample 13398 from Horsetooth Reservoir. (B)Photomicrograph showing (under crossed nicols) thegray facies of the Lyons Sandstone. Dolomite (D) andanhydrite cements (white arrows) fill pore spaces that

are coated with organic matter (black arrows). Organicmatter coats quartz grains (Q) and overgrowths. Scale

bar = 100 m. Sample 13390-6 from the Calco 1 Ferchwell. (C) View under a scanning electron microscope athigh magnification showing the pre-cement surface of adetrital quartz grain. The quartz is lightly pitted. Sample14435-21 from the Troy 1 Jones well.


6/21

High-altitude runoff of the Rocky Mountains in theTertiary and present-day have 18O values as low as14 (Taylor, 1974; Lander and Anderson, 1989).Since this part of the North America craton was notin far northern latitudes during the past 300 m.y.(Irving, 1979), it is unlikely that 18O values of mete-oric waters in this area could range as low as 26.The 18O contents of the cements appear incompati-ble with an early diagenetic origin, and therefore,preclude (barring the possibility of later isotopicreequilibration) the possibility that the cementsformed in a sabkha environment.

Formation temperatures of about 100C would beexpected in the Lyons Sandstone at present burialdepths of about 3 km (Table 1), assuming a normalgeothermal gradient of about 30C/km. Groundwaterhaving 18O values between 13 and about 0would be required to precipitate dolomite at this tem-perature. Because Tertiary rainfall in this area couldhave 18O values down to 14, infiltration of mete-

oric water deep into the basin could explain even thelightest 18O compositions determined. The scatter inthe 18O data (Figure 6) could arise from mixing invarying proportions of meteoric water with a basinbrine. If a brine had a 18O value of about 0, forexample, mixing it with meteoric water having a 18Ovalue of about 13 would produce fluids spanningthe required range in isotopic composition.

Carbon in Carbonate Cements

The small CO2 contents of many groundwaterslead many investigators to assume that the l3C com-position of a carbonate cement simply reflects thel3C of its precursor mineral (Mattes and Mountjoy,1980; Meyers and Lohmann, 1985). This assumptionmay not be true, however, if the sediment is open toflow of a groundwater charged with CO2. Especiallyin sandstone, where carbonate minerals occupy asmall fraction of the rock volume, carbonate dis-solved in the fluid can represent a second importantreservoir of carbon.

In the red facies, the l3C of calcite cementsranges widely from 26 to 2 (Figure 6). Thesesamples likely represent calcite from different originsand generations, as already discussed. The depletedl3C values (< 20) might result from oxidation oforganic matter, whereas the heavier l3C compositionssuggest that these calcite samples could have formed

from seawater or from isotopically heavy fluids inthe deep basin.On the other hand, dolomite cements of the gray

facies span a smaller range and are enriched in l3Ccompared to many of the calcite samples. It is diffi-cult to argue from these data that the dolomitereflects the carbon isotopic compositions of a calciteprecursor. The composition is also inconsistent withoxidation of organic matter. The dolomite can be


Table 1. Isotopic Composition () of Carbonate and Anhydrite Cements in the Lyons Sandstone

Sample Number* Well Depth (ft)** 18OSMOW** l3CPDB**

34SCDT

Gray Facies11326-11 Calco 1 Winder 8903 +11.413390-6 Calco 3 UPRR 9073 +9.5 1.3 +12.414385-200 Calco 1 Ferch 9136 +19.6 0.5 +11.514385-236 Calco 1 Ferch 9206 +9.2 2.1 +9.6

14431-14 Calco 1 Brunner 8997 +21.2 5.5 +11.814433-25 Priddy 4 9219 +18.9 3.3 +9.614435-21 Troy 1 Jones 9279 +11.5 4.6 +12.514435-23 Troy 1 Jones 9235 +l2.5 3.1 +11.614441-13 Venable Fee-1 8997 +8.8 1.9 +11.7

Red Facies13398 Horsetooth Reservoir +16.4 15.5 14674-6 Owl Canyon Outcrop +15.6 4 0 13452-2 Hummel 1 Allison 78297834 +15.4 11.0 14883-23 Calco 1 Hayes 91489149 +6.7 26.2 14883-5 Calco Hayes 90899090 +15.7 7.5 14875-1 Nebraska Drillers Ball 93959399 +8.1 3.7 14875-3 Nebraska Drillers Ball 94059410 +6.5 2.0 14875-7 Nebraska Drillers Ball 94259430 +3.4 16.6

*Numbered by Chevron Oil Field Research.**Levandowski et al. (1973).This study.Outcrop samples.Sample contains insufficient anhydrite for analysis.


7/21

interpreted to reflect the composition of a deep CO2-charged groundwater that has l3C values similar tothat of seawater, which ranges from 4 to +4. Thenormal carbon-isotopic composition of thedolomite cements likely reflects in large part thecomposition of marine carbonate that dissolved intoa brine deep within the basin, rather than carbonderived from early calcite cements or from oxidationof hydrocarbons in the Lyons Sandstone. The carboncomposition of dolomite is compatible with a fluid-mixing origin suggested by the oxygen composition;in a fluid mixture, the dissolved carbon derived froma CO2-charged brine overwhelms that from meteoricwaters; hence, the carbon isotopic composition ofdolomite that forms from such a mixture will largelyreflect the l3C values of species in the brine.

Sulfur in Anhydrite Cement

We analyzed the sulfur isotopic composition ofanhydrite cements from nine Lyons core samplesfrom various wells, using the method of Westgate

and Anderson (1982) for extraction and analysis. Thecements have sulfur isotopic compositions (Table 1)ranging from +9.6 to +12.5 relative to the CDTstandard. These values span almost exactly the rangein isotopic composition of sulfate derived from Per-mian evaporites worldwide, which is +9.6 to +13.0(Claypool et al., 1980). For reference, the composi-tion of sulfate from modern seawater is about +20.

The sulfur isotopic composition of the anhydrite

cements suggests that the anhydrite could haformed from seawater during the Permian or frsulfate that was derived from evaporite rocks in basin and imported into the Lyons Sandstone. Tanhydrite is unlikely to have formed from sulderived from H2S, organics, or from sulfide minethat oxidized to form sulfate. Sulfur isotopes strongly fractionated between sulfide and sulfa

mostly as a result of the bacterial reduction of sulin the near surface; consequently basin sulfide revoirs are typically depleted in 34S (Thode et al, 19The 34S values of reduced sulfide dissolved in natwaters normally range from 32 to 0 on the Cscale (Hartmann and Nielsen, 1969). Sulfide treoxidizes does not fractionate significantly, so asulfate derived from the oxidation of sulfide wobe depleted in 34S. The anhydrite cements are enriched in 34S to be consistent with such an origi

The anhydrite is more likely a late diagenephase, because the cement precipitated after oil fmigrated into the Lyons. Oil usually forms late ibasins history because source beds must be deeburied long enough to become thermally matuWe show in the next section that most Paleozoicin the Denver basin did not form until the Late Ctaceous or early Tertiary, long after the sandstowas deposited. In addition, even seawater evapoed to the point of anhydrite saturation would contonly about 14 g/kg of dissolved sulfate to accofor 0.7% of the sandstones compacted volummuch less than is present in the formation. For threasons, anhydrite cements in the Lyons likformed long after burial in an environment opengroundwater flow, rather than precipitated from swater or as pore fluid deposited with the sedimen

PETROLEUM GENERATION AND MIGRATION

Reservoirs in Cretaceous strata provide mostthe oil and gas produced from the Denver basin.this oil, most samples can be traced by geochemcorrelation to source beds, including the CarliGraneros, and Mowry shales and the GreenhLimestone, which also lie in the Cretaceous sect(Clayton and Swetland, 1980). Some oil in the Creceous section has migrated as far as 150 km from deep basin toward the basins eastern flank. Oil pduced from the Lyons, on the other hand, is a Pal

zoic variety that is chemically distinct from oils comon in younger rocks (Clayton and Swetland, 198Unlike overlying strata, the Lyons Sandstone is known to have served as a carrier bed for long-dtance migration.

Several theories explaining the source of Lyons oils have been proposed: the oil formed fr(1) adjacent Permian and Pennsylvanian sha(Levandowski et al., 1973), (2) the Permian Satan

Lee and Bethke

Figure 6Isotopic compositions of calcite (opensquares) and dolomite cements (solid squares) in theLyons Sandstone. Dolomite falls near the normal range

of13

C for marine carbonate, but is depleted in18

O.


8/21

Formation (Berman, 1978), or (3) the Permian Phos-phoria Formation near the Idaho-Wyoming border(Dimelow, 1972; Momper, 1978). The third theoryseems unlikely because the oils differ in carbon iso-topic composition from that of Phosphoria oils in theBig Horn basin (Clayton et al., 1987). Middle Penn-sylvanian black shales, however, are rich in organicmaterial and are considered to have excellent source

potential over much of the northern Denver basin(Clayton and Ryder, 1984; Clayton, 1989); the Lyonsoils could well have been derived from these beds.In a Cretaceous reservoir, Clayton (1989) found oilfrom a Middle Pennsylvanian source, demonstratingthat oil has migrated vertically upward through thePaleozoic and into the Mesozoic section. There is noreason to believe that oil could not have migratedvertically from a Pennsylvanian source to Permianreservoirs.

Dolomite and anhydrite cements in the grayLyons Sandstone, as already discussed, formed afteroil first migrated into the formation. We can place alimit on the age of the cements by calculating whenoil generation began in Pennsylvanian rocks.According to kinetic theory, time and temperaturecontrol the generation and preservation of oil in sed-imentary basins. For our calculations, we use thesimple model of Lopatin (Lopatin, 1971; Waples,1980). In the model, source-bed maturity isdescribed by a time-temperature index (TTI). TheTTI of a sediment is given as the sum of maturitiesdeveloped in each temperature interval

Index nvaries from the lowest to the highest temper-ature interval encountered by the sediment. Theinterval from 100110C was chosen as the baseinterval and given the index value n= 0. For each10C increase or decrease in temperature, the indexvalue increases or decreases by 1. tn is the length oftime (in million years) that the sediment spends with-in each temperature interval. The factor 2n assumesthe doubling of the maturation rate with each 10Crise in temperature. According to the method, sourcebeds are immature until they attain TTI values ofabout 15. Oil forms when TTI falls between 15 and160, which corresponds to vitrinite reflectance mea-

surements (Ro, in percent reflectance in oil) from 0.65to 1.3. When the TTI exceeds 160, oil begins to breakdown to form natural gas.

For a given burial history, models of thermal mat-uration depend on the choice of thermal conductivi-ty and basal heat flow into the basin. We assumethat sediment thermal conductivity (K, in cal/cmsC) increases during compaction according to thecorrelation with porosity:

K=(5.35 4.4) 103, (2)

which is taken from the data of Sclater and Christie(1980) for North Sea shales. We calibrate heat flow,assuming that it remained constant through time,against the maturity in Cretaceous source beds deter-mined by Clayton and Swetland (1980) by vitrinitereflectance measurement. Heat flow values, rangingfrom 1.2 to 1.8 HFU (1 HFU = 106 cal/cm2s), repro-duce individual measurements made near our studyarea. Much of this range, however, likely reflects theuncertainty in estimating the depth of past burial andthe timing of uplift and exhumation for samplestaken from outcrop in the western basin. We findthat a heat flow of about 1.5 HFU, near the averagefor continental crust, gives the best fit to the patternof maturity basinwide in Cretaceous strata.

Basin strata were rather shallowly buried duringmost of the Paleozoic and Mesozoic, until thick shale

sequences were deposited in the Cretaceous. Forthis reason, even the deepest strata did not becomethermally mature until late in the basins history. Fig-ure 7 shows how the thermal maturity of Pennsylva-nian source beds from the deepest basin evolvedthrough geologic time, as calculated for differingheat flows. According to the calculation, the deepestsources began to generate oil between 78 and 50Ma, during the Late Cretaceous or the early Tertiary.


Figure 7Evolution of thermal maturity in Pennsylvani-an source beds beneath the Lyons Sandstone, calculatedassuming differing heat fluxes of 1.2, 1.5, 1.8 HFU (50,63, 76 mW/m2, respectively) from basement. The oil

window is the area between the two dashed lines. The

time-temperature index (TTI) was calculated usingLopatins method (Waples, 1980). As the buried sedi-ment passes from one temperature interval to the next,the index valuen in equation 1 increases suddenly by 1,

which causes the small kinks on the TTI curves.

(1)


9/21

Assuming our best-fit value for heat flow of 1.5 HFU,the first oil probably formed about 68 Ma, near theend of the Cretaceous.

Most Paleozoic oil in the basin, according to theseresults, migrated from source to reservoir in the Ter-tiary, after infilling of the basin was complete. Thistiming provides a further argument that the cementsof the gray Lyons Sandstone formed late in thebasins history. The inferred timing also suggests thatthe cements did not precipitate from fluids remobi-lized by compacting sediments, as some have sug-gested, because sediment compaction was completeor nearly so by the time oil generation began.

BASIN PALEOHYDROLOGY

We used BASIN2, a numerical model of groundwa-ter flow in sedimentary basins, to simulate fluidmigration through the Denver basin in the geologicpast. The model calculates the groundwater flow thatarises from sediment compaction and topographicrelief, the transfer of heat by conduction and advect-ing groundwaters, the maturity of petroleum sourcebeds through time, the evolution of porosity and per-meability, and cementation by migrating groundwa-

ter. The mathematical basis of the calculation tech-nique is described in detail elsewhere (Bethke, 1985;Bethke et al., 1988, 1993; Corbet and Bethke, 1992).

We show the results of two reconstructions of thebasins paleohydrology. The first simulation depictshow the basins groundwater flow regime evolved asthe basin subsided and infilled with sediments; the sec-ond portrays flow resulting from Eocene uplift of theFront Range along the basins western margin. Model

results provide quantitative estimates of various pareters needed to evaluate diagenetic models, includpast rates and directions of fluid migration as welpressure and thermal gradients along flow paths.

Table 2 shows the hydrostratigraphy assumedthe simulations. The basin surface in our calculatiis held at a constant temperature of 10C and atmspheric pressure. The sides of the basin remainhydrostatic pressure and are open to groundwaflow. The bottom of the cross section is the con

with crystalline Precambrian basement, which take as a barrier to fluid flow and the source oconstant heat flux of 1.5 HFU.

Each stratigraphic unit in the calculations is coposed of varying fractions of four rock types: sastone, carbonate, shale, and evaporite-rich siltstonWe calculate the evolution of porosity and permability as each rock type is buried, using correlatishown in Table 3; thermal conductivity is set acco

Lee and Bethke

Table 2. Hydrostratigraphy Assumed in the Hydrologic Models

System Rock Stratigraphy

Cretaceous Shales Cretaceous Pierre Shale, Niobrara Formation, Carlile Shale, Greenhorn Limestone,Graneros Formation

Lower CretaceousSandstones Cretaceous Dakota Group

Triassic-Jurassic Jurassic Morrison Formation, Sundance Formation, Ralston Creek FormationTriassic upper Lykins Formation, Jelm Formation

Permian-Triassic (1) Triassic lower Lykins Formation

Permian-Triassic (2) Permian Lyons Sandstone

Permian-Triassic (3) Permian Owl Canyon Formation, Satanka Shale

Pennsylvanian-Permian Permian Ingleside FormationPennsylvanian Fountain Formation

Table 3. Correlations used in the Hydrologic ModelCalculate Porosity and Permeability

Porosity* Permeability**0 b(km

1) 1 A B kx/

Sandstone 0.40 0.50 0.05 15 3

Carbonate 0.40 0.55 0.05 6 4

Shale 0.55 0.85 0.05 8 7 1

Evaporite 0.55 0.85 0.05 8 7 1

* = oexp(-bZ) + 1, expressed as a fraction; Z is burial depth (km).**log kx (m2) = A + B; kX 1 m2; 1 m2 1 darcy.


10/21


Figure8Crosssectionusedinsimulationsofthecompaction-drivengroundwaterflowsho

wnattheendofCretaceous.Sectio

nextends400km

acrossthenorthernD

enverbasin(BB').Hydrostratigraph

icunitsaredescribedinTable2.


11/21

ing to equation 2. The correlations for porosity andpermeability of all rock types, except the evaporiticsiltstones, are taken from previous studies of basinsfrom the cratonic interior (Bethke et al., 1991);because of a lack of data, correlations for the hydro-logic properties of shales were assigned to the evap-oritic siltstones.

Flow Driven by Sediment Compaction

The first paleohydrologic model simulatesgroundwater flow resulting from sediment com-paction as the basin infilled through most of thePaleozoic and Mesozoic. The simulation followscross section BB' (Figure 8), which traverses thenorthern Denver basin. Figure 9 shows the results ofthe calculation for the final time slice in the calcula-tion, at the end of the Cretaceous. In the calculation,fluids migrate from compacting shale and carbonatestrata upward and downward, primarily into the Cre-

taceous aquifer complex and, to a lesser extent, intothe Lyons and Fountain sandstones. These aquifersact as drains that carry fluid laterally from deep stratato the basin margins. Because groundwater flowhere is too slow to alter the thermal structure of thebasin, temperatures in the calculation fall along aconductive gradient that corresponds to the assumedthermal conductivity correlation and heat flux frombasement.

Contours in Figure 9 are equipotentials, which rresent the drive for fluid migration; groundwamigrates toward areas of lower hydraulic potentThe equipotentials in this case (where there is topographic relief and the reference elevation is level) also show the extent to which pressure excehydrostatic. The calculation predicts that excess psures developed in the Cretaceous shales were vsmall, about 1 atm (0.1 MPa), near the basidepocenter. Greater pressures result when permeaity is set smaller, but it seems unlikely that sedimein this basin were impermeable enough to allow guine overpressures, such as those observed in Gulf of Mexico basin (e.g., Bethke, 1986) to hdeveloped. Overpressure in the Gulf basin resufrom burial rates sometimes exceeding 10,000 m/m(Harrison and Summa, 1991), about 100 times mrapid than rates of burial in the Denver basin.

The role that fluids displaced by sediment copaction could have played in precipitating cemeis limited by the modest flow rates predicted by

calculation. The estimated flow velocities (true ratthan Darcy) through the Lyons Sandstone are evewhere less than about 2 cm/yr or 20 km/m.y. Tsmall flow rates reflect the slow rates of burial acompaction in the basin, which infilled over 30 mat a rate less than 100 m/m.y. Such slow rates of ssidence, infilling, and fluid expulsion from comping sediments are common to basins of the cratointerior (e.g., Bethke et al., 1991).

Lee and Bethke

Figure 9Calculated groundwater flow driven by sediment compaction at the end of Cretaceous. Contours are culated excess pressures (atm) in the basin. Dashed lines show predicted temperature distributions. Arrows incate the flow directions in the Dakota Sandstone, Lyons Sandstone, and Fountain Formation. Vertical componeof cross section are exaggerated by about 40:1, as shown by the scale bar.


12/21


Figure10Crosssectionusedinsimulationsofthepastgro

undwaterflowduringtheEocene.Sectionextends350kmacrosstheno

rthernDenverbasin

(CC').HydrostratigraphicunitsaredescribedinTable2.s.l.

=sealevel.


13/21

Flow Driven by Topography

The second paleohydrologic model simulatesgroundwater flow driven by Eocene uplift of theFront Range in the western basin (Figure 10). On thebasis of the thicknesses of the Laramie Formationand the Dawson sandstone, which represent materi-als eroded and deposited during the Eocene, Trim-ble (1980) estimated that the Front Range was uplift-ed during the Laramide orogeny to between 2500and 6000 m above sea level. As the Front Rangeemerged during the Laramide orogeny, basin fluidsmigrated eastward in response to the hydraulic gra-dient created by the slope on the water table.Groundwater recharged at high elevation along the

basins uplifted western margin and dischargedtoward the eastern margin. Groundwater flow con-tinues in this regime today (Belitz and Bredehoeft,1988), but modern flow rates are presumably some-what smaller now than in the past because the FrontRange has eroded over time.

In Figure 10, we assume that the Front Range inthe Eocene was uplifted 2000 m above the elevationof the basins eastern margin. The topographic relief

drives groundwater through the Cretaceous aquiat velocities of tens of meters per year (Figure 1The velocities predicted, however, reflect two pooconstrained assumptions: the past elevation of western basin and the aquifer permeabilities. Psent-day flow rates in these aquifers are not knowbut a model calculated by Belitz and Bredeho(1988) predicted flow velocities of about 3 m/yr.this standard, the predicted velocities shown hare perhaps optimistically high. The predicted flvelocities are somewhat slower (


14/21

overpressured. Potentials in the Cretaceous aquifercomplex are slightly overpressured to the west andunderpressured to the east; the complex is under-pressured over broad areas in the present day (Belitzand Bredehoeft, 1988). The present-day underpres-suring in these shallow aquifers likely reflects ero-sion of the uplifted basin margin and poor hydrauliccontinuity across Front Range faults.

The Lyons Sandstone serves as an aquifer for east-ward flow. Along the basin axis, the aquifer is also azone of dispersive mixing. Fluids migrating alongthe underlying Fountain Formation dischargeupward across stratigraphy where the sandstonegrades into less permeable facies near the basin axis.As shown in Figure 11, the drive for cross-formation-al flow is provided by the hydraulic potentials in theFountain which, along the basin axis, are greaterthan the potentials in overlying strata. In the calcu-lated flow regime, Fountain fluids along the basinaxis discharge upward across the lower Satanka For-mation and into the Lyons, mix with Lyons ground-waters, and then continue to migrate eastward.

Simple Models of Cementation

Groundwater migrating through a sedimentarybasin alters the sediments through which it flowsbecause mineral solubility varies with temperatureand pressure along the flow path. Quartz, for exam-ple, is increasingly soluble with rising temperature.Groundwater saturated with quartz precipitatesquartz cement as it migrates along a path of decreas-ing temperature, whereas it dissolves quartz as itmigrates toward higher temperatures. It is possible,therefore, to calculate the rates at which variousminerals dissolve and cements precipitate through-out a basin from knowledge of mineral solubilityand the groundwater flow pattern; Appendix 1 givesthe mathematical basis for the calculation. We referto this type of calculation as a simple model ofcementation because it accounts only for the effectsof changing temperature and pressure along flowpaths.

We applied this type of calculation to predict theamounts of anhydrite and quartz that could formand be consumed as a result of (1) compaction-driv-en groundwater flow occurring as the basin subsid-ed during the Paleozoic and Mesozoic, and (2) flow

driven by uplift of the western basin during theEocene. Appendix 2 gives the correlations we usedto compute mineral solubility. We did not attempt tocalculate the distribution of dolomite cementation,because the solubility of this mineral varies morestrongly with pH than with temperature or pressure.

The calculation results for the compaction-flowregime (Figure 12) show the cumulative amounts ofanhydrite and quartz formed or dissolved within the

Lyons Sandstone in this hydrologic regime by theend of the Cretaceous. In the hydrologic models,groundwater flowed upward relative to the subsid-ing strata, but downward over much of the basin rel-ative to fixed elevation. For this reason, groundwaterin the deep basin warms with time and, because ofthe retrograde solubility of anhydrite, precipitatesanhydrite cement. The cumulative amount of anhy-drite precipitated in this regime, however, is lessthan about 0.1% of the formations volume. Not onlyis this volume too small to account for the origin ofthe sandstones gray facies, but the cement is dis-tributed across the sandstone, not just in deep stratawhere the gray facies occurs.

Cementation rates near the Lyons oil fields calcu-

lated for groundwater flow in the Eocene (Figure 13)are small (


15/21

ideas of cementation by flow through temperatureand pressure gradients cannot explain the origin ofthe gray facies. These simple models of cementation,however, do not account for other potential causesof cementation in the Eocene flow regime, such asthe mixing of groundwaters of varying compositionor common-ion effects in which species producedby dissolution of one mineral might drive the precip-itation of another.

REACTIONS ACCOMPANYING FLUID MIXING

From the results of paleohydrology modeling, weinterpret that, in the Eocene, the Lyons Sandstone atdepth contained a zone of fluid mixing. The locationof the mixing zone corresponds generally to the areain which rocks of the gray facies are found. In thissection, we use quantitative modeling techniques topredict which chemical reactions occurred and howstable isotopes fractionated during the mixing pro-cess. The modeling results show that fluid mixing

could have driven a reaction by which calcite dsolved and dolomite and anhydrate precipitated, cating the gray Lyons facies as an alteration prod

of the red facies.

Reaction Modeling

We used the computer program REACT (Beth1992) to solve for the overall reaction that woaccompany fluid mixing in the deep Lyons Sastone. REACT is one of a class of reaction path mod(Helgeson, 1968; Helgeson et al., 1970; Wole1979; Reed, 1982; Plummer et al., 1983) that trthe chemical evolution of systems open to mass fles. Calculations here are based on version R46 of thermodynamic database compiled at Lawrence Lermore National Laboratory (Delany and Lunde1990; Johnson et al., 1991), and employ the exteed Debye-Hckel method (Helgeson, 1969) for cculating activity coefficients.

The conceptual basis of our calculation is shoin Figure 14. A packet containing 1 kg of dilute flmigrates along the Lyons sandstone, maintainequilibrium with calcite and quartz in the formatiThe packet, which originated as meteoric rechaalong the Front Range, is initially dilute. A secofluid discharges upward into the Lyons and miinto the fluid packet. The second fluid is msaline, having reacted with evaporite beds in Pennsylvanian and Permian section. As the flumix over the course of the calculation, any minethat become supersaturated precipitate.

We used an analysis of modern Lyons groundwter to set the fluids initial composition (McConaget al., 1964, sample C5-69-18bbcc). We assumed tthe mixing occurred at 100C. Since McConaghyal.s (1964) analysis was made at 51C, we usREACT to correct it to the temperature of the calcu

Lee and Bethke

Figure 13Calculated cementation rates (dX/dt) ofanhydrite and quartz across the Lyons Sandstone dur-ing the Eocene. Anhydrite, which has a retrograde solu-

bility, precipitates where the fluid is descending anddissolves where the fluid is ascending. Quartz shows anopposite pattern due to its prograde solubility. Thecementation rate is expressed as a percentage of the for-mations bulk volume per unit time (%/m.y.). Again,this simple model of cementation poorly predicts thedistribution of anhydrite cement in the gray LyonsSandstone.

Figure 14Conceptual model for tracing chemical retion during fluid mixing in the Lyons Sandstone.


16/21

tion by heating the analyzed fluid in the presence ofcalcite and quartz. We assumed that the compositionof the Fountain groundwater reflects equilibriumwith minerals in the evaporite strata that lie beneaththe Lyons. The Fountain groundwater in our calcula-tions is a 3-molal NaCl solution that has equilibratedwith dolomite, anhydrite, magnesite, and quartz.

The choice of NaCl concentration represents theupper limit of validity of the activity coefficient correla-tions we used. We fixed pH by setting the CO2 fugaci-ty to reflect a partial pressure of 50 atm. The resultingLyons groundwater is predominantly a sodium-bicar-bonate solution, whereas the Fountain brine is mostlya sodium-chloride-bicarbonate solution (Table 4).

In the model, we set the Ca2+ concentration in thebrine by assuming equilibrium with dolomite, andfixed the fluids Mg2+/Ca2+ activity ratio to an end-member value of about 20 by specifying saturationwith magnesite. In fact, we do not know whethermagnesite occurs in the evaporite strata, so the actu-al brine from these rocks could have been undersat-urated with respect to this mineral and hence have asmaller activity ratio than predicted. Fortunately, themineralogic results of the fluid mixing reaction varylittle over a broad range of activity ratios. As long asthe Mg2+/Ca2+ activity ratio is greater than about0.06, calcite is undersaturated in the brine, as wouldbe expected in a saline groundwater from evaporitebeds, and we found that mixing the Fountain brineinto the Lyons Sandstone causes dolomite and anhy-

drite to precipitate at the expense of calcite.The choice of CO2 fugacity for the Fountain brinereflects our interpretation of the isotope data, as pre-viously explained. We might have chosen a differentfugacity, but the degree to which calcite is undersat-urated in the brine, the critical variable in the calcu-lation, is independent of pH or CO2 fugacity. Thefluid has equilibrated with dolomite and magnesite,so the saturation state is fixed by the reaction,

CaCO3 CaMg(CO3)2 MgCO3

calcite dolomite magnesite

As a result, calcite has a saturation index (log Q/K,where Q is the activity product for the dissolutionreaction, and K is the equilibrium constant) of about1.3, regardless of pH or CO2 fugacity. Hence, thereis no petrographic basis for determining the pH ofthe Fountain brine.

The reaction model (Figure 15) traces the chemi-cal consequences of progressively mixing the Foun-tain brine into the Lyons aquifer at 100C. As calcitedissolves into the undersaturated brine, the Ca2+ andHCO3added to solution drive precipitation of anhy-drite and dolomite by a common-ion effect. Theoverall reaction predicted by the model is

5 CaCO3 + 2 SO42

+ 52 Mg2+

calcite 2 CaSO4 + 52 CaMg(CO3)2 + 12 Ca2+

anhydrite dolomite

This reaction reduces porosity in the sandstone

because the volume of anhydrite and dolomite pro-duced is greater than the volume of calcite dissolved(Figure 15). The predicted reaction explains the ori-gin of anhydrite and dolomite cements, as well as thelack of calcite in the gray facies. It does not explain,however, the slight predominance of anhydrite overdolomite cements observed in the facies; dolomite isa more voluminous reaction product than anhydritein the calculation results (Figure 15). This apparent


Table 4. Chemical Compositions (mg/kg) of EndmemberGroundwaters used in Mixing Calculations

Lyons Fountain

Na+ 110 55,000Ca2+ 19 510K+ 4.6Mg2+ 0.6 3400

SiO2 49 22HCO3 310 28,000Cl 55 85,000S04

2 36 13,800pH (100C) 6.7 4.5

Figure 15Predicted overall reaction resulting frommixing of the ascending Fountain brine into the LyonsSandstone. The sandstone is initially filled with the infil-trating meteoric water. The plot shows changes in min-eral volume as brine mixes into the formation by dis-persion. Positive volume changes indicate precipitation,

and negative volume changes show dissolution.


17/21

discrepancy might be accounted for by slower pre-cipitation rates of dolomite, or a brine with a higherSO42/Mg2+ ratio than assumed in the calculation.

Reaction models, such as the one we used, arequite useful because they trace the effects among allof the reactions that occur in a system in order topredict the overall reaction. Treating reactions indi-vidually without concern for their interactions can

be misleading. For example, the solubility of anhy-drite varies with temperature and fluid salinity. Thesolubility vs. salinity curves are convex upward (seeFigure 6 of Blount and Dickson, 1969), indicatingthat mixing a dilute and a saline fluid should dis-solve, not precipitate, anhydrite. The flaw in this rea-soning, which would seem to negate the analysispresented above, is that the solubility curve or func-tion represents the single reaction,

CaSO4 Ca2+ + SO4

2,

anhydrite

in isolation, whereas the reaction model correctlyaccounts for the effects of other reactions occurring inthe system. An analysis that considers only anhydritesolubility would incorrectly predict that fluid mixingcould not account for the cementation observed.

Isotope Fractionation

To test whether the reaction model presented inthe previous section can explain the isotopic compo-sitions of cements in the gray facies of the LyonsSandstone, we traced how the stable isotopes 13Cand 18O fractionate during the predicted reaction.The calculation technique employs mass balanceequations similar to those derived by Bowers andTaylor (1985) and Bowers (1989). The technique, asimplemented in REACT, accounts for fractionationamong the solvent, dissolved species, gases, andminerals. Table 5 shows the fractionation factorsassumed in the calculation. Instead of assuming thatminerals maintain isotopic equilibrium with the fluid,as is commonly done in this type of model, the cal-culation segregates minerals from isotopic exchange.Only the increments of minerals precipitated duringa reaction step are in isotopic equilibrium with thefluid, and only the mass of minerals dissolved over astep affects the fluids isotopic composition. In this

way, the calculation procedure models the fractiona-tion that results from the reaction process in theabsence of isotope exchange.

We assumed that the initial system contained ameteoric water having an initial l8O of 13(SMOW), which matches estimates for Tertiary rain-fall in the region, as already discussed. The l8O ofcalcite was set to +11, the mean 18O compositionof this cement. The l8O composition of calcite

assumed in the model, however, has very little efon the results since the amount of oxygen in the stem provided by water overwhelms the smamount obtained by dissolving calcite. We furtassumed that quartz was in equilibrium with the tial fluid, but so little of this mineral reacts thatisotopic effect can be neglected. The l3C of the tial fluid was 12 (PDB), which was set to refthe mean isotopic composition of calcite ceme(11) found in the red facies of the Lyons Sastone. We further assumed that the Fountain briwhich migrates into the Lyons during the reactprocess (Figure 14), is isotopically heavier. The brhas l8O and 13C values of 0, as might be expeed in a sedimentary brine.

Figure 16 shows how the isotopic compositionsdolomite cements (assuming varying CO2 fugacitiesthe Fountain brine) evolved as the Fountain brmixed into the Lyons Sandstone. Results show thatrange in l8O values observed for the dolomcements can be explained by mixing in differing pportions of Lyons groundwater with Fountain brinewe previously hypothesized. The 13C values preded by the model depend on the CO2 fugacity assumfor the Fountain brine. As previously noted, the vachosen for CO2 fugacity has little effect on the minalogic results of the reaction path, and hence is constrained by petrographic observations.

When we assumed small values for the Cfugacity of the Fountain brine, the calculation pdicted dolomite compositions similar to the 13Cthe precursor calcite. Assuming CO2 fugacities in range of about 25 to 100, however, gives dolomcompositions near the observed range. In this cathe dolomite 13C more closely reflected the isotocomposition of carbon species in the brine than cbon from the precursor calcite. A reaction mo

Lee and Bethke

Table 5. Isotope Fractionation Factors Assumed in Calculations, Expressed as 1000 ln. Factors DescrFractionation at 100C of18O Relative to Water, andRelative to CO2*

18O 13C

Anhydrite 18.3 Calcite 17.1 3.38

Dolomite 21.5 1.91Quartz 21.1 CO2(aq) 28.0 0HCO3 0** 3.5CO3

2 0** 3.5SO

42 17.1

*Fractionation factors calculated from regression curves compiled byBohlke from the following sources: Northrop & Clayton (1966), Malinin (1967), Bottinga (1968), ONeil et al. (1969), Sheppard and Schwarcz (19Clayton et al. (1972), Richet et al. (1977), Chiba et al. (1981).

**Assumed.


18/21

that calls on migration of a CO2-charged brine intothe Lyons, therefore, explains both the petrographyand isotopic compositions of cements in the Lyonsgray facies.

The fugacities suggested by the above analysis aresomewhat greater than those measured in the GulfCoast, where CO2 pressures are less than 5 bars at100C (Smith and Ehrenberg, 1989). The fugacities

suggested for the Denver basin might better be com-pared to CO2 concentrations and temperatures mea-sured in fluid inclusions in late dolomite cementsthroughout the Ozark region (Leach et al., 1991). TheCO2 contents in fluid inclusions there are greater than0.35 molal, indicating that CO2 pressures at about120C could range up to at least 40 bars.

DISCUSSION

In this study, we integrated quantitative models ofgroundwater flow and chemical reaction with petro-graphic and isotopic observations to explain the ori-gin of the petroliferous gray facies of the LyonsSandstone. Our results show how past groundwaterflow in the Denver basin drove diagenetic reactionsin the Lyons that are intimately related to the accu-mulation of petroleum into present-day reservoirs.

We argue that the gray facies of the Lyons formedas an alteration product of the sandstones red faciesin zones where fluids, including groundwater andpetroleum, migrated into the formation from under-lying strata. In our model, Laramide uplift of theFront Range along the basins western margin, whichreached its peak in the Eocene, set up a regime ofeastward groundwater flow. Meteoric waterrecharged into the Lyons and moved deep into thebasin. Flow in the underlying Fountain Formationwas much more restricted because sandstones in thisformation pinched out to carbonate and shale bedsalong the basin axis. The slowly flowing Fountaingroundwater became saline by reaction with miner-als in the evaporite beds.

Oil in the Lyons Sandstone probably derived fromsource beds in underlying Pennsylvanian strata. Oilgeneration began toward the Late Cretaceous or theearly Tertiary, before the peak of the Laramide oroge-ny. The orogeny likely produced fractures along thebasin axis along which the oil could migrate upwardby buoyancy across aquitards and into the Lyons

Sandstone. The regime of groundwater flow set upby the orogeny drove brines along the Fountainsandstone. Where Fountain sandstone facies pinchout, the brines migrated upward across confiningaquitards, likely along the same fractures that trans-mitted the oil, and discharged into the Lyons.

Calcite dissolved as the brine mixed into theLyons because the upwelling brine was undersatu-rated with this mineral after having reacted with

dolomite and evaporite minerals. In a common-ioneffect, calcium from the dissolving calcite reactedwith magnesium, sulfate, and CO2 carried in thebrine to form the dolomite and anhydrite cements ofthe gray facies. This reaction accounts for the iso-topic compositions of the cements. The broad rangeof18O in dolomite requires mixing of an isotopical-ly heavy fluid, such as the Fountain brine, withmeteoric water. 13C in the dolomite is consistent withderiving the CO2 by dissolving Paleozoic marine car-bonate into the brine, and 34S in the anhydritematches the composition of sulfate from Permianevaporite beds.

Hematite stains of the red facies were also con-sumed in the mixing zone by reduction in the pres-

ence of the migrating petroleum, and perhaps by thebrine itself. Kilgore and Elmore (1989) documenteda similar reaction in which hydrocarbons removedhematite cements from red beds of the TriassicChugwater Formation in Montana. By this process offluid mixing, rocks of the red facies of the LyonsSandstone were transformed locally along the basinaxis into the gray facies.

The association in time and space of oil migration


Figure 16Calculated isotopic composition of dolomiteresulting from the mixing of ascending brines into theLyons Sandstone, as shown in Figure 15. Initial isotopic

composition of the Lyons groundwater is l8O = 13and13C = 12; the brine has l8O = 0 and13C =0 (). The precursor calcite cement () has l8O =+11 and13C = 11, reflecting the minerals meancomposition. Each curve shows, for a given CO2 fugacityof brine (on an atmosphere scale), how the isotopiccomposition of dolomite evolves over the course of thereaction. Tick marks along the center curve show water-rock ratios (by mass). Calcite and dolomite composi-tions (boxes) are from Levandowski et al. (1973).


19/21

and groundwater mixing explains the fact that allknown petroleum reservoirs in the formation arefound within the gray facies, and the observation thatsome dolomite and anhydrite cements overlie bitu-men stains left behind by migrating oil. The modelalso explains why oil in the Lyons apparently did notcontinue to migrate eastward through the sandstone,as occurred higher in the section along the Creta-

ceous aquifers. Since the migration and cementationprocesses were so closely linked, the cements proba-bly began to seal the oil into reservoirs as it accumu-lated, preventing farther migration.

APPENDIX 1: SIMPLE MODEL OF CEMENTATIONBY GROUNDWATER FLOW

To test the relationship between regional groundwater flow andcementation in the basin, we calculated the amounts of anhydriteand quartz that would have precipitated from groundwaters migrat-ing along temperature and pressure gradients. The solubility ofquartz, for example, increases with temperature. Quartz cementsform where groundwater flows in a direction of decreasing temper-

ature and hence decreasing solubility. Quartz dissolves wheregroundwater migrates toward higher temperatures because thefluid must acquire silica to remain in equilibrium with quartz.

Anhydrite has a retrograde solubility, on the other hand, and thuswould precipitate and dissolve in an opposite pattern.

If a minerals solubility depends on temperature and pressurealone, and the fluid can be assumed to remain in local equilibrium,the rate at which the mineral precipitates or dissolves can be com-puted from the velocities of groundwater flow and the temperatureand pressure fields. For simplicity, we represent solubility by thedimensionless ratio Vi of the volume of mineral i that can be dis-solved in per unit volume of groundwater. Similarly, we representcement volume by the ratio Xi of the volume of mineral ipresentper unit volume of the formation.

Under these conditions, the cementation rate dXi/dtfor mineraliis given directly by

(3)

where, Tand Pare temperature and pressure, is porosity, and 'xand 'z are lateral and vertical groundwater velocities (cm/yr) incurvilinear coordinates. Dxand Dzare coefficients of hydrodynam-ic dispersion (cm2/yr), which account for diffusion of the solute as

well as mechanical mixing that occurs in a groundwater flowingthrough a porous medium.

The time derivatives d/dtin equation 3 are Lagrangian, that is,taken in the reference frame of the sediments as they are buried oruplifted. The first terms on the right side of equation 3 give thecementation rate caused simply by changing temperature and pres-sure within a sediment, as would occur when a formation is buriedor exhumed. Terms on the next line account for the effects of diffu-sion and dispersion of the dissolved mineral within the basin. Thefinal terms, which are the most significant even at modest flowrates, account for the effects of groundwater advection.

APPENDIX 2: SOLUBILITIES OF ANHYDRITE ANDQUARTZ

The equilibrium solubility of anhydrite depends on tempture, pressure, and fluid salinity. To calculate the role that flmigrating across temperature and pressure gradients play in adrite precipitation, we correlated solubility to values measuredBlount and Dickson (1969) for a 2-molal NaCl solution. BlountDicksons measurements span the range of about 100 to 250C

5 to 1350 bars, so the correlation serves to extrapolate the datemperatures found in shallow strata. The resulting regress

which gives molal solubility as a function of temperature Tand pressure P(bars), is

lnmanh = 2.33 1.59 102T+ 6.1 104P+ 1.82 105T2

7.05 108P2 + 3.14 107TP.

By this correlation, anhydrite solubility decreases with increatemperature but increases with increasing pressure, especiallower temperatures. The net effect is that anhydrite is less soluat depth in a basin than at surface conditions.

Quartz solubility increases rapidly with increasing temperatbut depends only weakly on pressure and fluid salinity overtemperature range considered in this study (Kitahara, 1960; Waand Helgeson, 1977). To calculate quartz solubility, we assumethe mineral is soluble as SiO2(aq), as is generally the case ex

under alkaline conditions, and that the species has an activity cficient of about 1. In this case, the correlation

log mqtz = 1.881 2.028 103Tk1560/Tk

of Rimstidt and Bames (1980) gives the molal solubility of quara function of absolute temperature Tk(in kelvins).

The dimensionless solubilityVi, the volume of a mineral solper unit volume of groundwater, can be calculated from a minemolal solubility, mi, as

where MVi is the minerals molar volume in cm3/mole, f is the density in g/cm3, and TDSis the total dissolved solids (in mg/k

the fluid. Because the largest uncertainty in calculating cementrates is the estimation of groundwater flow rates, it is sufficienour calculations to carryfas 1 and TDSas 0 in equation 6.

REFERENCES CITED

Achauer, C. W., 1982, Sabkha anhydrite: the supratidal facies of cdeposition in the Upper Minnelusa Formation (Permian) Rfields area, Powder River basin, Wyoming, inC. R. Handford, RLoucks, and G. R. Davies, eds., Depositional and diagenetic spof evaporites: SEPM Core Workshop p. 193209.

Adams, J., and J. Patton, 1979, Sabkha-dune deposition in the LyFormation (Permian) northern Front Range, Colorado: The Motain Geologist, v. 16, p. 4757.

Back, W., B. B. Hanshaw, L. N. Plummer, P. H. Rahn, C. T. Rightmand M. Rubin, 1983, Process and rate of dedolomitization: mtransfer and 14C dating in a regional carbonate aquifer: Geocal Society of America Bulletin, v. p. 14151429.

Badiozamani, K., 1973, The dorag dolomitization modelapplicato the Middle Ordovician of Wisconsin: Journal of SedimenPetrology, v. 43, p. 965984.

Lee and Bethke


20/21

Banner, J. L., G. N. Hanson, and W. J. Meyers, 1988, Water-rock interac-tion history of regionally extensive dolomites of the Burlington-Keokok Formation (Mississippian): isotopic evidence, inV. Shuklaand P. A. Baker, eds., Sedimentology and geochemistry of dolo-stones: SEPM Special Publication 43, p. 97 114.

Belitz, K., and J. D. Bredehoeft, 1988, Hydrodynamics of Denverbasin: explanation of subnormal fluid pressures: AAPG Bulletin,

v. 72, p. 13341359.Berman, A. E., 1978, Permian stratigraphy and paleotectonics, Bell-

vue-Livermore area, Larimer County, Colorado: M.S. thesis, Col-

orado School of Mines, Golden, Colorado, 86 p.Bethke, C. M., 1985, A numerical model of compaction driven

groundwater flow and heat transfer and its application to thepaleohydrology of intracratonic basins: Journal of GeophysicalResearch, v. 90, p. 68176828.

Bethke, C. M., 1986, Hydrologic constraints on the genesis of theUpper Mississippi Valley mineral district from Illinois basin brines:Economic Geology, v. 81, p. 233249.

Bethke, C. M., 1992, The geochemists workbench, a users guide toRXN, ACT2, TACT, REACT, and GTPLOT: University of Illinois, Urbana,Hydrogeology Program, 174 p.

Bethke, C. M., and S. Marshak, 1990, Brine migrations across NorthAmericathe plate tectonics of groundwater: Annual Review ofEarth and Planetary Science, v. 18, 287315.

Bethke, C. M., W. J. Harrison, C. Upson, and S. P. Altaner, 1988,Supercomputer analysis of sedimentary basins: Science,

v. 239, p. 261267.Bethke, C. M., J. D. Reed, and D. F. Oltz, 1991, Long-range petroleum

migration in the Illinois basin: AAPG Bulletin, v. 75, p. 9251945.Bethke, C. M., M.-K. Lee, H. A. M. Quinodoz, and W. N. Kreiling,

1993, Basin modeling with BASIN2, a guide to using BASIN2,B2PLOT, B2VIDEO, and B2VIEW: University of Illinois, Urbana,Hydrogeology Program, 225 p.

Blood, W. A., 1970, Upper portion of the Fountain Formation andLyons formation at Morrison, Colorado: The Mountain Geologist,

v. 7, p. 3348.Blount, C. W., and F. W. Dickson, 1969, The solubility of anhydrite

(CaSO4) in NaCl-H2O from 100 to 450C and 1 to 110 bars:Geochimica et Cosmochimica Acta, v. 33, p. 227245.

Bottinga, Y., 1968, Calculation of fractionation factors for carbon andoxygen exchange in the system calcitecarbon dioxidewater:

Journal of Chemical Physics, v. 72, p. 800808.Bowers, T. S., 1989, Stable isotope signatures of water-rock interac-

tion in mid-ocean ridge hydrothermal systems: sulfur, oxygen,and hydrogen: Journal of Geophysical Research, v. 94,p. 57755786.

Bowers, T. S., and H. P. Taylor, Jr., 1985, An integrated chemical andstable-isotope model of the origin of Midocean Ridge hot springsystem: Journal of Geophysical Research, v. 90,p. 1258312606.

Butler, G. P., 1969, Modern evaporite deposition and geochemistry ofcoexisting brines, the sabkha, Trucial Coast, Arabian Gulf: Journalof Sedimentary Petrology, v. 39, p. 7089.

Chiba, H., M. Kusakabe, S.-I. Hirano, S Matsuo, and S. Somiya, 1981,Oxygen isotope fractionation factors between anhydrite and

water from 100 to 550C: Earth and Planetary Science Letters, v.53, p. 5562.

Claypool, G. E., W. T. Holser, I. R. Kaplan, H. Sakai, and I. Zak, 1980,The age curves of sulfur and oxygen isotopes in marine sulfate andtheir mutual interpretation: Chemical Geology, v. 28,p. 199260.

Clayton, J. L., 1989, Geochemical evidence for Paleozoic oil in LowerCretaceous O sandstone, northern Denver basin: AAPG Bulletin,

v. 73, p. 977988.Clayton, J. L., and R. T. Ryder, 1984, Petroleum source-rock potential

of Pennsylvanian black shales in Powder River basin, Wyoming,and northern Denver basin, Nebraska (abs.): AAPG Bulletin, v.68, p. 935.

Clayton, J. L., and P. J. Swetland, 1980, Petroleum generation andmigration in Denver basin: AAPG Bulletin, v. 64, p. 16131633.

Clayton, R. N., J. R. ONeil, and T. K. Mayeda, 1972, Oxygen isotopeexchange between quartz and water: Journal of GeophysicalResearch, v. 77, p. 30573067.

Clayton, J. L., J. D. King, C. N. Threlkeld, and A. Vuletich, 1987, Geo-chemical correlation of Paleozoic oils, northern Denver Basinapplications for exploration: AAPG Bulletin, v. 71, p. 103109.

Corbet, T. F., and C. M. Bethke, 1992, Disequilibrium fluid pressuresand groundwater flow in the Western Canadian sedimentarybasin: Journal of Geophysical Research, v. 97,p. 72037217.

Delany, J. M., and S. R. Lundeen, 1990, The LLNL thermochemicaldatabase: Lawrence Livermore National Laboratory Report UCRL-21658, 150 p.

Dimelow, T. E., 1972, Stratigraphy and petroleum, Lyons Sandstone,northeastern Colorado: M.S. thesis, Colorado School of Mines,Golden, Colorado, 127 p.

Downey, J. S., 1982, Geohydrology of the Madison and associatedaquifers in parts of Montana, South Dakota, and Wyoming: U. S.Geological Survey Open File Report 82-914, 124 p.

Gregg, J. M., and K. L. Shelton, 1989, Minor and trace-element distri-bution in the Bonneterre Dolomite (Cambrian), southeast Mis-souri: evidence :for possible multiple basin fluid sources andpathways during lead-zinc mineralization: Geological Society of

America Bulletin, v. 101, p. 221230.Harrison, W. J., 1991, Modeling fluid-rock interactions in sedimentary

basins, inT. A. Cross, ed., Quantitative dynamic stratigraphy:New York, Elsevier, p. 195231.

Harrison, W. J., and L. L. Summa, 1991, Paleohydrology of the Gulf ofMexico Basin: American Journal of Science, v. 291,p. 109176.

Hartmann, M., and H. Nielsen, 1969, 34S-Werte in rezenten Meer-essedimenten und ihre Deutung am Beispeil einiger Sediment-profile aus der westlichen Ostsee: Geologische Rundschau, v. 58,p. 621655.

Helgeson, H. C., 1968, Evaluation of irreversible reactions in geo-chemical processes involving minerals and aqueous solutionsI.Thermodynamic relations: Geochimica et Cosmochimica Acta, v.32, p. 853877.

Helgeson, H. C., 1969, Thermodynamics of hydrothermal systems atelevated temperatures and pressures: American Journal of Sci-ence, v. 267, p. 729804.

Helgeson, H. C., T. H. Brown, T. H. Nigrini, and T. A. Jones, 1970,Calculation of mass transfer in geochemical processes involvingaqueous solution: Geochimica et Cosmochimica Acta,

v. 34, p. 569592.Hoyt, J. H., 1962, Pennsylvanian and Lower Permian of northern

Denver basin, Colorado, Wyoming and Nebraska: AAPG Bulletin,v. 46, p. 4659.

Hubert, J. F. 1960, Petrology of the Fountain and Lyons formations,Front Range, Colorado: Colorado School of Mines Quarterly, v.55, no. 1, 242 p.

Irving, E., 1979, Paleopoles and paleolatitudes of North America andspeculations about displaced terrains: Canadian Journal of EarthScience, v. 16, p. 669694.

Izett, G. A., 1975, Late Cenozoic sedimentation and deformation innorthern Colorado and adjoining areas, inB. F. Curtis, ed., Ceno-zoic history of the southern Rocky Mountains: Geological Societyof America Memoir 144, p. 179209.

Johnson, J. W., E. H. Oelkers, and H. C. Helgeson, 1991, SUPCRT92: Asoftware package for calculating the standard molal thermodynamicproperties of minerals, gases, aqueous species, and reactions from 1to 5000 bars and 0C to 1000C: Lawrence Livermore LaboratoryEarth Sciences Department, 101 p.

Kilgore, B., and R. D. Elmore, 1989, A study of the relationshipbetween hydrocarbon migration and the precipitation of authi-genic magnetic minerals in the Triassic Chugwater Formation,southern Montana: Geological Society of America Bulletin, v. 101,p. 12801288.

Kitahara, S., 1960, The solubility of quartz in the aqueous sodiumchloride solution at high temperatures and high pressures: The



21/21

Review of Physical Chemistry of Japan, v. 30, p. 115121.Land, L. S. 1973, Contemporaneous dolomitization of middle Pleis-

tocene reefs by meteoric water, north Jamaica: Marine ScienceBulletin, v. 23, p. 6492.

Lander, R. H., and T. F. Anderson, 1989, Isotopic variations in theTertiary waters of the northern Rockies and Great Plains asrecorded by authigenic calcite (abs.): Geological Society of Ameri-ca Annual Meeting Abstracts, v. 21, p. 221222.

Leach, D. L., G. S. Plumlee, A. H. Hofstra, G. P. Landis, E. L. Rowan,and J. G. Viets, 1991, Origin of late dolomite cement by CO2-satu-

rated deep basin brines: evidence from the Ozark region, centralUnited States: Geology, v. 19, p. 348351.

Levandowski, D. W., M. E. Kaley, S. R Silverman, and R. G. Smalley,1973, Cementation in Lyons Sandstone and its role in oil accumu-lation, Denver basin, Colorado: AAPG Bulletin, v. 57, p.22172244.

Lopatin, N. V., 1971, Temperature and geologic time as factors incoalification [in Russian]: Akademiia Nauk SSSR Izvestiya SeriyaGeologiches-Kaya, v. 3, p. 95106.

Malinin, S. D., O. I. Kropotova, and V. A. Grinenko, 1967, Experimentaldetermination of equilibrium constants for carbon isotope exchangein the system CO2(g)-HCO3(sol) under hydrothermal conditions:Geochemistry International, v. 4, p. 764771.

Martin, C. A., 1965, Denver basin: AAPG Bulletin, v. 49,p. 19081925.

Mattes, P. W., and E. W. Mountjoy, 1980, Burial dolomitization of theUpper Devonian Myette Buildup, Jasper National Park, Alberta, inD. H. Zenger, J. B. Dunham, and R. L. Effington, eds., Conceptsand models of dolomitization: SEPM Special Publication 28, p.259297.

McConaghy, J. A., G. H. Chase, A. J. Boettcher, and T. J. Major, 1964,Hydrogeologic data of the Denver basin, Colorado: ColoradoGroundwater Basic Data Report 15, 224 p.

Meyers, W. J., and K. C. Lohmann, 1985, Isotope geochemistry ofregional extensive calcite cement zones and marine componentsin Mississippi limestones, New Mexico, inN. Schneidermann andP. M. Harris, eds., Carbonate cements: SEPM Special Publication36, p. 223239.

Momper, J. A., 1978, Oil migration limitations suggested by geologi-cal considerations: AAPG Short Course Notes 8, p. B1B60.

Morse, D. G., 1981, Sedimentology and tectonic implication ofOligocene Castle Rock Conglomerate, southern Denver basin,Colorado (abs.): AAPG Bulletin, v. 65, p. 962.

Northrop, D. A., and R. N. Clayton, 1966, Oxygen-isotope fractiona-tion in systems containing dolomite: Journal of Geology, v. 74, p.174196.

ONeil, J. R., R. N. Clayton, and T. K. Mayeda, 1969, Oxygen isotopefractionation in divalent metal carbonates: Journal of ChemicalPhysics, v. 51, p. 55475558.

Patterson, R. J., and D. J. J. Kinsman, 1977, Marine and continentalgroundwater sources in a Persian Gulf coastal sabkha, inS. H. Frost, M. P. Weiss, and J. B. Saunders, eds., Reefs and relat-ed carbonatesecology and sedimentology: AAPG Studies inGeology 4, p. 381397.

Plummer, L. N., D. L. Parkhurst, and D. C. Thorstensen, 1983, Thedevelopment of reaction models for ground-water systems:Geochimica et Cosmochimica Acta, v. 47, p. 665686.

Reed, M. H., 1982, Calculation of multicomponent chemical equilibria

and reaction processes in systems involving minerals, gases,an aqueous phase: Geochimica et Cosmochimica Acta, v. 46513528.

Richet, P., Y. Bottinga, and M. Javoy, 1977, A review of hydrocarbon, nitrogen, oxygen, sulphur, and chlorine stable isotfractionation among gaseous molecules: Annual Review of Eand Planetary Sciences, v. 5, p. 65110.

Rimstidt, J. D., and H. L. Barnes, 1980, The kinetics of silica-wreactions: Geochimica et Cosmochimica Acta, v. p. 16831699.

Sclater, R. K., and P. A. F. Christie, 1980, Continental stretchingexplanation of the postmid-Cretaceous subsiding of the ceNorth Sea basin: Journal of Geophysical Resea

v. 85, p. 37113739.Sheppard, S. M. F., and R. L. Schwarcz, 1970, Fractionation of car

and oxygen isotopes and magnesium between coexisting mmorphic calcite and dolomite: Contributions to MineralogyPetrology, v. 26, p. 161198.

Smith, J. T. and S. N. Ehrenberg, 1989, Correlation of carbon dioabundance with temperature in clastic hydrocarbon reservrelationship to inorganic chemical equilibrium: Marine Petroleum Geology, v. 6, p. 129135.

Sonnenberg, S. A., and R. J. Weimer, 1981, Tectonics, sedimentaland petroleum potential, northern Denver basin, Colora

Wyoming, and Nebraska: Colorado School of Mines Quarterl76, p. 3536.

Taylor, H. P. Jr., 1974, The application of oxygen and hydrogentope studies to problems of hydrothermal alteration anddeposition: Economic Geology, v. 69, p. 843883.

Thode, H. G., H. Kleerekoper, and D. McElcheran, 1951, Isofractionation in the bacterial reduction of sulfate: Research (don), v. 4, p. 581582.

Trimble, D. E., 1980, Cenozoic tectonic history of the Great Plcontrasted with that of the southern Rocky Mountains: a synsis: The Mountain Geologist, v. 17, p. 5969.

Walker, T. R., and J. C. Harms, 1976, Eolian origin of flagstone Lyons Sandstone (Permian), Boulder County, Colorado: ColoSchool of Mines Professional Contribution, v. p. 110122.

Walther, J. V., and H. C. Helgeson, 1977, Calculation of the therdynamic properties of aqueous silica and the solubility of quand its polymorphs at high pressures and temperature: Amer

Journal of Science, v. 277, p. 13151351.Waples, D. W., 1980, Time and temperature in petroleum forma

application of Lopatins method to petroleum exploration: ABulletin, v. 64, p. 916926.

Weimer, R. J., 1973, A guide to uppermost Cretaceous stratigracentral Front Range, Colorado: deltaic sedimentation, grofaulting and early Laramide crustal movement: The MounGeologist, v. 10, p. 5397.

Westgate, L. M., and T. F. Anderson, 1982, Extraction of varforms of sulfur from coal and shale for stable sulfur isotope ansis: Analytical Chemistry v. 54, p. 21362139.

Wolery, T. J., 1979, Calculation of chemical equilibrium betwaqueous solution and minerals: the EQ3/6 software packLawrence Livermore National Laboratory Report UCRL-52658p.

Lee and Bethke

Documents

Groundwater Flow, Sandstone, Denver Basin_AAPG1994