Unit 21

Unit 21 is a breccia-topped aphyric basalt flow. The uppermost few centimeters of Unit 21 are a smooth pahoehoe surface that has been highly fractured. This is followed by a breccia exhibiting a wide range of curious features. In interval 183-1138A-88R-2, 0-30 cm, the breccia is composed of at least seven originally >5-cm clasts that have been broken into 1- to 3-cm angular fragments that can be fit to reconstruct the original clasts. One of these clasts is recognizable as a >8-cm-diameter pahoehoe lobe with a 3-mm-thick glassy rind. The lobe has 20%-30% vesicularity, with slightly elongated <0.5-mm-diameter vesicles. The vesicle sizes become slightly smaller toward the margins. In interval 183-1138A-88R-2, 33-45 cm, a >12-cm dense (7% vesicularity) lobe with 0.5- to 1.5-cm elongate angular vesicles is present (Fig. F51). The top of this lobe has an irregular fluidal arm that the elongated vesicles fan into. There is no evidence of baking or welding of this clast to the surrounding clasts. However, a clast at interval 183-1138A-88R-2, 50-60 cm, does have a welded margin on one side (Fig. F52). Another margin of this same clast is glassy and fragmented, and a third margin has a fluidal shape with small angular fragments filling the void underneath. We find a clast at interval 183-1138A-88R-2, 74-75 cm, that is molded around another clast and has formed a welded and chilled margin. In general, fragmentation decreases with depth in this breccia. However, recovery of the breccia is poor at the base of Core 183-1138A-88R with several 10-cm pieces alternating between small fragments and coherent vesicular lava. The breccia continues in the top of Core 183-1138A-89R, but with a large zone of dense lava running up one side of the core (right side of the archive half) from interval 183-1138A-89R-1, 9-53 cm. This dense lava has a set of dipping joints that open in a fanning fashion on the opposite side of the core. This lava contains 3% irregular vesicles <1 mm in diameter. Between 53 and 58 cm, Section 183-1138A-89R-1 has only breccia, mostly composed of fragments of the dense lava. In interval 183-1138A-89R-1, 58-62 cm, the dense lava returns and has a square 3 cm × 4 cm vesicular clast in the center with a chill margin surrounding it. The pieces below this appear to be 5- to 10-cm clumps with 15%-20% vesicularity with 1-mm spherical vesicles surrounded by a matrix of lava with 10%-15% vesicularity with 0.5- to 5-mm round vesicles. From here, recovery drops and isolated pieces were recovered. Section 183-1138A-89R-1 (Piece 5) has 15% irregular, >1-cm angular vesicles. Piece 6 from the same section has 5% angular, <1-mm vesicles. Piece 7 is the bottom of Unit 21 and consists of a handful of highly vesicular pebble-sized fragments with 20%-30% round, ~1-mm vesicles.

Unit 22

Unit 22 is the final flow drilled at Site 1138 and is another breccia-topped aphyric basalt flow with some unique features. Parts of the breccia are composed of ~3-cm clasts surrounded by smaller angular fragments, much like the other breccias, but large portions are composed of intact pahoehoe lobes. In interval 183-1138A-89R-2, 0-19 cm, a lobe with 15% irregular but rounded 1- to 10-mm vesicles has partially enveloped and chilled against another lobe with ~40% small (<1 mm) vesicles. Interval 183-1138A-88R-2, 26-40 cm, contains the edge of a lobe with a near-circular cross section. This lobe is fractured in both radial and concentric directions and contains 25% round, ~1-mm vesicles. Interval 183-1138A-88R-2, 43-53 cm, contains a single coherent lobe with a fragmented upper margin. Section 183-1138A-88R-2 (Piece 4, 53-80 cm) has a lobe that partially enveloped and is quenched against another lobe. Below, the lava appears to be coherent with 10% rounded, occasionally elongated, 2- to 3-mm vesicles. A 3-cm clast with the surrounding lava chilled against it is present in Section 183-1138A-88R-2 at 100 cm. Below this, the lava contains 2%-5%, subrounded, 0.5- to 3-cm vesicles elongated in a generally horizontal direction.

Lithologic Units

Unit I: Foraminifer-Bearing Diatom Ooze

The unconsolidated upper part of this interval was disturbed during coring. The sediments are variable mixtures of clay and pelagic material. The disseminated volcaniclastic components indicate contributions from primary felsic and mafic pyroclastic fall deposits as well as a reworked lithic fraction (Table T10).

Units I, II and III: Pyroclastic Fall Deposits

Intervals with a concentration of volcanic ash material (Tables T10, T11) have been divided into (1) disseminated felsic glass and pumice in Cores 183-1138A-1R to 10R (<3.8 Ma), (2) felsic (trachytic?) tephra material in Core 183-1138A-11R (~3.8-3.2 Ma), (3) a suite of mixed (basalt + trachyte?) tephras in Cores 183-1138A-14R to 17R (10.6-8.9 Ma), and (4) a concentration of basaltic ash in Cores 183-1136A-30R to 34R (~31-26.5 Ma).

Previous work on Legs 119 and 120 (Bitschene et al., 1992) identified tephra layers preserved in similar age intervals to those at Site 1138 (Tables T10, T11). In Hole 120-736A, volcanic components with ages of 0.4, 0.7, and 0.8 Ma were recognized. Ash of these ages fall within the zone of disturbance during coring at Site 1138 (see [1] above), so correlation is not possible. However, at Site 736, a 2.8-Ma trachytic debris flow was identified. This age may correspond with feldspar-bearing pumiceous ash found in Core 183-1138A-8R.

Site 120-747 lies 150 km south-southeast of, and is the closest hole to, Site 1138 (Fig. F1), so ash horizons may be correlated with some confidence. Felsic ash with ages of 3.7, 3.8, and 4.3 Ma was recovered at Site 119-737, which indicates considerable activity in the region at the time that the ~3.8- to 3.2-Ma basaltic to bimodal ash (see [2] above) (Core 183-1138A-11R) was deposited. Volcanic material described as mixed basaltic and trachytic ash in Hole 119-737B does not have an equivalent at Site 1138, but we observe ~6.4-5.6 Ma trachytic ash in Core 183-1138A-12R and a ~10.6-8.9 Ma basaltic to bimodal interval in Cores 183-1138A-14R and 15R (see [3] above). These ages bracket the Hole 120-737B 7.8-Ma age and may or may not be related. Site 737 is north of Heard Island and close to the Kerguelen Archipelago, so it may be more strongly influenced by proximal volcanism. An older series of mafic volcanic events (28.3 Ma, 31.4 Ma, and 32 Ma) at Site 737 has the same age range (31-26.5 Ma) as basaltic ash in Core 183-1138A-30R to 34R (see [4] above). Two trachytic ash layers (2.2 Ma and 4.3 Ma) in Hole 120-745B on the Southern Kerguelen Plateau do not have equivalents at Site 1138.

Late Pliocene (3.7 Ma and 3.8 Ma) trachytic ashes and early Oligocene (30 Ma) basaltic ashes in Hole 120-747A correlate well with similar age intervals at Site 1138 (see [2] and [4] above). However, Cretaceous to Paleocene volcanic material observed in Hole 120-747A was not found at Site 1138.

Unit 1: Aphyric Flow-Banded Dacite Cobbles

Unit 1 consists almost entirely of rounded, aphyric, flow banded dacite (see "Igneous Petrology"). These cobbles represent a layer of dacitic rubble at the top of basement. Flow banding and the fine-grained internal texture suggest that they are eroded from a lava or a densely welded tuff. Dacitic lava is not preserved elsewhere in the basement succession cored at Site 1138.

Unit 2: Bedded Pumice Lithic Breccia, Lithic Breccia, Ash-Fall Deposits, and Volcanic Clay

Unit 2 is dominated by pumice lithic breccias together with reworked volcaniclastic sediments and altered pyroclastic fall deposits. Subunit 2A is an interval of dark gray massive clay and lithified fine-grained altered rock that does not appear to be related to the underlying volcaniclastic succession and may be rubble that has fallen downhole.

Subunit 2B is a reworked succession of alternating pumiceous and lithic-rich beds, each a few centimeters thick, which show a variety of grainsizes, sorting, and grading features (Figs. F18, F19). These intervals are interpreted as water-settled deposits and low-angle, clast- to matrix-supported flow deposits. The reworked materials have identical textures, range of clast types, and alteration features to the underlying pumice and lithic breccia deposits. They have probably been reworked in a low- to moderate-energy shallow-water environment. The degree of alteration of different sedimentary packages in the succession varies considerably, which relates to the porosity and permeability of the intervals. Generally, more pumiceous intervals are pale green and clay-rich, reflecting more intense alteration.

Overall, Unit 2 is dominated by pumice-lithic breccias (Subunits 2D, 2G, 2H, 2I, and 2J) (Fig. F14) and a lithic breccia (Subunit 2E) (Fig. F15). These breccias exhibit good hydraulic sorting and could be primary subaerial pyroclastic flow deposits. One of the breccias (Subunit 2E) is considerably more lithic rich than the others and is interpreted to either be the basal part of a thick flow, or may be more proximal to the source. Internally, the flow units show limited stratification, and contacts with overlying flow units are usually sharp. Subhorizontal alignment of elongate pumice clasts may be related to flow alignment during emplacement or load alignment after emplacement, but there is no evidence of welding. Clasts are variably altered, but generally the pumice is pale green and soft, indicating partial replacement by clay minerals (Fig. F16). In more indurated parts of the profile, relatively fresh pumice can be observed in thin section. Commonly, lithic fragments are less altered than pumice clasts. The matrix in all flow units is dominated by medium to very coarse sand-sized fragmented pumice material (Fig. F17).

The Subunit 2K lithic breccia is reverse graded and has clasts that are similar to the lithic component in the pumice lithic breccias. This interval may be the basal layer of the overlying pyroclastic flow (Subunit 2J) and shows reverse grading because of tractional processes at the base of the flow.

Two intensely weathered intervals (Subunits 2C and 2N) (Figs. F20, F21) retain little evidence of primary features. These may have been zones of enhanced fluid percolation, or were exposed at the land surface for some time to become so intensely altered. Both have evidence of clay pseudomorphs after pumice. One interpretation is that these are pumice fall deposits.

Accretionary lapilli were found in two places in the stratigraphy (Subunits 2F and 2L) within normally graded medium to fine ash deposits (Figs. F22, F23, F24). These are pyroclastic fall deposits, and the presence of accretionary lapilli indicates interaction with water in the eruption cloud. Two other highly altered volcanic clay intervals (Subunits 2M and 2O) (Fig. F25) may also be fine-grained volcanic ash deposits, but the presence of scattered lithic fragments in these deposits suggests they have been reworked.

Unit 2/Unit 3 Boundary

The appearance of basalt marks the boundary between Units 2 and 3. However, it is unusual to find relatively unaltered, unoxidized lava at the top of the lava sequence. This suggests an erosional unconformity or that we did not recover any altered material from the top of the lava pile.

Unit 3: Transitional (Slab-Pahoehoe?) Lava Flow

The first rocks from Unit 3 that we recovered have vesicularity consistent with being the lowermost part of the vesicular upper crust. The bulk of Unit 3 that was recovered is from the massive interior of a flow. Although the round vesicles in the upper part of the flow are consistent with pahoehoe, the elongated subangular shapes further down are unusual and suggest motion and shearing of a highly viscous, partly crystallized lava. The formation of megavesicles, but no sheets of segregated material, also suggests a lack of complete stagnation before crystallization. A large number of small vesicles in the lower chill zone suggests a near-vent location. With time and distance from the vent, bubbles will coalesce, forming larger vesicles. The angular breccia at the base of the flow suggests that Unit 3 might have been emplaced as a slab pahoehoe flow.

Slab pahoehoe flows are usually characterized by relatively rapid motion of relatively fluid lava. They are particularly common in steeper areas (3°-10° slopes) or where the lava has surged, breaking up the earlier more coherent top. In Hawaii, they are commonly emplaced as 10- to 100-m-wide sheets but can also form along the sides of channeled flows.

Unit 3/Unit 4 Boundary

A highly altered breccia makes a visually clear break between Units 3 and 4. However, there really is no definitive evidence that this altered material is not a basal breccia for Unit 3. The high alteration probably results from the extreme permeability of the breccia, but it is surprising that there is no obvious gradation from the more fresh rock into this clay-rich zone. We suspect that the partially altered, but less clay-cemented rocks were not recovered. Also, the high degree of alteration might suggest a significant time break between Units 3 and 4, but it is likely that much (if not all) of the alteration followed emplacement of Unit 3. It is therefore possible that Units 3 and 4 are simply different flows from the same eruption.

Unit 4: Aa Flow

On the basis of vesicle shapes (Fig. F26) and the suggestion of irregular clasts within the clay-rich material, it is likely that Unit 4 is a typical aa flow. The various vesicular clots and patchy zones are interpreted to be entrained clasts (Fig. F27), diagnostic of a disrupted flow top, but not necessarily an aa flow. The apparently higher vesicularity of the flow interior as compared to the upper and lowermost parts suggests that the vesiculation during crystallization was trapped in place. Also, lack of complete degassing of an aa flow suggests that the lava was not transported a great distance. The lack of vesicle sheets and similar features indicates that major segregation of late-stage vesicle-rich material did not happen, though a few 1- to 3-cm megavesicles were able to coalesce. The lowermost vesicular zone is interpreted to be a lower vesicular crust. In a typical aa flow, we would expect a substantial basal breccia below the massive interior, but no basal breccia was recovered with Unit 4.

Aa flows are most common where relatively viscous lava is subjected to high-strain rates. This can result from high eruption rates, steeper slopes, or topographic channeling. The longest aa flow known is the ~50-km-long 1859 Mauna Loa flow, so it is unlikely that this aa flow traveled more than a few tens of kilometers from its vent.

Unit 4/Unit 5 Boundary

The dark gray altered breccia underneath a coherent lava makes another visually obvious boundary. However, it is possible that this breccia is the basal portion of Unit 4. The most likely scenario is that Unit 4 had a basal breccia and Unit 5 had a brecciated top. However, given that no pieces showing the relationship between the altered breccia and the coherent lavas above and below were recovered, it was deemed simplest to place the unit boundary at the visually obvious location. Given the equivocal nature of this boundary, we are not confident in interpreting a major time break at this location.

Unit 5: Aa(?) Flow

The vesicular clots within the interior of Unit 5 indicate that it was entraining vesicular material, presumably from a disrupted brecciated flow top. The fining-upward sequence in vesicle size within the coherent lava can be explained by a reduction in the rate of crust growth with time, allowing vesicles to coalesce to a greater degree as they were trapped at the base of the upper crust. The elongated, irregular vesicle shapes are consistent with these vesicles forming in a shearing mush zone at the base of a crust while vigorous active flow persisted underneath. The sudden transition to rounder bubbles could indicate stagnation of the flow. The formation of mesostasis blebs and a megavesicle horizon are also consistent with the deeper interior of the flow crystallizing under more stagnant conditions. The loose pieces at the base of the flow could be from a basal chill zone. Given that it is not clear that the breccia at the top of Unit 5 belongs to the coherent lava from Unit 5 and the evidence for relatively fluid stagnant lava in the interior of the flow, we cannot confidently conclude that this is an aa flow. A slab pahoehoe or other type of transitional flow would also be consistent with the observations.

Irrespective of whether the flow is aa or a transitional lava flow, the observed morphology is strongly indicative of relatively fluid lava subjected to high strain rates. High flow rates through a confined area or higher slopes are the most common cause for high strain rates in a lava flow.

Unit 5/Unit 6 Boundary

The oxidized and slightly brecciated pahoehoe top is a clear visual break and is the first reasonably strong evidence for a significant hiatus in lava flow emplacement at Site 1138 (Fig. F28). However, the oxidized zone is quite thin and the flow top is not highly modified, suggesting a short time interval between the emplacement of Unit 6 and the arrival of Unit 5.

Unit 6: Inflated Pahoehoe Flow

Unit 6 has the features of an inflated pahoehoe flow. The top and bottom have smooth, undisrupted pahoehoe surfaces and fine-grained chills. The vesicle distribution changes from a vesicular flow top to a dense interior and back to a vesicular base, exactly as in the idealized models (see Fig. F9, in the "Explanatory Notes" chapter). The lack of evidence for extensive pooling of segregated material at the base of the upper crust suggests that this was a relatively thin inflated flow, probably of typical Hawaiian dimensions (3-5 m thick). The steep angle of the basal chill of Unit 6 indicates that the flow moved over local topography on the order of tens of centimeters. The possible internal chill zone at the section break could indicate that this pahoehoe flow was compound, with smaller lobes at the base overrun by the lobe that was able to inflate to a few meters thickness. In Hawaii, these relatively thin inflated pahoehoe flows are commonly found on slopes of 0.5°-4°.

Unit 6/Unit 7 Boundary

The contact between the relatively unoxidized pahoehoe lobe and the more oxidized underlying breccia is evident (Fig. F29) and indicates both a significant time break and a change in the style of eruptive activity. Also, fragmentation at the base of Unit 6 might suggest some quenching against a small amount of water. Alternatively, it could result from concentrated alteration from the fluids passing through the permeable breccia underneath. The rounding of the upper clasts in the breccia suggests some reworking by perhaps fluvial processes, again implying an extended time break between Units 6 and 7.

Unit 7: Aa(?) Flow

Isolated domains in the breccia at the top of Unit 7 show chill features and in situ brecciation textures (Fig. F30) that might be consistent with quenching of magma in water-saturated sediment (peperite). However, the upper part of the breccia with rounded clasts, many with only partially preserved chill margins, looks more like a volcaniclastic conglomerate. It is likely that the locally peperitic textures are an artifact of the reworking and partial sediment infilling of this flow-top breccia.

The shapes of the vesicles in the clasts and the coherent interior suggests that this flow was originally an aa flow. Angular protrusions on the aa clasts are readily broken by erosion and reworking processes. There is evidence of localized transport of glassy material within the sediment (Fig. F31), and rounding of clasts in the upper part of the profile implies more intense reworking near the top of the interval. The remainder of the structures within the coherent lava are consistent with a relatively thick aa flow that entrained clasts during emplacement and whose deep interior was able to undergo a small amount of segregation of late-stage liquids. This segregation might also suggest that the lava was more fluid than typical aa and that the flow was more similar to slab pahoehoe than aa.

Unit 7/Unit 8 Boundary

The contact between Units 7 and 8 was not recovered. The last recovered pieces in Core 183-1138A-81R appear to be from deep in the interior of Unit 7. The first recovered rocks in Core 183-1138A-82R are highly altered breccia that grades into coherent lava. Given the rounded clasts, slight oxidation, and sediment fill in the top of Unit 8, it is likely that Units 7 and 8 are distinct lava flows and that some time passed between their emplacement.

Unit 8: Aa(?) Flow

The similarities between Units 7 and 8 are rather striking and it is likely that they formed in similar manners. Thus, the relatively well-recovered base of Unit 8 might also provide information about the unrecovered basal portion of Unit 7. The patches of vesicular material at the base of Unit 8 are probably welded clasts of a basal breccia. Such a basal breccia would be rapidly covered by the bulk of the flow and would not be able to cool effectively. Although the weight of the overlying lava would help to meld the individual clasts into a coherent mass, the basal breccias of aa flows are commonly not welded. This suggests that Unit 8 (and presumably Unit 7) might have been more akin to slab pahoehoe where the disrupted clast is more fluid and plastic than typical aa clinker. It is also possible that the breccia clasts are actually spatter and were recovered extremely close to the vent. The presence of welding is strong evidence that the base of Unit 8 did not significantly interact with water or wet sediments.

Unit 8/Unit 9 Boundary

The change from the highly altered and welded breccia at the base of Unit 8 to the unwelded, moderately oxidized breccia of Unit 9 is a clear boundary (Fig. F32). The breccia on top of Unit 9 is only partially filled with sediments and is not highly oxidized, eroded, or weathered. This suggests that the time break between Units 8 and 9 was not very large.

Unit 9

The slabs of ropy pahoehoe (Fig. F33) and other pahoehoe fragments leave no question that flow was primarily emplaced as a slab pahoehoe. The prominent slab and many other pieces show one strongly quenched margin that cooled at the surface of the flow while the other sides cooled more slowly within the breccia. This is consistent with the clasts being slab pahoehoe that was actively disrupted from within by the flowing lava. The zone that has a mix of sediment and loose pieces of lava has the appearance of a peperite (Figs. F34, F35), but the absence of chill margins surrounding the clasts suggests that the sediments may have been deposited well after the lava flow had been emplaced. The material could be a slurry of fine-grained sediments (perhaps dominated by silt-sand sized glass shards from the flow breccia itself) and 1- to 2-cm vesicular lava clasts that flowed into a large open cavity in the breccia. The shapes of the voids at the bases of larger clasts also suggest that the sediments flowed in as a viscous slurry. However, these cavities might also have formed by trapped steam from water being driven off of the wet sediments by hot lava. Despite the appearance of this small portion of the breccia, the evidence for interaction between wet sediments and hot lava is equivocal, at best. The disaggregating dense zone (Fig. F36) has the appearance of a dense arm of aa-like lava pushing up into a breccia. This kind of disaggregation is identical in degree and morphology to that caused by subaerial chilling of microlite-rich lava in aa flows (see "Interpretation" in the "Explanatory Notes" chapter). The transition from irregular vesicle shapes to more spherical ones suggests a sudden cessation of shearing while the interior of the lava was still relatively fluid. This is supported by the presence of what appears to be a complicated set of sheets of segregated material at this boundary (Fig. F37). The increase in vesicularity at the base of the flow suggests that the last recovered rocks were from near the base of the flow, although a basal breccia might also have existed.

Unit 9/Unit 10 Boundary

The pebbles found at the top of Core 183-1138A-83R have been placed in Unit 9 but are likely to have fallen into the hole from anywhere above Unit 10. No contact between Units 9 and 10 was retrieved, but the change from the coherent lava interior of Unit 9 to the breccia of Unit 10 is a clear boundary. What is not as clear is whether any of the Unit 10 breccia might be part of a basal breccia for Unit 9. The minimal oxidation, only partial infilling of Unit 10 breccia with sediments, and limited erosion and weathering of the Unit 10 breccia suggest that a relatively short time passed between the emplacement of Units 10 and 9.

Unit 10: Transitional Flow

In many respects, the interpretation of Unit 10 is similar to that of Units 7-9. Whereas the breccia originally formed during emplacement of the flow, it was probably modified to some extent by sedimentary processes, including the deposition of sediments that filled the lower part of the breccia. The chill at the top of the coherent lava and the disaggregating dense arm both suggest that the flow top-breccia was quite cold near the end of the emplacement of Unit 10. The entrained clast within the top of the coherent lava indicates active mixing between the disrupted crust and coherent interior of the flow during emplacement. The irregular vesicles in the upper part of the flow indicate substantial shearing of a relatively viscous lava early in the emplacement history. The rounder vesicles below a megavesicle horizon and the formation of a sheet of segregated material and a dense interior suggest a more quiescent emplacement of relatively fluid lava before the flow stagnated. The irregular vesicular pods at the base of Unit 10 (Fig. F38) are partially welded clasts of a basal breccia. The subhorizontal shape of many of these suggests some flattening by the weight of the overlying lava. This suggests that the leading elements of Unit 10 were hot and plastic and not a cold aa-like breccia, implying that Unit 10 was again more similar to a slab-pahoehoe flow than an aa flow. Alternatively, the welded basal breccia could be welded spatter from a nearby vent.

Unit 10/Unit 11 Boundary

The change from a welded breccia to an oxidized coherent vesicular flow forms a clear boundary (Fig. F39). The contact between the two units may be the glassy chill pressed into the top of the 2 cm of vesicular breccia at the top of Unit 11. These two angular fragments could easily be the loose top of a slightly weathered pahoehoe flow top. Such surfaces are very common on Hawaiian pahoehoe flows that are several decades to a few hundred years old. This and the oxidation of Unit 11 suggest a significant time break between Units 10 and 11.

Unit 11: Transitional Flow

The large elongate vesicles in the top and bottom of Unit 11 seem to require that the lava was viscous enough to keep nonspherical shapes while being fluid enough to flow in a very laminar and plastic fashion. This is not typical of either pahoehoe or aa flows, but smaller elongate bubbles are common in pahoehoe flows. The irregularly shaped vesicles in the interior of the flow are somewhat similar to vesicles found in aa flows. The 1- to 3-cm vesicular clots could be welded breccia clasts or spatter. The high density of small vesicles strongly suggests a near-vent facies, making it likely that the clasts were originally pieces of spatter, not breccia. This requires proximity of the fountain or fissure vent, probably within a few tens to hundreds of meters.

Unit 11/Unit 12 Boundary

The break between Units 11 and 12 is defined by a single piece of slightly welded breccia (Fig. F41). However, the change from a relatively thick coherent flow with large vesicles to a mass of small lobes is visually obvious and a volcanologically important distinction.

Unit 12: Compound Pahoehoe Flow

The vesicular domains in Unit 12 are interpreted to be spongy pahoehoe lobes (S-type pahoehoe of Walker, 1989) that were welded together to produce a coherent rock. This level of welding requires very rapid accumulation and is most common very close to a lava source. Although the source of lava might be the eruptive vent, it can also be an overflowing skylight (also known as ephemeral vents or boccas) or an open lava channel. However, the very high density of extremely small vesicles trapped in the margins of the lobes strongly suggests that we are close to the primary vent. Based on examples in Hawaii, the eruptive vent is unlikely to be more than a few kilometers from Hole 1138A.

Unit 12/Unit 13 Boundary

The change from a welded stack of spongy pahoehoe lobes to an oxidized breccia is a clear break (Fig. F42). Whereas oxidation can form very quickly in a near-vent environment, the partial sediment infilling of the Unit 13 breccia suggests that this boundary reflects a significant time break. However, the angular clasts at the top of the breccia (which the Unit 12 lobes flowed onto) suggest quite minimal erosion and weathering.

Unit 13: Transitional (Aa?) Flow

The mix of aa and pahoehoe clasts suggests that this flow was sampled at a point shortly after it had made the transition from pahoehoe to aa. The vesicular domains at the top of Core 183-1138A-84R-2 (Fig. F43) are interpreted to be large folded pieces of pahoehoe, of the sort that typically forms on open lava channels where the lava is subjected to high shear but is too fluid to form aa or even slab pahoehoe. Pieces of such pahoehoe can be rafted many kilometers on the top of an aa flow. The pahoehoe pieces that have been dragged into the deeper parts of the breccia are rapidly broken into the smaller angular clasts such as those observed in the Unit 13 breccia. It is interesting that alteration-driven rounding of the clasts is more prevalent in the sediment-filled part of the breccia. Also, the loose pieces of recovered breccia suggest that zones with large void spaces can be found deep within this kind of breccia. The dense arm of lava pushing into the breccia from the flow interior is strong evidence of aa-like emplacement. The base of the flow appears to have incorporated and welded clasts (or possibly spatter), with the degree of welding decreasing downward. If the clasts are indeed spatter, they imply proximity to the vent.

Unit 13/Unit 14 Boundary

The highly altered breccia beneath the relatively unaltered welded breccia of Unit 13 is a clear visual marker. However, it is not clear if this alteration indicates a significant time break.

Unit 14: Transitional(?) Lava Flow

Given the alteration and relatively poor recovery of Unit 14, detailed interpretations are difficult. However, it appears generally similar to the other units with brecciated flow tops and is likely to be an aa or slab pahoehoe flow.

Unit 14/Unit 15 Boundary

The change from the welded base of Unit 14 to the oxidized, unwelded breccia at the top of Unit 15 is a clear change (Fig. F44) and is likely to signify some time break. However, given the poor recovery of Unit 15, it is not possible to rule out the possibility that some of Unit 15 might be unwelded basal breccia from the flow comprising Unit 14.

Unit 15: Transitional(?) Lava Flow(s)

The few pieces of breccia at the top of Unit 15 appear generally similar to the breccias found farther up the hole. The lone clast with a chilled margin might be from the top of a pahoehoe lobe or just be a clast within the breccia. Although it is clear that Unit 15 contains a change from breccia to coherent vesicular lava, what happens below the coherent lava is unclear. The loose pieces are too large to have been brought out of the hole in the wrong order, so it is likely that there are several dense and brecciated horizons within the lower part of what we have called Unit 15. These might be pieces of breccia from the base of the Unit 15 lava flow, parts of an entirely different lava flow, or some pieces from the top of the breccia on Unit 16.

Unit 15/Unit 16 Boundary

This boundary is completely arbitrary and reflects the return to better recovery where we can actually distinguish some internal features of the lava flows.

Unit 16: Lava Flow (Type Unknown)

The recovered rocks from Unit 16 are from the interior of a lava flow. It is reasonable to suppose that some of Unit 15 represents material from the top of Unit 16. Given the limited amount of lava recovered and the high degree of alteration and fracturing, it is hard to make any interpretation of this flow. It is possible that the poor recovery and the high alteration of this unit are related to the highly fractured nature of the lava.

Unit 16/Unit 17 Boundary

The change from the interior of a lava flow to a breccia is a clear boundary (Fig. F45). However, given the alteration and poor recovery, it is not possible to determine if this boundary indicates a significant time break.

Unit 17: Transitional Lava Flow (Unnamed Type)

Although the breccia on Unit 17 shares many features common to breccias higher up in Hole 1138A, two characteristics stand out. First is the lack of any evidence of aa-like irregular shaped clasts. Second is the appearance of 1- to 5-cm angular clasts of dense lava. Previously, the dense lava has been in the form of large coherent arms reaching into the breccia. The clasts within the breccia of Unit 17 seem to be derived from both the upper and the central parts of the pahoehoe flows. The simplest explanation is that this breccia is a redeposited eroded pahoehoe flow, but it is difficult to reconcile this with the lack of evidence for significant weathering, rounding, or sorting of the clasts. Our favored, but speculative, interpretation is that this is a pahoehoe flow brecciated by remobilization (see "Physical Volcanology" in the "Explanatory Notes" chapter).

Unit 17/Unit 18 Boundary

The change from the interior of a flow to a breccia is a clear boundary. However, the contact was not retrieved and the breccia at the top of Unit 18 is heavily altered, so it is not possible to determine any relationship between the two flows. If the breccias result from erosion and redeposition, this implies a major time break. However, if the breccias are primary, there is no reason to suppose that these two units were not closely spaced in time and could possibly be from the same eruption.

Unit 18: Transitional(?) Lava Flow

Given the poor recovery and altered state of Unit 18, we can only suggest that it is similar to Unit 17. Mesostasis blebs suggest only limited segregation of late-stage liquids inside these flows.

Unit 18/Unit 19 Boundary

The change from the coherent interior of Unit 18 to the breccia of Unit 19 is a clear boundary. However, no contact was recovered, making it difficult to interpret this change. The boundary separates two packages of lava that may be separate flows or two lobes of a single compound flow.

Unit 19: Transitional Lava Flow (Unnamed Type)

The excellent recovery of key parts of Unit 19 sheds light on the formation of the breccias on Units 17, 18, and 19. Of greatest importance is the observation that the dense lava is (in at least one case) welded to vesicular pieces (Fig. F46). This requires hot lava from the interior of a flow to have mingled with a solidified vesicular pahoehoe crust. The dark clay rims (which we interpret to have been glassy chill margins) on the other dense clasts supports this idea and indicates that at least part of the solid crust was cold relative to the solidus of the lava. The tearing and intense stretching of the top of the coherent lava is the final proof that the liquid interior of the flow was moving against the breccia. This allows us to rule out resedimentation for the primary mechanism for the formation of the breccia on at least Unit 19. Both the reactivated pahoehoe and hyaloclastite models are viable, but there are no observations diagnostic of water quenching. The various loose pieces in Section 183-1138A-87R-3 are difficult to place, much less interpret. It does appear that the coherent part of Unit 19 has elongated vesicles at its top and bottom and more rounded vesicles in the center. Also, somewhere between the coherent interiors of Units 19 and 20 there is a breccia where some clasts interacted with at least a thin (1 cm) layer of sediments.

Unit 19/Unit 20 Boundary

The location of the boundary between Units 19 and 20 is arbitrary and primarily reflects the return to more complete recovery. As noted above, a number of enigmatic pieces are placed at the base of Unit 19, including a breccia and some altered glass that could be related to nearby explosive activity.

Unit 20: Spiracle(?)

The unusual features within Unit 20 are not consistent with aa, pahoehoe, or transitional lava flows. The only scenario we have been able to construct that is consistent with all the observations is that we have recovered the side of a spiracle, a feature formed by the escape of steam from beneath a lava flow. The formation of spiracles and rootless cones requires at least groundwater or very wet sediments. The flow would also have to be emplaced rapidly in order to cover the wet ground before the steam could escape along the margins of the flow. The lack of sediments at the base of Unit 20 requires the underlying flow to have been water saturated.

Unit 20/Unit 21 Boundary

The chill margin between Units 20 and 21 and the change from coherent lava to a breccia make a clear boundary. If Unit 20 does include a spiracle, then time was sufficient between the emplacement of the two units for Unit 21 to cool and become saturated with water. The smooth pahoehoe surface immediately below Unit 20 is likely to be the surface of a pahoehoe lobe in the breccia of Unit 21.

Unit 21: Transitional Lava Flow (Unnamed Type)

While the breccia of Unit 21 has similarities to those in the overlying flows (e.g., Units 17-19), there are some interesting differences. The pahoehoe lobes at the top are relatively intact, whereas farther down there is evidence for the lobes fragmenting while their interiors were still hot enough to weld to adjacent lobes (Fig. F52). The dense arm (Fig. F51) that pushes into this breccia has the small angular vesicles more typical of aa flow, rather than the long elongate vesicles in the dense material intruding Units 17, 18, and 19. The fanning joints in the dense arm in Unit 21 are common where hot, viscous lava forms an overhanging margin and gravity attempts to pull the lava apart. Such features are quite commonly seen on the dense arms of aa flows. The incorporated vesicular clast with chill margin is proof that this dense arm actively pushed into the bottom of this breccia. The sporadic recovery of the coherent interior of Unit 21 makes interpretation difficult, but the observations suggest multiple domains with a variety of vesicle morphologies, as has been seen in most of the overlying units. The high degree of fracturing and poor recovery of the interior and base of Unit 21 might be consistent with quenching by water in the later part of its cooling history.

Unit 21/Unit 22 Boundary

There must be a boundary between the coherent interior of Unit 21 and the breccia on the top of Unit 22. However, the position we have selected is arbitrary and marks the first appearance of well-recovered breccia from Unit 22. It is likely that some of the pieces included in Unit 21 are actually from the top of Unit 22. Given that we did not recover a contact, we cannot interpret the relationship or time gap between Units 21 and 22.

Unit 22: Transitional Lava Flow (Unnamed Type)

The intact pahoehoe lobes mixed with fragmented pahoehoe lobes in the breccia on Unit 22 suggest that it formed by the repeated intrusion of vesicular pahoehoe lobes into the breccia. This process is identical to that envisioned for the formation of the breccia on Unit 2 at Site 1137 (see "Physical Volcanology" in the "Site 1137" chapter). The intense fracturing of some of the lobes may suggest the presence of water during a brief part of the formation of the breccia. The vesicles from the interior of Unit 22 suggest shearing of a lava that was too viscous to allow the bubbles to return to a spherical shape before freezing.

The internal characteristics of lava flows in the basement of Hole 1138A are diverse. However, all the flows are consistent with subaerial emplacement on slopes of 1°-4°. Much greater flow thicknesses are required to achieve the stresses (i.e., velocities) needed to brecciate the top of a basaltic flow on slopes <1°. Inflated pahoehoe flows such as Unit 12 have not been seen to have formed on slopes >4°. The repeated circumstantial evidence for running water suggests intermittent small rivers or lakes. The preponderance of various brecciated flow tops suggests that the eruptions had moderate eruption rates. Also, these types of flows are unlikely to have traveled more than tens of kilometers from the vent. The large proportion of very small vesicles and possible welded spatter suggest that Hole 1138A was probably no more than a few kilometers from the vents for at least a few of the flows.