b36378.pdf

This paper is published under the terms of the CC-BY license.

GSA Bulletin

;

Month/Month 2022; 0; p. 1–13;

https://doi.org/10.1130/B36378.1; 7 figures; 2 tables; 1 supplemental file.

published online 8 June 2022

1

New age and lake chemistry constraints on the Aptian

pre-salt carbonates of the central South Atlantic

M. Lawson

1,†

, J. Sitgreaves

2

, T. Rasbury

3

, K. Wooton

3

, W. Esch

2

, V. Marcon

2

, S. Henares

2

, A. Konstantinou

2

,

E. Kneller

2

, D. Gombosi

2

, V. Torres

2

, A. Silva

4

, R. Alevato

4

, M. Wren

2

, S. Becker

2

, and J. Eiler

5

1

Aker BP, Stavanger 4020, Norway

2

ExxonMobil, Springwoods Village Parkway, Spring, Texas 77027, USA

3

Department of Geosciences, Stony Brook University, Stony Brook, New York 11794, USA

4

ExxonMobil Exploração Brasil, Rua Lauro Muller, Rio de Janeiro, Rio de Janeiro 22290-160, Brazil

5

Division of Geological and Planetary Sciences, California Institute of Technology, East California Boulevard, Pasadena,

California 91125, USA

ABSTRACT

The Cretaceous lacustrine carbonates of

the offshore Brazilian and West African pre-

salt basins represent some of the most exten

-

sive non-marine carbonates discovered in the

geologic record. Despite being intensively

studied over the past decade, the age of these

carbonates and the overlying regional salt

sequences is highly controversial. Similarly,

the conditions under which these carbon

-

ates were deposited remains poorly under

-

stood. Here, we provide the first integrated

geochronology-thermometry study of these

carbonates to develop an improved under

-

standing of when and under what conditions

they formed. We utilize carbonate clumped

isotope and

87

Sr

/86

Sr geochemistry along-

side traditional petrographic techniques to

identify samples minimally altered from

burial diagenesis that may yield reliable age

and lake chemistry constraints. Carbonate

clumped isotope apparent temperatures for

the studied carbonates range from 36

°

C to

91

°

C, which we infer to represent a range

in sample preservation from minimally

altered depositional temperatures through

to those that have been overprinted by burial

diagenesis.

87

Sr

/86

Sr values of our samples

are consistent with those of previous stud

-

ies for Cretaceous pre-salt carbonates that

have not experienced significant alteration

from hydrothermal fluids. Through this

approach, we measured the first high resolu

-

tion isotope dilution U-Pb age constraint of

115.83

±

1.56

Ma (2

σ

) on a well preserved

carbonate. Combined with overlapping

lower resolution laser ablation U-Pb ages for

time-equivalent stratigraphy on two separate

carbonate platforms of 114.46

±

4.72

Ma

and 109.73

±

9.26

Ma, these ages pr

ovide

the first robust direct age calibration for

pre-salt carbonates deposited on either side

of the South Atlantic during the final stages

of the break-up of Gondwana in the Early

Cretaceous. These ages also provide the first

calibration for a combined

87

Sr

/86

Sr-facies-

log based relative age framework within the

Santos Basin, offshore Brazil. We further

utilize

δ

18

O constraints on samples that yield

depositional clumped isotope apparent tem

-

peratures to constrain the

δ

18

O of the water

in these ancient lakes to between 1.9 and

4.9‰

Vienna standard mean ocean water

. Such heavy

values reveal a picture of a hot and arid envi

-

ronment. This is consistent with prior bio

-

stratigraphic studies of the carbonates that

show a decrease in faunal diversity in these

lakes prior to marine ingress and the devel

-

opment of open marine conditions in the

South Atlantic Ocean.

INTRODUCTION

The Early Cretaceous breakup of Gondwana

created basins offshore Brazil with conjugates

in West Africa (

Fig. 1

). A key feature of these

basins is the presence of one of the most region

-

ally expansive salt sequences in the geologic

record, occurring in the final stages of conti

-

nental break-up, and extending

∼

2200

km from

the Sergipe Basin in the north to the Santos

Basin in the south (Szatmari et al., 2021). Until

recently, these evaporites masked the full extent

of the pre-salt geologic archive and the important

regional temporal and environmental conditions

it records of the terminal stages of continental

break-up and early seafloor spreading. Follow

-

ing improvements in seismic imaging over the

past decade and calibration from exploration

drilling, a picture is emerging of a vast lake sys

-

tem that produced extensive lacustrine carbon

-

ates prior to the development of open marine

conditions in the South Atlantic Ocean. Indeed,

the Barra Velha Formation (Fm) carbonate reser

-

voirs of the Santos Basin form part of the largest

lacustrine carbonate system discovered in the

Phanerozoic record (Wright, 2022).

One of the main challenges in developing a

robust understanding of these basins is the total

absence of reliable time markers to place any

environmental constraints into temporal con

-

text. This has led to significant uncertainty on

the age of key sequences within the stratigraphy.

The Ariri Fm that represents the Santos Basin

equivalent of the regional evaporite sequence,

for example, has been proposed to be as young

as 111

Ma based on an

Ar-Ar age of sylvinite

of 110.64

±

0.3

Ma

from the Sergipe Basin

(Szatmari et al., 2021). Further, evaporites

unconformably overly volcanic rocks dated at

113.2

±

0.1

Ma in the northern Pelotas Basin

(Davison, 2007; Dias et al., 1994). However,

biostratigraphic data from the Deep Sea Drill

-

ing Project Site 364 (offshore Angola) suggest

that the first marine deposition above these

evaporites occurred during the late Aptian prior

to 113

Ma (K

ochhann et al., 2013). This has

been further supported by the study of Sanjinés

et al. (2022), who also report upper Aptian aged

fauna in the stratigraphy immediately overly

-

ing the evaporites in numerous wells within the

M. Lawson

https://orcid.org/0000-0003-

2743-3773

†

Previous address: ExxonMobil, Springwoods

Village Parkway, Spring, Texas 77027, USA;

miclawson@gmail.com.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/doi/10.1130/B36378.1/5627154/b36378.pdf

by California Inst of Technology user

on 14 June 2022

Lawson et al.

2

Geological Society of America Bulletin

,

v. 130, no. XX/XX

Santos, Campos, and Espírito Santo basins off

-

shore Brazil. Alternatively, these evaporites have

been proposed to be as old as 120–125

Ma and

associated with the Aptian Ocean Anoxic Event

(OAE) 1a (Tedeschi et al., 2017). In addition to

the challenges associated with absolute age cali

-

bration, extraction of important environmental

conditions is hampered by the fact that the differ

-

entiation of the Barra Velha Fm and underlying

Itapema Fm is largely based on lithostratigraphic

correlation and log responses. However, given

the time transgressive nature of deposition from

the proximal to distal regions of the evolving

Santos Basin during this period, there is poten

-

tial for significant lithostratigraphic variability

that limits the robustness of this approach alone.

The aim of this study is to investigate the

potential for

87

Sr/

86

Sr isotopes to provide a

robust relative age framework for the Early

Cretaceous lacustrine carbonates of the South

Atlantic, and to calibrate this with absolute age

constraints of the uppermost sequence. Addi

-

tional information provided through the integra

-

tion of carbonate clumped isotope geochemistry

allows us to identify samples that have been opti

-

mally preserved for age dating, and develop the

first comprehensive picture of both the timing

and paleo-environmental conditions associated

with this remarkable epi-continental carbonate

factory. While recent studies have attempted to

investigate environmental conditions from anal

-

ysis of these carbonates, they have all relied on

characterization from one well and extrapolated

to the entire pre-salt system. Furthermore, none

of these studies have been able to move beyond

a relative age framework for the stratigraphy

on both the Brazilian and the conjugate West

African margin with any degree of confidence.

Here, we provide characterization of strata from

A

C

B

Figure 1. (A) Well locations in the Santos Basin, offshore Brazil, relative to the full extent of the Brazil salt basins. (B) Reconstructed base

-

ment at 115

Ma for the South Atlantic based on the GTS 2020 timescale. (C) Schematic cross section showing correlation of the pre-salt

stratigraphy between different isolated carbonate platforms penetrated by wells 1 and 3. Fm.—Formation.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/doi/10.1130/B36378.1/5627154/b36378.pdf

by California Inst of Technology user

on 14 June 2022

Brazilian Pre-Salt Carbonate Evolution

Geological Society of America Bulletin

,

v. 130, no. XX/XX

3

four wells, each from separate isolated carbonate

platforms located 15–80

km apart that together

span an area of

∼

6400 km

2

in the Santos Basin

(

Fig. 1

). We focus our characterization predomi

-

nantly on the upper section of the pre-salt stra

-

tigraphy because of the 15 million year uncer

-

tainty on the timing of the regionally important

salt deposition. Any age constraint would also

provide the first robust upper bracket for the pre-

salt stratigraphy which subsequent studies can

further refine through additional constraints on

key geologic boundaries deeper within the strati

-

graphic column.

GEOLOGIC SETTING

Situated offshore central Brazil, the Santos

Basin covers an area of

∼

350,000 km

2

(Ariza

Ferreira et al., 2019), and developed as a result

of the extension and breakup of Gondwanaland

in the Early Cretaceous ca. 135–140

Ma (Karner

et al., 2021; Rangel et al., 1994). Like all the

South Atlantic salt basins, the pre-salt of the

Santos Basin is dominantly lacustrine in origin,

with the Florianopolis High thought to provide

a palaeogeographic barrier against marine intru

-

sions from the South Atlantic Ocean. The initial

sediments filling the basin are volcanics and

volcaniclastics of the approximately Hauteriv

-

ian aged Camboriú Fm underlying alluvial fans

and lacustrine shales of the Piçarras Fm that is

thought to be Barremian in age (Moreira et al.,

2007). This was followed by the Itapema Fm,

which comprises lacustrine shales and coquina-

dominated carbonate reservoirs that represent

the first significant deposition of carbonate in

the pre-salt stratigraphy. The penultimate phase

of deposition comprises shales and carbonates

of the Barra Velha Fm. Finally, marine ingress

drove the deposition of evaporites of the Ariri

Fm (Moreira et al., 2007; Rodriguez et al.,

2018). Basin to sub-basin scale volcanic events

punctuate the entire pre-salt record. Despite this,

there is a paucity of detail in the literature on

the nature and age of these events. As a result,

there is considerable uncertainty surrounding

the absolute timing of Itapema through Ariri

deposition, which could range from Barremian

to Aptian-Albian in age (Moreira et al., 2007;

Szatmari et al., 2021).

While absolute age correlation is challenged

given the non-marine setting and the lack of

direct age constraints on the carbonate intervals

within the upper section of the stratigraphy, bio

-

stratigraphic constraints have been shown to be

able to subdivide the lacustrine record on both

sides of the Atlantic (Bate, 1999). The Itapema

Fm, for example, is characterized by the ostracod

genera

Hourcquia

and some species of

Petrobra

-

sia

, while the Barra Velha Fm is characterized

by a lower diversity of ostracods and dominated

by species from the genera

Kroemmelbeincpris

and

Pattersoncpris

(Poropat and Colin, 2012a,

2012b). However, discrepancies within the

ostracod zones between the Brazilian and the

West African basins, coupled with the absence

of absolute age calibration, creates challenges

in using biostratigraphy to extrapolate globally

beyond the Brazilian salt basins.

METHODS

Four wells made available by the Brazilian

National Agency of Petroleum, Natural Gas

and Biofuels were sampled, each from a dif

-

ferent carbonate platform in the Santos Basin

(

Fig. 1

). The wells were drilled in water depths

of

∼

2000

m, and encountered the pre-salt car

-

bonates at depths in excess of 5000

m total

measured depth (3000

m subsea, see

Table 1

).

Twenty samples were collected per well across

the Barra Velha Fm and the Itapema Fm, where

present, yielding a total of eighty samples for

this study. Rock samples were obtained from

conventional core plugs or from sidewall cores.

Samples were selected for isotopic analysis fol

-

lowing petrographic characterization to identify

well preserved depositional fabrics and qualita

-

tively assess the extent to which samples have

experienced any diagenetic overprint. Thin sec

-

tions of each sample were characterized using

optical and scanning electron (JOEL 7600 with

an Oxford Aztec energy dispersive spectros

-

copy (EDS) system) microscopy at ExxonMo

-

bil Upstream Integrated Solutions Company.

Further, point counts were performed follow

-

ing the Gazzi-Dickinson method (Dickinson,

1985; Zuffa, 1980) to quantify the detrital and

diagenetic mineralogy in ten samples (n

=

300

points per thin section). These analysis captured

the different lithologies and resulted in a para

-

genesis defined for each sample that comprised

depositional and diagenetic fabrics based on

observed superposition and cross-cutting rela

-

tions between the various diagenetic phases.

87

Sr/

86

Sr analysis was performed at Rutgers

University (New Jersey, USA) for most samples,

with a subset analyzed by Stony Brook Univer

-

sity (New York, USA) to complete the data set.

U-Pb age dating was performed at Stony Brook

University applying the procedures described in

Parrish et al. (2019). Carbonate clumped isotope

analysis was performed by Isotomics Ltd, yield

-

ing

δ

13

C,

δ

18

O, and

Δ

47

for each sample.

Strontium Isotope Analysis

Sample preparation and procedures fol

-

low that described by Gamboa et al. (2019).

Isotopes were run on a ThermoScientific Nep

-

tune Plus multicollector–inductively coupled

plasma–mass spectrometer (MC-ICP-MS)

at Rutgers University. Measurement of NBS

987 was included with all analyses and aver

-

aged 0.710245

±

0.000006 (2

σ

, n

=

40) by

thermal ionization mass spectrometry (TIMS)

or 0.710270

±

0.000004 (2

σ

, n

=

13) by MC-

ICP-MS. Analyses by MC-ICP-MS were then

corrected based on those values of NBS 987

values to match those of the TIMS instrument.

The Sr was separated from the matrix in the rock

samples by ion chromatography using 80 uL of

Sr-spec resin (Eichrom) using 2

M nitric acid to

remove major elements, 7

M to remove Ba in the

rock samples, and recover Sr in the water. A sub

-

set of

87

Sr/

86

Sr analyses were run on the samples

used for U-Pb dating at Stony Brook University.

Sr was separated using Sr-spec resin, washing

with 2

M nitric acid and eluting with water. No

correction was made since the TIMS analyses

for NBS 987 averaged 0.710248

±

0.000011

(2

σ

external, n

=

12) which is identical to the

recommended value (McArthur et al., 2001).

U-Pb Analysis

Core samples were slabbed and polished for

laser ablation–inductively coupled plasma–mass

spectrometry (LA-ICP-MS). A New Wave 213

laser system was coupled to an Agilent 7500cx

for the LA-ICP-MS analyses using settings that

are shown in Supplemental Table S1

1

. The NIST

612 soda glass standard was used for element

analyses as well as for Pb isotope fractionation

correction. The WC-1 calcite reference mate

-

rial (Roberts et al., 2017) was used for the U/

Pb fractionation correction. Small element maps

were made across fabrics to examine U/Pb in the

context of mineralogy using major elements (Ca,

Mg, and Si). Based on the element maps, areas

with favorable U/Pb ratios were analyzed by spot

LA-ICP-MS analyses using a 120

μ

m spot size.

Spots on NIST 612, WC-1, and an internal sec

-

ondary standard, Barstow, were analyzed at the

beginning, between every 10 spots, and at the

end of the analytical sessions. U-Pb isotope dilu

-

tion (ID) analyses using a

205

Pb-

236

U spike was

performed on a set of four samples with favor

-

able LA-ICP-MS U-Pb results. LA-ICP-MS data

were reduced in Iolite (Paton et al., 2011) and

ID data were reduced in PbDat (Ludwig, 1993).

All of the U-Pb data (LA-ICP-MS and ID) were

plotted on Tera Wasserberg isochrons using Iso

-

Excel (Ludwig, 2003) or IsoplotR (

Vermeesch,

1

Supplemental Material. Methods, models, and

all data and supporting figures. Please visit

https://

doi .org /10 .1130 /GSAB .S.19699822

to access the

supplemental material, and contact editing@

geosociety.org with any questions.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/doi/10.1130/B36378.1/5627154/b36378.pdf

by California Inst of Technology user

on 14 June 2022

Lawson et al.

4

Geological Society of America Bulletin

,

v. 130, no. XX/XX

2018). All ages included and discussed in this

study are isochron ages reported at the two stan

-

dard deviation confidence level.

Carbon and Oxygen (

δ

13

C and

δ

18

O)

Isotope Analyses

Forty powdered carbonate samples span

-

ning the Barra Velha Fm and, where present,

the Itapema Fm were analyzed at the California

Institute of Technology in spring 2019. Samples

were examined by binocular microscope to make

an initial estimate of the proportions of carbonate

and contaminant phases. Isotopic analyses (

δ

13

C,

δ

18

O, and

Δ

47

) were then performed on a custom-

made automatic vacuum line attached directly to

a Thermo MAT 253 mass spectrometer (Passey

et al., 2010). All analyses were performed by

phosphoric acid digestion using an automated

common acid bath held at 90

°

C. Samples were

loaded into silver capsules, which were placed

into an automated multi-sample chamber; a step

-

per motor dropped them individually into the

heated acid. Evolved CO

2

was passed through

a glass trap held in a dry-ice/ethanol slush to

remove co-evolved H

2

O and other condens

-

able contaminants, and collected in a glass trap

immersed in liquid nitrogen, after which CO

2

was entrained in helium and passed through a

chilled, packed column of Porapak resin, held

at 253

K. Concurrent extractions using the same

equipment and methods on two intra-laboratory

calcite standards (“CIT Carrera” and “TV04”)

were used to test consistency of the data set with

long-term averages for these standards.

Each aliquot of CO

2

was analyzed on a Thermo

Fisher 253 isotope ratio mass spectrometer

for the ratios, (

13

C

16

O

2

+

12

C

17

O

16

O)/

12

C

16

O

2

,

(

12

C

18

O

16

O

+

13

C

17

O

16

O

+

12

C

17

O

2

)/

12

C

16

O

2

, (

13

C

18

O

16

O

+

12

C

18

O

17

O

+

13

C

17

O

2

)/

12

C

16

O

2

and (

12

C

18

O

2

+

13

C

18

O

17

O)/

12

C

16

O

2

,

and were subjected to standard ion correc

-

tion schemes to

δ

13

C

Vienna Pee Dee belemnite (VPDB)

,

δ

18

O

Vienna standard mean ocean water (VSMOW)

of CO

2

and

the

Δ

47

value of CO

2

(here reported relative to

the “absolute reference frame”; Dennis et al.,

2011). Samples were evaluated for their

Δ

48

values to assess the purity of extracted, cleaned

gases. All values are reported with

±

1 standard

error, based on the reproducibility of several

periods of mass spectrometric integration that

contributed to the reported values;

Δ

47

errors

also include propagated errors in the reference

frame of heated and equilibrated gases.

We calculated the

δ

18

O

VPDB

value of reactant

based on the measured value of evolved CO

2

using an acid digestion fractionation factor

TABLE 1. S

r

CONCENTRATION AND

87

S

r

/

86

S

r

DATA ALONGSIDE THE STABLE CARBON AND OXYGEN ISOTOPE DATA

ACQUIRED FOR THE PRE-SALT CARBONATES OF THE SANTOS BASIN, OFFSHORE BRAZIL

Field

Depth

(m)

Depth from

top BV

(m)

Sr

(ppm)

±

1SE

87

Sr/

86

Sr

±

1SE

δ

13

C

(PDB)

±

1SE

δ

18

O

(VPDB)

±

1SE

Δ

47

±

1SE

Temperature

( ̊C)

Minus

( ̊C)

Plus

( ̊C)

1

5115.6

16.0

308.1

4.0

0.7135

0.00000

1.22

0.003

0.15

0.006

0.5731

0.0141

72

7

1

5117.2

1 7. 6

1227.7

19.0

0.7135

0.00000

3.52

0.003

0.30

0.008

0.6596

0.0210

36

7

8

1

5120.5

20.9

2836.9

36.2

0.7136

0.00000

nd

1

5121.1

21.5

2359.6

21.2

0.7135

0.00000

3.07

0.003

–0.40

0.008

0.5588

0.0194

79

10

1

5130.6

31.0

1870.2

24.4

0.7139

0.00001

2.84

0.004

2.97

0.006

0.6065

0.0135

57

6

1

5136.9

37.3

1984.9

24.1

0.7141

0.00000

4.04

0.004

0.35

0.006

0.5976

0.0129

60

6

1

5140.4

40.8

2402.2

32.6

0.7135

0.00000

nd

1

5192.6

93.0

2819.0

42.7

0.7134

0.00000

3.31

0.002

2.44

0.005

0.6041

0.0219

58

9

10

1

5241

141.4

1609.2

1 7. 0

0.7137

0.00000

3.26

0.015

1.70

0.004

0.6138

0.0166

53

7

1

5289

189.4

2000.2

28.6

0.7135

0.00001

2.55

0.003

1.66

0.012

0.5707

0.0204

73

10

1

5385

285.4

2690.9

22.0

0.7132

0.00000

nd

2

5325.5

9.6

3652.4

27.1

0.7133

0.00000

nd

2

5338

22.1

1520.9

1 7. 2

0.7132

0.00000

1.02

0.004

–0.62

0.007

0.5749

0.0140

71

7

2

5345.5

29.6

3330.1

32.3

0.7136

0.00000

2.81

0.002

0.42

0.005

0.5720

0.0125

72

6

2

5363

47.1

2475.6

30.7

0.7133

0.00000

0.62

0.004

–0.84

0.007

0.5733

0.0182

72

9

2

5370.5

54.6

2690.0

27.9

0.7136

0.00001

2.23

0.006

–0.82

0.009

0.5984

0.0143

60

6

2

5373

57.1

2701.2

46.9

0.7132

0.00001

1.53

0.002

2.41

0.005

0.5843

0.0161

66

7

8

2

5383

67.1

2390.6

29.3

0.7131

0.00000

2.08

0.003

0.52

0.005

0.5933

0.0140

62

6

2

5480.5

164.6

2483.2

39.8

0.7133

0.00000

0.66

0.005

–2.39

0.006

0.5492

0.0125

84

7

2

5495.5

179.6

2505.9

27.5

0.7136

0.00000

–0.29

0.007

–2.87

0.007

0.5786

0.0125

69

6

2

5495.5

179.6

2505.9

27.5

0.7136

0.00000

–0.27

0.003

–2.59

0.003

0.5817

0.0128

68

6

2

5591

275.1

1583.6

9.2

0.7115

0.00001

nd

2

5722

406.1

794.9

7. 3

0.7112

0.00000

nd

2

5740

424.1

508.9

5.3

0.7118

0.00000

–1.13

0.005

–6.27

0.009

0.5867

0.0202

65

9

10

2

5740

424.1

508.9

5.3

0.7118

0.00000

–1.07

0.002

–6.15

0.004

0.5944

0.0136

62

6

2

5745

429.1

1063.2

12.5

0.7111

0.00000

1. 1 6

0.003

0.50

0.005

0.5909

0.0143

63

6

7

2

5753

437.1

1984.1

1 7. 4

0.7113

0.00000

1. 1 5

0.011

–0.44

0.005

0.5744

0.0151

71

7

8

3

5406

13.6

1170.6

9.6

0.7132

0.00001

1.70

0.013

–0.14

0.004

0.5753

0.0136

71

6

7

3

5420

27.6

nd

0.7134

0.00001

1.84

0.002

1.08

0.005

0.6348

0.0123

45

5

3

5500

107.6

1589.2

19.9

0.7128

0.00000

1.49

0.004

3.59

0.008

0.6014

0.0145

59

6

3

5549.3

156.9

2607.5

16.8

0.7129

0.00000

2.23

0.004

0.76

0.007

0.6152

0.0191

53

8

3

5550.4

158.0

1047.6

12.4

0.7130

0.00000

1.21

0.003

–3.07

0.006

0.6532

0.0131

38

5

3

5562.8

170.4

156.9

1. 0

0.7121

0.00000

–1.62

0.004

0.91

0.009

0.6311

0.0112

47

4

3

5563.5

171.1

45.2

0.7

0.7125

0.00001

nd

3

5564.3

171.9

1510.7

16.2

0.7125

0.00000

nd

3

5565.4

173.0

1662.3

18.8

0.7119

0.00000

nd

3

5581

188.6

1244.6

9.1

0.7122

0.00000

2.38

0.003

1.82

0.006

0.5306

0.0138

94

8

3

5581

188.6

1244.6

9.1

0.7122

0.00000

2.70

0.002

2.46

0.004

0.5366

0.0141

91

8

3

5619

226.6

984.5

9.5

0.7121

0.00000

1.42

0.004

–0.49

0.006

0.5955

0.0137

61

6

3

5708

315.6

967.5

11. 2

0.7121

0.00000

nd

3

5729.5

337.1

2067.5

15.1

0.7118

0.00001

1. 7 6

0.003

0.70

0.004

0.5620

0.0133

77

7

4

5336.5

4.9

678.0

5.0

0.7128

0.00000

1.51

0.002

0.19

0.005

0.5762

0.0150

70

7

4

5367.7

36.1

1021.1

10.4

0.7124

0.00000

1.98

0.003

0.42

0.005

0.6119

0.0103

54

4

5370.5

38.9

1256.2

1 7. 0

0.7113

0.00000

nd

4

5381.7

50.1

1875.2

20.6

0.7126

0.00000

1.78

0.004

0.94

0.008

0.6002

0.0147

59

6

7

4

5387.6

56.0

2300.0

21.5

0.7118

0.00001

nd

4

5444.4

112.8

1878.4

20.8

0.7122

0.00001

nd

4

5476.9

145.3

1869.9

20.8

0.7122

0.00001

1. 1 7

0.003

–0.43

0.006

0.5644

0.0126

76

6

7

4

5509.6

178.0

2767.4

29.1

0.7120

0.00000

1.86

0.001

0.23

0.005

0.6008

0.0182

59

8

4

5514.6

183.0

2690.1

34.4

0.7121

0.00001

1.36

0.004

–0.20

0.008

0.5725

0.0134

72

6

7

4

5516.7

185.1

911.3

11. 4

0.7114

0.00000

2.33

0.004

1.82

0.008

0.5927

0.0143

63

6

7

4

5546.3

214.7

2322.1

21.8

0.7111

0.00000

nd

4

5570.9

239.3

1217.3

11. 2

0.7113

0.00000

2.67

0.001

1.48

0.006

0.6327

0.0130

46

5

Notes:

BV—Barra Velha Fm.; SE—standard error; PDB—Pee Dee belemnite; VPDB—Vienna Pee Dee belemnite; nd—no data.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/doi/10.1130/B36378.1/5627154/b36378.pdf

by California Inst of Technology user

on 14 June 2022

Brazilian Pre-Salt Carbonate Evolution

Geological Society of America Bulletin

,

v. 130, no. XX/XX

5

appropriate for calcite at our reaction tempera

-

ture of 90

°

C. We report apparent temperatures

corresponding to the corrected

Δ

47

values using

the experimental temperature calibration of

Bonifacie et al. (2017). We analyzed 26 carbon

-

ate reference standards over the course of this

study; results of these measurements agree with

long-term accepted average values, and there

-

fore, no corrections were made to this data set

to account for systematic methodological errors.

Similarly, 27 gases equilibrated at known tem

-

peratures were analyzed to establish the abso

-

lute reference frame used for correction for mass

spectrometric artifacts. Standard reproducibility

of a Carrara marble was 2.091‰

±

0.061‰,

−

2.002‰

±

0.091‰, and 0.399‰

±

0.011‰

for measured

δ

18

O,

δ

13

C, and

Δ

values, respec

-

tively, based on one standard deviation error, as

compared to the long-term

Δ

47

average for this

standard of 0.352‰ (in the Ghosh et al., 2006

or “Caltech” intralaboratory reference frame).

Errors in apparent temperature for each sample

are asymmetrically distributed about the reported

value because apparent temperature is a nonlin

-

ear function of clumped isotope composition.

RESULTS

Petrographic Observations

The carbonate samples obtained for this

study are heterogeneous, with a great diversity

of lithologies observed in the four carbonate

platforms. The primary fabrics include shrub

boundstones (or boundstone dominated fabric,

see

Figs. 2A

and

2B

), spherulite-bearing mud

-

stones (mud-dominated facies,

Figs. 2C

and

2D

), and grainstone dominated fabrics (

Figs. 2E

and

2F

) with a significant range of diagenetic

overprint (none to completely overprinted, see

Fig. 3

). Abundant Mg-rich clays occur in both

the shrub boundstone and mudstone lithologies

and are typically limited within the grainstone

dominated lithologies.

Primary fabrics: Shrub boundstone litholo

-

gies (

Figs. 2A

and

2B

) typically display calcic

shrubs that are sectored and can bear small circu

-

lar voids (100–500

μ

m) and occasional sinuous

to straight, parallel-sided channels. The shrubs

have a fascicular optical habit and are occasion

-

ally recrystallized to bladed and or equant sparry

calcite (500–1000

μ

m) (

Fig. 2B

). A vertically

stacked, repeated sequence occurs within some

thin sections of shrub boundstone laminations of

calcite dominated precipitation, dolomite domi

-

nated precipitation, and finally Mg-smectite

dominated precipitation.

The mudstone-dominated fabrics typically

contain spherulites that occur as spherical forms

(

Fig. 2C

) and can have a radial-fibrous optical

habit in clay-rich lamination where diagenetic

overprints are limited (

Fig. 2D

). Spherulites

typically have a more elongated, asymmetrical

form in lamina with lower clay volumes. These

asymmetrical forms are similar in shape to

shrubs and can be difficult to distinguish in sam

-

ples that have been significantly diagenetically

altered (e.g.,

Fig. 2E

). The clay groundmass in

the spherulitic mudstones commonly has fabrics

consistent with post-depositional physical com

-

paction (e.g., stylolites). If the spherulite frac

-

tion is great enough, the spherulitic mudstones

become spherulite-supported packstones.

The grainstone-dominated fabrics are largely

formed from shrub and spherulite allochems

(

Figs. 2E

and

2F

), but occasionally contain chert

fragments, mud intraclasts and/or rare volcanic

lithic fragments. The grainstone lithologies pro

-

vide evidence for current transport and sedimen

-

tation of allochems derived from shrub bound

-

stone and spherulitic mudstone environments of

deposition. Transport appears to effectively win

-

now the clay-size fraction from the mobilized

sediments.

Porosity: Average helium porosity in the stud

-

ied samples range from 9.4% to 15% depend

-

ing on lithology: shrub boundstones 15

±

3.3%,

mudstone dominated 9.4

±

4.0%, and grain

-

stone dominated 14

±

4.4% (mean

±

2

σ

).

Shrub boundstones and grainstones typically

have macro-porosity occurring between grains

or shrubs that can be enhanced by secondary dis

-

solution. Mudstones and clay-rich shrub bound

-

stones have significant micro-porosity. The pres

-

ence of corroded inter-shrub clay particles and

jagged margins along the calcite shrubs suggest

that dissolution of mud and calcite has generated

secondary porosity.

Diagenesis: The first diagenetic process we

observe is the dissolution of primary fabrics, in

particular Mg-smectite and to a lesser degree

dissolution of the carbonate components. Sig

-

nificant vuggy and moldic porosity is present in

samples where this has occurred (

Fig. 3A

). The

first diagenetic products we observe are one of

or a combination of dolomite, calcite or, in some

cases, silica cements filling secondary porosity.

Dolomite is microcrystalline to rhombic and fills

secondary porosity that develops as a result of

dissolution of Mg-clay (

Figs. 3B

–

3D

). The ear

-

liest form of diagenetic calcite we observe is a

continuous isopachous calcite cement that coats

individual grains in some grainstones, and in rare

cases, may cement multiple grains together to

form small grapestone clasts. In many bound

-

stone and mudstone dominated fabrics, we

observe recrystallization of shrub boundstones

and spherulites (

Figs. 3A

–

3C

). Finally, silica

cement is also observed to immediately follow or

be concomitant with dissolution of Mg-clays or

carbonates in some samples. The earliest phase

of silica cementation is in the form of micro

-

crystalline silica (i.e., chert), which precipitates

in secondary porosity in the same way as the

microcrystalline dolomite and isopachous cal

-

cite cements above. This is followed by fibrous

chalcedony that can be repetitively zoned, and is

observed to rim secondary porosity that has not

been replaced by the initial dolomite, calcite, or

silica cements. In some cases, silicification has

completely destroyed the primary fabric (e.g.,

Figs. 3E

and

3F

). While it is difficult to defini

-

tively reconcile the relative timing of these spe

-

cific dolomite, calcite, and silica cements, they

clearly occur prior to other crystal habits of the

same minerals and as such represent the earliest

forms of preserved diagenetic products.

In some samples, we observe later diagenetic

phases of dolomite, calcite, and silica, and addi

-

tional distinctive minerals. The latest diagenetic

dolomite is in the form of a pore filling and

fracture lining saddle dolomite. This form of

dolomite displays a characteristic larger crystal

size and curved crystal habit that allows it to

be differentiated from the microcrystalline and

rhombic dolomite described above. Similarly,

we observe a transition to equant sparry calcite

replacing earlier forms of calcite. This calcite

is translucent in plain light and at times occurs

after dolomitization. Finally, in some samples

we observe a transition to mega quartz with a

larger crystal size compared to the earlier micro

-

crystalline and fibrous chalcedony (

Fig. 3F

). In

core, mega quartz is often found within vugs that

are lined with chalcedony and terminate in open

pore space, and in thin section it is observed to

fill remaining pore space (i.e.,

Figs. 2B

and

3F

).

In addition to the transition in the character of

these diagenetic products, these later phases

are occasionally associated with metal sulfides

(i.e., pyrite/maracasite and sphalerite), fluorite,

an aluminum-phosphate-sulfate (APS) mineral

that we found to be compositionally consistent

with goyazite, barite/celestite, and dawsonite

(

Figs. 3G

–

3I

, composition confirmed by EDS).

δ

13

C,

δ

18

O,

Δ

47

, and

87

Sr/

86

Sr Geochemistry

The

δ

13

C and

δ

18

O (VPDB) values of the car

-

bonate samples display significant variability. In

general, we observe a range in

δ

13

C from

−

1.6

to

+

4‰

VPDB

and a range in

δ

18

O from

−

6.3 to

+

3.6‰

VPDB

(n

=

39). The

Δ

47

values for the

entire data set range from 0.537 to 0.660‰, with

a mean of 0.592‰. These values correspond to

a range in apparent temperature from 36–91

°

C,

with a mean apparent temperature of 63

°

C. The

apparent temperatures display no systematic cor

-

relation with

δ

13

C and

δ

18

O (see Supplemental

Figs. S32 and S33; see footnote 1). The

87

Sr/

86

Sr

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/doi/10.1130/B36378.1/5627154/b36378.pdf

by California Inst of Technology user

on 14 June 2022

Lawson et al.

6

Geological Society of America Bulletin

,

v. 130, no. XX/XX

values of samples from all four wells in this

study range from 0.7111 to 0.7141. All isotopic

data is presented in

Table 1

.

U-Pb Dating

A subset of 19 samples, selected based on

their stratigraphic positions and diagenetic histo

-

ries, were analyzed by LA-ICP-MS to determine

if they had favorable U/Pb ratios. Multiple maps

were made across samples with complex fabrics

to obtain information on all phases represented

by the sample. Based on this initial screening, 15

samples were selected for more comprehensive

LA-ICP-MS spot analyses in an effort to define

isochron ages for the samples. This comprised

of 11 samples from the Barra Velha Fm and four

samples from the Itapema Fm based on the chro

-

nostratigraphic framework defined by

87

Sr/

86

Sr,

facies and log correlations described above. Of

the samples analyzed, a total of nine samples

from the Barra Velha Fm and two samples from

the Itapema Fm yielded ages (see Supplemental

Material for more details). This analysis pro

-

duced seven isochron ages, all from samples

from the Barra Velha Fm, with a range from

114

±

4.72 to 58.85

±

10.74 Ma (Supplemental

Table S3). Based on the spread of U/Pb and the

mean square weighted deviation of the data, only

three of these samples were selected for parallel

additional analyses via isotope dilution thermal

ionization mass spectrometry (ID TIMS).

Most of the samples analyzed by LA-ICP-MS

are easily interpreted as non-depositional. How

-

ever, a sample from well 1, taken from a depth

of 37.3

m below the Ariri Fm, gives an isochron

age from LA-ICP-MS of 114.46

±

4.72 Ma.

Similarly, a sample from well 3 taken from a

depth of 27.6

m from the base of the Ariri Fm

gives a LA age of 109.73

±

9.26 Ma. Of the

samples selected for ID TIMS, only one sample

yielded a robust isochron age. This was a sample

taken from well 3 at 27.6

m below the base of

the Ariri Fm, and yielded an isotope dilution

(ID) isochron age of 115.83

±

1.56 Ma (

Fig. 4

).

This is in good agreement with the isochron age

obtained via LA-ICP-MS from the same sample

of 109.73

±

9.26 Ma described above.

Figure 2. Dominant primary

fabrics of pre-salt carbonate

analyzed in this study from

Santos basin, offshore Brazil:

boundstone dominated fabric

(A and B), spherulite-bearing

mudstone (C and D), and grain

-

stone dominated fabric (E and

F). (A) Weakly stylolitized, very

coarse shrub boundstone with

primary fasicular optical habit

and moderate vuggy porosity

in plain light. (B) Very coarse

shrub boundstone and medium

spherulite-bearing mudstone

with pore filling mega quartz

(cross polar). (C) Spherulite-

bearing mudstone in plain light

and cross polar (D) showing the

spherical form and radial fi

-

brous nature of the spherulites.

(E) Coarse to gravelly, shrub

and spherulite grainstone with

extensive interparticle porosity

in plain light and (F) cross po

-

lar. (Mg—Mg-smectite; Dol—

dolomite; S—calcite shrub;

sphr—calcite spherulite; Si—

as chert, chalcedony, and/or

mega-quartz). All thin sections

are stained with alizarin red so

calcite is presented as pink in

the images.

AB

CD

EF

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/doi/10.1130/B36378.1/5627154/b36378.pdf

by California Inst of Technology user

on 14 June 2022

Brazilian Pre-Salt Carbonate Evolution

Geological Society of America Bulletin

,

v. 130, no. XX/XX

7

DISCUSSION

Petrographic Assessment of Primary

Depositional Fabrics and Insights on

Diagenetic Overprints

The samples of the Barra Velha Fm studied

here display great variability in their composi

-

tion, with facies that include shrub boundstones,

re-worked shrub (dendritic) grainstones, spheru

-

litic packstones, and fine-grained carbonate mud

-

stones. These observations are consistent with

previous studies that focused on this interval (e.g.,

Gomes et al., 2020; Wright, 2022; Wright and

Barnett, 2020, 2015). Furthermore, these facies

are consistent with those described by Saller et al.

(2016) for the time equivalent stratigraphy of the

conjugate Kwanza Basin, offshore Angola.

The petrographic observations described

above suggest that the samples studied here have

experienced a broad range in both the extent and

the timing of diagenetic overprint. There is a

clear distinction in the samples from those that

experience relatively minimal alteration of dep

-

ositional fabrics to those that have experienced

almost pervasive replacement of primary fabrics

by late-stage diagenesis. Diagenesis begins with

an initial stage of dissolution, which is followed

by or is concomitant with the precipitation of

early diagenetic products that are dominated by

A

BC

DE

F

G

H

I

Figure 3. Range in diagenetic alteration of pre-salt carbonates from Santos basin, offshore Brazil. (A) Rhombic dolomite within Mg-

smectite and recrystallized calcite shrubs. (B) Dolomitized shrub boundstone with evidence of replaced spherulites and dolomitization. (C)

Extensively dolomitized fine to medium crystalline dolomite with vuggy to moldic to intercrystalline porosity. (D) Quartz and dolomitized

shrub boundstone with extensive dissolution leading to vuggy porosity. (E) Completely silicified dolostone with limited to no visible poros

-

ity with all three forms of silica: chert, chalcedony (chalc), and mega quartz (MQ) in plain light and (F) cross poles. (G) Scanning electron

microscope (SEM) backscatter image of sample containing fluorite (F) and an aluminum-phosphate-sulfate (APS) mineral compositionally

similar to goyazite (APS). Fluorite and APS are commonly found at the margins of pores and especially dissolution enhanced pores. (E)

designates epoxy-filled pore space. (H) SEM backscatter image of barite. (I) SEM backscatter image of sample containing dawsonite and

rhombic dolomite (Mg—Mg-smectite; Cc—calcite; Do—dolomite; sphr—spherulite; dis—dissolution; chalc—chalcedony; MQ—mega

quartz; B—barite; APS—goyazite; F—fluorite).

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/doi/10.1130/B36378.1/5627154/b36378.pdf

by California Inst of Technology user

on 14 June 2022

Lawson et al.

8

Geological Society of America Bulletin

,

v. 130, no. XX/XX

either dolomite, calcite, or silica cementation.

This is followed by the precipitation of saddle

dolomite, equant calcite, and/or mega quartz that

in some cases occur approximately concurrent

with minerals such as metal sulfides, dawsonite,

and APS that are diagnostic of high temperature

diagenesis or direct precipitation from hydro

-

thermal fluids (e.g., Dill, 2001). While all sam

-

ples were analyzed, samples that only display

limited diagenetic overprint (e.g., primary depo

-

sitional fabrics such as shrub boundstones and

early silica or dolomite diagenetic phases) are

considered favorable for age dating. In contrast,

samples that included evidence of pervasive

dolomitization, silicification, or APS minerals

are interpreted to have been impacted by burial

diagenesis or hydrothermal alteration and were

considered less favorable to address the objec

-

tives of this study.

Constraints on the Conditions during the

Terminal Stage of Lacustrine Carbonate

Deposition

The lowest temperatures we report here of

36

±

7

°

C and 38

±

5

°

C measured from sam

-

ples taken at depths of 17.6

m and 158

m from

the top of the Barra Velha Fm in wells 1 and 3,

both within the Barra Velha Fm, are consistent

with the average clumped isotope temperature

of carbonates reported for the time equivalent

Codó Fm (Bahniuk et al., 2015). We suggest that

these apparent temperatures record the original

physico-chemical conditions of the giant pre-salt

lake waters during deposition of these ancient

lacustrine carbonates. From these apparent tem

-

peratures and independent measurement of

δ

18

O

of the carbonates, we are able to calculate the

δ

18

O of the Santos Basin pre-salt lake water to

be between

+

1.9 and

+

4.9‰

(VSMOW)

(

Fig. 5

)

using the relationship of Kim and O’Neil (1997).

These values are similar to the fluid

δ

18

O of

+

6‰

(VSMOW)

estimated for the time-equivalent

conjugate pre-salt lakes from the Kwanza Basin

of Angola (Saller et al., 2016). Such

δ

18

O fluid

values are typical of modern lakes that have been

subject to evaporation in a relatively hot, semi-

arid to arid climate (Horton et al., 2016).

To quantify the extent of lake water evapo

-

ration needed to give rise to this range in

δ

18

O

(VSMOW)

values, we developed a box model

(see Supplemental Material for details) that

calculates lake water

δ

18

O at a nearly constant

lake level maintained by two river sources

relative to constant evaporation. Our modeling

suggests that such

δ

18

O values of lake water

require a range in the extent of evaporation of

40%–56% at a semi-arid humidity of

∼

50%.

Such evaporative conditions are consistent with

an increasingly stressed environment that is

also recognized by decreasing biodiversity in

the lake fauna records (Antonietto et al., 2012;

Bate, 1999). These authors concluded that the

ostracod fauna identified in the Barra Velha Fm

and time equivalent pre-salt stratigraphy in West

Africa display a decrease in ornamented and

robust shells. This impact on faunal morphology

was also associated with a clear decrease in the

diversity and abundance of the ostracod fauna

in the period leading up to the deposition of the

regional salt sequences. Our isotopic constraints

of paleo lake chemistry are therefore consistent

with observations in the ostracod record. Taken

together, these results suggest that the pre-salt

lakes became increasingly salinity stressed as a

result of the prevailing climatic

conditions

during

Figure 4. U-Pb isotope dilution (ID) isochron age obtained from a sample of the Barra Velha

Formation in well 3 in the Santos Basin, offshore Brazil. MSWD—mean square weighted

deviation; Incl.—including.

Figure 5.

δ

18

O

fluid

against ap

-

parent clumped isotope tem

-

perature for the Barra Velha

Fm and Itapema Fm carbon

-

ates analyzed in this study.

The two coldest temperatures

(

<

40

°

C) in the blue box that

we consider to reflect depo

-

sitional conditions provide a

range in lake water

δ

18

O

fluid

of

∼

2–5‰

Vienna mean ocean water(VSMOW)

.

Clumped isotope temperatures

above 40

°

C are interpreted to

reflect diagenetic alteration,

and as such, the calculated

δ

18

O

fluid

is not considered to be representative of true lake chemistry for those samples.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/doi/10.1130/B36378.1/5627154/b36378.pdf

by California Inst of Technology user

on 14 June 2022