Petrographical and mineral chemistry evidence to delineate the source/sources of the Central Indian Ocean Basin pumices

1. Introduction

Pumices have a low specific gravity and their dispersal around a volcanic region is strongly controlled by prevailing winds, topography, ocean currents and sub-bottom bathymetry. Because of these factors the pumice fall deposits would be dispersed in different directions in contrast to pyroclastic flows and remobilised deposits (Sigurdsson et al., 1980). Based on an extensive study carried out on thousands of pumice clasts, Iyer and Sudhakar (1993) suggested the existence of a 600 000 km² of pumice field to be present in the Central Indian Ocean Basin (CIOB). A few of the CIOB pumices were studied for whole rock chemistry and mineralogy by various workers (Iyer and Sudhakar, 1993; Martín-Barajas and Lallier-Verges, 1993; Mudholkar and Fujii, 1995; Iyer, 1996; Pattan et al., 2008). Detailed investigations of the morphology and petrography of tens of pumices were reported by Kalangutkar et al. (2011).The different types of pumices and their sources were later attested by a comprehensive investigation of the CIOB pumices by considering their chemical compositions (Kalangutkar, 2012).

The Indian Ocean pumices in general and particularly those in the CIOB have been suggested to be largely derived from the eruption of Krakatau (in 1883) and drifted to the basin (Frick and Kent, 1984; Iyer and Karisiddaiah, 1988; Mudholkar and Fujii, 1995; Mukherjee and Iyer, 1999; Pattan et al., 2008). Nevertheless, Iyer and Sudhakar (1993) had postulated a possible in situ submarine silicic volcanism to account for the derivation of the pumices.

Based on a study of the ferromanganese (FeMn) oxides that occur as coating over the CIOB pumices it was suggested that the pumices are between late Miocene and early Pleistocene and perhaps derived from the Indonesian Volcanic Arc (IVA) (Pattan et al., 2013, 2016). But by using empirical formulae based on Fe/Mn ratio it was opined that the ages of the oxides are between 7 Ma and 0.04 Ma which correspond to late Miocene to late Pleistocene periods (Kalangutkar et al., 2015). Furthermore, some of the pumices that form substrate for the FeMn oxides are older than late Miocene (Kalangutkar et al., 2015). Therefore, it is apparent that despite the availability of evidence for and against an in situ or IVA or Krakatau source, yet there is no single or more accepted source/s for the CIOB pumices.

It can be noted from the above information that there is a controversy concerning the derivation of the pumices and their occurrence in large numbers in the CIOB. Our objective is to document and classify a rich array of pumices in the CIOB. Based on mineralogical and petrographical variations we attribute the pumices to various volcanic sources that are within and around the CIOB. To our knowledge this is the first time that such parametres have been used to identify the possible sources of pumices to the basin.

2. Regional setting

The pumice samples for this study were recovered 9°−16°S, 72°−80°E in the CIOB and from an average water depth of 5 000 m (Fig. 1, Table 1). The FeMn nodules, encrustations and crusts were collected using free-fall grabs, Petterson grabs and dredges. Along with these materials, pumices that were invariably present were collected and preserved. The study area includes a number of fracture zones (FZ) such as the 73°E (Vishnu FZ), 76°30′E (Triple Junction Trace- TJT) and 79°E (Indrani FZ) (Fig. 1). In addition, there are eight chains of seamounts of different morphology are located along N–S trending propagative fractures (Das et al., 2007) and isolated seamounts some of which have cratered tops (Kodagali, 1998). The petrology of the seamounts and seafloor indicated their contemporaneous formation since 60 Ma (Iyer et al., 2018; Mukhopadhyay et al., 2018).

Figure  1.  Station location map for the studied Central Indian Ocean Basin pumices. A to H seamount chains along propagative fracture zones are shown. IOTJ: Indian Ocean Triple Junction; FZ: fracture zone. Contour intervals are in 100 m. Base map after Das et al. (2007). Mineral groups are as shown in Table 2. A23, A24, etc. are magnetic anomalies.

DownLoad: Full-Size Img PowerPoint

Table  1.  Station location, water depth and morphology of pumices used in this study

No.	Station	a/cm	b/cm	c/cm	Shape	Latitude	Longitude	Water depth/m
1	F1/42 A	−	−	−	−	10.02°S	74.01°E	NA
2	AAS 26−5	7	5.5	3	oblate	12.52°S	75.69°E	NA
3	AAS 26−16	5	3	2.5	prolate	12.70°S	75.76°E	5 410
4	AAS 22−13	4	3.5	3	equant	12.08°S	75.80°E	5 220
5	F1/48D	−	−	−	−	13.08°S	73.01°E	5 114
6	F1/47D	−	−	−	−	13.50°S	73.99°E	5 088
7	SK 23/277D	−	−	−	−	11.00°S	78.74°E	5 110
8	SS13/696	3.2	2.8	2.1	equant	12.75°S	79.50°E	5 070
9	SS13/696	3.2	1.5	1	equant	12.75°S	79.50°E	5 070
10	AAS 27/12a	−	−	−	−	12.75°S	75.82°E	5 260
11	AAS 27/12b	−	−	−	−	12.75°S	75.82°E	5 260
12	SS6/372	−	−	−	−	11.00°S	78.00°E	NA
13	SS13/704C	7.5	5.3	2	oblate	12.45°S	79.30°E	5 250
14	AAS 26−9	15	11	3	oblate	12.50°S	75.67°E	5 260
15	AAS 3/B15/65a	6.8	6	4.8	equant	12.60°S	75.5°E	5 320
16	SK 23/255 B	7	5	4	equant	9.99°S	78.97°E	5 260
17	AAS 26−33	13	9	9	equant	12.93°S	75.44°E	5 220
18	AAS 26−7	10	7	5	equant	12.54°S	75.67°E	5 180
19	AAS 26−46 B	9	6	4	equant	10.53°S	78.74°E	5297
20	AAS 3/B15/57	13	7.7	5.5	prolate	10.54°S	75.21°E	5 281
21	SS 23/255 CD	7	5	4	equant	9.98°S	78.97°E	5 235
22	AAS 26−44	4	2	1	bladed	12.96°S	75.49°E	5 280
23	AAS 26−19	9.5	6	5	equant	12.74°S	75.75°E	5 350
24	SS7/435C	4.4	3.5	2.5	equant	10.23°S	78.77°E	5 400
25	AAS 26−14	9	5	4	prolate	12.52°S	75.52°E	5 230
26	SS13/683	7	5	5	equant	13.51°S	76.00°E	5 250
27	AAS 27/12c	−	−	−	−	12.75°S	75.82°E	5 260
28	AAS 26−43	18	10.5	4	bladed	12.98°S	75.51°E	5 265
29	AAS 26−27	10.5	5	3	bladed	12.71°S	75.72°E	5 360
30	AAS 26−43	11	10	6	oblate	12.97°S	75.51°E	5 265
31	AAS 26−31	10	5	1	bladed	12.89°S	75.39°E	5 210
32	AAS 26−30	10	6	3	bladed	12.88°S	75.39°E	NA
33	AAS3/B15/69	7.1	5.2	3	oblate	12.73°S	75.70°E	5 305
34	SS10/650C	5.6	3.1	1	bladed	12.46°S	76.89°E	5 350
35	AAS 22−16	11.5	11	6.2	oblate	12.49°S	76.02°E	5 280
36	SS13/691C	8.5	7	5	equant	13.51°S	76.00°E	5 250
37	SS13/691C	6.5	4	2.8	prolate	13.51°S	76.00°E	5 250
38	AAS 26−36	5	3	2	prolate	12.97°S	75.47°E	5 270
39	AAS 26−19	9.5	6	5	prolate	12.74°S	75.75°E	5 350
40	SS13/683 A	4	2	1	bladed	13.51°S	76.00°E	5 250
41	SS1/21B	4.4	2.7	2.4	equant	11.99°S	81.03°E	5 250
42	SS7/435F	4.4	3	2.7	equant	10.22°S	78.76°E	5 400
43	SS7/469 D	5.9	2.6	2	prolate	13.24°S	79.50°E	5 000
44	SS2/104C	3	2.4	1.6	equant	12.00°S	74.99°E	5180
45	SS7/464E	4.8	2.9	1.4	bladed	12.97°S	78.75°E	5 100
46	AAS3/B15/67	4.2	3.1	1	prolate	12.73°S	75.69°E	5 299
47	AAS3/B15/65b	5.6	4.2	3.2	equant	12.62°S	75.57°E	5 320
48	AAS3/B15/59	8	5.3	4.5	prolate	13.49°S	76.99°E	5 356
49	F1/41B	−	−	−	−	13.49°S	76.99°E	5 356
50	SS12/676V	6.2	4.4	5	equant	12.53°S	79.01°E	5 270
51	SS13/704C	7.5	5.3	2	oblate	12.45°S	79.31°E	5 250
52	SS13/691A	5.5	5.1	4.1	equant	12.17°S	79.20°E	5 440
53	AAS3/B15/60	9.7	7.8	3.1	oblate	10.59°S	75.12°E	5 250
54	AAS3/B15/62	5.1	3.2	2.9	prolate	12.867°S	75.69°E	5 329
Note: a: length; b: width; c: height; NA: not available; − represents no data.

 | Show Table

DownLoad: CSV

3. Material and methods

Eighty pumice samples of different sizes and textures were ultra-sonically cleaned; oven dried and were observed under a binocular microscope (Kalangutkar et al., 2011). The shape and of the studied pumices were also determined (Table 1). Fifty four thin sections were prepared and examined for mineralogical and textural variations and on each thin section 500 counts of the minerals were made to determine the modal percentage. The grain size and percentage of each phenocryst (plagioclase, ortho-pyroxene, clino-pyroxene, hornblende and oxide) and total phenocrysts were noted. The vesicularity, glass percentage and the presence FeMn oxides in vesicles (if present) were also noted.

The surface features of the pumices were examined with an SEM-EDS at the Council of Scientific and Industrial Research-National Institute of Oceanography (CSIR-NIO), Goa. The system used was an JSM 5800 LV, JEOL fitted with an INCA Energy Dispersive Spectrometer (EDS). For Scanning Electron Microscopy (SEM) and EDS studies small pieces were chipped off from the pumice specimens and were placed on a carbon stub affixed with carbon tape and coated with carbon to help improve the sample conductivity during the analysis.

Based on petrographical features, grain size and freshness of the minerals, twenty polished thin sections were made for mineral analysis. The phenocrysts to be probed were identified and marked on the sections and analyzed using an JEOL SEM-EDX 840A probe at the Queensland University of Technology, Australia. Backscatter electron (BSE) images were obtained of each polished section to understand the compositional variation within the mineral grains. The data were processed after applying the ZAF correction.

To determine the mineralogy, twenty pumice samples were powdered in a Retsch PM 400-ball mill. X-ray diffraction (XRD) studies were made by using a Rigaku Ultima IV machine with Cu Kα as target. The analytical conditions were as follows: Scanning range between 8° 2θ and 45° 2θ; scan speed: 0.02° 2θ/s (or 1.2° 2θ/min) and the chart speed was 10 mm/2θ. By referring to the JCDPS cards, the minerals were identified by determining the “d” values of the intensity peaks produced by the phases present in the sample.

4. Results

4.1 Physical descriptions

An initial examination of 676 specimens (from which the presently studied 80 samples were selected) revealed two types of pumices: uncoated (75%) and coated with FeMn oxides (25%). The colour of the uncoated pumices ranges from white to light grey, and size varies from 0.25 cm to 12 cm in diameter. Most of the pumice clasts are well rounded to sub-rounded in shape. The shapes of the CIOB pumices are inconsistent with 32% being oblate, 31% equant, 21% prolate and 21% bladed (Kalangutkar et al., 2011) (Table 1). Vesicles in the pumices are both circular and elongated with diameters varying between 5 μm and 15 μm. A few well-developed crystals are present within the vesicles (Kalangutkar et al., 2011). We did not find any direct correspondence between shape and size amongst the studied pumices.

4.2 X-ray diffraction (XRD)

The XRD profiles of the CIOB pumice indicated them to be highly silicic and glassy as evident from the hump between 20° and 30° (Poulsen et al., 1995; Kalangutkar et al., 2011) (Fig. 2). The major minerals are quartz and plagioclase (labradorite) while the less commonly occurring minerals are augite, hornblende, quartz, fayalite and hypersthenes (Fig. 2).

Figure  2.  X-ray diffraction of pumice showing the presence of different mineral phases: fayalite, augite, labradorite, and quartz.

DownLoad: Full-Size Img PowerPoint

4.3 Petrography and mineralogy

Binocular microscopy of the pumices shows fibrous and oval to spherical vesicles and glassy webs. A few of the vesicles contain drusy crystals of quartz, phillipsite crystals, biota and FeMn micronodules (Iyer and Karisiddaiah, 1988; Kalangutkar et al., 2011). Examination of thin sections reveals phenocrysts of plagioclase, biotite, clinopyroxenes, hornblende, quartz and opaque minerals in a glassy and vesicular groundmass (Fig. 3). The vesicle shapes (spherical, oval, elongated) and sizes (20–400 µm) show intra- and inter-clast variations.

Figure  3.  Photomicrographs of pumices vesicles in glass (a), plagioclase and hornblende (b), biotite (c), glomeroporphyritic texture, plagioclase and clinopyroxenes (d), zoned plagioclase (e), hornblende (f).

DownLoad: Full-Size Img PowerPoint

Based on the mineral assemblages and modal percentages, the 54 thin sections were classified into eight groups and in all these groups glass is omni-present (Table 2). The eight groups are as follows.

Table  2.  The sub-groups of pumice based on mineralogy within the major eight groups

Group 1	Group 2		Group 3	Group 4	Group 5		Group 6	Group 7	Group 8
Plag	Plag +Cpx		Plag+Opx	Plag+biotite	Plag+Hbl		Plag+Opx+ Cpx	Plag+Hbl+ Pyx+Qtz	Plag+Cpx+ Qtz
−	Cpx rich	Cpx poor	−	−	Hbl rich+Plag poor	Plag rich+ Hbl poor	−	−	−
AAS 26−44	AAS 26−19	SK 23/255B	SS10/650C	AAS 26−36	AAS 26−5 oblate	SS6/372	AAS 26−30	SS13/691C	SS1/21B
AAS 26−19	AAS 27/12	F1/47D	AAS 22−16	AAS 3/B15/57	AAS3/B15/67 prolate	SS13/683	SS10/650C	−	−
AAS 27/12	−	SS13/704C	AAS 26−31	SK 23/255 CD	−	SS13/696	−	−	−
F1/48D	−	SS7/435C	−	AAS 26−43	−	−	−	−	−
SS7/464E	−	F1/42A	AAS 26−43	AAS3/B15/59	−	−	−	−	−
SS7/435C	−	−	SS7/469D	AAS3 B15/65	−	−	−	−	−
−	−	−	SS13/704C	AAS 27/12	−	−	−	−	−
−	−	−	SS12/676V	−	−	−	−	−	−
Note: Plag: plagioclase; Cpx: clinopyroxene; Opx: orthopyroxene; Hbl: hornblende; Qtz: quartz; − represents no data.

 | Show Table

DownLoad: CSV

Group 1: (plagioclase) (17%)

Group 2: (plagioclase+clinopyroxenes) (19%)

Group 3: (plagioclase+orthopyroxenes) (19%)

Group 4: (plagioclase+biotite) (19%)

Group 5: (plagioclase+hornblende) (14%)

Group 6: (plagioclase+clinopyroxenes+orthopyroxenes) (6%)

Group 7: (plagioclase+hornblende+pyroxene+quartz) (3%)

Group 8: (plagioclase+clinopyroxenes+quartz) (3%)

It should be noted that since sizes of the pumice are less than 12 cm, the different mineral assemblages may occur in the different fragments from the same source. Hence, we wish to show that based on the abundance of ferromagnesian minerals at least four major groups can be identified. These are as follows:

Group 2 (plagioclase+clinopyroxenes); Group 3 (plagioclase+orthopyroxenes); Group 4 (plagioclase+biotite) and Group 5 (plagioclase+hornblende). Within these groups, sub-groups (Table 2) were made considering the abundance of the phenocrysts. Group 2 has two sub-groups (plagioclase+clinopyroxenes rich) and (plagioclase+clinopyroxenes poor). Also, Group 5 has two sub-groups (hornblende rich+plagioclase poor) and (hornblende poor+ plagioclase rich). The major groups along with the subgroups showing different mineral assemblages were plotted on a bathymetry map to examine if there was any tendency of the groups to occur at a particular site (Fig. 1). Incidentally, no preferential affinity of the pumices with any of the seafloor features was noted. This observation rules out the possibility of a particular volcanic site to have produced pumices with unique and of a single mineral assemblage.

4.4 Mineral and glass chemistry

Textural and chemical zoning patterns of crystals have been used as guides in understanding magmatic processes. This is because the mineral phases have a tendency to record the processes occurring in the magmatic system during their growth (Allègre et al., 1981; Iyer and Banerjee 1998; Ginibre et al., 2007). The diagnostic signatures of such processes manifest in the form of compositional zones and twin crystals and, therefore, these can be used as tracers of the magmatic history. Microscopy studies show unzoned (Figs 3b, d) and compositional zoned plagioclase (Fig. 3e) while BSE images show plagioclase as phenocrysts and inclusions in some of the pyroxene and biotite grains (Fig. 4).

Figure  4.  Back scatter electron micrographs. a. Biotite crystals with oxide inclusion; b. plagioclase crystal with biotite inclusion; c. hornblende crystal; d. 3: plagioclase, 2: pyroxene, 1: iron oxide; e. 1: zoned plagioclase, 2: non-zoned plagioclase; f. zoned plagioclase.

DownLoad: Full-Size Img PowerPoint

The mineral compositions of plagioclase (zoned and unzoned), pyroxene, hornblende, biotite, magnetite and glass were determined using an EDS. The mineralogical data of the CIOB pumice samples are compared with those of Krakatau, IVA and Andaman. This is because these areas, that border the CIOB, are the postulated sources for the pumices in the CIOB (Iyer and Karisiddaiah 1988; Mudholkar and Fujii, 1995; Frick and Kent, 1984; Mukherjee and Iyer, 1999; Pattan et al., 2008) and for the silicic glass shards that occur in the CIOB sediments (Mascarenhas et al., 2006).

Plagioclase: Plagioclase was observed to be the most dominant phenocryst phase and range from 1% to ~10% (avg 4.7%). The plagioclase crystals range in size from 0.1 mm to 1.6 mm (minor axis) and 0.2 mm to 2.8 mm (major axis). Plagioclase phenocrysts and lath shaped microlites show lamellar twinning. A few phenocrysts are fresh, euhedral, have well-defined oscillatory zones (Fig. 3e) and do not have inclusions. There are two populations of plagioclase phenocrysts: twinned (and “fritted”) and unzoned ones.

The average composition of plagioclase is SiO₂~57 wt%, Al₂O₃~27 wt%, CaO~8 wt%, and Na₂O~4 wt% (Table 3). The anorthite (An) content indicates the presence of labradoritic, bytownitic to anorthitic plagioclase composition (Fig. 5). It is observed that most of the phenocrysts have calcic cores (An₅₄−An₈₈) and sodic rims (An₄₄−An₈₇) (Table 3).

Table  3.  Microprobe analysis of core and rim of plagioclase grains in the Central Indian Ocean Basin pumices

Analysis of core of plagioclase grains
Phenocrysts	Sample	SiO2	TiO2	Al2O3	FeO	MnO	BaO	CaO	Na2O	K2O	Total	An	Ab	Or
n=6	A42	56.17	bd	29.18	0.42	0.09	0.20	10.03	1.80	0.28	98.04	83.73	13.74	2.54
n=9	A39	57.75	0.11	27.80	0.42	0.11	0.07	8.70	2.15	0.13	97.39	78.98	19.82	1.20
n=3	S2	58.25	bd	26.60	0.20	0.38	bd	9.81	5.51	0.15	100.66	63.42	35.61	0.97
n=4	S1	54.36	0.07	26.03	0.47	0.04	0.04	9.43	5.23	0.18	97.13	63.75	35.33	0.92
n=2	A13	58.31	0.12	24.89	0.47	bd	0.02	7.82	6.18	0.34	98.67	54.46	43.21	2.34
n=1	S15	54.98	bd	25.82	0.47	0.11	0.03	9.69	5.21	0.16	97.27	64.34	34.59	1.06
n=4	A33	53.43	0.08	30.12	0.22	0.02	0.06	11.54	3.84	0.10	100.24	74.49	20.07	1.00
n=5	S7	61.24	bd	25.85	0.22	0.04	0.07	6.11	4.29	0.12	97.97	58.06	40.78	1.16
n=1	A15	57.05	bd	32.56	0.45	bd	0.17	9.85	1.25	0.05	101.80	88.34	11.21	0.45
Analysis of rim of plagioclase grains
Phenocrysts	Sample	SiO2	TiO2	Al2O3	FeO	MnO	BaO	CaO	Na2O	K2O	Total	An	Ab	Or
n=5	A42	56.43	bd	29.11	0.43	0.07	0.24	10.13	1.44	0.26	97.92	87.34	10.37	2.28
n=8	A39	57.88	0.07	27.70	0.47	0.18	0.08	8.47	2.43	0.12	97.30	76.79	22.12	1.09
n=4	S2	61.71	0.13	22.93	0.19	0.29	bd	5.99	6.98	0.45	98.33	44.98	52.51	2.52
n=3	S1	55.32	0.06	25.56	0.41	0.11	0.13	9.19	5.40	0.16	97.03	62.33	36.61	1.06
n=3	A15	57.20	bd	32.49	0.44	0.08	0.12	9.72	1.31	0.08	101.69	87.46	11.85	0.69
n=1	S9	57.83	0.02	25.56	0.21	bd	0.11	7.74	6.32	0.24	98.80	54.13	44.20	1.68
n=2	A13	57.73	0.06	25.30	0.46	bd	0.05	8.36	5.94	0.21	98.47	57.60	40.95	1.46
n=1	S15	57.62	bd	25.51	0.41	0.03	bd	8.58	5.71	0.06	98.39	59.79	39.79	0.42
n=2	A33	57.85	0.08	27.29	0.17	0.05	bd	8.31	5.97	0.13	100.63	57.71	41.43	0.87
n=1	S9	57.83	0.02	25.56	0.21	bd	0.11	7.74	6.32	0.24	98.80	54.13	44.20	1.68
n=2	A30	57.28	bd	28.03	0.47	bd	0.18	8.87	2.01	0.27	97.49	79.55	18.03	2.42
n=2	S7	61.85	bd	25.58	0.19	0.03	0.03	5.75	4.65	0.17	98.29	54.31	44.10	1.61
n=1	A31	58.36	0.03	28.00	0.47	bd	0.03	8.55	2.83	0.26	98.52	73.45	24.31	2.23
Note: bd: below detection.

 | Show Table

DownLoad: CSV

Figure  5.  The plot displays the rim and core compositions for the plagioclases of the Central Indian Ocean Basin, plagioclases of Krakatau pumices (Camus et al., 1987) and plagioclase composition of Toba pumices (Chesner, 1998). YTT: Younger Toba Tuff; MTT: Middle Toba Tuff; OTT: Older Toba Tuff; HDT: Haranggoal Dacite Tuff. Or: orthoclase.

DownLoad: Full-Size Img PowerPoint

We compared the plagioclase compositions of the CIOB pumices with those of Toba (Chesner, 1998) and Krakatau (Camus et al., 1987) (Fig. 5, Table 3). It is evident that the CIOB pumices do not belong to Younger Toba Tuff (YTT), Middle Toba Tuff (MTT) and Older Toba Tuff (OTT), but can be correlated with Haranggoal Dacite Tuff (HDT) and Krakatau 1883 eruption.

Andesite (age 1.8 Ma) from the Barren Island, Andaman shows the groundmass plagioclase to have An₃₅−An₈₅ with phenocryst up to An₉₁ (Ray et al., 2013) and these are compositionally similar to those of the CIOB pumices. The plagioclase phenocrysts from the Narcondam Island vary in composition from An₄₄ to An₈₁ while the groundmass plagioclases have An₄₇−An₈₀ (Ray et al., 2011). The Andaman andesites host labradoritic, bytownitic and andesinic plagioclases while our pumices are devoid of andesinic but have anorthitic plagioclases.

Pyroxene: Glomeroporphyritic aggregates of plagioclase and clino- and/or ortho-pyroxenes were observed in most of the pumice samples (Fig. 3d). The modal analysis indicates clinopyroxenes to be marginally more (0.2% to 1.8%, avg 0.8%) than ortho-pyroxenes (0.2% to 1.2%, avg 0.55%). Mudholkar and Fujii (1995) had reported an abundance of clinopyroxenes over orthopyroxenes in the CIOB pumices. Pattan et al. (2013) reported the presence of orthopyroxenes in pumices that were either fully or partly coated with FeMn oxides but clinopyroxenes were absent. Probably this may be due to the fact that coated and uncoated pumices are not from the same source and are of different ages (Kalangutkar et al., 2015).

The clinopyroxene crystals in our samples range in size from 0.1 mm to 0.2 mm (minor axis) and 0.2 mm to 0.6 mm (major axis) while orthopyroxene crystals are 0.1 mm (minor axis) and 0.15 mm to 0.3 mm (major axis). A few of the pyroxene crystals has inclusions of iron-oxides while clinopyroxenes occur as inclusions in orthopyroxenes. We identified high Ca-rich clinopyroxene i.e., augite and diopside and also Ca-poor pigeonite and orthopyroxene (Table 4). Pigeonite has not been reported from the CIOB pumices and their present compositions do not correspond with the Andaman dacite-andesites reported by Pal et al. (2007). Several of the clino- and ortho-pyroxenes from the CIOB pumices are similar to the HDT and Krakatau pyroxenes (Fig. 6). The average composition of the CIOB pyroxenes is SiO₂~53 wt%, Al₂O₃~4 wt%, FeO~15 wt%, MgO~17 wt%, MnO~1.1 wt%, CaO~7 wt%, and Na₂O~0.3 wt%.

Table  4.  Microprobe analysis of pyroxene grains in the Central Indian Ocean Basin pumices

Phenocrysts	Sample	SiO2	TiO2	Al2O3	FeO	MnO	MgO	CaO	Na2O	K2O	Cr2O3	Total	Wo	En	Fs
n=25	A42	54.01	0.23	0.83	18.64	1.27	24.27	1.82	0.04	0.03	0.06	101.10	4.06	54.25	41.69
n=17	A30	53.22	0.28	0.62	19.28	1.22	19.71	4.64	0.10	0.02	bd	99.03	10.71	45.14	44.15
n=10	A31	53.62	0.43	0.63	21.19	1.21	22.97	1.04	0.04	0.50	bd	100.98	2.30	50.82	46.88
n=4	S7	53.29	0.62	2.53	25.70	1.29	14.24	2.12	0.23	0.03	bd	99.69	5.04	33.85	61.11
n=16	A39	53.56	0.38	1.00	15.93	1.40	21.03	6.91	0.10	0.10	0.26	100.34	16.06	47.78	36.16
n=6	S11	52.52	0.22	0.48	21.71	1.36	21.76	1.20	0.06	0.01	bd	99.26	2.68	48.72	48.60
n=1	S2	55.43	0.08	bd	16.72	2.56	27.06	0.60	bd	bd	bd	102.44	1.35	60.97	37.68
n=1	A15	54.89	0.54	2.01	7.82	0.53	17.59	17.81	0.15	bd	0.27	101.60	41.21	40.69	18.10
n=1	S1	54.09	0.25	0.74	21.15	1.52	23.37	1.30	0.04	bd	bd	102.45	2.84	51.01	46.15
n=5	A25	55.03	0.56	2.59	23.87	1.26	14.27	1.92	0.30	bd	bd	99.79	4.79	35.63	59.58
n=1	A2	49.99	1.99	13.56	7.93	1.19	15.73	9.81	0.95	0.32	bd	101.47	29.31	47.00	23.69
n=2	A13	52.48	0.05	1.09	20.95	1.27	21.99	0.35	bd	1.50	bd	99.66	0.81	50.80	bd
n=2	S15	52.90	0.04	0.44	26.61	1.49	17.97	0.29	bd	1.62	bd	101.33	0.64	40.05	59.32
Note: Wo: wollastonite; En: enstatite; Fs: ferrosilite; bd: below detection.

 | Show Table

DownLoad: CSV

Figure  6.  Pyroxene classification diagram. Pyroxenes of the Central Indian Ocean Basin pumices plotted along with pyroxenes from Krakatau pumices (Camus et al., 1987) and pyroxenes from Toba eruptives (YTT, MTT, OTT) (Chesner, 1998). HDT: Haranggoal Dacite Tuff; OTT: Older Toba Tuff ; MTT: Middle Toba Tuff; YTT: Younger Toba Tuff; HDT: Haranggoal Dacite Tuff.

DownLoad: Full-Size Img PowerPoint

Hornblende: Hornblende grains are uncommon in the CIOB pumices and are reported for the first time (Table 5). Hornblende crystals have dimensions of 0.1 mm to 0.2 mm (minor axis) and 0.2 mm to 0.6 mm (major axis) and occur as microlites and phenocrysts. Some of the phenocrysts are slightly altered. Hornblende forms 0.4% to 2% (avg 1.2%) and co-occur with plagioclases (Fig. 3b) and/or biotites (Fig. 3f). The composition of the CIOB hornblende is SiO₂~50 wt%, TiO₂~1.5 wt%, Al₂O₃~6.6 wt%, FeO~12 wt%, MgO~14 wt%, MnO~0.2 wt%, CaO~10.2 wt%, Na₂O~1.7 wt%, and K₂O~0.2 wt%. Hornblende has not been reported from either Krakatau or Toba eruptives; hence the source for hornblende-bearing pumices may be different, as would be discussed later (Table 5).

Table  5.  Electron Dispersive Spectroscopy analysis of hornblende grains in Sample A33 (AAS26−43)

Phenocrysts	Sample	SiO2	TiO2	Al2O3	FeO	MnO	MgO	CaO	Na2O	K2O	Total
n=5	A33	50.52	1.4	6.636	12.602	0.265	14.65	10.285	1.72	0.266	96.27

 | Show Table

DownLoad: CSV

Biotite: Biotite grains are present in a few of the pumices (Fig. 3c) and depict typical pleochroism in shades of brown. The grains are euhedral and are present along with plagioclase and/or hornblende as phenocrysts or as individual grains. The size of the biotite grains ranges from 0.05 mm to 0.4 mm (minor axis) and 0.1 mm to 0.6 mm (major axis) and the modal percentage is 0.2% to 0.4% (avg 0.35%). Biotite grains were also observed as inclusions in plagioclase while those of oxides occur in most of the biotite grains. The biotites in the CIOB pumices have a very low FeO/MgO (0.77) and higher values of MnO (0.36 wt%) and TiO₂ (5 wt%) (Table 6).

Table  6.  Microprobe analysis of biotites grains in Sample A25 (AAS 3/B15/57)

Phenocrysts	Sample	SiO2	TiO2	Al2O3	FeO	MnO	MgO	CaO	Na2O	K2O	Total
n=5	A33	50.52	1.4	6.636	12.602	0.265	14.65	10.285	1.72	0.266	96.27

 | Show Table

DownLoad: CSV

Quartz: Free quartz was observed in a few thin sections with a modal percentage of 0.4%. No free quartz has been reported from earlier studies of the CIOB pumices.

Iron-titanium (Fe-Ti) oxides: The Fe-Ti oxides form aggregates with pyroxenes, hornblende, biotite and plagioclase and were also present as inclusions within other minerals or as individual grains (Fig. 4a) and has a modal percent of 0.2% to 1.8% (0.33) (Table 7). The Fe-Ti oxides with TiO₂ (6–47 wt%), FeO (48–87 wt%) and very low CaO (<1 wt%), fall in the ulvospinel zone while a few plots close to wustite and ilmenite fields (Fig. 7). Ulvospinel is being reported for the first time in the CIOB pumices while titano-magnetite (Oba et al., 1983) and ilmenite (Camus et al., 1987) are present in the Krakatau pumices.

Table  7.  Microprobe analysis of iron-titanium oxides in the Central Indian Ocean Basin pumices

Phenocrysts	Sample	SiO2	TiO2	Al2O3	FeO	MnO	MgO	CaO	NiO	Cr2O3	Total
n=4	A42	0.02	12.20	2.90	81.54	0.85	2.33	0.07	0.02	0.11	99.98
n=3	A31	0.20	13.59	2.48	80.98	0.55	1.99	0.03	0.01	0.26	99.99
n=2	A39	0.12	17.29	2.39	77.01	0.75	2.02	0.04	0.01	0.21	99.72
n=2	S11	0.15	16.80	2.37	77.58	0.75	1.99	0.04	0.01	0.21	99.59
n=3	A32	0.15	14.83	2.38	79.44	0.95	1.99	0.04	0.03	0.18	99.30

 | Show Table

DownLoad: CSV

Figure  7.  Oxides-ilmenite-ulvospinel plot after Kimura and Muan (1971). The Fe-Ti oxides from the Central Indian Ocean Basin pumices plot within the ilmenite-ulvospinel field.

DownLoad: Full-Size Img PowerPoint

Glass: The glassy groundmass is the most abundant phase in the CIOB pumices and has a modal percentage ranging from 33% to 54% (avg 43%). The glass either shows isolated, non-connected, oval, augen-shaped vesicles or has highly fibrous, well-connected vesicles and exhibits a swirled texture (Fig. 3a). In several samples, the glass is quite fresh and has SiO₂ from 52 wt% to 80 wt%, TiO₂ 0.6 wt% to 1.0 wt%, FeO 0.8 wt% to 12.6 wt%, MgO 1 wt% to 3 wt% and K₂O 0.7 wt% to 4.3 wt% (Table 8).

Table  8.  Electron Dispersive Spectroscopy analysis of glassy matrix of the Central Indian Ocean Basin pumices

Sample	SiO2	TiO2	Al2O3	FeO	MnO	CaO	Na2O	K2O	MgO	Total
SS7-432D	52.07	1.04	17.57	12.63	0.45	8.90	3.54	0.69	3.12	100.01
SS7-437F	68.94	bd	11.16	8.24	0.52	2.53	4.13	1.40	3.07	99.99
SS7-474C	78.95	bd	12.05	0.84	bd	0.77	3.14	4.25	bd	100.00
SS7-474C	78.92	bd	12.00	0.81	bd	0.65	3.42	4.19	bd	99.99
SS2-66BG	63.26	1.30	13.35	8.82	bd	4.36	4.94	2.23	1.74	100.00
SS7-460C	69.26	0.77	11.60	4.58	bd	2.38	7.76	2.58	1.07	100.00
SS1-57A	68.09	0.61	12.25	3.90	bd	2.41	9.00	2.55	1.19	100.00
SS2-106C	59.55	1.40	15.14	9.12	0.47	5.16	4.27	2.77	2.12	100.00
SS2-106C	59.84	1.13	15.49	9.59	0.65	4.15	3.91	2.89	2.35	100.00
SS4-216 E	77.67	bd	12.17	1.65	bd	0.69	3.61	3.81	0.40	100.00
SS4-216 E	77.67	bd	12.17	1.65	bd	0.69	3.61	3.81	0.40	100.00
SS3 131 A,B	75.32	bd	12.00	3.02	bd	2.03	3.93	3.70	bd	100.00
AAS 26−19	79.80	bd	11.03	1.89	bd	0.53	2.91	3.84	bd	100.00
Note: bd: below detection.

 | Show Table

DownLoad: CSV

The major oxides of the CIOB pumices are as follows, SiO₂ 60 wt% to 75 wt%, TiO₂ 0.14 wt% to 1.06 wt%, Al₂O₃ 12 wt% to 17 wt%, FeO 1 wt% to 8 wt%, MnO 0.05 wt% to 3 wt%, MgO from 0.22 wt% to 2 wt%, CaO 0.8 wt% to 6 wt%, Na₂O 3 wt% to 5 wt%, K₂O 0.81 wt% to 5 wt% and P₂O₅ 0.02 wt% to 0.4 wt% (Kalangutkar, 2012). We have initiated detailed geochemical studies of the various pumices that will help to confirm their sources in the CIOB. Similar studies have been reported for the Okinawa Trough by Shinjo and Kato (2000), Guo et al. (2018), Chen et al. (2020).

5. Discussion

We now discuss the results that were obtained through XRD analysis, microscopy observations, petrographical features and mineral compositions of the CIOB pumices. Further, a comparison of the characteristics of the CIOB pumices is made with those that erupted in the vicinity of the basin (Table 9). Finally, we provide evidence that either supports or rules out some of the postulated eruptive source/s to have contributed drift pumices to the CIOB.

Table  9.  Comparison of the mineral assemblages of the Central Indian Ocean Basin (CIOB) pumices with those in the vicinity of the basin

Samples	Mineral assemblage	Reference
CIOB pumices	plagioclases 1.1% to 9.6% (4.7%)	present work
	clinopyroxene 0.2% to 1.8% (0.8%)
	orthopyroxene 0.2% to 1.2% (0.55%)
	biotite 0.2% to 0.4% (0.35%)
	hornblende 0.4% to 2% (1.2%)
	quartz (0.4%) (±fayalite+oxides)
	vesicles 33% to 45.4% (38.2%)
	glass 32.8% to 53.86% (42.6%)
Krakatau 1883 pumice	white band grey band	Camus et al. (1987)
	plagioclase (12%) (7.5%) clinopyroxene (rare) (2%)
	orthopyroxene (rare) (0.5%)
	oxide (rare) (1%)
	hornblende, biotite, quartz (not present)
Krakatau 1883 pumice	plagioclase (5.7%)	Oba et al. (1983)
	clinopyroxene (0.4%)
	orthopyroxene (0.6%)
	oxide (1.1%)
	vesicles (92%)
Anak 1927–1979 pumice	plagioclase (29%)	Oba et al. (1983)
	olivine (2.5%) (sometimes jacketed by Opx)
Anak 1981 pumice	clino- and ortho- pyroxene (3%)	Oba et al. (1983)
	magnetite oxide (1%)
	plagioclase (23%)
	clinopyroxene (3%)
	orthopyroxene (3.5%)
	magnetite and ilmenite oxide (2%)
Andaman, Barren and Narcondam andesite and dacite rocks	calcic and sodic plagioclase, forsteritic olivine, clinopyroxene, orthopyroxene, hornblende, minor quartz and apatite, andesites and dacites (not pumice), hence modal proportions is not given	Ray et al. (2011)
Narcondam pumice	plagioclase, amphibole, pyroxene	Pal et al. (2007)
HDT Toba pumice	plagioclase, ortho- and clinopyroxene, magnetite, ilmenite	Chesner et al. (1998)
OTT Toba pumice	quartz, plagioclase, sanidine, biotite, amphibole, orthopyroxene, magnetite, ilmenite, allanite, zircon
MTT Toba pumice	quartz, plagioclase, sanidine, biotite, amphibole, orthopyroxene, magnetite, ilmenite, allanite, fayalite, zircon
YTT Toba pumice	quartz, plagioclase, sanidine, biotite, magnetite, ilmenite, allanite, fayalite, zircon, orthopyroxene
Taupo, New Zealand pumice	plagioclase, orthopyroxene, quartz, hornblende, magnetite (as inclusions in orthopyroxene, plagioclase, hornblende or as microphenocrysts), ilmenite, acessory zircon and apatite (as inclusions in orthopyroxene, plagioclase and magnetite and as microphenocrysts)	Wilson et al. (2006)
South Sandwich pumice	plagioclase (andesine-labradorite), corroded feldspar	Risso et al. (2002)
	(sanidine-anorthoclase), clinopyroxene (diopside or augite), orthopyroxene (hypersthene), red-brown hornblende, magnetite,vesicle (50%−60%)
Tonga pumice	calcic plagioclase, augite, pigeonite, titanomagnetite	Bryan et al. (2004)
Kerguelen pumice lapilli layer in basaltic and silicic ash	fresh sanidine, Ab-rich plagioclase, hedenbergitic clinopyroxene, few quartz	Bitschene et al. (1992b)

 | Show Table

DownLoad: CSV

XRD profiles revealed the presence of plagioclase, quartz, hornblende, augite and fayalite (Fig. 2) and these were confirmed through petrographic examinations. As stated earlier, based on the presence of ferromagnesian minerals, the CIOB pumices form eight groups and of these the dominant groups are: Group 2: plagioclase+clinopyroxenes, Group3: plagioclase+orthopyroxenes, Group 4: plagioclase+hornblende and Group 5: plagioclase+biotite. Depending on the group, plagioclase, quartz, hornblende and augite occur as phenocrysts of variable sizes. The modal percentage in the studied samples show up to 12.4% phenocrysts, glass 33% to 54% and vesicles 29% to 54%.

The observed range of phenocryst chemistry and zoning of plagioclases are indicative of their formational conditions and variable magma compositions. Perhaps the source magma erupted under polybaric conditions and the pumices represent composite products of more than one petrogenetic stages and/or sources. The phenocrysts of plagioclase show fine concentric zoning along the rim and suggest changing pressure (P), temperature (T) and/or water conditions. This could be due to an increase in water pressure during ascent of magma (cf. Vance, 1965; Almeev and Ariskin, 1996) which led to the crystallisation of sodic plagioclase, perhaps due to an influx of Ca-rich magma or due to magma mixing. Since zoning in plagioclase is inversely related to ascent rate of magma (Fisk, 1984), the CIOB pumices with zoned plagioclases indicate a rapid cooling of the lava and insufficient time for the crystals to equilibrate with the host melt. We suggest that the lowering of pressure during magma ascent may have led to fracturing and resorption and recrystallisation of the plagioclase grains.

The plagioclases in the CIOB pumices are labradoritic, bytownitic and anorthitic in composition (Fig. 6) and most of the normal zoned phenocrysts (Samples S9, S7, S15, A33) show compositional differences and have calcic cores (An₅₀−An₉₆) (Table 3) and sodic rims (An₄₁−An₈₈) (Table 3). From the observations it can be said that these compositional changes could have been caused by rhythmic changes in the magma chemistry.

Some of the studied pumices show resorbed margin of quartz grains which again indicate fluctuating magma temperature and this is supported by the presence of zoned plagioclase that suggest variable magma composition or mixing of magmas (cf. Hibbard, 1981; Müller et al., 2005). Magma mixing could also result in rapid crystallisation and build-up of volatiles in the magmatic mush (Cox and Bell, 1972) which can trigger vesiculation and explosion (Clynne, 1999). This fact is reinforced by the occurrence of hydrous minerals like biotite and hornblende in the CIOB pumices and implies their formation in the presence of high contents of volatiles or interaction with water during crystallisation (cf. Beard et al., 2004).

5.1 Comparisons with other pumices

We now compare the mineral assemblages of the CIOB pumices with those of the submarine silicic volcanics surrounding the study area, so as to identify if any of the terrestrial volcanoes may have contributed pumices to the CIOB (Table 9).

Andaman-Narcondam: Pal et al. (2007) identified vesicular glass-rich pumices in the Andaman. These had resorbed plagioclase and sometimes clinopyroxenes and amphibole within a glassy mass that produced a sieve texture in the xenoliths. But no further data about the pumices are available for comparison.

“Fritted” plagioclase, similar to those observed by Ray et al. (2011) in the Andaman rocks, has also been observed in a few of our CIOB samples. Labradoritic, bytownitic and anorthitic composition of plagioclase in the pumices are akin to those from Narcondam volcanic that are a product of mixing of basic (andesitic) and acidic (rhyolitic) magmas (Pal et al., 2007; Ray et al., 2011). Silicic magma produces foam (Clynne, 1999) and the formation of such foam at Andaman-Narcondam perhaps resulted in the breaking of several pumice blocks due to instability and their subsequent drift to the CIOB; but this possibility needs further confirmation.

South Sandwich Island: The micro phenocrysts of plagioclase in the South Sandwich (Southern Atlantic Ocean). Pumices occur as individual phenocrysts or in an ophitic arrangement and form less than 5% while vesicles make up 50%–60%. The plagioclases (andesine-labradorite) are mostly euhedral in shape and are zoned. The associated minerals are corroded K-feldspar (sanidine-anorthoclase), clinopyroxene (diopside or augite), orthopyroxene (hypersthene), magnetite and red-brown hornblende (Risso et al., 2002). Though labradoritic plagioclase occurs in the CIOB pumices but alkali feldspars are absent. Hence, the South Sandwich Island as a source for the CIOB pumices seems to be doubtful.

Tonga volcano: The Tonga volcano pumices have calcic plagioclase, augite, pigeonite and titano-magnetite (Bryan et al., 2004). This mineral assemblage is similar to that observed in most of the examined CIOB pumices. Some of the pigeonite composition of our samples is akin to those from the 1964 pumice eruption of Tonga. Though the orthopyroxenes from the CIOB pumices show similarity to 1964 Tonga pumice but lie between the compositions of oldest (3.5–1.8 ka) and youngest (September 2001) eruptions of Tonga (Bryan et al., 2004).

Kerguelen Plateau: The Kerguelen Plateau is known to have erupted voluminous basaltic lavas (Bitschene et al., 1992a, b) but phases of silicic volcanism occurred in late Miocene and Pliocene with the major volcanic activity in the northern Kerguelen Plateau (Bitschene and Schmincke, 1990). These volcanics contain Na and Fe-rich pyroxenes, alkali feldspar, apatite and Ti-magnetite with subordinate biotite and zircon. Pumice has also been reported from Ocean Drilling Program (ODP) and Deep Sea Drilling Project (DSDP) Legs 119 and 120 and comprise Site 736 (Leg 119) from the Kerguelen Plateau. The younger pumices are of Quaternary age while the older ones are of Oligocene to Pliocene age (Kidd et al., 1992).

Frey et al. (2000) recovered a 20 m thick volcano-clastic succession containing six trachytic pumice-lithic breccias deposited by pyroclastic flows from the Central Kerguelen Plateau. Bitschene et al. (1992b) reported the occurrence of 90%–95% of blocky to well-rounded, silicic, highly vesicular pumice lapilli layer containing fresh sanidine, Ab-rich plagioclase, hedenbergitic clinopyroxene, quartz and biotite along with basaltic and silicic ash from the Kerguelene Plateau. Such mineral assemblages are absent in our pumices hence, Kerguelene is ruled out as a source for the CIOB pumices.

Krakatau (in 1883): It is observed that the eruption of Krakatau (in 1883) produced pumices with minerals and modal assemblages that are largely comparable to some of our samples though hornblende and biotite have not been reported from Krakatau. This suggests that the hornblende and biotite containing pumices of the CIOB may have a different source. We also rule out the younger eruptions from Krakatau, the Anak 1927–1979 and Anak 1981 as sources because these have higher modal percentage 29% and 23%, respectively (Oba et al., 1983) than our samples (3%–19%).

Toba: The Toba volcano in the IVA had four major eruptions: YTT, MTT, OTT and HDT and these have ages of 0.074 Ma, 0.50 Ma, 0.84 Ma and 1.20 Ma, respectively (Chesner, 1998). The pumices of these eruptions have quartz, plagioclase, sanidine, biotite, magnetite, ilmenite, allanite, fayalite, zircon and orthopyroxene (Chesner, 1998). Of these, allanite, zircon and sanidine are absent in the CIOB pumices. The HDT bears plagioclase, ortho- and clinopyroxenes, magnetite and ilmenite and this assemblage is comparable to that found in our pumices.

Taupo, New Zealand: The Taupo volcanics of New Zealand shows plagioclase, orthopyroxene, quartz, hornblende, magnetite, ilmenite, zircon and apatite (Wilson et al., 2006). Although apatite is reported as an accessory mineral in ortho- and clino-pyroxenes in only two pumices (Mudholkar and Fujii, 1995) but none of the 54 specimens we examined have apatite. Hence, Taupo may not be the source for the CIOB pumices under investigation.

In the light of the above findings, it is clear that some of the external sources could be ruled out as contributors for the CIOB pumices. This leaves us with three probable sources: Krakatau volcano (Iyer and Karisiddaiah, 1988; Mudholkar and Fujii, 1995), the several eruptions of Toba (Chesner, 1998) and possibility of intraplate silicic volcanism (Iyer and Sudhakar, 1993; Iyer, 1996; Mukherjee and Iyer, 1999).

Based on the mineral chemistry of our investigations we confirm the Krakatau volcano as one of the sources for the CIOB pumices. This is also borne out by the fact that the August 26–27, 1883 eruption of Krakatau volcano (Indonesia) resulted in pumice blocks that drifted across the Indian Ocean and were even found along the beaches of the East coast of Africa (Frick and Kent, 1984). However, it was reported that in the absence of favourable surface current circulation patterns, Krakatoan volcanic products may not occur in any significant amount in the recent sediments east of 80°E (Svalnov, 1981).

Although the YTT as a source was suggested (Pattan et al., 2008, 2013), we negate this view because the YTT eruption largely produced huge amounts of silicic glass shards and hardly any pumices (Chesner and Rose, 1991). There are several studies concerning the extensive occurrence of glass shards in the CIOB sediments but there is no mention about the find of YTT pumices (Amonkar et al., 2020 and references therein). It appears that the HDT eruption may have contributed some of the pumices to the CIOB. There is no reported occurrence of biotite in pumice from Krakatau 1883 eruption (Camus et al., 1987; Oba et al., 1983). Similarly, hornblende has not been reported from Toba (Chesner, 1998) and Krakatau (Camus et al., 1987; Oba et al., 1983). Hence, biotite-and hornblende-bearing CIOB pumices may be from unknown terrestrial sources or from in situ eruption.

5.2 The explosive eruptions in the deep sea

Pumice clasts recovered from abyssal depths are considered to be exotic because of the constraints in their formation yet; there are several reports of pumices in the deep sea. For example, in the Atlantic Ocean pumice was reported over submarine volcanoes at depths of 1 500 m to 2 000 m (Neumann von Padang, 1967), near Tonga depths at >1 500 m ( Fouquet et al., 1991) Okinawa Trough, South of Japan (Halbach et al., 1989) and trachytic pumice flows (with hydrothermal crusts) occur at intraplate volcanoes in the Society and Austral hotspot regions (water depth=3 500 m; Binard et al., 1992). Silicic volcanism at an off-axis geothermal field in the Mariana Trough resulted in rhyodacitic pumice between 2 600 m and 3 600 m water depth. A possible suggested source for the silicic magma was ascribed to melting of the sea floor sediments during volcanisms and the sediments include altered volcano clastic debris from the adjacent island arc and siliceous hemipelagic muds (Lonsdale and Hawkins, 1985).

Hédervari (1982) reported the presence of pumices at water depths >1 500 m in the Indian Ocean. He opined the pumices to have originated from an inferred submarine volcano near the central part of the Ninety East Ridge. Studies by Iyer and Sudhakar (1993, 1995) led to the postulation of a seemingly possible in situ submarine silicic volcanism as a source of pumice in the CIOB.

Explosive eruptions could be possible at water depths in excess of 3 000 m (Kalangutkar and Iyer, 2012), leading to the production of a wide range of pyroclastic deposits (Head and Wilson, 2003). Deposits of fragmented rhyolitic pumice and basaltic scoria that erupted subaqueously have been found in several places in the world ocean, at depths up to 2 000 m (Kato, 1987; Cashman and Fiske, 1991; Fiske et al., 1998). Studies by various researchers suggest potential submarine explosive eruption styles that include Plinian (sustained discharge of volatile-rich magma; Kokelaar and Busby, 1992; Kano, 2003), vulcanian (short-lived explosion resulting from failure of a solid plug in the vent), strombolian (bursting of large bubbles in low-viscosity magma), and fire fountaining (eruption of low viscosity magma in the form a spray; Mueller and White, 1992; Head and Wilson, 2003).

The paradigm of deep sea silicic volcanism has gained ground considering the recent studies of silicic volcanic glass shards of the CIOB. Detailed investigations revealed that although Toba glass shards of 74 ka are ubiquitous in the basin, however there are strombolian-type shards that seemingly have no age correlations with the Toba event. Based on ²³⁰Th_excess values and radiolarian stratigraphy, Sukumaran et al. (1998) opined that the shards were deposited in situ as a consequence of suboceanic volcanic activities during periods of elevated glaciation. Considering the morphology, occurrence and age of the shards; localized, in-situ phreatomagmatic episodes have been proposed to account for the origin of the shards (Iyer et al., 1997; Mascarenhas-Pereira et al., 2006; Amonkar et al., 2020).

Collectively, the above discoveries strongly suggest that silicic volcanic products may have recently erupted at great depths from submarine volcanoes. Interestingly, large pumice clasts that have been frequently recovered from the CIOB may indicate that they had short flotation times and derived from intraplate volcanoes (Iyer and Sudhakar, 1993) because locally derived hot pumices have a tendency to sink faster and near to the volcanic source (cf. Whitham and Sparks, 1986).

5.3 The in situ origin, transport and deposit mechanism

The shapes of the CIOB pumices are variable: 32% oblate, 31% are equant, 21% prolate and 21% bladed (Kalangutkar et al., 2011). It is believed that the clasts with other than equant size may be derived from a source nearer to the site of deposition, which could be within or near the CIOB. The eight groups of pumice that were identified based on mineralogy have the following shapes.

Group 1: either equant/bladed/prolate

Group 2: equant/prolate and oblate

Group 3: equant followed by oblate, prolate/bladed

Group 4: either equant/prolate/bladed

Group 5: oblate and prolate

Group 6: bladed

Group 7: equant

Group 8: equant

The above mineral groups when correlated with the Zings classification (Kalangutkar et al., 2011), indicates that most of the groups have clasts which are either oblate, prolate or bladed while equant is dominant in a few groups, e.g., Groups 2 and 3 (Table 2). This suggests that equant pumice clasts perhaps travelled from a distant source and this is also supported by the mineral composition studies that are comparable with those of the pumices from HDT of Toba and Krakatau (Figs 5 and 6). The non-equant clasts that are oblate or prolate may have settled rapidly near the source of formation, since hot pumice tends to sink near the source of formation (Whitham and Sparks, 1986).

The drift of the pumices into the present site from the nearby possible volcanic sources can be explained by water circulation pattern as shown in Fig. 8.

Figure  8.  A schematic representation of currents observed during the months January–February and July–August in the Indian Ocean. These currents may have helped to transport pumice clasts across the Indian Ocean from the nearby terrestrial volcanoes while the water-laden clasts may have sunk in the Central Indian Ocean Basin. The currents identified are: South Equatorial Current (SEC); South Equatorial Counter Current (SECC); Northeast and Southeast Madagascar Current (NEMC and SEMC); East Africa Coastal Current (EACC); Somali Current (SC), West India Coastal Current (WICC); Lakshadweep High (LH); East India Coastal Current (EICC); Northeast Monsoon Current (NMC); South Java Current (JC); and Leeuwin Current (LC). The Indo-Pacific Throughflow, which enters from the east, influences both the SEC and the LC. The base figure is modified from Schott and McCreary (2001). The volcanoes marked are Narcondum (a), Barren (b), Toba (c), Krakatau (d), and Tambora (e).

DownLoad: Full-Size Img PowerPoint

Silicic shards of rhyolitic composition of in situ origin (phreatomagmatic eruption) along flanks of a seamount and in the vicinity of 76°30′E FZ, assumed to be formed due to possibility of reactivation of tectonic activity in the faulted area (Mascarenhas-Pereira et al., 2006) and also in other regions of the basin (Amonkar et al., 2020). The evidence perhaps indicates sporadic intraplate silicic volcanism near seamounts and FZ in the CIOB.

6. Conclusions

(1) Based on the mineral assemblages we have identified eight groups of pumices in the CIOB. All the pumices were not derived from Krakatau volcano and nor from YTT, MTT, and OTT eruptions of the IVA, as was suggested by earlier researchers.

(2) The pumices of Group 2 (plagioclase+clinopyroxenes) and Group 3 (plagioclase+orthopyroxenes) are comparable to those from HDT of Toba and Krakatau in terms of mineralogy and are equant clasts in size indicating distant source. Hence, it can be inferred that perhaps HDT and Krakatau pumices drifted and reached the CIOB.

(3) Pumices similar to Group 1 (plagioclase), Group 4 (plagioclase+biotite), Group 5 (plagioclase+hornblende) and Group 6 (plagioclase+orthopyroxe+clinopyroxene) have not been reported from Toba, Krakatau and from Andaman-Narcondam. These are mostly the non-equant clasts which may have settled rapidly near the source of formation. Therefore, the pumices of these groups are ascribed to plausible silicic eruptions in the CIOB.

(4) Group 7 (plagioclase+hornblende+pyroxene+quartz) and Group 8 (plagioclase +clinopyroxene+quartz) are from unknown source and require a more detail study.

(5) Pigeonite, fayalite and ulvospinelare noted for the first time in the CIOB pumices.

It is concluded that the mineralogy and mineral compositions together with morphology can be used to study the source of pumices in the CIOB. The study indicates that the pumices have been derived from multiple terrestrial sources and also from intraplate volcanism. The present findings would be validated (manuscript under preparation) by taking recourse to the geochemical characteristics of the CIOB pumices and their comparison with their suggested sources of derivation.