rossi_2001

7/29/2019 Rossi_2001

1/20

Diagenetic and oil migration history of the Kimmeridgian AsclaFormation, Maestrat Basin, Spain

C. Rossia,*, R.H. Goldsteinb, R. Marla, R. Salasc, M.I. Benitod, A. Permanyerc,J.A. de la Penaa, M.A. Cajaa

aDepartamento de Petrologa y Geoqumica, Facultad de Ciencias Geologicas, Universidad Complutense, 28040 Madrid, Spain

bDepartment of Geology, University of Kansas, Lawrence, KS 66045-7613, USAcDepartamento de Geoqumica, Universidad de Barcelona, 08028 Barcelona, Spain

dDepartamento de Estratigrafa, Universidad Complutense, Madrid, Spain

Received 17 November 2000; received in revised form 14 January 2001; accepted 19 January 2001

Abstract

The marine limestones of the Kimmeridgian Ascla Formation in the Maestrat Basin reached more than 3500 m in burial depth during the

Cretaceous era. Despite containing organic-rich intervals, mature in parts of the basin, its potential as oil source-rock has been either

overlooked or questioned. A petrographic, geochemical and uid-inclusion (FI) study of the cements of the Ascla was performed in order to

unravel its diagenetic and thermal evolution. We particularly sought evidence of oil migration and its timing. Three sequences of cement

were distinguished. Sequence 1 lls the primary porosity and began with Fe-poor calcites with geochemistry and FIs consistent with

precipitation from marine-derived waters during shallow burial. These calcites were followed by burial cements, including ferroan calcite,

dolomite, and minor celestite and barite. Sequence 2 consists of Mg-rich, fracture-lling calcite cement zones. The earlier ones are ferroan

and contain primary aqueous and oil FIs with homogenization temperatures suggesting precipitation at temperatures as high as 1178C.

Sequence 3 is dominated by fracture-lling calcites with geochemistry and FIs indicating precipitation at low temperatures (less than,508C)

from meteoric waters. Cross-cutting relationships with compressional microstructures indicate that Sequence 3 formed after the Eocene

Oligocene tectonic inversion of the basin. Oil FIs in Sequence 2 provide evidence that light oils migrated through the Ascla Formation viafractures and microfractures. These oils were likely generated in the organic-rich marls of the basal part of the Ascla. The paragenetic

sequence and burial history are consistent with oil generation when the Ascla was at or close to maximum burial depth, but before the Eocene

Alpine tectonism, which likely formed the structural traps in the basin. Oil generation and migration occurred long before this event.

Therefore, it is probable that early traps were breached by the Alpine structures and that potential in this basin sector is low. q 2001 Elsevier

Science Ltd. All rights reserved.

Keywords: Burial diagenesis; Fluid inclusions; Oil migration

1. Introduction

The Kimmeridgian Ascla Formation was deposited in

open marine environments in the Salzedella sub-basin,which was the main depocenter of the Maestrat Basin (E

Spain) (Fig. 1). This formation contains intervals of lami-

nated marly limestones with enough organic matter to be

considered as potential oil source rocks (Permanyer, Marl,

de la Pena, Dorronsoro, & Rossi, 1999; Salas, 1989). The

source-rock potential of the Ascla was not recognized when

the Maestrat basin and adjacent offshore areas were

explored for hydrocarbons during the seventies. Onshore,

exploration was unsuccessful. Offshore, in the Tarragona

basin, several oil elds were discovered. Although most of

the oils discovered in the Tarragona basin undoubtedly

originated from Miocene source rocks (Albaiges, Algaba,Clavell, & Grimalt, 1986), the oil from Amposta eld,

which is located only 40 km to the west of La Salzedella,

has unique geochemical characteristics and is not correlated

with the other oils (Albaiges et al., op. cit.).

It has been postulated that the oil of the Amposta eld

could have been generated from the Ascla Formation

(Seifert, Carlson, & Moldowan, 1983). Albaiges et al.

(1986) discarded the Ascla as a potential source for the

Amposta oil, arguing that the Ascla (1) does not have

source-rock potential, and (2) it was overmature for oil

generation at the time of trap formation (the Miocene).

Marine and Petroleum Geology 18 (2001) 287306

0264-8172/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved.

PII: S0264-8172(01) 00008-3

www.elsevier.com/locate/marpetgeo

* Corresponding author. Fax:134-91544-2535.

E-mail address: [email protected] (C. Rossi).

7/29/2019 Rossi_2001

2/20

However, Seemann, Pumpin, and Casson (1990), found that

the Ascla Formation is still in the oil generation phase and

has some residual oil generation potential in a nearby

offshore well (Cenia-1). Moreover, they found somegeochemical similarities between source-rock extracts and

the Amposta crude. The Ascla formation is also still in the

oil generation phase in outcrops near la Salzedella

(Permanyer, Salas, & Marl, 2000a). Here, the formation

may have entered the oil window during the Late Cretac-

eous, when it was buried at depths from about 3000 to

3700 m (Permanyer et al., op. cit).

The purpose of the present study was to investigate the

diagenetic history of the Ascla Formation in outcrops of the

Salzedella sub-basin, with special emphasis on the possible

existence of indications of petroleum generation and migra-

tion. A combination of petrographic (transmitted light, cath-

odoluminescence, and mineral epi-uorescence) andgeochemical (stable isotopic and microprobe analyses) tech-

niques was used. This combination has provided detailed

constraints on the entire diagenetic sequence, with epi-uor-

escence uniquely able to show the relative timing of petro-

leum migration. Fluid inclusions (FIs), petrography, and

geochemical composition of diagenetic phases were used

to constrain thermal history and the evolution of pore-

uid composition.

Special emphasis was made on the distribution and prop-

erties of oil inclusions, as they may provide valuable infor-

mation about oil migration events and their timing relative

to the diagenetic, burial, tectonic, and thermal evolution of

the basin (e.g. Burruss, 1989; Karlsen, Nedkvitne, Larter, &

Bjrlykke, 1993; McLimans, 1991; Wilkinson, Lonergan,

Fairs, & Herrington, 1998). Oil inclusions may also revealrelict migration pathways (Mann, 1994) and oilwater

contacts (Lisk, Eadington, & O'Brien, 1998; Oxtoby,

Mitchell, & Gluyas, 1995), and conrm the effectiveness

of a potential source-rock as an oil generator. The use of

oil inclusions as geothermometers may be difcult, because

of the variable and commonly high differences between

their homogenization temperatures (Th) and their entrap-

ment temperatures. However, if the oil inclusions were

trapped as single homogeneous liquid phases in equilibrium

with a gas phase, their Th may reect true temperatures of

entrapment and thus represent excellent geothermometers

(Goldstein & Reynolds, 1994, p. 40).

2. Geological setting

The Maestrat Basin is located in the eastern part of the

Iberian rift system, that developed during the Mesozoic

opening of the western Tethys in the eastern part of the

Iberian Peninsula. The Maestrat Basin (Fig. 1) contains up

to 5.8 km of Mesozoic strata dominated by carbonates. It

experienced two major rifting stages (Salas & Casas, 1993;

Salas & Guimera, 1996; Salas, Guimera, Gimenez-

Montsant, Martn-Closas, & Roca, 1997; Salas, Guimera,

C. Rossi et al. / Marine and Petroleum Geology 18 (2001) 287306288

Fig. 1. Simplied map of the Maestrat Basin (from Salas et al., 1995).

7/29/2019 Rossi_2001

3/20

Mas, Martn-Closas, Melendez, & Alonso, 2001). The rst

stage occurred during the Triassic and resulted in high-angle

normal faults in the Palaeozoic basement. The second stage

occurred during the latest JurassicEarly Cretaceous in

association with the opening of the North Atlantic, and

created an extensional fault system which divided the

basin into several sub-basins. The Salzedella sub-basin,

which was the main depocenter of the Maestrat basin,

C. Rossi et al. / Marine and Petroleum Geology 18 (2001) 287306 289

Fig. 2. Restored cross-section of the Maestrat Basin during late Albian time, showing the Salzedella and nearby sub-basins (after Salas et al., 2001). See Fig. 1

for location.

Fig. 3. Post-Hercynian restored stratigraphy of the Salzedella area, showing the calculated magnitude of the post-Eocene erosion and the location of the

sampled outcrop section. The dashed lines represent the restored pre-Tertiary stratigraphy. The Alpine structure is not considered.

7/29/2019 Rossi_2001

4/20

represents a large, northwards-tilted block bounded by a

deep-seated fault (Fig. 2) (Salas & Guimera, 1996).

The late Kimmeridgian Ascla Formation was deposited in

the depocenter of the Salzedella sub-basin, recording the

onset of the second rifting stage (Salas et al., 2001).

The Ascla Formation is composed of calcareous mudstones

and marlstones, deposited in a relatively deep platform

setting (Salas, 1989), and thickens towards the centre of

the sub-basin, where it reaches 800 m in thickness (Fig. 2).

It is overlain by a thick succession (up to 3500 m in the

depocenters) of Tithonian-to-Albian synrift strata (Fig. 3)

dominated by platform carbonates and deltaic siliciclastics,

and up to 500 m of late Cretaceous, post-rift shallow-

marine carbonates (Salas, 1987). During the Eocene

Oligocene, the extensional faults were inverted as a

result of Alpine compressional tectonics, resulting in the

formation of a fold-and-thrust system known as the Iberian

chain, which crosses eastern Spain in a NWSE direction

(Guimera & A lvaro, 1990; Salas et al., 2001). Post-Eocene

erosion removed signicant portions of the JurassicCretaceous sedimentary section of the earlier depocenters

of the Salzedella sub-basin (Fig. 3). During the Late Oligo-

cene to Early Miocene, extensional tectonics linked to the

opening of the offshore Valencia trough led to the develop-

ment of large grabens in nearby inshore and offshore areas

(Vegas, 1992; Bartrina, Cabrera, Jurado, Guimera, & Roca,

1992).

3. Methods

The Ascla Formation was sampled in outcrops in its typelocality near La Salzedella (Figs. 1 and 3) (Salas, 1989).

From an initial set of 320 samples (Permanyer et al.,

1999), 31 samples were selected on the basis of the presence

of visible cements. Samples were treated using `cold

preparation techniques' (Goldstein & Reynolds, 1994) to

avoid reequilibration of FIs. Double-polished thin and

thick sections were prepared from each sample. In the

thick sections, FI analyses were performed before any elec-

tron microprobe, cathodoluminescence or staining work to

protect the FIs. In addition to transmitted light microscopy,

cathodoluminescence (CL) microscopy of calcite cements

was accomplished using a Technosyn MK4 system. Inci-

dent-light uorescence microscopy employed OlympusBX-60 and Diastare epiuorescence systems. A 490 nm

excitation lter and a 520 nm barrier lter were used for

most photographic work. Most epiuorescence photomicro-

graphs were performed on very thin (,20 mm), polished

sections in order to improve denition of mineral uores-

cence. The polished thin sections were stained by immer-

sion in an acid solution of alizarin red S and potassium

ferricyanide (using the recipe of Lindholm & Finkelman,

1972) for two minutes, which allowed for the optical detec-

tion of Fe-enriched growth bands in calcite cements with as

little as 0.3 mole% FeCO3. BSE petrography on a JEOL

JXA-8900 electron microprobe was used to further observe

chemical variability of cements.

Detailed petrography and microthermometry of oil and

aqueous FIs were performed on selected unheated thick

sections using a Fluid Inc gas-ow heating and freezing

stage. Its thermocouple was calibrated using synthetic FI

standards. Thick sections were then cut into small chips,

and a single eld of view was selected for each chip for

performing microthermometry. Low heating rates were

used for measuring homogenization temperatures (Th),

which were recorded in the order of increasing Th.

The chemical composition of the different carbonate-

cement zones were determined by wavelength dispersive

X-ray spectrometry using a JEOL JXA-8900 electron

microprobe (15 kV accelerating voltage, 2.047E-08 A

beam current, 5 mm beam size, 100^ 3.5% totals

accepted). Detection limits are approximately 150 ppm for

Ca, 100 ppm for Mg, 300 ppm for Fe, 275 ppm for Mn, and

250 ppm for Sr. The results were normalized to 100 mol%

CaCO3, MgCO3, FeCO3, MnCO3, and SrCO3. Selectedcalcite-cement zones were micro-sampled using a micro-

drill mounted on a microscope, and their oxygen and carbon

isotopic compositions were determined at the University of

Michigan using a Finnigan MAT 251 mass spectrometer.

Isotopic enrichments were corrected for acid fractionation

and the values are reported in notation relative to VPDB

standard.

4. General texture and composition

The analyzed samples are predominantly calcareousmudstones, with rare wackestones and grainstones,

more common towards the top of the formation.

Bioclasts include serpulids, arenaceous foraminifera, and

bivalves, with rare echinoderms, miliolids, and sponge

spicules (cf. Salas, 1989). The micritic matrix is cathodolu-

minescent and uorescent, and is less brightly luminescent

in haloes bordering fractures and stylolites. The matrix typi-

cally contains small amounts of illite, silt-sized quartz, and

disseminated hematitic pseudomorphs after authigenic

pyrite.

The original intraparticle porosity in bioclasts is lled by:

internal sediment, hematitic pseudomorphs after authigenicpyrite, drusy calcite cement, and, in places, cloudy calcite

with dolomite and hematite inclusions. Several generations

of fractures, in part cut by vertical stylolites, are lled by

generations of sparry calcite cements.

5. Cement stratigraphy and geochemistry

Three cement sequences were distinguished. The petro-

graphic and geochemical data of the different cement

generations are summarized in Table 1.


7/29/2019 Rossi_2001

5/20

Table 1

Summary of the petrographic, geochemical, and FI data of the different cement generations. CL, cathodoluminescence; FL, uorescence; NL, non luminescent; DL

non uorescent

Sequence Cement Mineralogy CL FL CO3Mg CO3Fe CO3Mn CO3Sr d13C (VPDB) d1

Sequence 1 Pyrite Replaced by hematite

C1.1 Non-ferroan calcite NL NF to yellow 1.76 0.00 0.02 0.00

C1.2 Non-ferroan calcite BL orange Yellow-green 1.96 0.04 0.01 0.04 1109 to 12.2 2

C.12 Non-ferroan calcite NL sectors NF 0.07 0.01 0.01 0.41

C1.3a Non-ferroan calcite DL dark orange Weak 0.82 0.04 0.02 0.03 11.9 2

C1.3a Non-ferroan calcite NL sectors NF 0.85 0.05 0.01 0.14

C1.3b Non-ferroan calcite DL dark orange Weak 1.49 0.13 0.02 0.00

C1.4 Ferroan calcite DL dark orange NF 1.37 0.37 0.02 0.01

Dolomite Replaced by calcite

Celestite Replaced by calcite

Barite Replaced by calcite

Sequence 2 C2.1 Ferroan calcite NL to DL orange NF 1.83 0.36 0.02 0.01 11.2 to 11.3 2

Fluorite Fluorite BL blue NF

C2.2 Non-ferroan calcite BL orange NF 1.82 0.02 0.02 0.02 20.7 to 20.2 2

C2.2 Non-ferroan calcite NL sectors NF 0.40 0.02 0.00 0.19

Sequence 3 C3.1 Fer./non-fer. calcite DL/BL yellow NF to green 0.80 0.22 0.02 0.01 10.8 to 11.4 2

C3.2 Non-ferroan calcite NL to BL yellow Bright green 0.84 0.05 0.01 0.00

C3.3 Ferroan calcite DL orange NF 0.59 0.52 0.02 0.07

C3.4 Slightly ferroan calcite DL orange-BL NF to green 0.49 0.20 0.03 0.00

C3.5 Non-ferroan calcite BL(yellow)-NL Weak, green 0.59 0.04 0.05 0.00 26.3 2

7/29/2019 Rossi_2001

6/20

5.1. Cement sequence 1

The rst cement sequence lled the primary porosity and

some vertical fractures. Pyrite, which is now replaced by

hematite, was the rst cement to precipitate, followed by

three generations of non-ferroan calcite (C1.1 to C1.3), one

generation of ferroan calcite (C1.4) and one generation of

dolomite, which is currently replaced by cloudy calcite.


Fig. 4. Transmitted-light, uorescence and CL petrography of sequence-1 calcite cements lling intraparticle porosity in serpulids. A: epiuorescence

photomicrograph showing an initial predominantly non-uorescent generation of brous calcite (C1.1), which predated some internal sediment, followed

by brightly uorescent C1.2 with non-uorescent sectors and nally, weakly uorescent C1.3. B: CL photomicrograph showing the predominantly bright

orange luminescent C1.2 with prominent non luminescent sectors, followed by subordinate darker orange luminescent C1.3. Notice that compaction breakage

of grains predates precipitation of C1.2 (arrow). C: transmitted-light photomicrograph of C1.1C1.3, which appear as a drusy mosaic of spar crystals with

abundant twin lamellae. D: CL photomicrograph showing C1.1C1.3 lling primary porosity. Note (arrow) the presence of a thin fracture lled by

luminescent, oil-FI-bearing C2.1 calcite. E: same eld of view as D in blue-light epi-illumination. Both C1.2 and C1.3 show non-uorescent sectors. Note

the presence of small brightly uorescent oil FI enclosed in the non-uorescent C2.1 fracture (arrow). F: uorescence photomicrograph showing relict brous-

fan textures (center-left) in C1.2. Pore is nally lled by non-uorescent ferroan calcite of C1.4 (center right). Note the presence of abundant oil FIs (bright

spots) trapped along microfractures in both C1.2 and C1.4.

7/29/2019 Rossi_2001

7/20

C1.1 is typically very thin and has a brous texture, which

commonly is only visible under blue-light epi-illumination

(Fig. 4A). It is non-cathodoluminescent and non- to yellow-

ish uorescent. It is best developed towards the top of the

formation, where it rims the primary porosity. C1.1 has a

mean composition of (Ca0.982Mg0.0176Fe0Mn0.0002Sr0)CO3.

C1.2 is predated by `grain-breakage' compaction of

bioclasts (Fig. 4B). It is predominantly yellow- to green

uorescent and bright orange in CL with sector zoning.

Non luminescent sectors show prominent triangular crosssections (Fig. 4BE) and are highly enriched in Sr and

depleted in Mg, with a mean composition of (Ca0.995Mg0.0007-Fe0.0001Mn0.0001Sr0.0041)CO3. Luminescent sectors have mean

composition of (Ca0.979Mg0.0196Fe0.0004Mn0.0001Sr0.0004)CO3. In

places, C1.2 shows a relict brous-fan texture which can be

seen under blue-light epi-illumination (Fig. 4F) or after

slight etching. C1.2 has moderately positive d13C values

(11.9 to 12.2) and d18O close to 0 (20.2 to 10.2).

C1.3 is also sector zoned, but less bright orange under CL

(Fig. 4B and D), weakly uorescent (Fig. 4E), and termi-

nated by a distinctive greenish band in blue-light epi-

illumination (Fig. 4F). In the lower part of the studied inter-

val, C1.3a has a mean geochemical composition of

(Ca0.991Mg0.0082Fe0.0004Mn0.0002Sr0.0003)CO3 in the brighter

sectors, with the darker luminescent sectors slightly enriched

in Sr ((Ca0.989Mg0.0085Fe0.0005Mn0.0001Sr0.0014)CO3). In the upper

part of the formation, C1.3 (C1.3b) is signicantly richer in

Mg and Fe, having a mean composition of (Ca0.984Mg0.0149-Fe0.0013Mn0.0002Sr0)CO3. C1.3a has an isotopic composition

d13C 11:9; d18O 20:3 practically identical to C1.2.

C1.4 consists of ferroan, non-uorescent (Fig. 4F), dull

orange cathodoluminescent calcite with a mean composi-

tion of (Ca0.982Mg0.0137Fe0.0037Mn0.0002Sr0.0001)CO3.

Cements of C1.1 to C1.4 are cross-cut by subvertical

fractures. Cement sequence 1 is terminated by a distinctive

generation of cloudy calcite, which reduces the fracture and

the remaining primary porosity. Its cloudiness results from

abundant uid, dolomite and hematite inclusions (Fig. 5A).

These inclusions typically dene ghosts of rhombs, which in

places, show the curved cleavages typical of saddle dolo-

mite. The cloudy calcite is thus interpreted as dedolomite,i.e. a replacement product after original dolomite. The

cloudy calcite typically has a bright, patchy luminescence

in CL (Fig. 5B), but in places it shows a less cloudy, dull

cathodoluminescent band which may represent a later

calcite cement which overgrew the dedolomitized material.

The mean composition of the cloudy calcite is

(Ca0.988Mg0.0108Fe0.0008Mn0.0001Sr0)CO3. Although the more

inclusion-rich areas were avoided for analysis, the Mg and

Fe concentrations obtained probably do not re ect true

concentrations in the calcite, because they must be inuenced

by the abundance of small dolomite and hematite inclusions.

Calcitic pseudomorphs after dolomite rhombs are also foundenclosed in the micritic matrix, isolated or more commonly in

the vicinity of fractures (Fig. 6A). These pseudomorphs are of

dark-cathodoluminescent, non-ferroan calcite with a micro-

spar texture. They typically have small inclusions of dolomite

and Fe-oxides. Thus, they are interpreted as calcitized matrix-

replacive dolomite, with timing of calcitization probably

different from that of the dolomite cements.

In some fractures, celestite and minor Sr-rich barite

cement postdated the original dolomite precipitation. In

these fractures, both celestite and barite are preserved only

as inclusions, and the original crystals have been replaced

by ferroan calcite (Fig. 6BC) with a geochemical compo-

sition identical to that of generation C3.3 (see below). TheMg and Fe contents of sequence-1 calcites are summarized

in Fig. 7.


Cement sequence 2 is virtually only found in fractures,

which usually cut across cement sequence 1.C2.1 appears to

include several subgenerations, all consisting of relatively

Mg-rich ferroan calcite usually containing oil FIs. In places

C2.1 contains a prominent corrosion surface which is post-

dated by later C2.1 cements containing primary oil FIs


Fig. 5. A: transmitted light photomicrograph of primary pore in the upper

part of the Ascla Formation rimmed by brous calcite (C1.1 1 C1.2) and

lled by cloudy calcite. This calcite contains dolomite (arrows) and hema-

tite inclusions, which dene ghosts of rhombs. B: CL photomicrograph

showing the brous and non luminescent nature of pore-rimming C1.1

and the bright, patchy luminescence of the cloudy calcite, due to the highinclusion density. Note the presence of a less cloudy, darker luminescent

band postdating the inclusion-rich patchy core.

7/29/2019 Rossi_2001

8/20

(Fig. 8A and B). In some fractures, C2.1 is altered and

partially replaced by subsequent generations (Fig. 8C).

C2.1 is slightly more magnesian than C1.4, having a mean

composition of (Ca0.977Mg0.0183Fe0.0036Mn0.0002Sr0.0001)CO3.

C2.1 has d13C from 11.2 to 11.3 and d18O from 25.3

to 25.9. Minor amounts of uorite have been observed

postdating C2.1. Where present, it is predated and postdated

by corrosion surfaces.

C2.2 is predated by fractures which cut across at low

angles C2.1-bearing fractures. C2.2 reduces signicantamounts of porosity and is composed of non-ferroan pris-

matic calcite. C2.2 is bright orange cathodoluminescent and

shows prominent, darker-luminescent, irregularly shaped

sectors (Fig. 8C and B). In the bright sectors, the mean

composition is (Ca0.981Mg0.0182Fe0.0002Mn0.0002Sr0.0002)CO3.

As in C1.2 and C1.3, the dark sectors in C2.2 are depleted

in Mg and enriched in Sr, having a mean composition of

(Ca0.994Mg0.004Fe0.0002Mn0Sr0.0019)CO3.


Sequence 3 postdates a phase of compressive tectonics

that produced vertical stylolites and extensive deformation

structures, which cut across sequence 2 cements. Deforma-tion includes twin lamellae, abundant planes of secondary

aqueous FIs (Fig. 8C and D), and fractures which are in

places parallel to the stratication. Sequence 3 cements

ll those fractures. Deformation appears to have continued


Fig. 7. Plot of Mg versus Fe for sequence-1 calcites. Earlier cements (C1.1, C1.2) are Mg- and Sr-rich and Fe-poor. Late sequence-1 generations (C1.3b, 1.4)

have higher Fe and lower Mg contents (NL, non luminescent).

Fig. 6. A: calcite pseudomorphs after dolomite rhombs (arrows) are present on the borders of this fracture. The pseudomorphs are of microspar, have small

inclusions of dolomite and Fe-oxides, and are interpreted as a replacement after matrix-replacive, inclusion-rich dolomite rhombs. The reactivation of this

fracture allowed for the precipitation of oil-FI-bearing ferroan calcite (C2.1) and later non-ferroan calcite (C2.2). B: purple-coloured, K-ferricyanide-stained

prismatic pseudomorphs after celestite. The pseudomorphs consist of ferroan C2.2 calcite, and are enclosed in brighter-coloured (unstained by K-4ferricya-

nide) non-ferroan C3.3 calcite. Note that the celestite-replacive C3.3 calcite contains corroded celestite inclusions (arrows), shows a mottled texture, and

overgrowths, and hence postdates, C2.2. The interpreted paragenetic sequence in this fracture is thus celestite-C2.2 calcite-C3.3 calcite. C: BSE image of

another view of the same fracture shown in A. Note the patchy aspect of the celestite-replacive ferroan calcite. The bright spots represent the sites of

microprobe analyses.

7/29/2019 Rossi_2001

9/20

during precipitation of sequence 3 calcite cements. The

earliest growth zones of sequence 3 contain extensive defor-

mation twins and aqueous FI-bearing microcracks, whereas

the younger growth zones are relatively clean of such

features. In one sample, sequence-3 cements ll a re-openedvertical stylolite, indicating precipitation after a local

compressionextension cycle. These cements are in turn

affected by pressure solution along the borders of the re-

opened stylolite, indicating its later-stage compressional

reactivation.

Cements are alternating ferroan and non-ferroan calcite.

The non-ferroan cement zones typically luminesce greenish

with concentric zones in blue-light epi-illumination, and

non luminescent or brightly subzoned in CL. The ferroan

zones typically do not luminesce in blue-light and

luminesce dully in CL. Growth zones of sequence 3 are

difcult to correlate between different samples, but ferroan

zones typically give way to non-ferroan zones (Fig. 9F). The

early part of this sequence consists of three zones, C3.1 to

C3.3. C3.1 is commonly formed by a ne-scale alternation

of ferroan and non-ferroan bands (Fig. 9AC). The transi-tion between late C3.1 (ferroan), and the next generation

(C3.2; non-ferroan), is commonly marked by an iron-

hydroxide-rich surface, which appears bright yellow under

transmitted light (Figs. 9DE). C3.3 is predominantly

ferroan and dully cathodoluminescent.

The later part of sequence 3 is formed by two genera-

tions (C3.4 and C3.5). C3.4 is slightly ferroan, dull to

bright yellow under CL, and dark green and zoned

under blue-light epi-illumination. C3.5 is non-ferroan,

predominantly non cathodoluminescent but showing several

bright hairlines, green to dark and zoned under blue-light


Fig. 8. Petrography of sequence-2 calcites. A: In this CL photomicrograph, the earlier calcites are ferroan, dull luminescent, oil-FI bearing, and affected by a

prominent corrosion surface. This surface is covered by a layer of dark luminescent calcite (late C2.1, arrows) bearing primary oil inclusions, which ispostdated by non-ferroan and lighter luminescent C2.2 calcite and nally by C3.3 ferroan calcite. B: Transmitted light photomicrograph after staining. Note

the differences between the two darker-coloured, ferroan generations: the rst ferroan calcite (2.1) is more rich in inclusions and forms the central parts of the

crystals, whereas the last ferroan calcite (C3.3) has a cleaner aspect, forms the outer parts of the crystals, and replaces celestite. C: CL photomicrograph

showing dark luminescent, prismatic C2.1 ferroan calcite postdated by thicker and non-ferroan C2.2 calcite with, relatively bright and dull orange lumines-

cence and sector zoning. Note that C2.1 is altered and partially replaced. Note also that C2.2 has abundant aqueous FIs which appear as trails of bright spots

that terminate against a growth zone boundary (arrow). The formation of these pseudosecondary inclusions was postdated by the precipitation of sequence-3

calcites, which lled the residual porosity of this fracture (lower right). D: CL photomicrograph showing sector and diffuse concentric zoning in C2.2 calcite.

Note the presence of trails of bright spots (upper right) representing aqueous FIs trapped along microcracks.

7/29/2019 Rossi_2001

10/20

epi-illumination, and typically includes one or two intervals

of uorescent microsparitic internal sediment. In places,

these cements ll fractures that postdate C3.3.

There is a clear tendency for the sequence-3 cements to

be less magnesian than sequence 2 (Fig. 10), but with a wide

variation in Fe content, reecting the alternation of ferroan

and non-ferroan subzones. Sr contents are invariably below

the detection limits, and the MnCO3 is typically below

0.3 mole% except in some bright luminescent bands,

where the mole percent MnCO3 can be as high as 1.4. The

earliest cements of sequence 3 (3.1) have d18O from 28 to

28.6 and d13C from 10.8 to 11.4. The latest cements


Fig. 9. Thin-section petrography of sequence-3 calcites. AB: transmitted light photomicrograph ofne-scale alternations of ferroan and non-ferroan zones in

early C3.1. The ferroan zones appear darker as a result of selective staining by K-ferricyanide. As detailed in B, the non-ferroan zones host primary all-liquid

aqueous FIs (arrow). C: CL photomicrograph of early C3.1 showing the bright luminescence of the non-ferroan subzones, and the dull luminescence of the

ferroan ones. D: transmitted light photomicrograph of iron hydroxide-stained surface (arrow) separating the predominantly ferroan C3.1 (lower part) from the

non-ferroan C3.2 (upper part). Plane polarized light. E: CL photomicrograph of approximately the same view as D, showing the predominantly dull-

luminescent C3.1 and C3.3, and C3.2, which is non- to bright yellow luminescent. F: CL photomicrograph showing a C3.3-lled vertical fracture crosscut

by a horizontal fracture lled by C3.4 (bright yellow luminescent and zoned) and C3.5 (predominantly non-luminescent with orange hair-lines and internal

sediment).

7/29/2019 Rossi_2001

11/20

of sequence 3 (3.5) have d18O 26.3 and

d13C 26.4.

6. FI petrography

6.1. Oil inclusions

Oil inclusions have been observed in seven samples,whose stratigraphic position is shown in Fig. 11. Oil inclu-

sions are particularly abundant in samples from the basal

part of the formation, which is the interval showing the best

source-rock characteristics (Fig. 11). Here, abundant

calcite-cemented fractures (less than 0.5 mm in thickness)

and groups of closely spaced microfractures are crowded

with small oil inclusions (Figs. 12 and 4F). In these fractures

and groups microfractures, hundreds of oil FIs per mm are

observed even in very thin sections (Fig. 12). In samples

located upsection, oil inclusions are signicantly more

scarce and, when present, they are only found in one or

two fracture llings per sample. These oil-bearing fractures

are typically thicker that 0.5 mm (e.g. Fig. 13A), and theiroil-inclusion density is usually very low (tens of inclusions

per mm in thick sections).

Abundant oil inclusions are present in microfractures that

cut across the cements of sequence 1 (Figs. 4F and 12), but

not in later cements, so they are interpreted as pseudose-

condary in origin (sensu Roedder, 1984). Some oil inclu-

sions, however, are primary in the ferroan calcites of

sequence 2 (C2.1) (Figs. 4E, 12 and 13A). In this cement

generation, oil FIs are restricted to the last growth zone and

elongate in the direction of growth (Fig. 13A). The corro-

sion surface that predates late-stage C2.1 cement is, in

places, marked by abundant oil inclusions, further indicat-

ing a primary origin (Fig. 8A).

Most oil inclusions uorescence blue in U.V. epi-illumi-

nation, which is commonly characteristic of light oils

(McLimans, 1991). In samples from the basal part of the

formation, inclusions with a yellow uorescence, more typi-

cal of lower gravity oils (McLimans, op. cit.), are present in

small amounts. The yellow-uorescent inclusions are the

only oil inclusion type in some microfractures, while inothers, they coexist with the blue ones. The relative timing

of the entrapment of these two oils is unclear. The yellow-

uorescent oil inclusions are typically very small and all

liquid, or contain very small vapour bubbles. The blue-uor-

escent ones, however, are typically larger, most have the

appearance of two-phase oil-gas inclusions, and ratios of

gas to liquid appear relatively consistent within a given FI

assemblage (i.e. a single microfracture: Fig. 13B), but vary

widely from one assemblage to another.

In one sample from the middle part of the formation,

C2.1 contains primary blue-uorescent oil inclusions

which show extremely variable gas-to-oil ratios, ranging

from all liquid to small amounts of liquid with largebubbles. Some of these inclusions also contain a visible

aqueous phase (Fig. 13C). The oil-bearing zones in this

sample are crowded with possible fungal or bacterial

lament moulds (Fig. 13B and C) that cut across the

cement zones. These laments have not been observed

in other cement zones or in other samples. In places,

the laments are seen to penetrate oil inclusions, appar-

ently causing leakage of the oil inclusion and partial

lling of the bacterial or fungal lament mould by oil

(Fig. 13D). We interpret the variability in phase ratios

of the oil inclusions as a consequence of endolithic biologic


Fig. 10. Plot of Mg versus Fe content in sequence-2 and sequence-3 calcites. In this diagram, the calcites of both sequences are distinguished by their different

Mg contents.

7/29/2019 Rossi_2001

12/20

penetration of the FIs, probably at outcrop, leading to their

alteration.

6.2. Aqueous inclusions

Cement C1.2 typically contains very small primary

aqueous inclusions, which are predominantly all liquid,

although locally, they contain small bubbles, suggesting

an originally low-temperature origin and later thermal re-

equilibration of variable magnitude or metastability of the

all-liquid inclusions (cf. Goldstein & Reynolds, 1994).

These inclusions are too small to use for microthermometry.

Most of the aqueous inclusions are trapped in micro-

fractures. Some of the microfractures are clearly lled

with a cement, so inclusions trapped within them, but

not within the surrounding calcite, are best interpretedas primary. In C2.1, these `primary' aqueous inclusions are

extremely rare, but locally, they are associated with primary

oil inclusions in the same cement. These aqueous inclusions

are generally very small, and in places, they have small

bubbles.

Aqueous inclusions are abundant in healed microfrac-

tures in C2.2. Commonly, the microcracks terminate against

the base of sequence 3 cements (Fig. 8C). Therefore, the

inclusions formed during the end of C2.2 precipitation or

after, but before C3.1, and are thus interpreted as pseudose-

condary in origin (sensu Roedder, 1984). These inclusions

are predominantly all liquid, but in places, they have

bubbles and highly variable vapour-to-liquid ratios, suggest-ing heterogeneous entrapment of a liquid and gas phase at a

low temperature. Overall ratios of vapour to liquid are too

high for the variable phase ratios to originate from necking

down of homogeneously entrapped liquid inclusions after

appearance of the vapour bubble.

In sequence 3, some of the earliest growth zones contain

very small primary all-liquid inclusions, concentrated along

cloudy growth zones and typically elongated in the direction

of growth (Fig. 9A and B). In addition, the early sequence 3

cements contain either secondary or pseudosecondary FIs

similar to those in C2.2. They are predominantly all liquid,


Fig. 11. Stratigraphic distribution of the studied samples, showing the loca-

tion of the oil-FI-bearing samples and their estimated relative abundance of

oil inclusions (RAOI). The total organic content (TOC) values are from

Permanyer, Salas, and Bitzer (2000b). Note that the samples with higher

RAOIs are from the basal part of the formation, which also has higher TOCs

and the best source-rock characteristics according to Permanyer et al.

(2000b).

Fig. 12. Epiuorescence photomicrograph showing abundant uorescent oil

inclusions (bright spots) in a very thin (,20 mm) section of a sample from

the basal part of the Ascla Formation. Most uorescent oil inclusions are in

microfractures that cut across C1.2 and C1.3. These microfractures are

aligned with fractures (arrows) lled by C2.1 and also contain uorescent

oil inclusions. The oil inclusions are therefore interpreted as primary in

C2.1 and pseudosecondary in C1.21.3.

7/29/2019 Rossi_2001

13/20

but in places, they have bubbles and highly variable vapour-

to-liquid ratios.

7. FI microthermometry

We obtained 61 homogenization temperatures, most of

them from oil inclusions, and 6 ice nal melting tempera-

tures on aqueous inclusions (Tm ice). Because the aqueous

inclusions of some generations were so small, it was

extremely difcult to get microthermometric measurements

from them.

We avoided microthermometry on oil inclusions

suspected to have been altered by biologic processes,

or FIAs with inclusions of highly variable vapour-to-

liquid ratio. The measured Th in oil inclusions are

from primary and pseudosecondary, blue-uorescent

inclusions trapped in microfractures lled with C2.1. In

these inclusions, the overall range of Th is very wide, from

45 to 115.98C (Fig. 14). Within individual FIAs, the range of

Th is narrower but also variable: ,118C (4 FIAs), between

18 and 248C (4 FIAs), and more rarely higher (2 FIAs)

(Fig. 14).The aqueous inclusions with measured Th are primary

inclusions trapped in the C2.1 fracture-lls along with oil

FIs. Data could be obtained only from two FIAs, one with

only one measurable inclusion (Th 1148C), and the other

showing a wide range ofTh variation, from 84.5 to 117.28C

(Fig. 14). Tm ice values measured in these inclusions are

22.7 and 23.78C. In the pseudosecondary aqueous inclu-

sions present in generation C2.2, measured Tm ice values are

high and relatively consistent from 0.0 to 20.28C. For those

in secondary or pseudosecondary inclusions in Sequence 3,

Tm ice values ranged from 0.0 to 20.28C.


Fig. 13. Photomicrographs illustrating oil FIs in thick (,

100 mm) sections using a combination of epiuorescence and transmitted light. A: assemblage ofprimary oilFIs deninga growthzone in a late growth zone of C2.1.B: uorescent FIs in C1.3. They are aligned along what wouldnormally be interpreted as a

healed microcrack, containing secondary FIs. In CL and blue-light epi-illumination, however, these oil inclusions occur within a narrow fracture, lled with

C2.1, indicating the inclusions are actually primary in origin and were trapped after precipitation of sequence 1 and during precipitation of C2.1. Note that all

FIs are two phase and contain similar ratios of vapour to liquid. C: Three-phase inclusion mostly containing uorescent oil, with a gas bubble and a small rim

of aqueous liquid visible in the uppermost tip. Note the presence of possible endolithic fungal or bacterial lament moulds (arrows). D: all-liquid oil inclusion,

surrounded by possible endolithic fungal or bacterial lament moulds partially lled by uorescent oil (arrows).

7/29/2019 Rossi_2001

14/20

8. Discussion

8.1. Interpretation of FI data

The primary aqueous inclusions in C1.2 are dominated by

all-liquid FIs. This suggests that C1.2 precipitated at low

temperature (less than ,508C; Goldstein & Reynolds,

1994).

The oil inclusions are primary in C2.1 and pseudo-

secondary in cement sequence 1. Thus, we conclude thatthe oil inclusions must have been trapped broadly during

the time of precipitation of C2.1. The variability in Th of the

oil inclusions needs to be explained. Some FIAs are

narrowly distributed and thus a reliable indicators of mini-

mum entrapment temperatures of 45498C, 55618C, 50

598C, and 63748C. Other FIAs are more wide-ranging and

could have been altered by thermal reequilibration.

However, if oil inclusions are buried after entrapment and

overheated above their entrapment temperatures, the inter-

nal pressures are less likely to rise signicantly above the

external pressure than aqueous inclusions, and thus oil

inclusions are less prone to thermal re-equilibration

(Burruss, 1987, 1989; Goldstein & Reynolds, 1994).Another explanation could be that the range in Threects true variation in entrapment temperature. This

is possible, but seems unlikely, given the apparent

restriction of oil entrapment to a single position in the

paragenesis. Another possibility is that the variation in

Th reects entrapment at similar temperatures but under

variable pressures, which seems likely given the basinal

lithology and complex fracturing history in the system.

The Ascla Formation is composed of low-permeability

marly limestones that could have developed overpres-

sure during C2.1 time. Episodic fracturing may have

allowed overpressure release. And nally, it remains possi-

ble that the variation in Th is merely the result of secular

compositional variation of the oils entrapped. This would be

consistent with the lightness of the trapped oils (as deduced

from their blue uorescence), because the liquid-vapour

phase envelopes for light oils often show major variation

for only small variations in composition (cf. Goldstein &

Reynolds, 1994, p. 39).

The meager data on Th in the aqueous inclusions

trapped along with the oil in C2.1 also are variable.These Th tend to coincide with the uppermost Th data

measured from the oil inclusions measured in the same

FIAs. The variability in Th could easily be explained by

variable amounts of thermal reequilibration; however,

pseudosecondary FIs in C2.2 are dominated by all-liquid

inclusions. This shows that between the end of C2.2 and the

beginning of C3.1 precipitation, temperatures had dropped

to very low values. Although a transient thermal event

cannot be disproved for C2.2 precipitation, the simplest

hypothesis would be that temperatures decreased after

C2.1, and that thermal reequilibration of the aqueous inclu-

sions was unlikely. Thus, as argued above, the variability in

Th is most easily explained by pressure or compositionalvariations.

If it is true that neither the aqueous nor petroleum inclu-

sions have reequilibrated, then the coincidence of the petro-

leum and aqueous, higher temperature Th at 1178C, indicates

a period of FI entrapment in which both phases were satu-

rated with respect to the gas phase. This would indicate that

at least some of C2.1 precipitated at 1178C in the presence

of gas-saturated oil.

Primary aqueous FIs in cements of sequence 3 are all-

liquid, which suggests that these cements precipitated at

relatively low temperatures.


Fig. 14. Homogenization temperatures of blue-uorescent oil and aqueous FIs in unaltered C2.1-lled microfractures.

7/29/2019 Rossi_2001

15/20

8.2. Interpretation of cement sequence 1: precipitation

during the rst burial stage (Kimmeridgian-Valanginian)

The rst calcite cements probably precipitated near the

surface or at shallow depths, as suggested by their parage-

netic position before and just after initial compaction, and

low-temperature FIs. Composition is consistent with preci-

pitation from marine water, especially the relatively high

overall Mg and Sr content, low Fe content (Fig. 15), and

isotopic composition (Fig. 16). Moreover, relict brous

texture of both C1.1 and C1.2 is common in modern and

ancient marine calcite cements. The presumably marine,

Mg- and Sr-rich, Fe-poor cements (C1.1, C1.2) were

followed by non-uorescent generations with higher Fe

and lower Mg contents (C1.3b, 1.4; Fig. 15), indicating

more reducing waters consistent with progressive burialand a different uid composition (cf. Meyers, 1991).

Sequence 1 ended with the precipitation of dolomite,

which lled the remnant primary porosity and some new

fracture porosity. The dolomite must have precipitated rela-

tively early in the diagenetic history, because it lled

remaining primary porosity. The abundance of dolomite

increases up section, where the original primary porosities

were higher due to the predominance of shallower-water

facies (Salas, 1989). This suggests that the original abun-

dance of dolomite was controlled by depositional facies.

The dolomite not only lled porosity but also partially

replaced the host limestone, especially towards the top of


Fig. 15. Summary of Mg and Fe contents of the different calcite generations. Rectangular elds represent mean values^ one standard deviation for each zone.

Fig. 16. Oxygen and carbon isotopic compositions of individual calcite cement zones. Data for meteoric calcites in the Catalan Coastal ranges are from Trave

et al. (1998).

7/29/2019 Rossi_2001

16/20

the formation. The dolomitization of Sr-rich marine carbo-

nates might have released the Sr necessary for the precipita-

tion of the celestite that postdated dolomite.

In the Salzedella area, the Ascla Formation grades

upwards into an extensive unit of dolomitized carbonate

(the Talaies Dolomite, sensu Salas, 1987) that replaces the

Tithonian to Berriasian Bovalar Formation and the top of

the Ascla Formation (Nadal, 2000). The location of dolomi-tization in the Talaies is related to normal faults. Timing has

been interpreted as predating deposition of the upper part of

the Valanginian, because the dolomites were partly eroded

during development of a late Berriasian-to-late Valanginian

unconformity (Fig. 3; Aurell, Mas, Melendez, & Salas,

1994; Nadal, 2000; Salas, Martn-Closas, Querol, Guimera,

& Roca, 1995). If the dolomites in the Ascla and Talaies

Formations are equivalent, then the dolomites in the Ascla

would have precipitated at burial depths of less than about

10001500 m (Fig. 3), which is consistent with its observed

paragenetic position (Fig. 17).

Nadal (2000) has reported in the Talaies dolomites the

presence of FIs with variable densities, from all liquid totwo-phase with inconsistent Th (801258C). If the reported

FIs are primary and have not suffered necking down after a

phase change, then their relatively high variability in density

could be the result of: (1) original entrapment at lower

temperatures followed by later incomplete re-equilibration

during further burial; or (2) multiple events of dolomite

precipitation at variable temperatures. The rst explanation

is consistent with the relatively small depth of burial for

dolomite precipitation inferred by Nadal (2000). However,

the second explanation (multistage dolomitization) is

consistent with the association with faults, which could

have acted as conduits for hot uids, and with the evidences

of zoning described by Nadal (op. cit). Unfortunately, the

relationship between the Ascla and Talaies dolomites could

not be evaluated further because the Ascla dolomites have

been almost completely calcitized.

8.3. Interpretation of cement sequence 2: main burial stage

and oil migration

Sequence 2 begins with the precipitation of syntectonic

(C2.1) ferroan calcite, which records oil generation and

migration through the Ascla Formation via fractures at

temperatures of at least 1178C as deduced from the FI

data (see above). The Tm ice data indicate that C2.1

precipitated from waters slightly more saline than sea

water (4.56 wt% NaCl eq.). Assuming a precipitation

temperature of 1178C, the oxygen isotopic data

(d18O 25.6VPDB) indicates that C2.1 precipitated

from a water with d18O 19.8SMOW (Friedman &

O'Neil, 1977), which is consistent with a basinal brine

enriched in 18O from evaporation or uidrock interaction.The relatively positive d13C values (Fig. 16) are also consis-

tent with such a uid.

C2.1 hosts primary oil inclusions. It postdates the dolo-

mite, which may predate the late part of the Valanginian, so

oil migration must have taken place during or after the late

Valanginian. The high-temperature FIs in C2.1 and abun-

dant fracturing during this stage is consistent with timing

beginning during rapid BarremianAlbian rift subsidence or

Late Cretaceous post-rift subsidence and ending with maxi-

mum burial during the latest Cretaceous or earliest Tertiary

(Fig. 17). Burial-thermal modelling shows that the Ascla


Fig. 17. Generalized paragenetic sequence of the Ascla Formation.

7/29/2019 Rossi_2001

17/20

Formation in the study area probably entered the oil window

during the Cenomanian (Permanyer et al., 2000a). Our data

are consistent with this, but the least equivocal FI data

suggest oil generation and migration continued during the

latest Cretaceous or earliest Tertiary, when the Ascla

reached its maximum burial depth and presumably maxi-

mum temperature. The FI data indicate that C2.1 records the

maximum temperatures reached by the Ascla, which are

about 1178C. Similar maximum temperatures (1118C)

have been obtained by Permanyer et al. (2000a) from vitri-

nite reectance values using empirical calibrations. In the

nearby Penagolosa sub-basin (Fig. 1), hydrothermal circula-

tion resulting in ZnPb mineralization (with local presence

of uorite), at temperatures as high as 1508C, occurred

during the early Paleocene (Grandia, Asmerom, Getty,

Cardellach, & Canals, 2000). In the Salzedella sub-basin,

ZnPb deposits are not known and there is no evidence that

such high temperatures had been reached.

Widespread vertical fracturing predated and postdated oil

migration. The resulting fractures were partly cemented byferroan calcites of C2.1 and lled with non-ferroan calcite of

C2.2, suggesting increasingly less reducing conditions or

limited Fe availability. C2.2 and the rst cements of

sequence 1 (C1.1C1.2) are geochemically similar. Both

are Mg-rich, Fe-poor and have Sr-enriched sectors (Fig.

15). As in C1.1C1.2, this trace-element composition

could be consistent with precipitation from marine-derived

waters. The FI data suggest that C2.2 might have been

preceded by the maximum temperatures and clearly were

postdated by low temperatures. Thus, it makes sense to

interpret that C2.2 represents part of the cooling path of

this unit. If one were to assume a temperature of 801008C for precipitation of C2.2, then the its d18O of

211.2VPDB would indicate precipitation from a water

with d18O from about 0 to 12 SMOW (Friedman &

O'Neil, 1977), i.e. consistent with precipitation from

marine-derived water. The carbon isotopic composition of

20.2 to 20.7VPDB is also compatible with marine-

derived waters.

C2.2 is deformed by compressional structures. The onset

of compressive tectonics and basin inversion has been

interpreted as Eocene in age (Salas et al., 2001). Therefore,

C2.2 precipitation, and the events of oil migration that

predate it must have taken place before the Eocene. If

C2.2 does represent cooling of the unit, and if it predatescompressional tectonism and its related unroong, one must

question why the cooling has occurred. One possibility

would be that some compression and uplift coincided

with C2.2, but there are no deformation features that

would indicate this change in stress eld until after C2.2

precipitation. Another explanation could be that the unit was

cooled by deeply circulating pore uids. Given the fractured

nature of the unit and the evidence for a change in pore

uid composition, this remains a possibility. Finally, the

cooling could easily be related to decreasing heat ow in

the basin.

8.4. Interpretation of cement sequence 3: Tertiary uplift

Sequence 3 typically lls fractures that cut the sequence-2

fractures at high angles, suggesting a radical change in the

stress eld. The cements of sequence 3 are syntectonic.

Because they ll re-opened vertical stylolites, they must

postdate at least one compressional and one extensional

tectonic event, at least at a local scale. These cements are

cut across by further vertical stylolitization, indicating

further compression. These relationships suggest that

sequence 3 precipitated during and after the EoceneOligo-

cene (Fig. 17), when the area suffered its rst major stage of

compressional tectonics. Some sequence-3 zones may have

precipitated during the Late Oligocene to Early Miocene, a

period which was characterized by extensional tectonics

leading to the development of large grabens in nearby

inshore areas and to the opening of the offshore Valencia

trough (Bartrina et al., 1992; Vegas, 1992). Later sequence-

3 compressive structures could be Middle Miocene, when

the area suffered a subordinate compressional tectonic pulse(the Betic compression of Calvet, Trave, Roca, Soler, &

Labaume, 1996; Trave, Calvet, Soler, & Labaume, 1998).

Sequence-3 cements are characterized by low Mg and Sr

contents relative to sequences 12 (Fig. 15). This suggests a

change from predominantly marine-derived waters in

sequences 1 and 2, to meteoric waters in sequence 3. Pseu-

dosecondary FIs that formed between the end of sequence 2

and the beginning of sequence 3 contain fresh waters that

were trapped at low temperature. Sequence 3 is typically

formed by alternating ferroan and non-ferroan calcites,

which suggests nely alternating reducing and oxidizing

conditions compatible with a meteoric origin. Thus, weinterpret that sequence 3 precipitated from low-temperature

meteoric water, after the inversion of the Maestrat basin and

its transformation into an emergent fold and thrust belt.

The earliest sequence-3 calcites precipitated at low-

temperature (less than about 508C), as indicated by the

presence of abundant all-liquid primary FIs (Goldstein &

Reynolds, 1994) (Fig. 9B). Assuming a precipitation

temperature of 30408C, then the d18O of C3.1

(28.3VPDB) indicates precipitation from a water with

d18O from about 25.2 to 23.3 SMOW (Friedman &

O'Neil, 1977), i.e. consistent with meteoric-derived water.

However, the relatively positive carbon isotopic composi-

tion (10.8 to 11.4VPDB) is not consistent with theincorporation of soil-derived carbon, and must then reect

uidrock interaction in a setting distal from the soil zone.

The low trace element content and moderately light d18O

(26.3) and d13C (26.4) of the last sequence-3 zone

(C3.5) are typical of meteoric calcites precipitated at low-

temperature in open systems. The low d13C is consistent

with the incorporation of soil-derived CO2. The overall

trend of the carbon and oxygen isotopic compositions of

sequence 3, from C3.1 to C3.5 (Fig. 16) is also compatible

with precipitation during a phase of erosional unroong, in

which cooler and more soil-inuenced waters progressively


7/29/2019 Rossi_2001

18/20

predominated. Overall, the isotopic compositions of

sequence-3 calcites are similar to the values reported by

Trave et al. (1998) for Tertiary fracture-lling calcites of

meteoric-water origin in the nearby Catalan Coastal Ranges

(Fig. 16). Some of the latest sequence-3 zones may even

have formed in the vadose zone, as suggested by the crystal

silt internal sediments coinciding with precipitation. The

earliest cements in sequence 3 contain the most abundant

fresh-water inclusions along healed fractures, which could

point to entrapment during late sequence 3. These inclusions

are essentially fresh water with extremely inconsistent gas

to liquid ratios, compatible with trapping in the vadose zone.

Precipitation of sequence 3 in some solution-enlarged frac-

tures and presence of some solutional discontinuities within

sequence 3 may also point to a late-stage, near-surface

origin.

8.5. Implications for oil potential

The oils observed as inclusions in the Ascla Formationwere likely generated in the organic-rich intervals of the

formation. This is suggested by (1) the absence of other

feasible source rocks in this part of the basin, and (2) the

higher abundance of oil inclusions in the basal part of the

Ascla, which contains the richer potential source rocks (Fig.

11). In samples from this part of the formation, oil inclu-

sions are abundant in fractures and especially in clusters of

microfractures (Fig. 12), and have variable compositions as

indicated by their uorescence (blue and yellow). These

features are typical in source rocks which have generated

oil (Norman Oxtoby, pers. comm.). However, the overall

abundance of oil FIs is perhaps lower than one shouldexpect for a very rich source-rock. Possible causes are: (1)

The studied samples are from limestones, and the richer

source beds are the interbedded marls. Therefore, the oil

inclusions likely represent trapping along a migration

route in close association with the source-rock. (2) the

source-rock potential, even in the richer intervals, is not

particularly high (Permanyer et al., 1999). In the middle

and upper parts of the Ascla, samples bearing oil inclusions

are scarce, the trapped oils are restricted in composition, and

the oil inclusions are only found in fractures showing low

oil-FI densities. These characteristics suggest trapping

along a migration route through fractures.

The possible contribution of oils generated in the AsclaFormation to the Amposta oil eld has been a matter of

debate (Albaiges et al., 1986; Seemann et al., 1990; Seifert

et al., 1983). This eld is located offshore, 40 km to the west

of Salzedella (Fig. 1), and reservoirs oil in Paleogene-aged

karst developed in Lower Cretaceous carbonates (Seemann

et al., 1990). The age of the trap and seal is Miocene, when

the area experienced an important rifting phase. The

Amposta oil has unique geochemical characteristics and is

not correlated with the other oils in the Tarragona offshore

basin, which have their source in Miocene rocks deposited

in the Tarragona Trough (Fig. 1) (Albaiges et al., 1986).

It has been argued that the Ascla Formation cannot be the

source for the Amposta oil, because (1) lacks source-rock

potential and (2) it was already overmature for oil genera-

tion in the Amposta area during the Miocene (Albaiges et

al., 1986). However, the Ascla formation is still in the oil-

generation phase and have some oil-generation potential at

least in two locations: an offshore well near Amposta Field

(Seemann et al., 1990) and in outcrops near la Salzedella

(Permanyer et al., 1999). Moreover, Seemann et al. (1990)

found geochemical similarities between Ascla-source-rock

extracts and the Amposta crude.

The results of the present study indicate that oil indeed

migrated from the Ascla Formation, at least in the sampled

location. This migration, however, took place when the

formation was at or near maximum burial depth and

temperature, very likely during the Cretaceous to earliest

Tertiary, and before the Eocene. Thus, oil migration

occurred just before Alpine tectonics, which created the

most obvious traps. Thus, the potential for oil accumulation

in this particular part of the basin is probably low, unlesspre-Alpine traps are preserved.

The outcrop locality (La Salzedella) in which the Ascla

has been found to have residual oil-generation potential is in

a basin sector which suffered elevated erosion rates during

the Paleogene (Fig. 3). It is therefore conceivable that simi-

lar conditions could have existed in the present-day offshore

near Amposta eld. Unlike the Salzedella sub-basin, in this

area, some Miocene subsidence could have caused a second

phase of oil generation from the Ascla Formation. Even in

this case, the effectiveness of the Ascla, as an oil generator,

must have been limited because (1) the organic content of

the Ascla source-rock intervals is relatively low, (2) its oil-generation capabilities could have been partially exhausted

during the Late Cretaceous to Earliest Tertiary, and (3) oil

generation predated formation of most of the traps.

9. Conclusions

Three distinct calcite cement sequences are distinguished

in the Ascla Formation by their CL, uorescence, trace-

element composition and cross-cutting relationships.

(1) Cement sequence 1 precipitated during the Kimmer-

idgian-Valanginian burial stage and mainly lled the

primary porosity. The rst sequence-1 zones precipitatedfrom marine-derived waters. Subsequent sequence-1

calcites precipitated from reducing waters during increasing

burial. Sequence 1 ended with the widespread precipita-

tion of dolomite. This dolomite is currently replaced by

inclusion-rich calcite.

(2) Cement sequence 2 precipitated exclusively in frac-

tures, very likely during the latest Cretaceous to earliest

Tertiary burial when the Ascla Formation reached its maxi-

mum burial depth. Sequence 2 begins with the precipitation

of ferroan calcite, which records oil generation and migra-

tion through the Ascla Formation via fractures. Both the FI


7/29/2019 Rossi_2001

19/20

and isotopic data can be interpreted to indicate that this

calcite records the maximum temperatures reached by the

Ascla, which are around 1178C. Widespread vertical frac-

turing postdated oil migration. The resulting fractures are

mainly lled by non-ferroan calcite yielding geochemical

data compatible with precipitation from marine-inuenced

waters at elevated but potentially decreasing temperatures.

(3) Sequence 3 is formed by alternating ferroan and non-

ferroan calcites, which precipitated exclusively in fractures

during the Tertiary. The rst sequence-3 zones precipitated

during or after the EoceneOligocene stage of compres-

sional tectonics, whereas subsequent sequence-3 zones

likely precipitated during the Neogene. The petrographic

and FI characteristics, and the trace-element and isotopic

compositions of sequence-3 indicate precipitation predomi-

nantly from meteoric waters in the low-temperature (less

than about 508C) phreatic zone. The ferroan calcites of

sequence 3 represent precipitation in relatively closed

systems characterized by low uidrock ratios and reducing

conditions, whereas the non-ferroan subzones represent theingress of oxidizing waters perhaps related to episodes of

fracture opening. Some of the latest sequence-3 zones may

have formed in the vadose zone.

Oil inclusions in sequence 2 provide evidence that light to

medium gravity oils were generated in the Ascla Formation

in the Salzedella sub-basin and migrated through fractures.

Oil migrated before the Eocene, very likely when the forma-

tion reached its maximum burial depth during the Late

Cretaceous to earliest Tertiary. As oil migration predated

the Alpine tectonics, the potential for oil accumulations in

this part of the basin is probably low, unless pre-Alpine

traps are preserved.

Acknowledgements

Funding was provided by research projects DGICYT

PB96-1236-C02-01/02 y PB95-1142-C02-01. Ruben Duro

is acknowledged for his eld work and Alfredo Fernandez-

Larios for his technical assistance with the electron micro-

probe. The helpful comments of Jorge Navarro are also

gratefully acknowledged. Norman Oxtoby and an anony-

mous reviewer are thanked for their careful reviews and

critical comments.

References

Albaiges, J., Algaba, J., Clavell, E., & Grimalt, J. (1986). Petroleum

geochemistry of the Tarragona Basin (Spanish Mediterranean off-

shore). Organic Geochemistry, 10, 441450.

Aurell, M., Mas, R., Melendez, A., & Salas, R. (1994). El transito Jurasico

Cretacico en la Cordillera Iberica: relacion tectonica-sedimentacion y

evolucion paleogeograca. Cuadernos de Geologa Iberica, 18, 369

396.

Bartrina, M. T., Cabrera, L., Jurado, M. J., Guimera, J., & Roca, E. (1992).

Evolution of the central Catalan margin of the Valencia trough (western

Mediterranean). Tectononophysics, 203, 219247.

Burruss, R. C. (1987). Diagenetic palaeotemperatures from aqueous FIs; re-

equilibration of inclusions in carbonate cements by burial heating.

Mineralogical Magazine, 51, 477481.

Burruss, R. C. (1989). Paleotemperatures from FIs; advances in theory and

technique. In N. D. Naeser & T. H. McCulloh, Thermal history of

sedimentary basins; methods and case histories (pp. 119131). New

York: Springer.

Calvet, F., Trave, A., Roca, E., Soler, A., & Labaume, P. (1996). Fractura-

cion y migracion de uidos durante la evolucion tectonica neogena en elSector Central de las Cadenas Costero Catalanas. Geogaceta, 20, 1715

1718.

Friedman, I., & O'Neil, J. R. (1977). Compilation of stable isotope fractio-

nation factors of geochemical interest. In M. Feisher (Ed.), Data of

Geochemistry (6th ed.). U.S. Geological Survey Professional Paper

440KK, (12 pp.)

Goldstein, R. H., & Reynolds, T. J. (1994). Systematics of FIs in diagenetic

minerals, SEPM short course, 31 (199 pp.).

Grandia, F., Asmerom, Y., Getty, S., Cardellach, E., & Canals, A. (2000).

UPb dating of MVT ore-stage calcite: implications for uid ow in a

Mesozoic extensional basin from Iberian Peninsula. Journal of

Geochemical Exploration, 6970 (13), 377380.

Guimera, J., & A lvaro, M. (1990). Structure et evolution de la compression

alpine dans la Chaine Iberique et la Chaine cotiere catalane (Espagne).

Bulletin de la Societe Geologique de France, 6, 339348.Karlsen, D. A., Nedkvitne, T., Larter, S. R., & Bjorlykke, K. (1993). Hydro-

carbon composition of authigenic inclusions; application to elucidation

of petroleum reservoir lling history. Geochimica et Cosmochimica

Acta, 57, 36413659.

Lindholm, R. C., & Finkelman, R. B. (1972). Calcite staining: semiquanti-

tative determination of ferrous iron. Journal of Sedimentary Petrology,

42, 239245.

Lisk, M., Eadington, P. J., & O'Brien, G. W. (1998). Unravelling complex

lling histories by constraining the timing of events which modify oil

elds after initial charge. In J. Parnell, Dating and duration of uid ow

and uid-rock interaction. Geological society special publications 144

(pp. 189203).

Mann, U. (1994). An integrated approach to the study of primary petroleum

migration. Geouids: origin, migration and evolution of uids in sedi-

mentary basins. J. Parnell. Geological Society Special Publication, 78,

233260.

McLimans, R. K. (1991). Studies of reservoir diagenesis, burial history, and

petroleum migration using luminescence microscopy. In C. E. Barker &

O. C. Kopp, Luminescence microscopy and spectroscopy; qualitative

and quantitative applications. SEPM short course notes 25 (pp. 97106).

Meyers, W. J. (1991). Calcite cement stratigraphy: an overview. In C. E.

Barker & O. C. Kopp, Luminescence microscopy: quantitative and

qualitative aspects. SEPM short course 25 (pp. 133148).

Nadal, J. (2000). Dolomas relacionadas con fallas durante la etapa de rift

Jurasico superiorCretacico inferior en la subcuenca de la Salzedella

(Cuenca del Maestrazgo, Cadena Iberica). Geotemas, 1 (2), 247252.

Oxtoby, N. H., Mitchell, A. W., & Gluyas, J. G. (1995). The lling and

emptying of the Ula Oileld: FI constraints. The geochemistry of reser-

voirs. In J. M. Cubitt & W. A. England, Geological Society Special

Publication, 86, 141157.

Permanyer, A., Marl, R., de la Pena, J. A., Dorronsoro, C., & Rossi, C.

(1999). Potencial petrolgeno de la Formacion Margas de Mas D'Ascla

(Jurasico superior) en la Cuenca del Maestrazgo. In L. Aires-Barros,

M. J. Matias & M. J. y Basto, Actas do II Congreso Iberico de geoqu-

mica (pp. 199202). Lisboa: Instituto Superior Tecnico.

Permanyer, A., Salas, R., & Marl, R. (2000). Organic geochemistry and

sequence stratigraphy: an example in the Kimmeridgian-Tithonian from

the Maestrat Basin (Eastern Iberian Chain, Spain). In 31st International

Geological Congress, Rio de Janeiro, Brasil.

Permanyer, A., Salas, R., & Bitzer, K. (2000). Thermal modelling and

geochemical constrains in the Late Jurassic of the Southern Iberian

Chain (NE Spain). In 31st International Geological Congress, Rio de

Janeiro, Brasil.


7/29/2019 Rossi_2001

20/20

Roedder, E. (1984). FIs, Reviews in mineralogy, Vol. 12. Mineralogical

Society of America, Washington, DC, (644 pp.).

Salas, R. (1987). El Malm I el Cretaci inferior entre el Masss de Garraf i la

Serra d'Espada. Analisi de conca. Unpublished PhD thesis, Universi-

dad de Barcelona, (345 pp.).

Salas, R. (1989). Evolucion estratigraca secuencial y tipos de plataformas

de carbonatos del intervalo Oxfordiense-Berriasiense en las cordilleras

Iberica oriental y costero catalana meridional. Cuadernos de Geologa

Iberica, 13, 121157.Salas, R., & Casas, A. (1993). Mesozoic extensional tectonics, stratigraphy,

and crustal evolution during the Alpine cycle of the eastern Iberian

basin. Tectonophysics, 228, 3355.

Salas, R., & Guimera, J. (1996). Main structural features of the lower

Cretaceous Maestrat basin (Eastern Iberian Range). Geogaceta, 20,

17041706.

Salas, R., Martn-Closas, C., Querol, X., Guimera, J., & Roca, E. (1995).

Evolucion Tectonosedimentaria de las cuencas del Maestrazgo y

Aliaga-Penyagolosa durante el Cretacico Inferior. In R. Salas & C.

Martn-Closas, El Cretacico Inferior del Nordeste de Iberia (pp. 13

94). Publicaciones de la Universidad de Barcelona, Barcelona.

Salas, J., Guimera, J., Gimenez-Montsant, J., Martn-Closas, A., & Roca, E.

(1997). Mesozoic rift structure, stratigraphy and Paleogene inversion of

the Maestrat Basin, Iberian Range (NE Spain). Field trip guide, IGCP

project 369, Comparative evolution of PeriTethyan Rift Basins, (37 pp.).

Salas, J., Guimera, J., Mas, R., Martn-Closas, A., Melendez, A., & Alonso,

A. (2001). Evolution of the Mesozoic Central Iberian Rift System and

its Cenozoic inversion (Iberian Chain). Peritethyan rift/wrench basins

and passive margins. W. Cavazza, A. H. F. Robertson & P. A. Ziegler.

Peri-Tethyan Memoirs, 6 (in press).

Seemann, U., Pumpin, V. F., & Casson, N. (1990). Amposta Oil Field. In

E. A. Beaumont, N. H. Foster (Eds.), Structural traps II, traps asso-

ciated with tectonic faulting. AAPG Treatise of Petroleum Geology,

Atlas of Oil and Gas Fields (pp. 1 20).Seifert, W. K., Carlson, R. M. K., & Moldowan, J. M. (1983). Geomimetic

syntesis, structure assignement and geocheminal correlation application

of monoaromatized petroleum steroids. In M. Bjory, Advances in

organic geochemistry 1981 (pp. 710724).

Trave, A., Calvet, F., Soler, A., & Labaume, P. (1998). Fracturing and uid

migration during Paleogene compression and Neogene extension in the

Catalan coastal ranges, Spain. Sedimentology, 45, 10631082.

Vegas, R. (1992). The Valencia trough and the origin of the western Medi-

terranean basins. Tectononophysics, 203, 249261.

Wilkinson, J. J., Lonergan, L., Fairs, T., & Herrington, R. J. (1998). FI

constraints on conditions and timing of hydrocarbon migration and

quartz cementation in Brent Group reservoir sandstones, Columba

Terrace, northern North Sea. In J. Parnell, Dating and duration of

uid ow and uid-rock interaction. Geological Society Special Publi-

cations 144 (pp. 6989).


rossi_2001

Documents