IODP Proceedings    Volume contents     Search


Igneous petrology and volcanology

The procedures for core description outlined here are essentially those used during Expedition 324 to Shatsky Rise (Expedition 324 Scientists, 2010). The goal of our shipboard studies was to produce an integrated picture of each drill site’s style of volcanism and environmental setting. This goal was achieved by systematically describing rocks and identifying key effusive, volcaniclastic, and igneous textures known to be diagnostic of specific modern physical volcanic processes. Physical description of the volcanic rocks and deposits recovered during Expedition 330 required a multistage process. First, lithologic unit boundaries were defined by visual identification of actual lithologic contacts or by inference of such contacts using observed changes in composition (e.g., phenocryst assemblages), volcanic characteristics, or volcaniclastic features. Then, lithology, phenocrysts, lithologic variation, and characteristic igneous textures and vesicle distributions were described. Finally, these macroscopic observations were combined with those from detailed thin section petrographic studies of key igneous units.

Physical characteristics of volcanic units

Subaerial lava eruptions

Aʻa lava flows

Aʻa flows are characterized by angular, spinose clinker at both flow tops and bottoms, in contrast to pahoehoe and pillow flows, which are often characterized by smooth surfaces and clear contacts between successive or adjacent units. Internal disruption during emplacement permits efficient degassing, and the crust, core, and base of aʻa flows are typically sparsely vesicular or nonvesicular. Transitional flow types (e.g., rubbly, slab, or toothpaste pahoehoe; Guilbaud et al., 2005) show some of the characteristics of both pahoehoe- and aʻa-type flows.

Pahoehoe lava flows

Pahoehoe flows are usually vesicular, often highly so, and can exhibit bulk porosities in excess of 20 vol% (typical range = 20–60 vol%). Thin pahoehoe lobes (<0.5–1 m) are often vesicular throughout and exhibit gradual coarsening in vesicle size from the lobe margins toward the interior (Wilmoth and Walker, 1993). During development of subaerial pahoehoe lava fields, gas bubbles can frequently become trapped and entrained in the magma body as it cools from both above and below. Thus, thicker inflated pahoehoe lobes, such as those documented in continental flood basalt provinces, are often characterized by a threefold structure of vesicular upper crust, dense core, and thinner vesicular lower crust (Aubele et al., 1988; Thordarson and Self, 1998). Very close morphological similarity exists between pahoehoe lava formed on land and pillow lava formed under water, though the former typically lacks, or else has only relatively thin, glassy rinds (i.e., glassy surface selvages).

Subaqueous lava flows

Subaqueous lava flows and hyaloclastite associations have been documented where flows from land-based eruptions entered water (e.g., into seas or freshwater lakes) or else erupted in shallow-submarine situations (e.g., ongoing eruption from Puʻu Oʻo, Hawaii; Surtsey eruptions of 1963–1967; and around the coasts of the Azores). On entering water or soft unconsolidated sediments, the eruptive style changes fundamentally, and depending on local conditions advancement of the lava front typically generates pillow basalt, hyaloclastites, or peperites (or any combination of these three). Hyaloclastites and peperites are volcaniclastic sediments that form by quench fragmentation (Carlisle, 1963). Pillow basalts are a common product of these environments and can usually be distinguished from pahoehoe flows by their much thicker glassy rinds (a consequence of more efficient aqueous quenching) and by the presence of lacustrine/marine sediments preserved between the pillows. However, although less frequently documented, larger subaqueous flows may also be massive and sheetlike if the local flow rate is high enough, as determined by slope and eruption rate (Walker, 1992; Gregg and Fink, 1995).

Submarine lava eruptions

Pillow lava flows

Pillow lavas are subaqueous extrusions resulting from individual budding, inflation, and separation of lava pods from point sources along the advancing lava front. They can accumulate in a variety of styles depending on effusion rate, number of point sources, and internal plumbing architectures within the growing volcanic edifice. However, pillow lavas typically consist of discrete subrounded units of relatively small size (~0.2–1.0 m diameter). Characteristically, their exteriors are entirely bounded by glassy rinds as a result of rapid cooling, and their interiors typically display internally radiating vesicle and joint patterns.

Lobate lava flows

Larger lobate flows or flow lobes (~1–2 m diameter) can also develop by the same endogenous inflation process as pillow lava flows. Although these extrusions broadly resemble pillow lavas, they differ in that they have massive, coarser grained, and sparsely vesicular cores, often with pipe vesicle domains. Because their larger size permits a slower rate of cooling, these inflation units are characterized by more effective degassing and maturation of internal vesicle patterns. Accordingly, vesicle zoning is more concentrated in the upper regions of the inflation unit and often occurs as a series of vesicle bands that develops through the inward migration of the cooling front. By contrast, the lower part of the inflation unit typically contains either sparse, poorly defined vesicle banding or teardrop-shaped vesicles at, or just above, the basal chill zone. Rather than being spherical or subspherical, these massive inflation units tend to have flattened, oblate, or tubular shapes and are most probably generated as a series of interconnected, semiadjacent inflation units fed simultaneously by internal plumbing. Their size and morphology are considered to be indicative of higher effusion rates than those that generate pillow lavas. Individual units can cover an area of several square meters (or significantly more where adjacent units merge either through internal plumbing or as a result of rupture and coalescence). In those instances where they merge into a single entity, they appear as a lava flow displaying a dominantly botryoidal surface morphology. Recognizing lobate flows in drill core is often very difficult or impossible, especially in the thick volcaniclastic successions drilled during Expedition 330. The terms “lava fragment” or “flow lobe” were used for >30 cm thick continuous intervals of basalt when features diagnostic of lobate flows were lacking.

Massive basaltic lava flows

Massive basaltic lava flows form thick (several meters or more), laterally extensive sheetlike units (van Andel and Ballard, 1979). They have a variety of surface features generated by deformation and disturbance of the solidifying lava crust while in a plastic/semiplastic state. In dimension and internal features they can resemble larger subaerially erupted flows. This type of flow may be fed by lava tubes and develop by internal inflation or marginal budding. These flows often have texturally uniform massive cores as thick as several meters, are characterized by sparse vesicle layering, or have vertical vesicle pipes containing late-stage melt segregation material. Massive or sheetlike flows may be a response to particularly high effusion rates or local slope and other controlling topographic conditions. These unit types have been called “sheet flows” in previous ODP and IODP reports, although this term implies a lateral extent that cannot be determined from core alone. Nevertheless, where recovered, these units are significantly thicker than the normal (~0.2–2 m) dimensions of pillow lava or lobate flows observed in drill cores.

Volcaniclastic deposits

Volcaniclastic deposits include a range of materials from rubbly, in situ volcanic debris to resedimented material such as volcanic sand and gravel. Volcanic material of all sizes may be a direct product of eruptive processes (pyroclastic) or an accumulation of processes involving transport, sorting, and deposition (epiclastic). Pyroclastic activity includes hydrovolcanic deposits formed by explosive interaction between magma and water, as well as nonexplosive quench fragmentation processes (e.g., hyaloclastite and peperite).


Hyaloclastite literally means “glass-fragment rock” and includes all glassy fragmental debris formed by eruptions involving water. Environments of hyaloclastite production relevant to the Louisville Seamounts may include shallow-water phreatomagmatic eruptions when the seamounts were emergent volcanic islands and, most probably, volcaniclastic materials associated with the development and advancement of submarine lavas at greater water depths (Clague et al., 2009). Phreatomagmatic conditions can result in the fragmentation of gas bubbles inflated by steam, producing thick layers of glass shards that may be sent into subaerial trajectories or else form suspended plumes within the water column. In either case, they then settle upon the surrounding land or seafloor surface (i.e., on newly formed volcanic material) or are transported and redeposited elsewhere. Depending on the nature of the eruption, welded and unwelded volcanic rubble and breccia may also occur in these environments (Fisher and Schmincke, 1984). Submarine lavas are almost invariably associated with hyaloclastites and a variety of other fragmental volcanic debris, including breccia (Staudigel and Schmincke, 1984; Batiza and White, 2000). Hyaloclastite is common on seamounts, where it may be the product of cooling contraction granulation and surficial spalling of lava flows. Seamounts are also characterized by volcanic breccia and are commonly associated with mass debris flows, especially where the eruption occurs on significant seafloor topography.


Peperite is a distinct volcanic sediment facies occurring where subaqueous basaltic lava flows interact with unconsolidated sediment as they erupt into water bodies on land or, more commonly, on the seafloor (Skilling et al., 2002). The mingling of basaltic lava and wet sediment produces distinct volcanic textures resulting from the physical interaction of lava and sediment (entrainment, baking, chilling, etc.), as well as physical and chemical fragmentation and alteration through steam-rock interaction during flash heating of the water. Common features include quenched lava flow margins of basaltic clasts surrounded by sediment, as well as basaltic clasts with a variety of fragment morphologies, including fluidal (e.g., where fragments remained plastic enough to deform after deposition) or jigsaw-fit (e.g., where fragments cracked apart in situ, forming blocky peperite). The presence and texture of peperite may provide important information regarding the nature of the eruptive environment. Importantly, the occurrence of peperite provides evidence that the volcanism and affected sedimentary environment existed contemporaneously.

Other lava-sediment interactions include features involving the mixing or mingling of molten or plastic lava with sediment (e.g., where lava injects or intrudes into a substrate). In some instances, there is a lack of bedding or lamination in sediment adjacent to pillow lavas or the material surrounding lava injections. A likely explanation is fluidization of the fine-grained sediment, attributable to flash heating and vaporization of sediment pore water during emplacement of the pillow lava (Kokelaar, 1982). Such momentary fluidization causes sediment reconstitution, localized transport, and redistribution and hence the destruction of any inherent bedding structures.

Definition of lithologic units and volcanic successions

In recording our observations of the drill core, it was important to avoid nomenclature or terminologies that imply particular styles of basaltic volcanism. Most lava type classifications are based on field observations that consider both lateral and vertical variations of eruptive units and their stacking relationships. Such refined classifications are neither realistic nor justifiable from core observation alone and can only become achievable when additional petrographic, geochemical, or downhole logging data are available. In most instances, only the smaller volcanic unit divisions (i.e., individual cooling units) can be identified in core material, so these are necessarily the basic units of description.

Descriptive nomenclature

Volcanic succession

To aid efficient core description and enable later collation of the observed volcanic elements into volcanically meaningful successions, we adopted a simple twofold numerical hierarchy. First, to define lithologic units we looked for lava contacts, chilled margins or identifiable flow tops, changes in phenocryst populations, vesicle distributions, and intercalated volcaniclastic or sedimentary horizons. When considered together, these features typically define the key volcanic cooling, inflation, and depositional units of various sizes and the scales within the eruptive succession. Lithologic units are given consecutive downhole Arabic numerals (i.e., lithologic Units 1, 2, 3, etc.), irrespective of whether they are pillows, massive flows, volcaniclastic deposits, or igneous intrusions. This procedure, therefore, provides a nongenetic cataloging system.

Stratigraphic units were defined where successions of consecutive cooling, inflation, or depositional units of similar or shared evolutionary characteristics could be identified, usually on the basis of phenocryst type. In effect, these stratigraphic units combine similar types of eruptive products and provide a first step toward considering the volcanic stratigraphy in terms of packages of petrologically and genetically related lithologic units and, ultimately, the evolving volcanism. These packages are given consecutive downhole Roman numerals (e.g., stratigraphic Units III, IV, V, etc.) that follow consecutively from the overlying sedimentary units (e.g., Units I and II). Although we made every effort to ensure that the identified stratigraphic unit boundaries reflect individual lava packages or eruptive units, the term “boundary” should not be considered synonymous with an individual lava flow or single eruptive event but rather as indicating a division of core into elements or sections displaying broadly similar volcanic and petrologic characteristics.

The best evidence for dividing core into lithologic units is the presence of flow/flow contacts. Unfortunately, these are not always preserved in low-recovery sections. Alternatively, the presence of glassy material or finer grained chill zones can be a useful proxy for determining the presence or estimating the position of lithologic unit boundaries. However, grain-size changes close to flow tops are often no greater than those within inflated lava flows, so interpretation of grain-size changes in core can prove difficult. Changes in vesiculation that occur through inflation or cooling units of all sizes are of particular use in sections of poor core recovery. To identify and estimate the position of lithologic unit boundaries using this method we paid special attention to vesiculation characteristics by measuring, estimating, and otherwise recording (1) the percentage volume of vesicles, (2) their modal size range, (3) their shape and sphericity, and (4) changes in vesicle size (i.e., fining- or coarsening-upward) at intervals appropriate for the variability shown in the core (typically 10–100 cm intervals). The volume fraction of vesicles within the unit or section under consideration, together with the modal shape and sphericity of vesicle populations, was estimated using appropriate comparison charts similar to those used by sedimentologists and derived from the Wentworth (1922) classification scheme (Fig. F5). Maximum and modal sizes were determined using a hand lens with a submillimeter graticule. It proved impractical to measure the minimum size because this was always at or below the limit of visual resolution.


Lithologic descriptions followed accepted conventions. Porphyritic basaltic rocks were named according to major phenocryst phase(s), but only where the total abundance of phenocrysts was >1% (e.g., plagioclase-phyric basalt or olivine-augite-phyric basalt). Phenocryst phases were always listed in order of decreasing abundance so that the dominant phase was listed first. The term “phenocryst” was used for any crystal that was (1) significantly (typically five times) larger than the average size of the groundmass crystals, (2) >1 mm, and (3) euhedral or subhedral. The term “microphenocryst” was used for crystals larger than the modal groundmass grain size but smaller than 1 mm. Where macroscopic observation was possible, these microphenocrysts were also described in the DESClogik application under a suitable heading in the phenocryst table. Aphyric rocks were not assigned any mineralogical modifier.

Volcaniclastic deposits

We used “volcaniclastic” as a nongenetic term for any fragmental aggregate of volcanic parentage containing >60% volcaniclastic grains and <40% other types of clastic or biogenic material. This definition is necessarily broader than that for pyroclastic deposits because the term “pyroclastic” strictly applies only to products of explosive volcanic activity, including hydroclastic deposits formed by explosive interaction between magma and water/sediment and nonexplosive quench fragmentation (i.e., hyaloclastite and peperite). Our adopted definition may also include the volcanic detritus produced by erosion or transport of volcanic material (i.e., epiclastic sediment). Accordingly, the term “volcaniclastic” does not necessarily imply active volcanism at the time of deposition.

In fragmental lava facies we documented volcaniclastic textures using standard sedimentologic techniques (e.g., clast versus matrix modal proportions, clast size, shape, sorting, and lithology). Of particular interest was the presence (or absence) of features diagnostic of viscous or quench fragmentation while the lava was hot. These include clasts engulfing fragments of other (earlier) clasts, welding, glassy margins surrounding the clasts, extensions of the interior of the flow into the breccia, entrained clasts within the interior of the flow, and the presence of basal breccia. We used changes in clast morphology and crystallinity as well as changes in vesicle abundance and shape to identify clast types within breccia.

Our subclassification of volcaniclastic sediment differs from the standard ODP classification (Mazzullo et al., 1988) in that we adopted a descriptive (nongenetic) terminology similar to that employed during Leg 197 (Shipboard Scientific Party, 2002). We simply described volcaniclastic sediment according to the classification scheme for clastic sediment, noting the dominance of volcanic grains. We followed the clastic textural classification of Wentworth (1922) to separate the various volcanic sediment types and sedimentary rocks (according to modal grain size) into the following:

  • Volcanic gravel (>2–8 mm; equivalent to fine gravel on the Wentworth scale),

  • Volcanic sand (>0.0625–2 mm),

  • Volcanic silt (0.0039–0.0625 mm), and

  • Volcanic clay (<0.0039 mm).

Volcaniclastic rocks with grain sizes in the silt and clay ranges were not found in any of the basement successions drilled during Expedition 330. For coarser-grained (modal grain size >8 mm) volcaniclastic rocks with angular clasts we applied the term “breccia” and used lithologic or genetic modifiers (e.g., volcanic, basalt, and hyaloclastite) for further description. Thus, we divided the volcaniclastic rocks into volcanic sandstone (0.0625–2 mm), gravel-size volcanic breccia (>2–8 mm), and volcanic breccia (>8 mm). For simplicity we used the expressions “volcanic sand” and “volcanic gravel” on the stratigraphic columns. We further classified volcaniclastic rocks by adding additional modifiers to the designated principal name. Relative proportions of vitric (glass), crystal (mineral), and lithic (rock fragment) components (where the proportion exceeds 5%) were used to determine these modifiers. For example, a volcanic sandstone composed of 75% glass, 5% feldspar crystals, and 20% lithic fragments would be named “vitric-lithic volcanic sandstone.”

Core and thin section descriptions

The first step in describing the recovered core was identifying unit boundaries on the basis of lithologic changes, color, grain size, the presence of volcaniclastic or sedimentary intercalations, volcanological features (e.g., presence of contacts or chilled margins), vesicle distribution patterns, changes in primary mineralogy (occurrence and abundance of plagioclase, augite, and olivine), structure, alteration, and veins. Unit boundaries of volcanic rocks were chosen to reflect different volcanic cooling or volcaniclastic units. In some cases, limited recovery required interpolation to position unit boundaries. Of fundamental importance to core interpretation was the additional use of digital image printouts of the archive core halves. During core processing, these printouts are generated from the archive half of the split core using the digital SHIL. Annotation of these printouts provides a working record onto which observation notes, sketches of key features, or indication of shipboard sampling or analytical requirements can easily be made. This visual record also proved invaluable during the preparation of site reports. These handwritten records were later scanned and are provided as supplementary data to this volume (see SCANS in “Supplementary material”).

Macroscopic visual core description

We used the DESClogik application to document each section of the igneous rock cores and upload these descriptions to the LIMS database (see DESClogik user guide). For consistency, our procedures and database templates closely followed the methods for core descriptions from recent ODP and IODP volcanic basement expeditions, in particular those to the Hawaii-Emperor Seamounts (Leg 197; Shipboard Scientific Party, 2002) and Shatsky Rise (Expedition 324; Expedition 324 Scientists, 2010).

VCD standard graphic reports were generated from data downloaded from the LIMS database to summarize each core section. An example VCD for igneous rocks is shown in Figure F6. The VCD readouts display the following items:

  • Depth in mbsf;

  • Scale for core section length (0–150 cm, or as appropriate);

  • Sample piece number;

  • Scanned digital image of the archive half;

  • Upward-pointing arrow indicating oriented pieces of core;

  • Sample type and position of intervals selected for different types of shipboard analytical studies;

  • Lithologic unit number;

  • Symbolized structural information;

  • Structural measurements of dip direction and dip angle;

  • Block chart displaying the percent vesicularity;

  • Stacked line chart displaying phenocryst percentage for plagioclase (pl; red line), olivine (ol; green line), and pyroxene (px; blue line);

  • Block chart displaying variation in crystal size of modal groundmass (in millimeters) and indicating presence of fresh glass (f), slightly altered glass (s), moderately altered glass (m), or completely altered glass (a);

  • Ornament depicting alteration intensity;

  • Chart displaying point magnetic susceptibility, including both raw (shaded) and filtered (solid) data; and

  • Chart displaying color reflectance, with total reflectance (L*), red (a*), and blue (b*) data arranged side by side, again showing both raw (shaded) and filtered (solid) data.

Additionally, the VCDs contain written descriptions to accompany the schematic representation of the core sections. This information includes the following:

  • Expedition, site, hole, core and core type, section number, and top of core section (measured according to the CSF-A depth scale, which directly corresponds to mbsf; see “Procedures”), shown at top of VCD;

  • Unit number(s) (numbered consecutively downhole) and piece numbers belonging to unit (and on which piece, or pieces, the description was based);

  • Lithology, rock description, and name;

  • Volcanic description based on type of (partial) unit (e.g., pillow lava, massive flow);

  • Rock texture: aphyric (<1% phenocrysts), sparsely phyric (1%–2% phenocrysts), moderately phyric (>2%–10% phenocrysts), or highly phyric (>10% phenocrysts);

  • Color determined on dry rock surfaces;

  • Phenocryst percentage and type based on minerals identifiable by eye, hand lens, or binocular microscope;

  • Groundmass grain size and texture: glassy, aphanitic (crystalline, but individual grains not discernible with a hand lens), fine grained (<1 mm), medium grained (1–5 mm), or coarse grained (>5 mm);

  • Vesicle percentage by volume;

  • Upper and lower unit contact relations and boundaries, as based on physical changes observed in retrieved core material (e.g., presence of chilled margins, changes in vesicularity, and alteration), and information regarding their position within the section. “Not recovered” was entered where no direct contact was recovered (if igneous contact was inferred using proxy observations such as a chilled margin or changes in vesicularity); and

  • Alteration and veins (see “Alteration petrology”) and structural features (described in “Structural geology”).

Microscopic visual core description

Thin section analyses of sampled core intervals were used to complement and refine macroscopic core observations. Typically, one thin section was examined and logged per defined lithologic unit, although this was not possible in volcaniclastic units with frequent alternations of massive and brecciated units. As far as was practically possible, the same terminology and nomenclature hierarchy employed for visual core compositional descriptions were also used for thin section descriptions. The textural terms used are those defined by MacKenzie et al. (1982). Phenocryst assemblages (and their modal percentages and sizes), groundmass, and alteration phases were determined, and textural descriptions were constructed. All observations were entered into the LIMS database via a special DESClogik thin section template. Downloaded tabular reports of all igneous thin section descriptions can be found in “Core descriptions.”

Assessing the probability that igneous rocks have remained in situ since emplacement within the volcanic seamount structures

Determining the paleolatitude of the Louisville Seamounts was one of the primary objectives of Expedition 330, so it was essential that a large number of individual in situ igneous cooling units were collected at each site. It was therefore important to establish that (1) these units were correctly identified and (2) that the identified intervals had not rotated about a horizontal axis since they acquired their remanent magnetization. Both of these facts are easy to establish in the case of subaerial lava flows, but problems arise when the recovered units form part of a submarine succession. Individual pillows and flow lobes may be identified where recovery is high, but establishing that they had retained their cooling orientation was less straightforward. For example, individual lava lobes or pillows in an eruptive stack will probably have remained in situ, but an isolated pillow or small lobe in a volcaniclastic sequence will probably have cooled and rolled before coming to rest. It was therefore necessary to construct a set of objective criteria with which to assess the probability of an individual igneous lithologic unit having remained in situ after emplacement. These criteria, set out below, allowed us to assign to each cooling unit (our lithologic units) of lava or intrusive igneous rock in the core a number (0, 1, 2, 3) or NA (not applicable) that we call the in situ confidence index:

  • 0 = unlikely to be in situ (individual isolated pieces <30 cm long; any piece <30 cm long separated from its neighbor by a selvage of hyaloclastite or breccia; any section of core of any length where there is compelling evidence that it is not in situ, such as large boulders in a conglomerate).

  • 1 = could be in situ (discrete lithologically similar pieces that form continuous sections of 30 cm to 1 m; continuity is broken if traces of hyaloclastite or breccia are identified between individual pieces).

  • 2 = probably in situ (discrete lithologically similar pieces that extend over recovered intervals of >1 m, provided there are no traces of hyaloclastite or breccia between any individual pieces; core recovery need not be high).

  • 3 = highly likely to be in situ (must satisfy at least one of the following requirements: [1] lava flows with identified flow tops or bottoms; [2] dikes or sills with at least one clearly identified margin; [3] recovered intervals of core >1 m with pieces that clearly belong together and where core recovery over the interval extending from above to below the unit was >60%; [4] pieces form part of an eruptive lobe or pillow stack within a continuous core showing individual cooling units in contact with each other; [5] clear vertical pipe vesicles or consistent horizontal magmatic foliation).

  • NA = not applicable (applied during Expedition 330 to volcaniclastic deposits made of clasts that possibly did not retain their orientation since cooling, although the deposit as a whole may have acquired the ambient magnetic field through alteration processes soon after eruption).