Tải bản đầy đủ - 0 (trang)
Chapter 5.4 Volcanology of the Barberton Greenstone Belt, South Africa: Inflation and Evolution of Flow Fields

Chapter 5.4 Volcanology of the Barberton Greenstone Belt, South Africa: Inflation and Evolution of Flow Fields

Tải bản đầy đủ - 0trang


Chapter 5.4: Volcanology of the Barberton Greenstone Belt, South Africa

early life. Furthermore, traces of bacterial alteration of submarine glass have recently been

discovered in these ancient rocks (Staudigal et al., 2006, for review). Understanding this

Early Archean ecosystem, a volcanic environment, is a crucial step to understanding the

conditions of life’s origin. On a more practical note, reconstructing the architecture of submarine flow fields is an important guide in exploration for volcanogenic ore deposits, for

example, komatiite-hosted nickel deposits (e.g., Hill, 2001; Beresford et al., 2002). Finally,

the ultramafic volcanics have no modern analogues, so we have the opportunity to extend

the principles of modern flow fields and reconstruct these ancient volcanic settings.

Volcanologists who describe the evolution of basaltic flow fields have provided us with

compelling analogues for komatiites (Walker, 1991; Hon et al., 1994; Rossi and Gudmundsson, 1996; Self et al., 1998; Keszthelyi and Self, 1998; Thordarson and Self, 1998).

In subaerial basaltic flow fields, lava spreads as a series of thin flow lobes, which under

some conditions form thick compound flows. In others, the lobes coalesce into sheets that

inflate to thick flows. Flows thicken or inflate when the crust holds incoming lava that lifts

the roof or upper crust. Thick sheet flows record episodes of inflation in textural and vesicle

layering in the upper crust. When flow is focused in channels, inflation rotates the upper

crust of lava channels, forming tumuli, craggy domes and ridges on flow surfaces. When

one area of a flow inflates more than another, the crust breaks and rotates, forming the edge

of a lava rise. When the crust breaks, lava can escape as a break out, a lobe that may flow

over the original flow surface before the flow’s interior has solidified. In addition to inflation by lava influx or endogenous intrusion, flows can also deflate if more lava drains out

of an area of a flow than comes in, forming collapse structures and withdrawal cavities. In

the inflation model, the lava flowing under a thickening crust remains thin and within the

laminar flow regime (versus turbulent) and is increasingly insulated and slowly cooled.

The question here is whether these observations of subaerial flow fields can explain features of submarine komatiites exposed as cross sections in greenstone belts. Basalt erupted

underwater cools rapidly as pillows, and the added pressure inhibits vesiculation, so pillowed flows and sparse vesicles distinguish submarine flows. Beyond these superficial

effects of water, the submarine environment plays a diminishing role because as the upper crusts of sheet flows thicken, lava is increasingly insulated and free to travel across

submarine lava plains the great distances that characterize komatiite flow fields.



Volcano-sedimentary sequences of the BGB record over 400 million years of crustal

evolution, a history punctuated by three TTG volcano-plutonic events and three phases

of deformation (de Ronde and de Wit, 1994; Lowe and Byerly, 1999). The volcanosedimentary component has three stages: (1) an early stage dominated by mafic and ultramafic, submarine volcanism of the Onverwatch Group, (2) the basin-filling sequence of

the Fig Tree Group, and (3) the unconformably overlying Moodies Group, molasse deposits associated with fold-and-thrust deformation (Fig. 5.4-1). When the 3.445 Ga TTG

5.4-2. Tectono-Volcanic History of Barberton Greenstone Belt (BGB)


Fig. 5.4-1. Geologic map of the southern Barberton Greenstone Belt, showing location of stratigraphic columns and maps.

suite intruded and erupted, the crustal section was thickened, deformed, and eroded near

sea level, a pivotal event that separates the Onverwatch Group volcanics into three distinct groups: (1) the Sandspruit-Theespruit section that hosts 3.445 Ga batholiths, (2) the

Komati–Hooggenoeg section, basement for the 3.445 Ga felsic volcanics and host to numerous subvolcanic sills and dikes, and (3) the Kromberg–Mendon volcanics that started

about 30 million years after the TTG event. This gap in volcanism is filled by the Buck

Ridge Chert, a 150–350 m thick sequence of silicified sediments that overlies the volcanic

expression of 3.445 Ga TTG intrusions, the Buck Ridge Dacite. This paper describes the

mafic and ultramafic, submarine lavas of the Komati–Hooggenoeg and Kromberg–Mendon

sections. Our work focuses primarily on the Komati Formation (e.g., Dann, 2000, 2001;

Grove et al., 1997; Parman et al., 1997, 2001, 2004; Parman and Grove, 2004), with lesser

work in the Hooggenoeg and Kromberg Formations.

The volcanic stratigraphy of submarine lavas divides into three magma types: basalt,

komatiitic basalt, and komatiite (Fig. 5.4-2), each with distinct flow morphologies and geochemical and petrological characteristics. Komatiite volcanism repeats in three formations

in association with komatiitic basalt at ca. 3.48, 3.46, and 3.33 Ga. The dominant stratigraphic trend within the Onverwatch Group is an increasing frequency of chert horizons


Chapter 5.4: Volcanology of the Barberton Greenstone Belt, South Africa

Fig. 5.4-2. (Next page.) Ages, thickness, rates, and composition of volcanism in the Onverwatch

Group’s Komati, Hooggenoeg, Kromberg, and Mendon formations: (A) The histogram of lava types

– basalt, komatiitic basalt, and komatiite – is plotted by stratigraphic thickness of members or formation. Kromberg and Mendon data from Lowe and Byerly (1999); (B) The rate of volcanism is

average over entire formation and does not include tectonic, stratigraphic, or analytic uncertainties.

U/Pb zircon ages are tabulated in de Ronde and de Wit (1994) and Lowe and Byerly (1999).

within the volcanic formations. The cherty sediments represent gaps in effusive volcanism

when the sediments were silicified on the ocean floor. The alteration process penetrated

the underlying volcanic rocks, producing silicified pillows and even converting komatiites into cherts and/or carbonates. Along with more chert and alteration, the duration of

the formations also increases. The oldest volcanic section – the Komati Formation – has

no cherty sediments or alteration beyond greenschist metamorphism: the komatiites are

serpentinites, and the komatiitic basalts are greenschists, both without penetrative fabrics.

Based on available zircon ages for a 2.3 km section, the Komati Formation erupted during

a 10 million year period between 3481 ± 2 Ma and 3472 ± 5 Ma for a rate of 0.25 km per

million years. In contrast the 2.7 km thick Hooggenoeg Formation with 4–5 chert layers

and seafloor alteration zones may record up to 27 million years of volcanism between 3.47

and 3.445 Ga – 0.1 km per million years. The Mendon Formation with the most cherts and

seafloor alteration may represent only 0.03 km per million years. In summary, the younger

formations record longer periods with less volcanism, presumably due to longer gaps between volcanic events (as opposed to actual rates of eruption). Although more work is

needed to rigorously constrain the timing of volcanism, the ultramafic formations erupted

at higher rates within each pair.


The volcanological component of research in the BGB has had three hurdles: (1) distinguishing the volcanic stratigraphy from intrusions, especially mafic and ultramafic sills,

(2) defining flow fields in vertical dipping sequences despite deformation and locally extreme alteration, and (3) seeing the dynamics of flowing lava recorded by discontinuous

outcrops. With interesting variations, the same problems occur in each volcanic formation,

and our research focuses on the Komati Formation, primarily because of its exquisite exposure of the oldest, well-preserved komatiites on Earth. Their composition extends the field

of volcanology into ultramafic realms with no modern analogues.

Submarine flows of spinifex-textured komatiite (>23% MgO) erupted at least three

times over 180 millions years from the ca. 3.48 Ga Komati Formation and overlying

Hooggenoeg Formation (<3.445 Ga) to the 3.3 Ga Mendon Formation (Fig. 5.4-2). This

repeated volcanic history within one crustal section implies that we might find equivalents

of the Mendon komatiites intruding the underlying formations and raises the question: Can

we distinguish massive komatiite flows in the Komati Formation from komatiite sills that

5.4-3. Volcanology of the Barberton Greenstone Belt



Chapter 5.4: Volcanology of the Barberton Greenstone Belt, South Africa

may be 130 million years younger and equally serpentinized? During this interval, the

crustal section evolved tectonically, especially during intrusion of ca. 3.445 Ga tonalite

batholiths, so we might expect the komatiite magmatic suites to reflect those changes. All

three magma types – basalts, komatiitic basalts, and komatiites – have the same possibilities for creating a complex, evolving magmatic architecture that we can unravel with

detailed mapping, coupled with geochemical and petrologic analysis. Confusion in both

the Komati and Kromberg Formations has centered on ultramafic sills – are they flows,

subvolcanic sills, or much younger intrusions unrelated to the volcanic stratigraphy?

In the Komati Formation the volcanic stratigraphy is intruded by a suite of wehrlite

dikes and sills, black serpentinites that locally blend with massive komatiite flows. Previous

researchers considered these wehrlites to be feeder dikes for the komatiite flows (Viljoen

et al., 1983; de Wit et al., 1987b), but the following features distinguish this intrusive

suite from the komatiites: (1) cross-cutting contacts, (2) co-intrusive gabbroic and diabasic

phases, (3) coarse-grained, poikolitic pyroxene and igneous hornblende, and (4) higher

TiO2 . This suite intrudes the overlying Hooggenoeg Formation but with less outcrop-level

ambiguity because the Hooggenoeg has no black, serpentinized komatiite lava flows. This

suite also intrudes some tonalite sills, a fact that indicates this intrusive suite is at least 30

million years younger than the komatiites in the Komati Formation. This widespread suite

includes several layered intrusions, and a comprehensive study including geochronology is

needed to understand its role in the evolution of the crustal section and to test the possibility

that a single magmatic event links structural blocks or stitches terrains.

In the second challenge, we need to divide the volcanic stratigraphy into distinct units

or flow fields. A flow field is an eruptive episode of geochemically coherent lava, a batch

of mantle melt that has undergone the same AFC processes, possibly to varying degrees.

Therefore, when reconstructing flow fields, we must exclude sill complexes, divide the

stratigraphy into compositional groups, and recognize any volcanic hiatus marked by sediments or horizons of seafloor alteration or weathering. In the Komati Formation, alternating

layers of komatiite and komatiitic basalt define distinct flow fields. In addition to compositional grouping in the Hooggenoeg Formation, distinct episodes of volcanic activity are

bound by gaps marked by chert beds and severe alteration of underlying volcanic rocks.

In the Kromberg Formation, a pyroclastic deposit of komatiitic basalt – possibly erupted

subaerially – is overlain by submarine basalt flows. In this case, both the composition and

style of eruption define distinct volcanic episodes. Since modern volcanism varies with

its tectonic environment, we assume that the geochemical evolution of flow fields of the

Onverwatch Group records the early tectonic evolution of the crustal section.

Besides the outcrop scale differences – komatiites are black, basalts are green – geochemistry is essential for defining flow fields. The volcanic rocks of Barberton divide into

three main groups – komatiites, komatiitic basalts, and basalts, listed in order of the number of available analyses. Komatiites and komatiitic basalts are strongly fractionated within

flows: olivine phenocrysts settled to form basal cumulates. As a result komatiitic basalts

with erupted liquid compositions around 15% MgO fractionated to make rock compositions with 6 to 22% MgO, a compositional range that suggests a continuum of compositions

between the basalt and komatiites. However, the gap in liquid compositions between ko-

5.4-4. Komati Formation


matiitic basalts (15% MgO) and komatiites (25% MgO) is real, making their association

truly bimodal in Barberton. In the larger context of all greenstone volcanics, basalts and

komatiites may be end members in a compositional continuum of mantle melts, but in

many greenstone belts, komatiites and komatiitic basalts occur together repeatedly with a

broad gap in parental compositions. Besides the bimodality, the komatiite-komatiitic basalt

association shows overwhelming evidence for fractionation within flows and very little for

pre-eruptive fractionation in magma chambers.

The third challenge – seeing the dynamics of flowing lava – attempts to move modern

volcanology into the Archean. The flow field concept provides an important perspective

while we interpret the cross sections of steeply dipping volcanic stratigraphy, the map view

in most orogenic belts. First, a flow starts at the vent and ends at its distal reaches, so lava

has passed through almost every cross section – our map views – whether we see evidence

of flow or not. In the old view, spinifex flows were viewed as closed systems of ponded lava

(e.g., Silva et al., 1997), but now they are understood to have been lava conduits that record

the flux of lava and inflation/deflation in their textures and zoning. Similarly, sheet flows

of komatiitic basalt have pyroxene-spinifex layers and mingling features, both recording

inflation of lavas in that compositional range. The contacts between pillowed and massive

basalt flows reveal complex systems of lava channelized in tubes with levees of pillows,

pillowed flows with prograding foresets, flows with in situ flow-top brecciation, and other

structures that freeze and record lava flow dynamics. What is interesting about flow fields

in submarine lava plains is how flows create topography that determines the course of

successive flows, a self-organizing system that we just start to appreciate in mapping some

of the best outcrops.


The Komati Formation has two members: (1) a lower member of alternating layers

of komatiite and komatiitic basalt, and (2) an upper member dominated by pillowed and

massive flows of komatiitic basalt (Fig. 5.4-3(A)). In the Lower Komati, five layers of

komatiite share the same zoning – spinifex komatiite overlying massive komatiite – yet

each layer is geochemically distinct (Fig. 5.4-3(B)). The repeated pattern – spinifex over

massive – along with the changing geochemistry defines komatiite flow fields within the

Komati Formation (Fig. 5.4-3(C)). Similar to komatiites, the flow fields of komatiitic basalt

are geochemically distinct and have two flow types – pillowed and massive – but they

do not show a repeated pattern of pillowed over massive flows (or visa versa). Thus, we

consider all the major layers in the Lower Komati to be distinct flow fields, defined by the

stratigraphy, flow morphology, petrology, and geochemistry. These flow fields represent

distinct batches of mantle melt because low-P, olivine fractionation can not account for

the difference in Al2 O3 /TiO2 . The Upper Komati is not as continuously exposed as the

Lower Komati, it lacks stratigraphic markers, and geochemical coverage is too sparse, so

we cannot divide this section of komatiitic basalt into flow fields without further work.


Chapter 5.4: Volcanology of the Barberton Greenstone Belt, South Africa

5.4-4. Komati Formation


Fig. 5.4-3. The Komati Formation: (A) Stratigraphic column of the Upper and Lower Komati Formation with komatiite flow fields labeled #1–5, solid brackets; (B) Plot of Al2 O3 /TiO2 vs MgO

distinguishes flow fields for both magma types and reveals shared trend; (C) Longitudinal section of

komatiite flow field #2 with massive, vesicular, and spinifex komatiites.

The Komati Formation only contains volcanogenic sediments produced directly from

endemic volcanic processes with one exception: the Komati Tuff, a dacitic airfall tuff

preserved between lava flows (Fig. 5.4-4). Several thin beds grade from greywacke to

porcelainite, and the greywacke contains volcanic fragments and crystals of quartz and

plagioclase. A single zircon age of 3481 ± 2 Ma (S. Bowring, pers. comm.) records a

nearby felsic volcanic event and provides the best age for Komati volcanism.

5.4-4.1. Lower Komati Formation: Komatiite Flow Fields

With the Lower Komati neatly divided into flow fields, we can focus on the architecture

of komatiite flow fields by examining the morphology and relationship between the three

flow types: massive, vesicular, and spinifex. The most important volcanological discovery

of our mapping is probably the thick vesicular tumulus (Fig. 5.4-5) because its textural

zones illustrate how channelized komatiite lava inflated to 50 m thick, creating relief on

the seafloor and potentially transporting lava long distances from a vent under water (Dann,

Fig. 5.4-4. Polished slab of the Komati Tuff, showing graded beds of dacitic airfall tuff, overlain by

flow of komatiitic basalt (black).


Chapter 5.4: Volcanology of the Barberton Greenstone Belt, South Africa

5.4-4. Komati Formation


Fig. 5.4-5. Thick vesicular tumulus or lava tube: (A) Outcrop map showing cumulate, spinifex, and

vesicular zones of 50 m thick, lenticular komatiite; (B) Thin section of a segregation vesicle, showing

a serpentine-filled compound vesicle (top) and segregation (bottom) of augite and glass; (C) Model

based on restoring rotated contacts to horizontal (1). Influx of lava rotated thick crust (2) and spinifex

zone crystallized (3).

2001). Vesicular komatiite occurs discontinuously within flow field #2 (Fig. 5.4-3(C)) between overlying spinifex flows and underlying massive flows and, in total, within three of

the five komatiite flow fields. Although vesicular komatiites are volumetrically rare, they

record stages in the emplacement of komatiite flows and link the massive and spinifex

flows within single komatiite flow fields.

5.4-4.1.1. Vesicular flows

The thick vesicular tumulus has three extraordinary textural zones: a lower olivinecumulate zone, a central olivine-spinifex zone, and an upper vesicular zone (Fig. 5.4-5(A)).

The cumulate, an orthocumulate dunite, consists of tabular olivine crystals (1 cm), tightly

packed to form an originally horizontal magmatic foliation. The interior spinifex zone

crystallized fist-sized, skeletal olivine with interstitial augite and glass. The vesicular zone

consists of large serpentine-filled segregation vesicles (Fig. 5.4-5(B)) that are distinctly

elongate near the lower contact, parallel to a cryptic vertical fabric of skeletal olivine.

Forming the tumulus’ lenticular shape, the vesicular carapace broke into blocks and rotated 35–45◦ outward, rotating also the originally vertical, linear fabric of vesicles. The

lenticular cavity is filled by the spinifex zone, indicating that it crystallized after the rotation of the upper carapace. As shown in the model (Fig. 5.4-5(C)), the flow thickened in

two stages: (1) vesicles accumulated beneath a downward crystallizing roof, and (2) the

roof broke and rotated with the influx of lava, forming the domed cross section or tumulus.

The tumulus is probably the cross section of a lava channel or tube that focused flow

like a river along a topographic low (the lower contact is not exposed). Long-lived flow

within a conduit explains the exceptional features of this tumulus. First, an extensive lavatube system was needed to transmit enough hydraulic pressure downstream to inflate the

20 m thick crust and form the 50 m thick tumulus. Second, the long-lived flux of lava

within a tube facilitated the growth and accumulation of both exceptionally large olivine

phenocrysts and vesicles in, respectively, the cumulate and upper crust. The 20 m thick

vesicular crust effectively insulated lava from seawater, potentially allowing a lava tube

of this size to rapidly deliver large volumes of lava long distances from the vent, prior to

crystallization of the spinifex zone. Similar lenticular units, thick and vesicular, occur in

the younger Kambalda komatiites of Australia, which Stolz and Nesbitt (1981) interpreted

as lava channels, key arteries within komatiite flow fields. What makes the tumulus exceptional is how rotation of the upper crust is recorded in textures and contacts, well exposed

in outcrops of the Komati Formation (Fig. 5.4-5).

Along with the thick tumulus in flow field #2, a vesicular sheet flow has inflation structures exposed in every outcrop for several km along strike (Fig. 5.4-3(C)). Locally, the


Chapter 5.4: Volcanology of the Barberton Greenstone Belt, South Africa

Fig. 5.4-6. Vesicular sheet flow: (A) Outcrop map of tilted flow top (x) and breccia with blocks

of spinifex flow with glassy flow base (y) and top (z); (B) Flow top and blocks restored to

horizontal; (C) Model of domino-style rotation with tilting; (D) Outcrop map showing vesicle-spinifex-layered crust intruded by cumulate B-zone lava, linking chilled, cross-cutting contacts

with cryptic, layer-parallel, intrusive contacts; (E) Possible drill core view of upper crust in (D) with

cryptic intrusive contacts; (F) Model of sheet-flow inflation: initial crust-forming stage (1), intermediate stage of differential inflation and breakout of flow lobes (2), and final stage that rotates flow

lobes in thick breccia (3).

upper crust is intruded by dikes from the flow’s interior, tilted 35◦ , and overlain by a thick

breccia (Fig. 5.4-6(A)). The breccia fills fissures that cut through the flow top into the vesicular crust. The breccia contains blocks of a spinifex flow with its glassy flow top, spinifex

zone, and lower cumulate zone, blocks that can be reassembled (Fig. 5.4-6(B)), indicating

that the spinifex flow fractured and rotated like dominoes as the crust tilted (Fig. 5.4-6(C)).

The lower part of the crust is intruded by a network of dikes emerging from the flow’s

lower cumulate zone (not shown here, see Dann (2000), Fig. 5), indicating that massive

komatiite remained mobile beneath the insulating crust, behavior not usually associated

with a cumulate.

Where cumulate komatiite intrudes the upper crust (Fig. 5.4-6(D)), the chilled margins

of dikes are well formed near the flow top (x) and progressively disappear downward (y), a

feature that illustrates (1) how cryptic intrusive contacts pose as layering and (2) how dikes

might fractionate lava. Without chilled margins or cross cutting contacts (e.g., ‘z’), layerparallel intrusions of this crystal-rich lava can not be distinguished from crustal layers or

zones, a problem that plagues research limited to drill core (Fig. 5.4-6(E)) and may explain

vesicular and massive zones, repeatedly layered in thick Kambalda vesicular komatiites

described by Beresford et al. (2000). Even thin sections of these cryptic contacts are

ambiguous. Furthermore, with massive komatiite making up 62% of the volcanic stratig-

Tài liệu bạn tìm kiếm đã sẵn sàng tải về

Chapter 5.4 Volcanology of the Barberton Greenstone Belt, South Africa: Inflation and Evolution of Flow Fields

Tải bản đầy đủ ngay(0 tr)