Sa ae · the less-porous shelf edge and slope carbonates are not. They envisioned reflux...

Saline dolomites can also replace and expand the volume of ear-lier forms of dolomite during later diagenesis under deeper buri-al conditions (mesogenesis) where basinal fluid flow is driven by hydrothermal gradients in temperature, pressure and salini-ty. These burial or hydrothermal dolomites are discussed in the fourth article of this series.

Brine Reflux dolomiteEvaporation of restricted bodies of seawater creates high-densi-ty saline brines that subsequently seep downward and seaward under the influence of gravity; this process is commonly known as brine reflux (Figure 1). A brine reflux model driven by den-sity (salinity) contrasts in saltern waters atop a carbonate plat-

IntroductionAncient saline dolomites display textures and extents that em-phasise widespread crystallisation via replacement of a carbon-ate precursor, typically a limestone, but sometimes an earlier syndepositional style of dolomite. Replacement generally oc-curs during early diagenesis (eogenesis) when the transforming carbonate is bathed in crossflows of relatively shallow subsur-face saline pore brine. Such flows tend to be driven by salinity/density contrasts in the depositional environment. This style of saline dolomite is termed brine reflux dolomite. It is best devel-oped in shallow groundwater zones directly below or adjacent to an accumulating bed of evaporites, most often bedded platform sulphates.

Salty MattersSaline Dolomites: Ancient - Part 3 of 4: Brine reflux dolomite

www.saltworkconsultants.com

John Warren -Monday September 30, 2019

PENESALINE VITASALINESALINE

Evaporation rapid Barrier Reef ForereefLagoonwater

surface353‰ 1.2138

NaCl427‰

NaCl

NaCl

CaSO4

CaSO4

CaCO3 CaCO3

Sea Level

520‰30‰ 1.00

199‰ 1.1264 CO 2

72‰ 1.05

35‰ 1.02

35°C

520‰ 50°C 9.5pH

OrganicLimestone

Reef Complex

CoolDeepWater

Seepage Re�uxion ZoneAragonite & Mg Calcite

AnhydriteHalite &Anhydrite

Halite (Few Bitterns)

Clastics

IMPERMEABLE SEA FLOOR

Warm water

Original Adams and Rhodes (1961) model has hydrologic problems by proposing simultaneous lateral salinity-density variation A single brine mass cannot support lateral salinity variation (this sets up potential shear but water is a �uid not a solid)

Landward

Basinward

Re�ux of dense brinesdolomitises adjacent carbonate aquifers

Platform evaporites Platform carbonates

Hydrographic isolation(exposed barrier)

Salinesea�oor seeps

Marine seepage

Re�ux brines100 m100s - 1,000s of km

Hypersaline and holomictic saltern seaway Normal marine

Mostly subaqueous saltern gypsum (lesser halite)

DOLOMITISATION

INTENSE DOLOMITISATION

Saltern must precipitate near monomineralogic phases when holomictic (same density) and driving brine re�ux(density strati�ed brines cannot drive brine re�ux)

Warren (2016)

A.B.

Figure 1. Brine reflux models driving dolomitisation in the hypersaline eogenetic realm A) Synopsis of the brine reflux model as first proposed by Adams and Rhodes (1960). B) Revised model utilising more realistic saltern hydrology modelling (from Warren, 2016).

form was first proposed by Adams and Rhodes in 1960 to explain dolo-mitization of Permian reefs in West Texas (Figure 1a). Since then, reflux circulation is widely acknowledged as the prime driver of dolomitization of ancient arid-zone carbonate plat-forms across much of the geological record. However, the original model of Adams and Rhodes (1960) requires moderate modification to better incor-porate more realistic notions of brine stratification and holomictic chemis-tries (Figure 1b: Warren, 2016; Chap-ter 2).

Brine reflux, as a fluid circulation mechanism, has much broader im-plications than just a driver for dolomitisation. Reflux occurs wherever ponded or concentrating holomictic brines sit atop the porous floor of an evaporitic seaway or lake and over time become dense enough to displace underlying pore fluids. The dense waters then percolate into the underlying succession so creating a sinking brine plume or curtain with a fluid thermal and ionic chemistry different to the original pore waters. The result-ing chemical interfaces can drive both prograde and retrograde reactions (Figure 1b).

The creation of contrasting chemical interfaces can also drive reflux induced dissolution, as well as mineralogical alteration and transformation. Descending warm brines can have chem-istries that with cooling become supersaturated with differ-ent prograde mineral phases compared to the hosting matrix through which they are flowing. Such brines tend to not only dolomitise but also backreact and pseudomorph earlier mineral phases (Warren, 2016; Chapter 2). Reflux can also drive the pre-cipitation of widespread authigenic phases such as K-feldspar in siliciclastic hosts at burial depths that are much shallower, and in authigenic stages much earlier than those of equivalent precipitates in non-evaporitic siliciclastics (Sandler et al., 2004).

It also promotes the formation of a variety of zeolite minerals in arid rift valley setting and a variety of low-temperature alter-ations and precipitation processes, such as the formation halite cement (prograde reaction) by the cooling of a NaCl brine, the creation of sylvite by the incongruent dissolution of carnallite, or the formation of polyhalite via interaction of a highly saline K-Mg-SO4 brine with earlier gypsum or halite. We shall discuss the broader implications of reflux across various saline systems, other than dolomitisation in a later article. For now, we will focus on most widely recognized outcome of brine reflux in a carbonate platform, namely formation of dolomite in limestone units that underlie or are adjacent to an aggrading evaporite in-terval (Table1).

Such hydrologies and the resulting dolomite geometries tend to transect primary depositional boundaries. To model such sys-tems requires more than the simple bounding unit assumptions inherent to many sequence stratigraphic models. When evapor-itic carbonate or gypsum precipitates in hypersaline holomictic saltern areas on such platforms, the surface brine densities gen-erally range from 1.1 to 1.2 gm/cc (Figures 1, 2). Concurrent Mg/Ca ratios of the remaining saltern brines rise, compared with their seawater seepage feed, brines and so the dense Mg-en-

riched warm solutions sink through the underlying platform carbonates. These reactive, generally cool-ing and sinking pore fluids displace underlying con-nate fluids, that can retain original marine densities of 1.03 gm/cc. The resulting brine cell slowly seeps basinward, with elevated flow rates through lime-stone aquifers, fractures and joints. In this way large volumes of Mg-rich brine pass through previously deposited shelf limestones, precipitating dolomite both as a pore fill and as a replacement in the former shelf limestones. Cooling along the brine-flow path-way can also drive halite cementation and the partial or complete alteration/dissolution of retrograde cal-cium sulphate.

Adams and Rhodes (1960) first devised the brine reflux model to explain widespread dolomitisation of the Permian Capitan reef limestones of West Tex-as; whereby backreef, shelf and lagoonal carbon-

Evaporite unit proximity, geometry and position

Close spatial relationship of dolomite to evaporite sequences that can reasonably be deduced as being subaqueous in origin.Reflux dolomitised units should exhibit dolomitisation gradients, with increasing volumes of dolomite toward the source of refluxing brines typically indicated by and anhydrite bed or its dissolution breccia and residues.The stratigraphic, tectonic, and eustatic setting of the associated evaporites must be appropriate for the development of a barred base or lagoon.

Textures and hydrological prepa-ration

The dolomitised reflux receptor units must have had high porosity-permeability, and suitable hydrologic continuity at the time of dolomitisation.Reflux dolomite crystal size are generally coarser than those found associated with sabkha sequences, because coarse receptor beds dictate fewer nucleation sites.

Geochemical gradient While geochemical information relating reflux dolomite to evaporative pore fluids might be lost by recrystallization, gross vertical geochemical gradients, such as increasing Sr, Na trace elements, and stable isotopes enriched in 18O toward the evaporative brine source might be preserved

Table 1. Features associated with reflux dolomites (after Moore, 2009, Warren, 2016)


FFFFFFFFFF

FF

FFF

FFFFFFFFFFFFF

FFFFFF FF F

FFFFFF

F FFFF

F

1.0

1.1

1.2

1.3

1.40 10 20 30 40 50 60 70 80 90 100

Den

sity

(gm

/cc)

Degree of evaporation

Gypsum onsetHalite onset

MgSO4K salts

Figure 2. Brine density increases with increasing salinity (seawater brine chemistry is replotted from data tables in McCaffrey et al., 1987).

ates are intensely dolomitised, while the less-porous shelf edge and slope carbonates are not. They envisioned reflux dolomitisation brines moving through porous sediments down to depths of several hundred metres be-low the saline lagoon brine source.

A perceived difficulty with the silled platform brine-reflux model is the lack of same-scale modern analogues. Although dolomites associated with salterns are common in the geologic past, 1same-scale counterparts do not occur today. Then again, as I detailed in Chapter 5 in Warren 2016, nor are low-amplitude 4th and 5th-order greenhouse eustatic cycles, nor the appropriate continent-continent prox-imity tectonic settings that favour formation of saline mega-evaporite basins. All these ancient evaporite settings, with no same-scale coun-terpart, favoured widespread reflux dolomitisation.

Modern salt ponds such as Pekelmeer lagoon on the island of Bonaire (Lucia and Major, 1994; Murray, 1969) and East Salina on West Caicos (Perkins et al., 1994) in the Caribbean have been proposed as settings for reflux dolomitization. These modern hypersaline ponds, however, are essentially small scale coast-al desiccation features, all occur in a high-amplitude icehouse eustatic context and all lack evidence of significant dolomitisa-tion. In 1965, Deffeyes et al. reported modern Mg-rich brines in the marine spring-fed gypsum salina that is Pekelmeer Lagoon (Figure 3). There, the hypersaline lagoon is separated from the sea by a beach-barrier composed of coralgal debris. From the lagoon water chemistry, they predicted that concentrated brines should be sinking and forming reflux dolomite in underlying sediments. Beneath gypsum crusts, they observed micritic dolo-mite in organic-entraining Holocene lagoon muds but, with only a soft-sediment coring device to work with, could not penetrate the cemented Pleistocene base to the lagoon. Outcropping on the opposite end of the island, well away from the present lagoon, they identified dolomite replacing what they interpreted as Pleis-tocene limestone, and inferred that this had formed during an earlier episode of brine reflux.

Lucia (1968) subsequently drilled beneath the Pekelmeer La-goon and found no widespread dolomites in the underlying Pleistocene carbonates. He also found that porewaters had nor-mal-marine salinity in areas where sinking brine was predicted by a seepage reflux model. A thin clayey ash layer of general-ly low permeability was found to form a hydroseal separating Pleistocene sediments below with normal salinity, from lagoonal

1 See the addendum to this article for a partial modern example of brine reflux instigated by anthropogenic changes to the hydrology of Owens Lake, California.

sediments with elevated salinity. Thus, reflux dolomitisation is not found in Pleistocene carbonates beneath the Holocene sedi-ments of the Pekelmeer Lagoon.

Salt springs supplying seawater into the Pekelmeer salina are lo-cated in areas where the ash bed is broken or missing and forms the terminations of flow pathways in the underlying Pleistocene/Pliocene (Murray, 1969). For most of the year, seawater rises into springs feeding into the Pekelmeer, driven by evaporative drawdown. However, during a short period in the late summer, the springs are inactive and the hydrostatic pressure on the land-ward side is higher than that on the seaward side. Murray con-cluded that return flow or reflux could occur at such times, but noted it did not appear to do so with sufficient volumes to form replacement dolomites.

The claim that the Pekelmeer area was forming extensive seep-age reflux dolomites beneath an evaporite lagoon was further challenged when Sibley (1980) concluded that the solutions re-sponsible for extensive dolomitisation in Tertiary sediments on Bonaire (the same dolomites that were earlier used as evidence for reflux dolomite) were probably fresh to brackish meteoric waters mixed with seawater, and were not derived by reflux of the modern Pekelmeer hydrology. Later work by Fouke et al. (1996) on dolomite in the Seroe Domi (Miocene), defines a unit more than 20 m thick that extends 115 km across the Nederlands Antilles. Their work reinforced Sibley’s conclusions that is is not a reflux dolomite

Until recently, most of this dolomite, hosted in Pliocene carbon-ates, was thought not to be mixing zone dolomites (as once con-cluded by Sibley, 1980), but rather formed from Mg-rich waters with circulations that are thermally driven. However, Lucia and Major (1994) continued to argue for a significant input of ma-rine/hypersaline waters during Neogene dolomitisation across the region. In a recent paper Teoh et al. (2018) have taken the


Figure 3. Pekelmeer Lagoon, Bonaire, Nederlands Antilles. View from the salt piles, looking seaward across the halobiota-stained salt pans

reflux argument full circle and based on their detailed mapping of diagenetic geometries, textural studies and isotope determina-tions of replacement zones in clinoforms in the Miocene Seroe Domi Fm, conclude that these contentious dolomites are indeed reflux dolomites.

There are other islands in the tropics where highly localized brine reflux, associated with hypersaline conditions, is inter-preted to form local-scale Holocene dolomites (Budd 1997). Müller and Teitz (1971) document brine-reflux dolomite replac-

ing earlier carbonate cement in skel-etal grainstones along the shoreline of Fuerteventura in the Canary Is-lands. Kocurko (1979) found similar brine-reflux dolomite a few metres above the high tide line, in the spray-zone pools of the shoreline of San An-dres, Columbia. Aharon et al. (1977) described small volumes of reflux dolomite precipitating in organic-en-riched Holocene sediments of Solar Lake, Sinai. Similar, localized thin brine-reflux dolomite cement sheets can also be found about the pres-ent-day edges of Ras Mohammed on the Sinai Peninsula.

It seems small-scale examples of brine reflux dolomites do exist, but the de-gree of dolomite development pales in comparison to extensive reflux dolomites of past saltern platforms. Further complicating any argument of the formative mechanism of many of the possible reflux dolomites in mod-ern sediment hosts is that they tend to precipitate in saline brines where sul-phate-reducing bacteria can flourish (see part 2 of this set of articles).

Hydrological modelling of the world's ancient reflux systems in times of greenhouse eustacy, first carried out by Shields and Brady (1995), used very conservative assumptions of brine head and drawdown. They con-cluded that reflux can explain ancient platform dolomites forming within geologically realistic time frames of burial (Figure 4). Their flow mod-el used flow lengths varying from 1 to 1000 km and the relative densi-ties of seawater and brines from two end-member compositions (Figure 4a): brine 1 is anhydrite-saturated (density ρ = 1120 kg/m3) and brine 2 is halite-saturated (ρ = 1200 kg/m3). Brine 1 is approximately 9.3% dens-er than seawater (ρseawater = 1024.5 kg/

m3), and brine 2 is 17.1% denser. Using a 100 m tall and 1 m wide flow cell (Figure 4b), brine 1 has approximately 9.3 m more hy-draulic head than a comparable column of seawater, while brine 2 has a differential hydraulic head of 17.1 m. Shields and Brady (1995) assumed a pure calcite precursor and a final product that is 100% dolomite with 7% porosity that would require 350 kg/m3 magnesium source. Their results (Figure 4c) show that reflux flow can circulate the Mg necessary to completely dolomitise a carbonate platform across radii of a few tens of km in a few mil-


Di�erential Brine Head

INFLOWOUTFLOW

Seawater head

Base level

Hsw

(�gh

) sw

Hbrine

No �ow boundary

�ow scale

FLOW DOMAIN

No �ow boundary

100m (�

gh) b

rine

B.

C.Pe

rmea

bilit

y (d

arci

es)

102

101

100

10-1

10-2

10-3

10-4

10-5

10-3

10-4

10-5

10-6

10-7

10-8

10-9

10-10

Hyd

raul

ic c

ondu

ctiv

ity (m

/sec

)

reef

s, br

ecci

as,

cave

s, co

llaps

e br

ecci

asw

acke

ston

epa

ckst

one

grai

nsto

ne

wac

kest

one

mud

ston

e

rare

frac

ture

sab

unda

nt

cem

ent

abun

dant

rare

coar

se, w

ell

sort

ed s

and

and

grav

el

sort

ed a

nd

unce

men

ted

sand

ston

es

unso

rted

or

cem

ente

dsa

ndst

ones

shal

es a

ndce

men

ted

sand

ston

es

Total �ow time (years)104 109108107106105

Lithologic parameters

1

2

3

4

Flow length(km)

Volume ofslice (m 3)

Required Mg(kg)

Required volume (m 3)

brine 1 brine 2

1 1 x 10 5 3.5 x 10 6 5.19 x 10 6 2.59 x 10 6

10 1 x 10 6 3.5 x 10 7 5.19 x 10 7 2.59 x 10 7

100 1 x 10 7 3.5 x 10 8 5.19 x 10 8 2.59 x 10 8

1000 1 x 10 8 3.5 x 10 9 5.19 x 10 9 2.59 x 10 9

1

2

3

4

A.

?

Figure 4. Dimensional analysis of seepage reflux (after Shields and Brady, 1995). A) Table of brine properties and amount of Mg required to covert calcite to a dolomite with 7% porosity. B) Dimension of slice model used in their calculations. Slices are similar except that flow length varies from 1 to 1000 km. Base level is 100 m in all cases and head drops are 9.3 m for anhydrite saturated brine 1 and 17.1 m for halite saturated brine 2. C) Dimensional analysis of Darcy’s Law for a regional brine reflux system with 100 m base level. Diagram illustrates the time necessary for the minimum volume of dolomitizing brine to flow through a rectangular slice. All slices are 100 m thick and 1 m wide with varying flow lengths (1-4). It is assumed 350 kg/m3 of Mg are required to dolomitise the block. The lower curve for each flow path represents brine at halite saturation with a Mg exchange efficiency of 47%. The upper curve represents brine at anhydrite saturation with a Mg exchange efficiency of 36%. Permeability conversions assume an average brine density of 1160 Kg/m3 and a viscosity of 8.904 x 10-3 kg/(m/sec) and a temperature of 40°C.

lion years. Larger platforms require more time or that the evap-orite recharge areas migrate with time. Modelling with realistic assumptions show carbonate platforms that are tens to several hundred kilometres across can be dolomitised by such epeiric hydrologies.

The relative abilities of rocks to transmit brine control where the bulk of fluid flux occurs within platform limestones of any age. It is unrealistic to think of reflux dolomitisation as a homog-enous process uniformly overprinting all platform limestones. The inherent permeabilities of the various subsurface precursor beds, along with rates of change of relative sea level, must play essential roles in defining dolomite intensity. As the brine seeps seaward, the most permeable lithologies beneath the evaporite lagoon will focus the bulk of density flow. Whatever unit acts as the aquifer at the time of reflux will also be the subject of the most intense dolomitisation. High permeability units within flow paths of tens to hundreds of kilometres can channel the volume of necessary brine in time frames ≈ 1 Ma. Such high permeabili-ty units also have the potential to be dolomitised in regions well removed from the brine source.

Jones et al. (2002) have shown that the effects of a single 100,000-year platform evaporite episode can generate a reflux plume that persists for as long as 10 million years in underlying carbonates, before it is once again displaced by geothermally driven brines (Figure 5). They name this long term effect “latent reflux;” it defines a hydrological situation where the effects of a sinking reflux brine persist long after the at-surface brine lake or

saltern has disappeared. Their modelling utilised a single deposi-tional episode and a surface brine at gypsum saturation (150 ‰). They went on to note that higher salinity waters and more pro-longed episodes of evaporite precipitation generate more wide-spread and longer-acting brine plumes in the underlying sedi-ments (Jones and Xiao, 2005). In this context, it is interesting that dolomite is forming today in the slope sediments of southern Australia in zones flushed by marine-derived reflux brines, with elevated salinities (Rivers et al. 2012).

Detailed mapping of dolomite distribution across the Permian Capitan shelf, the type area for the brine reflux model, clearly shows reflux works in ancient evaporitic platforms (Melim and Scholle, 2002). Dolomite extent is greatest beneath and adja-cent to the drawndown lagoon facies. In contrast, dolomite dis-tribution through the Capitan reef and into the forereef outlines the distribution of those brine aquifers that were more active at the time of brine reflux, namely fractures and fore reef debris aprons. Aquifers carrying the descending brines out into the ad-jacent basin are dolomitised, while tighter aragonite-cemented intervals between aquifers are largely isolated from the flushing brines and so remain undolomitised.

Garber et al. (1990) used a similar model of reflux but called upon bittern brines, derived from the very saline Salado Forma-tion, to explain magnesite, not dolomite, distribution in the same basin. They argued that when a brine has a Mg/Ca ratio > 40, it precipitates a magnesite replacement phase, rather than the more generally observed dolomite. Similar reflux processes, driven by


0

2

4

0

2

4

0

2

4

0

2

4

500 100 150Distance (km)

Salinity (‰)150

116

116

360

2

40

2

40

2

40

2

4

432 1

43

2

43

2 123

4

1

23

43

221

23

4

3 21

500 100 150Distance (km)

Dep

th b

elow

sea

leve

l (km

)

Log stream function(kg.yr-1.m-1) 36 150

0.1 m.y.

0.2 m.y.

1 m.y.

10 m.y.

1

1

1

1

7636

7636

7636

66

46

36

Figure 5. Evolution of groundwater circulation and salinity persists for as much as 10 m.y. after a single reflux event at gypsum saturation which lasted 0.1 m.y. Dashed line in the stream-function plots represents interface between geothermal and reflux and latent reflux. Only the upper 4 km of platform is plotted; vertical exaggeration = 9 (after Jones et al., 2002).

the deposition of Zechstein evaporites, may have precipitated authigenic magnesite cements in siliciclastic dune sands of Rot-liegende reservoirs in the North Sea (Purvis, 1989).

Reflux dolomitisation in a platform setting can often be dis-tinctive, even in core, as the intensity of total dolomitisation decreases the further down core or laterally one moves from the evaporite-carbonate contact (Figure 6). If not completely re-equilibrated during later burial dolomitisation, the oxygen isotope signatures in platform dolomites formed by reflux typi-cally follow a lightening trend away from the evaporite unit; that is, the further one moves from the contact, the less was the sed-iment’s isotopic signature influenced by the heavier reflux-de-rived brines. If the sediments are not extensively recrystallised by later burial dolomitisation, trace elements, along with strati-graphic position, can give another clue to a reflux association. In a study of dolomite from Neogene to Permian age, Sass and Bein (1988) found that evaporative dolomites associated with gypsum/anhydrite tend to show 50-57% CaCO3 and the highest

levels of sodium (up to 2700 ppm). Marine nonevaporitic sub-tidal dolomites have a similar range of compositions, but sodium concentrations of only 150-350 ppm.

Most brine reflux models published to date utilise a platform evaporite setting typically tied to saltern anhydrites. Examples include: the Cambrian Ouldburra Formation in the Officer Basin, Australia (Kamali et al., 1995); the Devonian Birdbear (Nisku) Formation of Canada (Whittaker and Mountjoy, 1996); the De-vonain Upper Stettler Formation of Alberta (Al-Aasm and Ray-mus, 2018); Pennsylvanian platform carbonates of West Texas (Dickson et al., 2001); the Lower Permian Abo-Tubb (Hartig et al., 2011); Permian San Andres Formation, USA (Leary and Vogt, 1986; Ruppel and Cander 1988; Elliott and Warren, 1989); the Permian Tansill and Yates Formations of the Central Basin Platform, USA (Andreason, 1992; Garber et al., 1990); the Up-per Permian Changxing Formation and the Lower Triassic Feix-ianguan Formation of the NE Sichuan Basin, China (Jiang et al., 2013); the Jurassic Smackover and Haynesville Formations of the Gulf of Mexico (Moore et al., 1988); the Lower Cretaceous Edwards Formation of Texas (Fisher and Rodda, 1969) and the Middle Palaeocene Beda Formation, of the Sirt Basin, Libya (Garea and Braithwaite, 1996).

Widespread reflux dolomite beneath salterns is typically eoge-netic, but the dolomitisation process can continue into the me-sogenetic realm and that these earlier reflux dolomites can in-teract with later burial hydrothermal dolomites (Adams et al., 2018). The present may be the key to the past, but it is simply not a good time to document depositional or diagenetic analogues for ancient platform and basinwide evaporites, nor their associ-ated reflux hydrologies (Figure 7). There are no modern deposi-tional counterparts to most types of ancient marine-platform and basinwide evaporites; hence there are no modern same-counter-parts to brine-reflux dolomites (Warren, 2016; Chapter 5). Holo-cene evaporites are accumulating within an “icehouse” climatic phase. The current “icehouse” mode, typified by waxing and waning polar icesheets, has dominated the earth’s climate for the last 12 million years. “Icehouse” times, with their high-frequen-cy sealevel oscillations, are not conducive to the formation of

Re�ux

BASINWIDE EVAPORITES

(SUBSEALEVEL ISOLATED SUMP)

Brine

Mud�at or saltern evaporites

Negligiblesurface connection

Basinal evaporite

Laminated or massive halitein isolated tectonic depression

Reworkedshallow-water

evaporites

Continentalgroundwater

5 - 3

00 mSlope evaporite

Elevation head> 1km

Saline springs

10 - 100s kmOpen marine

Marineseepage

Brine drawndown to whereseepage in�ow equals out�ow

Tend to form on tectonic plate edgesTend to be halite-dominant

Salt masses up to hundreds of metres thickMono or bimineralogic masses (NaCl & CaSO4)

Halokinetic (loading, extension and collision settings)

~5-20 mRe�ux

Open marine

Sea levelMud�at Saltern Brine ~5 - 10 kmExposed barrier

PLATFORM EVAPORITES

(SUBSEALEVEL LAGOONAL SUMP)

Carbonate shelf(intercalated)

Epeiric and epicontinentalTend to be sulphate-dominant

Evaporite cycles up to tens of metres thickIntercalated with marine shelf beds

Can be main bed type in intracratonic sagsNot halokinetic (unless in later collision belts)

Figure 7. Platform versus basinwide evaporites have no modern same-scale counterparts (after Warren, 2016)

0 100 0 30

Dolomite(%)

Porosity(%)

Ool

itic

Gra

inst

one

10,230 ft

10,250 ft

10,260 ft

10,240 ft

10,270 ft

10,280 ft

10,290 ft

Anh

ydrit

eS.P. RES (IL)

10,200 ft

10,500 ft

Humble #1 Beltex

Smac

kove

r For

mat

ion

Norphlet Formation

cored interval

Pelo

idal

Pack

ston

e

Buckner Formation(anhydrite) Base of Evaporite

Figure 6. Wireline signature and the interpreted core description of the Humble #1 Beltex Well, Bowie County, Texas. Shows a close relationship between Buckner anhydrite and dolomitisation in the underlying Smack-over Formation. Note changes in proportion of dolomite and porosity in terms of proximity to evaporite contact (after Moore et al., 1988).


thick platform evaporites. In contrast, “greenhouse” time, with their inherent lack of polar ice sheets and low amplitude sealevel changes, favour the formation of platform evaporites and hence widespread reflux dolomitisation of platform carbonates. Basin-wide evaporites and their associated dolomites form via tectonic controls (continent-continent proximity) on basin isolation that are independent of “icehouse versus greenhouse” sealevel sce-narios (Figure 8; Warren, 2016).

Some reflux dolomites make good hydro-carbon reservoirsAs we have seen so far, during saltern deposition and the pre-cipitation of reflux dolomites, there is a hydrological system of basinward-flushing dense saline brines (Figure 1b). Significant substrate flushing occurs in the early stages of saltern deposition and shallow burial (<100 m burial) and before the loss of intra-salt and subsalt permeability. This is especially obvious where a gypsum saltern is accumulating atop a marine platform car-bonate and prior to the burial alteration of gypsum to anhydrite. Circulation is driven by gravitational instability, set up by an updip dense brine plume or brine curtain below accumulating evaporites.

At its periphery, this refluxing plume mixes with and displaces less dense subsurface and connate waters to create or enhance the reservoir properties of adjacent subsalt limestones via reflux

dolomitisation. As dense Mg-rich brines seep downdip from the platform saltern or mudflat, they create intervals of porous dolo-mites in more distal zones, located seaward of overdolomitised evaporite-plugged areas (Figure 9). In some cases, the most ex-tensively dolomitised carbonates in these more distal zones are high-energy, shallow water grainstones and packstones. These were the sediments with the higher permeabilities at the time of dolomitisation. However, in areas of pervasive syndepositional marine cementation, some initially porous carbonates may have already lost much of their intergranular permeability via over-growths of isopachous rim cements. In this situation, the less

Saltern waters Typically hypersaline

highly supersaturated with respect to dolomite

Open marinebiogenic marine

sediments

nonporous limestonecompacted and

calcite cemented

nonporous dolomite(micritic, primary?)

excess dolomite reduces porosity

(overdolomitised)

porous dolomite

Seaward �uxing Mg-rich subsurface brine mixing with

marine phreatic waters gives reduced dolomite

saturation (large crystals)

sucrosic dolomitereplacing marine

platform carbonates

Basinal

Marine platform

Figure 9. Reflux brines drive dolomitisation in marine carbonates some distance downdip from intervals of more restricted shoreward saltern-sealed carbonate that tend to be overdolomitised and evapo-rite-plugged beneath the main evaporite mass (model after Saller and Henderson (1998) and Elliott and Warren (1989).


Intracratonic evaporite

seaway

Shallow shelf seawayconverts to saltern

with mud�at

Deepsea

Deepsea

Peri-platformevaporite

seaway

Former epeiric seawayconverts to saltern

with mud�at periphery

Creation of tectonic barrier(greenhouse or icehouse mode)

Creation of depositional eustatic barrier(favored by greenhouse eustacy)

Epeiric(epicontinental)

seaway(normal marine) Sh

elf (

peric

ontin

enta

l)

seaw

ay

Deepsea

0 500 km

500 m

Intrashelfdepression

Small salina & mud�at

evaporites

Shallow marine seaways(normal to slightly restricted salinities)

100 m

200 100 0

10 m

0200 100

Greenhouse mode: 4th order 100,000 year sealevel curve is low amplitude - less than 10 m change per 100 ky)

Icehouse mode: 4th order 100,000 year sealevel curve is high amplitude - more than 100 m change per 100 ky)

Hydrographicisolation

Hydrographic

isolatio

n

Ferry Lake AnhydriteHith Anhydrite

Khu� AnhydriteSeven Rivers Fm

Amazon BasinCanning BasinMichigan BasinWilliston Basin

Barrier

Barrier

Alternation between epeiric hydrographic and at-surface marine connection to the depositional basin leads to alternating

stacks of marine-carbonate and evaporite beds

ky ky

GREENHOUSE EUSTACY

DEPOSITION OF PLATFORM EVAPORITES

Figure 8. Epeiric (epicontinental) seaways that covered large areas of the continental interior with shallow marine waters are also known as intracra-tonic basins, while epeiric seaways that formed as very wide shelf edges along the continental margin as known as pericontinental seaways. Waters were very shallow over areas that were hundreds to thousands of kilometres across. Neither setting has a modern counterparts; pericontinental seaways that evolve into salterns and or evaporitic mudflats are more typical of marine-margin eustatic styles on the earth in greenhouse climate mode, when the lack of permanent polar icecaps, especially coupled with higher rates of seafloor spreading, meant the continental freeboard was much larger than it is today. Greenhouse meant relatively stable low-amplitude 4th-order sealevel and that there was time to build up a near continuous platform edge shoal Intracratonic basins tended to go evaporitic with tectonically driven hydrographic isolation, in both greenhouse and icehouse mode. With hydrographic isolation both these restricted marine seaways quickly became salterns or evaporitic mudflat systems (after Warren, 2010, 2016).

cemented platform mudstones act as aquifers to the refluxing brines and so are preferentially dolomitised.

Likewise, in more restricted portions of an epeiric carbonate platform where the only sediment beneath a salt seal is lagoonal mud, it is dolomitised and sometimes, in areas away from zones of immediate evaporite plugging, is converted into a reservoir with intercrystalline permeability (e. g. the Levelland-Slaugh-ter trend in the San Andres Fm, West Texas and New Mexico; Elliott and Warren, 1989). In some muddy platform carbonates, only the pelletal and sandy burrow fill retains sufficient perme-ability to be dolomitised, while their muddier surrounds are not (e.g. Red River Dolomite of North Dakota). Whatever the set-ting, the rule of thumb is that refluxing brines will preferential-ly dolomitise whatever carbonate facies is capable of acting as an aquifer at the time of reflux. Regions proximal to the brine source tend to be overdolomitised and evaporite-plugged. Re-gions somewhat outboard and more distal to the saltern edge

tend to the reservoir sweet-spots that remain as sucrosic porous dolomite that are not plugged with evaporite cements (Figure 9).

Reflux dolomitisation controls po-rosity/permeability distribution in the finer-grained parts of some Jurassic Arab Formation reservoirs in Saudi Arabia, as well as the Permian Khuff Formation in the Arabian Gulf, the Smackover Formation carbonates of the Gulf of Mexico, the Upper Perm-

ian Changxing Formation and the Lower Triassic Feixianguan Formation, NE Sichuan Basin, China, and many other oil and gas fields across the globe where dolomites occur beneath an evaporite seal (Figures 10, 11; Sun, 1995; Warren, 2000; Jiang et al., 2013).

Yet, in other depositionally similar evaporitic settings, some of the mesohaline laminites that underlie bedded salts are not dolo-mitised, typically because they were evaporite-cemented in the strandzone before the onset of reflux flushing. They were too tight to transmit the volumes of Mg-rich brine needed for do-lomitisation. Likewise, dolomites immediately below a saltern or mudflat brine source tend to be overdolomitised and evap-orite-plugged, there the aquifer limestones were located in the zone of most intense flushing by salt-supersaturated brines (Fig-ures 9, 10).

Syndepositional permeability in units deposited immediately


Anhydrite layer

Re�ux brine Anhydrite cement (nodules)Boundary of P3ch & T1f

Brine lagoon(saltern)

Margin barrier(shoals and banks)

T1f ooid-replaced dolomite

T1 over-dolomitised &evaporite plugged

P3ch reef dolomite

P3ch ooid-replaceddolomite

Tight limestone

Trough(basin)

Sealevel Slope

Brine escapesinto sea

Figure 10. Model of reservoir creation via saltern-sourced brine reflux, based on dolomite distribution in the Upper Permian Changxing Formation (P3ch) and the Lower Triassic Feixianguan Formation (T1f), NE Sichuan Basin, China (after Jiang et al., 2013).

Seal (drape)

Reservoir

Platform Bedded Seal (dolomite source) Basinwide Bedded Seal (dolomite source)

Seal penecontemporaneuswith reservoir dolomitisation

Seal postdates and drapesre�ux-dolomite reservoir

Seal

pen

econ

tem

pora

neus

with

rese

rvoi

r

Seal edge focuses upwelling hydrothermal �uids that overprint a variably re�ux-dolomitised reservoir

North Ward-Estes, Tx.Rotliegende, North Sea

Levelland-Slaughter trend, NMGhawar field, Saudi ArabiaFeixianguan fields, China

Miocene carbonates, Iraq, IranJurassic Arab cycles, Saudi, UAE

Winnipegosis and Michigan Basin

Yates field, USAPermian Khuff, Saudi

Devonian carbonates, BC, Canada

Carb

onat

eSi

licic

last

ic

Pinn

acle

sSa

nd s

hoal

s

Buria

l

Best eservoir (dolomitic) typically downdip of main saltern edge

Dolomitic reservoir sand thicks typically along main saltern edge either as salt-sealed erg edges or dissolution-induced eolian stacks transgressed by salt seal

Reservoir draped by sabkhaover elevated parts of reservoirseal facies (salina in lows)

Ø (dolomitised)

saltern barrier marine

Basinwide (intracratonic)Shale seal

Former shelf edge Former platform interior

Marine

MeteoricSeeps SeepsAnhydrite

Pinnacleout�ow Salt plugged

pinnacle

Halite

Ø fairway

Basinwide salts

Sand shoals mostly active prior to hydrographic isolation

Patches of stacked anhydrite-sealed subsalt re�ux-dolomitised shoals (evap. mud�at cap, saltern o� structure)

MarineIntracratonic basinwide

or epicontinental cycles

Figure 11. Summary of the various ancient platform and basinwide bedded evaporite settings where a combination of evaporite sealing, dolomi-tisation (both early and late) and focused evaporite related fluid flow create a variety of dolomitic hydrocarbon traps (Warren 2016).


after a transition into widespread evaporite deposition control the type and extent of the brine reflux alteration halo as does the position of subsalt aquifersat the time of reflux. So, according to the ambient hydrology, porosity evolution in subsalt sediments can vary widely. Worldwide, Phanerozoic saline dolomites define three of the four main dolomite reservoir associations (Figure 11; after Sun, 1995; Warren, 2000, 2016): (1) Evapor-itic mudflats in a carbonate platform, most often a ramp profile (roughly equivalent to peritidal-dominated dolomitic carbonate of Sun, 1995), (2) Saltern-sealed platform carbonates with sa-line mudflat development over palaeotopographic highs. This system typically develops across an epeiric platform behind a rimmed shoal profile and is roughly equivalent to subtidal dolomitic carbonate associated with evaporitic tidal flat/lagoon of Sun (1995), (3) Basinwide evaporite seal (subtidal dolomit-ic carbonate associated with basinal evaporites of Sun, 1995), and (4) Nonevaporitic dolomitic carbonate sequences associated with topographic highs/unconformities: platform-margin build-ups or burial related fault/fracture controls. That is, three of the four main worldwide dolomite-reservoir associations show an intimate association with evaporites (Figure 11).

Brine reflux is the dominant process in dolomite creation in all three, although it can be overprinted by later burial dolomites. Mixing-zone dolomitization once considered part of process set (4) has received enough criticism in recent years to suggest that it is no longer considered viable as a subsurface mechanism for making large amounts of dolomite (e.g. Moore 2009, p. 166). In its place Li et al. (2013) propose a somewhat more specific mod-el, which still requires fluid mixing whereby an ascending flow of freshwater mixes with evaporated seawater about the edge of a saline basin, leading to extensive dolomitization.

Dolomite reservoirs created by brine reflux can be subdivided using the dominant depositional style of the evaporite capstone, namely, platform and basinwide (Figure 11, Table 2). In plat-form-evaporite-sealed dolomite associations, the evaporite seal is typically ≈ 5 -10 metres thick (although it can be 20-30m thick) and is usually dominated by saltern or mudflat anhydrite interlayered with shoaling platform carbonates. This evaporite cap can be further subdivided into evaporitic mudflat or salt-ern-dominated. Basinwide seals to carbonate reservoirs tend to be thick (>100m), halite-prone in the basin centre and to cap car-bonate buildups about the basin edge. Such buildups typically grew prior to the saline giant stage. Ongoing or longterm region-al reflux beneath accumulating evaporites in both platform and basin settings can “overdolomitise” the reservoir and encourage evaporite plugging leading to occlusion of effective porosity.

In platform settings, such porosity-depleted overdolomitised zones tend to be most common in regions of restricted carbonate deposition beneath leaky evaporitic mudflats. These sediments tend to accumulate in the more updip evaporite-dominated por-tions of a carbonate platform, where underlying carbonates have been flushed for extended periods by hypersaline brines. With basinwide-sealed associations, a later burial overprint and per-vasive intra-reservoir karst/fracturing events seem to be needed to create and maintain economic porosity levels (Figure 11; War-ren, 2016, Chapter 10).

Prediction of saline reflux reservoir settingsIn my experience, most variation in subevaporite reservoir qual-ity is the end product of a combination of depositional facies and varying intensities of evaporite plugging, dissolution, reflux do-lomitisation and burial stage cementation (Table 2). Lateral and vertical variations in all but the latter are typically eogenetic and indicated by facies variations in the seal itself. Yet, for much of the oil industry, evaporite plugging and reflux dolomitisation are associations that geological and geophysical staff do not quan-titatively integrate into a reservoir model (other than via loosely controlled geostatistical formulations).

The relevance of a bedded evaporite to adjacent reservoir qual-ity at any scale, beyond the consideration of the evaporite's seal integrity during exploration or field development, is not part of most reservoir evaluation studies (The typical question to be asked and answered of the evaporite seal in a petroleum sys-tem is; Is it thick enough? Typical magic numbers worldwide for answering yes are; greater than 10 metres thick for a clean anhydrite seal, and greater than 5-10 metres thick for clean ha-lite). Once the integrity of an evaporite seal or cap to a poten-tial reservoir is considered established, further study of the seal properties or intraseal textures is not regarded as relevant, oth-er than hoping for, or establishing, the seals lateral persistence. Typically, in most subsequent study, measurement focuses on the properties of the reservoir itself, the seal is rarely cored or sampled and, if it is, this is usually the result of an error in pick-ing depth to top reservoir.

Attempting to better understand diagenetic intensity in the res-ervoir using signatures from wireline tool measurements taken across the seal intersection is typically considered too difficult, or the possibility is not even recognized. Yet, as any carbonate geologist will testify, diagenesis is what distinguishes properties in any carbonate reservoir from those in a sandstone. Pervasive, but variable, diagenetic intensity is what controls carbonate res-ervoir quality in every bedded evaporite-sealed petroleum sys-tem worldwide. Almost all of the matrix quality in the world’s giant and supergiant bedded evaporite-sealed carbonate fields was established during deposition and early burial. This is when surface topography controls the intensity of circulation in the di-agenetic hydrology (Table 2). Resultant poroperm plumbing can be locally enhanced by later fracture development and perhaps the formation of coarsely crystalline saddle dolomites (typically fed by fractures and faults). In petroleum systems exemplified by the examples of Yates and North Ward-Estes fields, reservoir quality is predictably tied to the style and position of the early evaporite hydrology and is indicated by evaporite textures in the seal. The same approach is also useful in other reflux-dominated carbonate reservoirs, as in the Arab and Khuff formations in the Middle East (Table 2; Figure 11).

See the addendum of this article for a partial modern example of brine reflux instigated by anthropogenic changes to the hydrolo-gy of Owens Lake, California.


ReferencesAdams, A., L. W. Diamond, and L. Aschwanden, 2018, Do-lomitization by hypersaline reflux into dense groundwaters as revealed by vertical trends in strontium and oxygen isotopes: Upper Muschelkalk, Switzerland: Sedimentology, v. 0.

Adams, J. F., and M. L. Rhodes, 1960, Dolomitisation by seep-age refluxion: American Association of Petroleum Geologists Bulletin, v. 44, p. 1912-1920.

Aharon, P., Y. Kolodny, and E. Sass, 1977, Recent hot brine do-

lomitization in the “Solar Lake”, Gulf of Elat: Isotopic, chemical and mineralogical study. Jour. Geology, vol.85, pp.27-48.: Jour-nal of Geology, v. 85, p. 27-46.

Al-Aasm, I. S., and S. Raymus, 2018, Reflux dolomitization and associated diagenesis of Devonian Upper Stettler Formation and Crossfield Member, south central Alberta: Petrologic and isoto-pic evidence: Bulletin of Canadian Petroleum Geology, v. 66, p. 773-802.

Andreason, M. W., 1992, Coastal siliciclastic sabkhas and relat-ed evaporative environments of the Permian Yates Formation,

Feature Outcome Recognising in the subsurface

Platf

orm

evap

orite

Anhydrite is the dominant seal mineralogy

Immediate subsalt region is typically evaporite plugged Require understanding of variations of intrasalt wireline signatures (gam-ma-neutron-density FMI or acoustic image logs) sabkha versus salina in seal?Depositional facies variations in subseal facies may be

indicated by textures in the seal itselfReflux dolomitisation in subsalt and salt-adjacent positions

Best reservoir is typically located downdip and basinward of thickest saltern interval

Resolve whether the current salt edge indicates the primary position, or is the result of salt dissolution.Reliably recognise signatures of marine flooding, subaerial exposure and salt dissolution horizons (core tied to gamma-neutron-density, FMI and acoustic image logs)

Hydrodynamics define stepout distance from saltern edge to likely economic reservoir poroperm

Basin

wide

evap

orite

Halite is the dominant seal mineralogy

Anhydrite proportions typically higher in intracratonic settings and/or atop former subsalt highs

Require understanding of variations of intrasalt wireline signatures (gam-ma-neutron-density FMI or acoustic image logs)

Non-intracratonic salt systems tend toward halokinesis Halokinetic reservoir associationsDiagenetic overprints can control poroperm distribu-tion (stable isotope and trace element values are useful discriminants of fluid source)

Shelf edge with buildups:Shelf edge is typically exposed during episodes of evaporative drawdown, and may be karstified

Definable patterns of drawdown-induced seepage fringe outflow and associ-ated alteration in basinal sediments

Downdip pinnacle reefs: Crests acting as seep outflows during drawdown

Indicated by reef fringe cap facies made up of seep indicators, such as pisolites and tepees

Problems related to salt plugging of reef further out in basin

Circum-basin poroperm fairway can be defined by combination of wireline and seismic mapping of basin architecture

Common features of note in both platform and basinwide settings

Intra

basin

de

posit

ion

Thick evaporite beds require stable long term aridity

Typically meteoric effects are minor in most parts of drawndown part of basin during time the evaporite seal is precipitated

Meteoric waters can be derived updip via deep meteoric circulation and outflow occurs in particular circum salt or salt-adjacent positions (stable isotope and trace element values are useful discriminants)

Diss

olutio

n is o

ngoin

g

Exposure surfaces Penecontemporaneous local intrabed dissolution can occur during short term freshening via increased marine or meteoric seepage (recognition of intra salt leachate styles)

Stable isotope and trace element values are useful discriminants using core or cuttings. If region can be inferred by anomalies in wireline or seismic

Early burial Brine reflux) Identified by subsalt dolomite intensity map - wireline derived tied to basin architecture

Later burial Hydrothermal brine crossflow alters poroperm distribution Occurs in particular positions in basin architecture (fault, fracture and folding controls, typically salt adjacent, but can be subsalt

Edge effects (Mesogenet-ic focus)

Presalt can create karst overprint in subsalt reservoir facies

Defined by combination of core study, wireline and seismic mapping of basin architecture

Synsalt, can created zones of non salt sediment thicks (eg increased thickness in fluvial sediments or eolian sandflats)Post salt, act as escape conduits for subsalt basinal fluids can be carrying hydrocarbons and/or metals

Uplift (Telogenetic focus) Dissolution reservoir associations are discussed in detail in Warren, 2016 Confusion between dissolution residue layers and marine flooding surfaces (different diagenetic associations)

Ambiguity can lead to errors in basin history, diagenetic history, maturation timing, etc

Distinction requires core, or wireline tied to seismic, or core-calibrated FMI interpretation.

Table 2. Methods for identification and prediction of diagenetic features, especially dolomitisation, associated with bedded evaporites that are useful in helping define distribution of quality in the subsalt or salt-adjacent reservoirs (from Warren, 2016).

North Ward-Estes field, Ward County, Texas: American Asso-ciation of Petroleum Geologists Bulletin, v. 76, p. 1735-1759.

Bischoff, J. L., J. P. Fitts, and K. M. Menking, 1993, Sediment pore waters of Owens Lake Drill Hole OL-92, in I. S. Smith, and J. L. Bischoff, eds., Core OL-92 from Owens Lake, south-east California, U.S. Geological Survey Open-File Report 93-683 (available online at http://pubs.usgs.gov/of/of93-683/report.html).

Budd, D. A., 1997, Cenozoic dolomites of carbonate islands; their attributes and origin: Earth-Science Reviews, v. 42, p. 1-47.

Dickson, J. A. D., I. P. Montanez, and A. H. Saller, 2001, Hy-persaline burial diagenesis delineated by component isotopic analysis, Late Paleozoic limestones, west Texas: Journal of Sed-imentary Research, v. 71, p. 372-379.

Elliott, L. A., and J. K. Warren, 1989, Stratigraphy and deposi-tional environment of lower San Andres Formation in subsur-face and equivalent outcrops; Chaves, Lincoln, and Roosevelt counties, New Mexico: Bulletin American Association of Petro-leum Geologists, v. 73, p. 1307-1325.

Fisher, W. L., and P. U. Rodda, 1969, Edwards Formation (Low-er Cretaceous), Texas: Dolomitization in a Carbonate Platform System: Bulletin American Association Petroleum Geologists, v. 53, p. 55-72.

Garber, R. A., P. M. Harris, and J. M. Borer, 1990, Occurrence and significance of magnesite in Upper Permian (Guadalupian) Tansill and Yates formations, Delaware Basin, New Mexico: AAPG Bulletin, v. 74, p. 119-134.

Garea, B. B., and C. J. R. Braithwaite, 1996, Geochemistry, isotopic composition and origin of the Beda dolomites, block NC74F, SW Sirt Basin, Libya: Journal of Petroleum Geology, v. 19, p. 289-304.

Hartig, K. A., G. S. Soreghan, R. H. Goldstein, and M. H. Engel, 2011, Dolomite in Permian paleosols of the Bravo Dome CO2 field, U.S.A.; Permian reflux followed by late recrystallization at elevated temperature: Journal of Sedimentary Research, v. 81, p. 248-265.

Jiang, L., C. F. Cai, R. H. Worden, K. K. Li, and L. Xiang, 2013, Reflux dolomitization of the Upper Permian Changxing Forma-tion and the Lower Triassic Feixianguan Formation, NE Sichuan Basin, China: Geofluids, v. 13, p. 232-245.

Jones, G. D., F. F. Whitaker, P. F. Smart, and W. E. Sanford, 2002, Fate of reflux brines in carbonate platforms: Geology, v. 30, p. 371-374.

Jones, G. D., and Y. Xiao, 2005, Dolomitization, anhydrite ce-mentation, and porosity evolution in a reflux system; insights from reactive transport models: Bulletin American Association Petroleum Geologists, v. 89, p. 577-601.

Kamali, M. R., 1995, Sedimentology and petroleum geochemis-try of the Ouldburra Formation, Eastern Officer Basin, Australia: Doctoral thesis, University of Adelaide, Adelaide, 165 p.

Kocurko, M. J., 1979, Dolomitization by spray-zone brine seep-age, San Andres, Colombia: Journal of Sedimentary Petrology, v. 49, p. 209-214.

Leary, D. A., and J. N. Vogt, 1986, Diagenesis of the San Andres Formation (Guadalupian), Central Basin Platform, Permian Ba-sin, in D. G. Bebout, Harris, P. M., ed., Hydrocarbon Reservoir Studies San Andres/Grayburg Formations, Permian Basin, v. 26, SEPM Special Publ., p. 67-68.

Lucia, F. J., and R. P. Major, 1994, Porosity Evolution through Hypersaline Reflux Dolomitization: Dolomites, Blackwell Pub-lishing Ltd., 325-341 p.

McCaffrey, M. A., B. Lazar, and H. D. Holland, 1987, The evap-oration path of seawater and the coprecipitation of Br(-) and K(+) with halite: Journal of Sedimentary Petrology, v. 57, p. 928-937.

Moore, C. H., A. Chowdhury, and L. Chan, 1988, Upper Jurassic Smackover dolomitization, Gulf of Mexico; a tale of two waters, in V. Skula, and P. A. Baker, eds., Sedimentology and geochem-istry of dolostones, v. 43: Tulsa, Okl., SEPM Special Publica-tion, p. 175-189.

Moore, C. H., and W. J. Wade, 2009, Carbonate Reservoirs (2nd Edition); Porosity and diagenesis in a sequence stratigraphic framework, Elsevier, 392 p.

Murray, R. C., 1969, Hydrology of South Bonaire, Netherlands Antilles; a rock selective dolomitization model: Journal of Sedi-mentary Petrology, v. 39, p. 1007-1013.

Purvis, K., 1989, Zoned authigenic magnesites in the Rotliegend Lower Permian, southern North Sea: Sedimentary Geology, v. 65, p. 307-318.

Rivers, J. M., T. Kurt Kyser, and N. P. James, 2012, Salinity re-flux and dolomitization of southern Australian slope sediments: the importance of low carbonate saturation levels: Sedimentolo-gy, v. 59, p. 445-465.

Ruppel, S. C., and H. S. Cander, 1988, Dolomitization of shal-low water carbonates by seawater and seawater-derived brines, San Andres Formation (Guadalupian), West Texas, in V. J. Shuk-la, and P. A. Baker, eds., Sedimentology and Geochemistry of Dolostones, v. 43: Tulsa Okla, SEPM Special Publ., p. 245-262.

Saller, A. H., and N. Henderson, 1998, Distribution of poros-ity and permeability in platform dolomites - Insight from the Permian of west Texas: AAPG Bulletin-American Association of Petroleum Geologists, v. 82, p. 1528-1550.

Sandler, A., Y. Harlavan, and G. Steinitz, 2004, Early formation of K-feldspar in shallow-marine sediments at near-surface tem-peratures (southern Israel): evidence from K-Ar dating: Sedi-mentology, v. 51, p. 323-338.

Sass, E., and A. Bein, 1988, Dolomites and salinity; a compara-tive geochemical study: Shukla, Vijai, Baker, Paul A. Sedimen-tology and geochemistry of dolostones, based on a symposium. Special Publication Society of Economic Paleontologists and

Mineralogists, v. 43, p. 223-233.

Shields, M. J., and P. V. Brady, 1995, Mass balance and fluid flow constraints on regional scale dolomitization, Late Devo-nian, Western Canada sedimentary basin: Bulletin of Canadian Petroleum Geology, v. 43, p. 371-392.

Sibley, D. F., 1980, Climatic control of dolomitization, Seroe Doni formation (Pliocene), Bonaire, N. A.: Spec. Pub., v. 28, SEPM, p. 247-258.

Sun, S. Q., 1995, Dolomite reservoirs; porosity evolution and reservoir characteristics: American Association of Petroleum Geologists Bulletin, v. 79, p. 186-204.

Teoh, C. P., J. C. Laya, F. Whitaker, T. Gabellone, M. Tucker, C. Manche, S. E. Kaczmarek, and B. Miller, 2018, Unravelling Reflux Dolomitization: Why Size Matters: AAPG Search and Discovery Article #51519 (Adapted from poster presentation given at AAPG 2018 AAPG Annual Convention and Exhibition, Salt Lake City, Utah, May 20-23, 2018).

Tietz, G., and G. Müller, 1971, High-magnesian calcite and ara-gonite cementation in recent beachrocks, Fuerteventura, Canary Islands, Spain, in O. P. Bricker, ed., Carbonate cement, The John Hopkins Press, Baltimore, p. 4-8.

Warren, J. K., 2000a, Dolomite: Occurrence, evolution and eco-nomically important associations: Earth Science Reviews, v. 52, p. 1-81.

Warren, J. K., 2010, Evaporites through time: Tectonic, cli-matic and eustatic controls in marine and nonmarine deposits: Earth-Science Reviews, v. 98, p. 217-268.

Warren, J. K., 2016, Evaporites: A compendium (ISBN 978-3-319-13511-3): Berlin, Springer, 1854 p.

Whittaker, S. G., and E. W. Mount-joy, 1996, Diagenesis of an Upper Devonian carbonate-evaporite se-quence: Birdbear Formation, south-ern Interior Plains, Canada: Journal of Sedimentary Research A: Sedi-mentary Petrology and Processes, v. 66, p. 965-975.

Addendum, does reflux dolomitisation real-ly work?Some have argued that brine reflux cannot be a geologically sig-nificant process because we do not see modern marine brines capable of displacing substantial volumes of pore waters at suf-ficiently rapid rates to drive brine reflux. Instead, the deposition of bedded salt can create a “hydroseal.” At shallow burial depths (>40-100 m) I would argue that most gypsum and halite beds are porous as they form and in many subaqueous evaporite settings do not lose their porosity until at least few tens of metres into the burial realm. Unlike much evaporite deposition in a pre-ice-house earth, marine-fed reflux dolomite is not a widespread pro-cess in today’s marine platforms but was a significant process in ancient evaporitic settings.

Laterally discontinuous buildup geometries characterise today's carbonate platforms. They are a response to high-amplitude high-frequency eustatic signals of the current icehouse climate mode whereby continuous reef rims and shoals do not form, and so saltern-covered carbonate platforms cannot develop. Subsea-level marine seepage inflows into an isolated platform lagoon and consequent reflux of dense brine is impossible in the hydrol-ogies of icehouse mode carbonate shelves. Likewise, there is no-where on the world’s surface today where large subsealevel ma-rine basin floors are hydrographically isolated from the ocean, so no basinwide evaporites and reflux dolomites are accumulating in modern examples of this setting.

But the viability of reflux hydrology and its ability to quickly drive brine descent into underlying strata can be seen in mod-ern continental systems where man has interfered with natural hydrologies and created saline hydrologies in regions where there were none before. This is perhaps scale-limited but offers

Figure 12. Owens Lake and Valley, California. View to North.


our best set of hydrological analogues to characterise the vast reflux hydrologies of ancient platform and basinwide evaporite systems.

One of the best examples comes from Owens Lake, California, where the water needs of the City of Los Angeles converted a moderately saline perennial lake (salinity ≈ 90‰) into a hypersa-line pan accumulating bedded halite and trona. Starting in 1913, the streams that fed the Owens River, which in turn fed Owens Lake, were diverted by Los Angeles Department of Water and Power to feed the Los Angeles Aqueduct. The lake level started to drop quickly and by 1924 the lake had completely dried out (Figure 12). Prior to this, the lake was saline, but perennial.

In order to study the Quaternary history of the lake, the US Geological Survey collected a 323 m long core from the lake (Bischoff et al., 1993). The core encompasses the last 800,000 years of basin fill and was analysed for water content, pore water salinity, sulphate and chloride values (Figure 13). Water con-tent, a measure of compaction, varies erratically down the core, generally decreasing from about 60 wt % at the top to about 20 wt % at 240 m (Bischoff et al., 1993). Below 1200 metres and to the bottom of the core, the water content sharply maintains levels of less than 50% by weight. This zone is characterized by an increasing abundance of sandy units, which are less subject to dewatering-related compression. Salinity varies with depth in a smooth pattern, with a minimum at 30 m, gradually increas-ing to a single broad maximum at about 150 metres depth, and sharply declining thereafter to steady low values at 210 metres and below.

The salinity of the modern lake before the diversion of the Ow-ens River, was about 9% (Bischoff et al., 1993). Assuming that similarly elevated salinity characterizes previous interglacial times when Owens Lake was the terminus to the Owens River and that freshwaters must have char-acterized the glacial periods of intense overflow, Bischoff et al. expect about 8 salinity oscilla-tions during the past 800 kyr span captured in the core. Such cycles are seen in the solid components of the sediments, particularly for carbonate and organic carbon content (Bischoff et al., 1993) indicating the lake did indeed experience such climate chang-es. Yet the salinity-depth profile has been drastically smoothed by post-depositional diffusion of dissolved salts. This is not a hy-drologically closed system.

Remnant water, a leftover from the last glacial is the likely ex-planation only for the first sa-linity minimum seen at 30-40 m depth (Bischoff et al., 1993). The

smooth and gradual increase of salinity in the older sediments below this depth to a maximum at about 150 m is likely the re-sult of diffusional smoothing of waters of older cycles. Diffusion should have had more than sufficient time, therefore, to smooth salinity gradients even lower in the core. The abrupt and erratic decrease of salinity from 150 to 210 m depth, and the erratic and generally low salinities from 210 to the bottom of the hole points to an open throughflow system for the basal pore fluids. The most likely explanation for this pattern is that deeply circu-lating fresher waters are actively moving through the sandy units below 200 m, diffusionally harvesting solute from the overly-ing fine-grained sediments in the process. Thus groundwater is moving at different velocities in the varyingly permeable sandy units. Neither diffusional steady state or smoothing of the salini-ty gradients has been achieved. This particular depression in the Basin and Range province of the USA, is in the type region for what has been argued are “hydrologically closed” lake systems, clearly is not. It is leaking waters vertically and laterally at a rate sufficient to smooth any former climatic layering.

But what is most impressive in the lake’s hydrology is what has happened in the upper 20 metres in the last 80-90 years (Figure 13). The pre-1900 waters of Owens Lake had a salinity less than 9%, the lake was alkaline and probably had a pH ≈9-10. The extremely high salinity of the top 20 metres is not a reflection of the pre-1912 natural lake. Rather, it is a consequence of riv-er diversion and almost complete desiccation of the lake basin since 1924, and the associated downward migration of dense re-sidual-brines (Bischoff et al., 1993). As the post-1924 lake dried out, Na-carbonate minerals precipitated to form a 1 to 2 m thick salt bed now covering the lower parts of the lake depression. Re-sidual brines from the salt bed were denser than the underlying 9% pore water and were also relatively enriched in Cl. Modern

Figure 13. Core OL-92 a 327 m core collected in Owens Lake, California. Shows depth plots of sand and water content (%), salinity (%), chloride and sulphate content (mmol) of pore waters squeezed from the core (replotted from data in Table 1 in Bischoff et al., 1993).

Dep

th (m

)

0

50

100

150

200

250

300

350

Content %0 50

Salinity (%)0 10

Concentration (mmol)0 1000

SandWater

ClSO4

Post 1912 re�uxhydrology

Ant

hrop

ogen

ical

ly u

ndis

turb

ed

laye

red

lake

geo

hydr

olog

y

Rela

tivel

y st

able

pore

�ui

d sa

linity

Upw

ard

decr

ease

inpo

re �

uid

salin

ity


brine percolated downward by gravity displacement and ionic diffusion, so that high salinity pore waters now dominate down to about 15-20 m.

Reflux brines derived by the accumulation of 2 metres of bedded salt have penetrated some 20 m of relatively fine-grained wa-ter-saturated sediment in 80 years. The original now-displaced pore waters had a salinity up to three times that of seawater and so would have been more difficult to displace than seawa-ter brines in an ancient carbonate platform undergoing reflux, which would have had salinities around 3-4%. Clearly, once a holomictic bedded salt system has formed, associated reflux is rapid and ongoing.

What is more, the total system leaks water and brine at all levels in the cored hydrology. It possesses an open hydrological base to the lake depression even though the underlying nonevaporite sediments are more than 90% clay. Reflux rules the upper por-tions of the hydrology, and deeper artesian circulation rules the lower part of the hydrology. It is a small, but temporally impres-sive, example of brine-driven hydrological processes that would be active in ancient drawndown platform or basinwide evaporite settings.

Saltworks Consultants Pty LtdABN 068 889 127Kingston Park,5049 South Australia

Email: [email protected]

Web Page: www.saltworkconsultants.com


Sa ae · the less-porous shelf edge and slope carbonates are not. They envisioned reflux...

Documents

Transcript of Sa ae · the less-porous shelf edge and slope carbonates are not. They envisioned reflux...