You are on page 1of 8

918

Singh, A.P., Mishra, D.C., Vijaya Kumar, V. and Vyaghreswara Rao, M.B.S. (2003) Gravity-magnetic signatures and crustal architecture along Kuppam-Palani geotransect, south India. In: Ramakrishnan M. (Ed.), Tectonics of Southern Granulite Terrain. Geol. Soc. India Mem. No. 50, pp. 139-164. Stern, R.J. (1994) Arc Assembly and continental collision in the Neoproterozoic East African orogeny - implications for the consolidation of Gondwana. Ann. Rev. Earth Planet. Sci., v. 22, pp. 319-351. Tsunogae, T. and Santosh, M. (2003) Sapphirine and corundum bearing granulites from Karur, Madurai Block, Southern India. Gondwana Res. (this issue).

Tucker, R.D., Ashwal, L.D., Handke, M.J., Hamilton, M.A., Le Grange, M. and Rambeloson, R.A. (1999) UPb geochronology and isotope geochemistry of the Archean and Proterozoic rocks of north-central Madagascar. J. Geol, v. 107, pp. 135-153. Wilde, S.A. (1999) Evolution of the western margin of Australia during the Rodinian and Gondwanan supercontinent cycles. Gondwana Res., v. 2, pp. 481-499. Windley, B.F., Razafiniparany, A., Razakamanana, T. and Ackermand, D. (1994) Tectonic framework of the Precambrian of Madagascar and its Gondwana connections: a review and reappraisal. Geol. Rundsch., v. 83, pp. 642-659.

Gondwana Research (Gondwana Newsletter Section) V. 6, No. 4, pp. 918-925. 2003 International Association for Gondwana Research, Japan.

GNL

CORRESPONDENCE

Facies Clustering in Turbidite Successions: Case Study from Andaman Flysch Group, Andaman Islands, India
Basab Mukhopadhyay1, Partha Pratim Chakraborty2 and Soumen Paul2
1 2

Geodata and Database Division, Geological Survey of India, 27, J.L.Nehru Rd., Kolkata - 700 016, India Department of Applied Geology, Indian School of Mines, Dhanbad - 826 004, India

(Manuscript received May 25, 2003; accepted July 26, 2003)

Abstract
Bed thickness data of two turbidite sections viz., Corbyns Cove section, South Andaman and Kalipur section, North Andaman those belong to Oligocene Andaman Flysch Group, a forearc submarine fan system, were assessed for facies clustering employing Hurst statistics. Both the sections show Hurst phenomenon and reveal clustering in terms of thick and thin beds. Forcing behind event (bed) depositions in either of the studied sections was assessed statistically and inferred to be non-random and with cyclicities of irregular physical length. The inferred paleogeography through Hurst criteria though worked well for distal fan setting i.e., basin floor sheet sandstones of Corbyns Cove section, its unequivocal application in proximal fan deposits remains to be tested. The mismatch in paleogeographic interpretation between Hurst test result (lobe-interlobe) and field observation (channel-levee) for the inner fan deposit is explained through differential facies stacking between fans grow in sea-level lowstand and highstand. Lower bed amalgamation, poor sand to mud ratio and subordinately present thick event deposits may be the result of active growth of Andaman Flysch fan in sea level highstand and expressed in lower Hurst K value for inner fan channel-levee association (Kalipur section) compared to many of the channel-levee deposits of lowstand fan systems observed world over. Key words: Hurst statistics, Andaman Flysch, lobe-interlobe, channel-levee, sea level highstand.

Introduction
Shanmugam and Moiola (1988) defined submarine fans as channel and lobe (or sheet sand) complexes formed from sediment gravity flows in deep-sea environments, commonly beyond the continental shelf. Documentation of these three components of submarine fan, i.e., channel, lobe and sheet sand, though common

in modern large-scale fan systems, their distinction in rock record as proximal or distal fan products demands appreciation of depositional and preservational bias besides detailed documentation of specific facies characteristics. Traditionally, workers on ancient turbidite systems relied heavily on identification of asymmetric cycles in bed thickness, thinning- or thickening-up, for
Gondwana Research, V. 6 , No. 4, 2003

919

identification of individual fan sub-environment and/or any deterministic factor of deposition (Mutti and Ricci Lucci, 1972; Stow et al., 1996). Reliance on identification of asymmetric cycles, however, heavily contested in recent times with reports from sandstone packets of submarine fan successions, are not characterized by either asymmetric upward thickening or thinning cycles (Murray et al., 1996; Chen and Hiscott, 1999a). This raised the necessity of identifying an alternate statistical proxy that can be used to test randomness and trends in turbidite successions, in its own right for subsurface data and in association with bed-scale facies characterization for outcrop-based data, for delineation of subenvironments in submarine fan systems. Chen and Hiscott (1999a) through their study on 286 sandstone packets selected from 28 turbidite sections evaluated a number of statistical procedures viz., rank correlation test (Kendall, 1969, 1976), non-parametric test (Waldron, 1987), median crossing test (Fisz, 1963) etc., for identifying null hypothesis of randomness but could not single out any universal method with sufficient power to give a definitive conclusion. In a follow-up paper, these workers (Chen and Hiscott, 1999b) advocated Hurst statistics as potential tool for delineating irregular, long-term clustering in certain physical attributes viz., bed thickness, grain size or coarse-division thickness percentage in a bedding couplet and recommended it as a reliable criterion in combination with facies characteristics for identifying submarine fan subenvironments. The observed clustering was interpreted as a result of lateral shifting in submarine fan sub-environments without any major control of sea level fluctuation on facies organization. Though active siliciclastic submarine fan growth is a common observation in periods of lowered sea level (Vail et al., 1977), reports of fan growth are also not uncommon in sea level highstand, particularly for tectonically active basins and basins with narrow shelves (Ito, 1998; Chakraborty and Pal, 2001). Certainly, the facies development pattern and their stacking motif differ between fans developed in sea-level lowstand and highstand. This paper presents the results of Hurst statistics applied on bed thickness distribution of a forearc submarine fan i.e., Andaman Flysch fan, grown in sea level highstand (Chakraborty and Pal, 2001) and discuss the constraints in using Hurst statistics as a universal tool for sub-environment identification in submarine fan systems. The upper Eocene-Oligocene Andaman Flysch Group was studied for the present study from the sections exposed at Kalipur, North Andaman and Corbyns Cove, South Andaman (Fig. 1). Chakraborty et al. (2002), through non-parametric Waldron test, observed failure
Gondwana Research, V. 6, No. 4, 2003

of both the sections in rejecting the null hypothesis of randomness and discarded presence of any cyclic pattern in either of these sections. In this paper, we apply Hurst statistics (Hurst K and H values) to original bed thickness measurements from both the sections and 300 shuffled sequences generated from the original data by MonteCarlo simulation method.

Fig. 1. Generalised geological map of Andaman Islands (after Pal et al., 2001) showing locations of studied sections.

920

Geologic Setting
Andaman Group of islands represents the subaerial part of subduction zone complex having active subduction of the Indian plate on its west along the Java trench. Thick (3150 m; Roy, 1983) sedimentary sequence exposed in these islands (viz., Mithakhari Group, Andaman Flysch Group and Archipelago Group; Table 1) offer unique opportunity to understand many critical aspects of deep marine sedimentation, particularly important for a subduction margin. Bounded between the Mithakhari Group below and Archipelago Group above (Table 1), the 750 m thick (Roy, 1983) siliciclastic Andaman Flysch Group is exposed as detached outcrops over a strike length of 200 km. A longitudinal submarine fan model with a detached slope fan lobe in an accretionary forearc has been proposed (Ray, 1982; Chakraborty and Pal, 2001) for this succession, which shows an onlapping relationship with the underlying Mithakhari Group of trench-inner slope origin (Chakraborty et al., 1999; Pal et al., in press). Unconformably the flysch sediments are overlain by the Archipelago group of sediments, Miocene-Pliocene in age, of outer shelf to open marine origin (Roy, 1983). The two sections chosen for the present study viz., Kalipur section of North Andaman and Corbyns Cove section of South Andaman though represent parts of single submarine fan system; they differ in lithology (grain size), clast composition, facies development motif and paleocurrent pattern and are inferred to be the product of two different paleogeographic settings viz., channellevee complex of inner fan and basin floor outer fan (Chakraborty and Pal, 2001). While the basin floor outer fan deposits of Corbyns Cove section reveal a southerlydirected paleocurrent and a substantial amount of quartz (max. up to 28.5%) in the clast population, the inner fan product of Kalipur section is distinctive in its eastward paleocurrent direction and absence of quartz (Chakraborty and Pal, 2001). Both the sections, however, document laterally persistent and stratigraphically selective soft sediment deformation features (slump folds, convolutions, flames, etc.) those were identified as signatures of
Table 1. Generalised stratigraphy of the Andaman group of Islands (after Ray et al., 1982). Stratigraphic Unit (Group) Age Archipelago Andaman Flysch Mithakhari Ophiolite Unconformity Unconformity/transitional Tectonic/Unconformable Pliocene to Miocene Oligocene to Upper Eocene Middle to Lower Eocene Cretaceous

synsedimentary seismic activity and assessed for active delta shedding from growing orogenic front in course of Andaman Flysch fan growth (Chakraborty and Mukhopadhyay, 2003). The measured sections of Corbyns Cove and Kalipur area reveal 28 and 18 numbers of such seismite layers respectively (see Chakraborty and Pal (2001) for detailed section measurement data).

Theoretical Background
In a turbidite succession that fails to reject null hypothesis of randomness, non-random bed thickness trends may still exist in terms of pronounced fluctuation or clustering of thin and thick beds. The clustering may be expressed either as trends or can generate wavy pattern (discussed in a later section). Such clustering in bed thickness, if identified, can serve as proxy for the conventional tendency of sedimentologists to diagnose submarine fan environments through recognition of facies packaging and stacking motif in long vertical sections. Run about mean (RAM) and Runs up and down (RUD) statistics carried on bed thickness data collected from the two sections using Waldron (1987) test failed to identify any cyclicity in thick or thin beds (Chakraborty et al., 2002). A Poisson-like stochastic process was interpreted for Corbyns Cove section from negative exponential bed thickness frequency distribution (Fig. 5 of Chakraborty et al., 2002) while bed thickness pattern of Kalipur section (Fig. 6 of Chakraborty et al., 2002) was found to follow power law distribution where N scales with an exponent 0.917. Such differential behavior among the two studied sections those belong to the same submarine fan (i.e., Andaman Flysch fan) was claimed by Chakraborty et al. (2002) as paleogeography driven based on qualitative field-based facies characterization, subjectivity of which can not be overruled. To overcome this subjectivity Hurst statistics (the K value; as defined in a later section) was employed to quantitatively assess individual fan subenvironment through bed clustering (thick or thin) and change in clustering pattern through time as documented in sedimentary facies packaging motif. Hurst (1951,1956) proposed a statistical application (termed as Hurst statistics) while working on long-term storage of reservoirs along river Nile and deduced a relationship that states R/S ~Nh where R is the maximum range of cumulative departure from mean annual river discharge, N is the year of observations, S is the standard deviation of river discharge. Hurst approximated the coefficient h by K where K is equal to log (R/S)/log10(N/2). To calculate K for N number of observations the observations are transformed logarithmically (log10). Mean and standard deviation of the data are calculated on these log-transformed values.
Gondwana Research, V. 6 , No. 4, 2003

921

R is calculated as the maximum cumulative departure from the mean value over N sequential observations. Natural sequences with large N follow Hurst phenomenon and yield a K value greater than 0.5 (Wallis and Matalas, 1971). Hurst (1951,1956) observed that a natural process occurring in irregular groups of high and low values show high K values. Alternately, where the number of observations is small, an estimator H is calculated by dividing the log-transformed (log10) observations into n number of subseries (Wallis and Matalas, 1971). For each sub-series the value of Rn/Sn is calculated. The slope of regression line in log (Rn/Sn)/log(n) plot gives the value of H. For large values of N, Chen and Hiscott (1999a) observed K value lying between 0.5 and 0.8.

hundred randomly shuffled data sequences were generated through the programme from the original sequence by using Monte-Carlo simulation method. Further, Hurst K and H values were calculated on original turbidite sections and 300 shuffled sections generated using the programme.

Application of Hurst Statistics


Bed thickness data for 199 numbers of beds of Kalipur section and 250 noumbers of beds of Corbyns Cove section along with their sedimentological characteristics were entered into an ASCII text format. Bed thickness data of each of the sections were log10 transformed and Hurst K values were calculated (Table 2). Both the sections reveal Hurst phenomenon with K value greater than 0.5. The Hurst K value calculated for Corbyns Cove section is 0.687 and that of Kalipur section is 0.781. Figure 2 illustrates clustering of high and low values of field measurements in a cyclic pattern where the cycles are of irregular physical length. In case of Corbyns Cove data (Fig. 2a) the thin-bedded clusters are long in the lower half of the section while in the upper half there is regular clustering of thick and thin beds in a rhythmic pattern with an overall trend of thickening towards top. The Kalipur data, in contrast, show wavy organization of thinand thick-bedded clusters (Fig. 2b) in the lower part of the section with a general trend of thick-bedded cluster except towards the top part of the section where the trend reverts to thin-bedded clusters. That the thick-bedded clusters are coarse-grained and mudstone poor and thin bedded clusters are fine-grained and muddy are confirmed through field-based facies characterization of the two studied sections: dominance of conglomerates and graded sandstones at Kalipur section and siltstone-shale at Corbyns Cove section (Chakraborty and Pal, 2001). To check whether the K values from original field data approximate the expected values for randomized turbidite successions, K values were generated on 300 shuffled sections of each of the field-based sections (Table 2). The frequency distributions of K values from shuffled sections of the two studied sections follow normal distribution (Fig. 3). K values varies between 0.49 and 0.71 with mean

Methodology
Bed thickness measurements were carried out in two sections, 175 km apart, of Andaman Flysch fan: 199 beds at Kalipur section, North Andaman and 250 beds at Corbyns Cove section, South Andaman. Special care was taken for demarcation of the top and bottom of the turbidite beds in order to avoid errors in recording the thickness. Besides grain size and sedimentary structures, positive weathering relief of sandstone and siltstone of turbidity origin relative to the autochthonous interturbidite shale units helped us to demarcate and measure the event beds, which were only used for further statistical analysis. The minimum thickness recorded in the Kalipur section and Corbyns Cove section is 4 cm and 10 cm respectively. Lithologically, normally graded pebble conglomerate, parallel-sided matrix-supported conglomerate, graded sandstone and laterally continuous shale on the scale of hundreds of metres represent the Kalipur section. The Cove section, on the other hand, displays monotonous interlayering of parallel-sided siltstone and shale laterally traceable on a scale of kilometres. Measurements made in one field season were verified in two successive field trips to avoid any bias before carrying out statistical operation. The bed thickness data were processed using the Fortran-77 programme Hurst.for (developed by Chen and Hiscott, 1999b). Three
Table 2. Hurst K and H values of original and shuffled sections. Name of turbidite section K of original sequence H of original sequence Max K of shuffled section Min Mean

Std dev

Max

H of shuffled section Min Mean

Std dev

Deviation from mean of shuffled section 1.5635 2.2258

Corbyns Cove Kalipur

0.6871 0.7815

0.7395 0.8014

0.7154 0.759

0.4963 0.4896

0.599 0.60

.056 0.059

0.801 0.882

0.347 0.099

0.5607 0.553

.0884 0.010

Gondwana Research, V. 6, No. 4, 2003

922

Fig. 2. Plots of bed number against cumulative departure from mean of bed thickness data for (a) Corbyns cove section (b) Kalipur section.

0.599 and standard deviation 0.056 for 300 shuffled sections of Corbyns Cove data and between 0.489 and 0.7599 with mean 0.6 and standard deviation 0.059 for 300 shuffled sections of Kalipur data.

Discussion
In both the sections the K values obtained from the original data (Ko) exceed the mean K values obtained from the shuffled sections (Km): for Corbyns Cove section K o = 0.687> K m = 0.599 and for Kalipur section Ko=0.781> Km= 0.6. This necessitated the assessment of whether the clustering shown by the studied sections (Ko > 0.5) is outcome of bias in section preservation or reflection of depositional process specific to fan subenvironment i.e whether the studied field sections belong to random distribution or the process of deposition involved was random. A two tailed test with null hypothesis is employed that presumes turbidite

depositional events to be random, and hence product of random process. Alternately, failure of this hypothesis signifies asymmetric trend in bed thickness distribution or a systematic process operative behind the deposition. The parameter Z is calculated where Z=(Ko Km)/Sk; Ko is the Hurst value of original section, Km is the mean of Hurst K values of shuffled sections and Sk is the standard deviation of K values of shuffled sections. If value of Z lies between 1.3, it can be inferred with 90% confidence that original section is taken from a population with random distribution or process of deposition is random. Z values for Corbyns Cove and Kalipur sections are 1.56 and 2.22 respectively (Table 2), which reject null hypothesis with 90% confidence. Involvement of nonrandom process specific to depositional environment is inferred. This inference, however, is in contradiction with the observation made by Chakraborty et al. (2002) that both the sections fail to reject null hypothesis of randomness under RAM (run about mean) and RUD (runs

Fig. 3. Histogram showing the distribution of K values of 300 shuffled sections generated from original data of (a) Corbyns cove section (b) Kalipur section. Gondwana Research, V. 6 , No. 4, 2003

923

up and down) test. It is important to note that Waldrons method tests only for dominance of a section by one or the other type of asymmetric cycles it does not detect a difference from randomness in case of (i) presence of both upward-thinning and upward-thickening bed thickness cycles or (ii) grouping of thin-bedded and thick-bedded units. Comparable negative and positive transitions (119 vs. 124 nos. in Corbyns Cove section and 99 vs. 96 numbers in Kalipur section respectively) under RAM test (Chakraborty et al., 2002; their Table 2) though accepted the null hypothesis of randomness with lack of any signal for dominance of thick or thin beds, was not in a position to infer grouping of thick or thin beds. In the absence of detailed chronology for the depositional history of Andaman Flysch fan, the present study is not in a position to evaluate the observed thick and thin bedded clusters of irregular physical length in terms of their possible forcing; whether it is orderly intrabasinal or hierarchical external one. Only clue that comes from field observation is the presence of 28 and 18 numbers of soft sediment deformation layers in Corbyns Cove and Kalipur section those register syndepositional seismic activity. Though process-product correlation between the earthquake seiches and frequency of event deposition is beyond the scope of present study, the irregular thick and thin clusters in the studied sections may be related to changes in the sediment supply from the source area (resulting large and small turbidity currents) triggered by earthquake tremors. Chen and Hiscott (1999b) categorized submarine fan systems into three major environments viz., channel-levee complex, lobe-interlobe deposits and basin-floor sheet sand based on degree of bed clustering (the K value) and departure of Ko from mean K of random sequences. From a plot of 19 turbidite sections for which field-based environmental characterizations are available, these workers showed definite segregation of environments when plotted in terms of their Hurst K values against deviations from mean K values (Fig. 9 of Chen and Hiscott, 1999b). While a channel-levee complex shows high K, strong clustering and large departure from mean K of random sequences, the basin floor sheet sand reveals lowest Hurst K and a small departure from mean K of random sequences. The lobe-interlobe deposits fall in between. Plotting of K values against their departure from mean value of shuffled sections for both the sections under study (Fig. 4) revealed that corbyns Cove section falls in the field of basin floor sheet sand while Kalipur turbidite section signifies a lobe-interlobe depositional motif. While the interpreted depositional environment (through Hurst K value) for Corbyns Cove section (i.e., basin floor sheet sand) firmly corroborates with field-based paleogeographic interpretation, the lobe-interlobe interpretation for the
Gondwana Research, V. 6, No. 4, 2003

Kalipur section under the present study does not corroborate the field-based interpretation of Chakraborty and Pal (2001). For Kalipur section, Chakraborty and Pal (2001) through facies characterization and documentation of stacking motif concluded in favor of a possible channellevee association in an inner fan setting. This discrepancy in paleogeographic interpretation between field observation and statistical modeling is explained through temporal variation in lithofacies organization in submarine fan systems in response to sea level stand. Though growth of submarine fans is characterized by the highest sand-to-mud ratio, it is in general, interpreted as a product of lowstand sea level. Active turbidite deposition is also reported in submarine fan environments at transgressive and highstand sea level stages in response to variations in tectonic setting, climate and physiography of sedimentary basins (Posamentier et al., 1991; Kolla and Perlmutter, 1993; Ito, 1998). In forearc setting the growth of Eocene-Oligocene Andaman Flysch fan is also visualized through tectonically triggered active sediment supply in rising or hightened sea level stand (Chakraborty and Pal, 2001). Lithofacies organization in submarine fan successions that develop in rising or high sea level stand will certainly vary from those of lowstand origin. Ito (1998) from his study on lower Kazusa Group, a forearc submarine fan succession, documented interbedded sandstones and siltstones as highstand depositional motif that gives way to thicker sandstone beds up-section with the progradation of highstand coastal depositional system. Understandably, bed amalgamation will be poor and ratio of thicker sandstone to thinner siltstone/shale will be low even in the proximal deposits of rising and highstand fans in comparison to the lowstand fan systems. The tendency

Fig. 4. Plot of Hurst K of original sequence against deviation from mean K of shuffled sections of both the turbidite sections.

924

of grouping in thicker beds will be minimal and Hurst K value will be low. The low K value (in the lobe-interlobe field of Chen and Hiscott, 1999b; their Fig.9) obtained for field-identified channel-levee association of Andaman Flysch fan can possibly be explained through reduced bed clustering as expected in a fan developed in heightened sea level. This difference in stacking motif between lowstand and highstand fan systems is likely to get minimized away from fan source, where either of the sea level stand will result the same i.e., grouping of thin beds. This is confirmed by the Hurst K value obtained from Corbyns Cove section that falls under the sheet sand field of Chen and Hiscott (1999b) and corroborated by the field observations.

Acknowledgment
The authors would like to acknowledge the help and co-operation extended by Tapan Pal in the course of data collection. We thank D.G, GSI and Dy. D.G, GSI for permission to communicate the paper. PPC acknowledges the financial help provided by DST.

References
Chakraborty, P.P., Pal, T., Dutta Gupta, T. and Gupta, K.S. (1999) Facies pattern and depositional motif in an immature trenchslope basin, Eocene Mithakhari group, Middle Andaman, India. J. Geol. Soc. India, v. 53, pp. 271-284. Chakraborty, P .P . and Pal, T. (2001). Anatomy of a forearc submarine fan: upper Eocene-Oligocene Andaman Flysch Group, Andaman Islands, India. Gondwana Res., v. 4, pp. 477-486. Chakraborty, P.P., Mukhopadhyay, B., Pal, T. and Dutta Gupta, T. (2002) Statistical appraisal of bed thickness patterns in turbidite successions, Andaman Flysch Group, Andaman Islands, India. J. Asian Earth Sci., v. 21, pp. 189-196 Chakraborty, P.P. and Mukhopadhyay, B. (2003). Soft sediment deformation as possible clue for sedimentation rate: a case study from Oligocene Andaman Flysch Group, Andaman Islands, India. J. Geol. Soc. India, v. 61, pp. 411-418. Chen, C. and Hiscott (1999a) Statistical analysis of turbidite cycles in submarine fan successions: tests for short term persistence. J. Sed. Res., v. 69, pp. 486-504. Chen, C. and Hiscott, R.N. (1999b) Statistical analysis of facies clustering in Submarine fan turbidite successions. J. Sed. Res., v. 69, pp. 505-517. Fisz, M. (1963) Probability theory and mathematical statistics. New York, Wiley, 326p. Hurst, H.E. (1951). Long term storage capacity of reservoirs: Amer. Soc. Civil Engineers, Transactions., v. 116, pp. 770-808. Hurst, H.E. (1956). Methods of using long-term storage in reservoirs: Institute of Civil Engineers, Proc., v. 5, pp. 519-590. Ito, M. (1998) Submarine fan sequences of the Lower Kazusa Group, a Plio-Pliestocene forearc basin fill in the Boso Peninsula, Japan, v. 122, pp. 69-94. Kendall, M.G. (1969) Rank correlation methods. London, Charles Griffin and Co., pp.1-9. Kendall, M.G. (1976) Time series 2nd ed. New York Hafner press (Macmillan) pp. 21-28. Kolla, V. and Perlmutter, X. (1993). Timing of turbidite sedimentation on the Mississipi fan. Amer. Assoc. Petrol. Geol. Bull., v. 77, pp. 1129-1141. Murray, C.J., Lowe, D.R., Graham, S.A., Martinez, P.A., Zeng, J., Carroll, A.R., Cox, T., Hendrix, M., Heubeck, C., Miller, D., Moxon, I.W., Sobel, E., Wenderboung, J. and Williams, T. (1996) Statistical analysis of bed thickness pattern in a turbidite section from the Great valley sequence, Cache Creek, North California. J. Sed. Res., v. 66, pp. 900-908. Mutti, E. and Ricci Lucci. (1972) Turbidites of the Northern Appenines: introduction to facies analysis. Int. Geol. Rev., v. 20, pp. 125-166. Pal, T., Chakraborty, P.P., Dutta Gupta, T. and Singh, C.D. Geodynamic evolution of the outer arc-forearc belt in the Andaman Islands, The central part of the Burma-Java subduction complex. Geol. Mag. (in press).
Gondwana Research, V. 6 , No. 4, 2003

Conclusion
The bed thickness distribution of two turbidite successions in the Oligocene Andaman flysch group, which differ in terms of their paleogeography, exhibit Hurst phenomenon and reveal thick- and thin- bedded clusters of irregular physical length. The non-random character of facies organization in either of the sections is also reflected in Z parameter of Hurst K value of original section and mean and standard deviation of K values of shuffled sections generated through Monte-Carlo simulation. Lithologically, the dominant muddy lithology constitute the thin-bedded clusters of Corbyns Cove section while the coarser lithologies represent thick bedded clusters of Kalipur section. Extrabasinal seismic forcing has been conceived behind such non-periodic, irregular length grouping of bed thickness in turbidite successions. The Hurst K values for Corbyns Cove and Kalipur section and their deviations from mean K of shuffled sections give clue for paleogeographic interpretation of individual sections. While the interpreted subenvironment for Corbyns Cove section under Hurst test i.e., basin floor sheet sand corroborates well with the field-based observation, the method does not work well for proximal fan setting i.e., the Kalipur section. The field-based channel-levee interpretation for the Kalipur section does not match with lobe-interlobe interpretation obtained through Hurst method. Active growth of Andaman Flysch fan in sea level highstand might had resulted poor bed amalgamation and low sand:mud ratio in proximal fan deposits of Andaman Flysch fan and biased Hurst K values towards the lower end. Unequivocal application of Hurst phenomenon in determination of submarine fan sub-environment from any turbidite succession should therefore be reconsidered.

925

Posamentier, H.W., Erskin, R.D., Mitchum, Jr.R.M. (1991) Models for submarine fan deposition within a sequence stratigraphic framework. In: Weimer, P. and Link, M.H. (Eds.), Seismic facies and sedimentary processes of submarine fans and turbidite systems. Springer-Verlag, New York. pp. 127-136 Ray, K.K. (1982) An overview of the geology of Andaman and Nicobar islands. Geol. Surv. India, Misc. Publ. v. 41, pp. 110-125 Roy, T.K., (1983) Geology and hydrocarbon prospects of Andaman-Nicobar. Petrol. Asia J., KDMIPE, ONGC, Dehradun, pp. 37-53. Shanmugam, G. and Moiola, R.J. (1988) Submarine fans: characteristics, models, classification and reservoir potential. Earth Sci. Rev., v. 24, pp. 383-428.

Stow, D.A.V., Reading, H.G. and Collinson, J.D. (1996) Deep seas. In: Reading, H.G. (Ed) Sedimentary environments: processes, facies and stratigraphy, 3rd ed., Oxford, U.K, Blackwell Sci., pp. 395-453. Vail, P.R., Mitchem, R.M. and Thompson, S. (1997) Seismic stratigraphy and global changes of sea level, part 3: relative changes of sea level from coastal onlap. In Payton, C.E. (Ed.), Seismic stratigraphy: application to hydrocarbon exploration. Amer. Assoc. Geol. Mem., v. 26, pp. 63-81. Waldron, J.W.F. (1987) A statistical test for significance of thinning- and thickening upward cycles in turbidites. Sed. Geol., v. 54, pp. 137-146. Wallis, J. R. and Matalas, N. C. (1971). Correlogram analysis revisited. Water Resource Res., v. 7, pp. 1448-1459.

Gondwana Research (Gondwana Newsletter Section) V. 6, No. 4, pp. 925-930. 2003 International Association for Gondwana Research, Japan.

GNL

CORRESPONDENCE

Sapphirine and Corundum-bearing Granulites from Karur, Madurai Block, Southern India
Toshiaki Tsunogae1 and M. Santosh2
1 2

Institute of Geoscience, University of Tsukuba, Ibaraki 305-8571, Japan Department of Natural Environmental Science, Kochi University, Kochi 780-8520, Japan

(Manuscript received May 22, 2003; accepted July 23, 2003)

Abstract
We report new occurrences of sapphirine- and corundum-bearing granulites intercalated within orthogneisses at Lachmanapatti and Malappatty in the northern part of Madurai Block. Sapphirine in these localities occurs either as needle-like intergrowth with cordierite and corundum in symplectites and medium- to fine-grained euhderal to subhedral crystals associated with cordierite and corundum (Lachmanapatti) or in association with plagioclase, corundum and gedrite (Malappatty). The sapphirine from Lachmanapatti is highly magnesian (XMg = 0.870.94) with higher Cr content (up to 0.9 pfu) as compared with those in other localities in the Madurai Block. The sapphirine-corundum association reported in this study has important bearing on the ultrahigh-temperature metamorphism and exhumation history of the northern Madurai Block as well as on the tectonic evolution of the continental deep crust in southern India. Key words: Sapphirine, corundum, ultrahigh-temperature metamorphism, Madurai Block, southern India.

Introduction
Sapphirine- and corundum-bearing aluminous rocks are of special interest to petrologists, because they sometimes carry the signatures of crustal metamorphism at extreme thermal conditions, and also preserve reaction textures suitable for construction of metamorphic P-T paths. Therefore, numerous petrological studies have been carried out on sapphirine-bearing metamorphic rocks leading to a number of publications in the last two
Gondwana Research, V. 6, No. 4, 2003

decades. Rarely, sapphirine and corundum coexisting with quartz have also been reported from some pelitic and quartzo-feldspathic rocks (e.g., Dallwitz, 1968; Motoyoshi et al., 1990) which provide robust evidence for ultrahightemperature (UHT) crustal metamorphism with T>950C (Harley, 1998). The Precambrian terrane of southern India, which is a classic example of exhumed middle and lower continental crust, is a mosaic of several crustal blocks dominantly composed of high-grade metamorphic rocks (Drury et al.,

You might also like