Contemporary sedimentation in the Cocos (Keeling) Islands, Indian Ocean: interpretation using settling velocity analysis

Se.a ataty ELSEVIER

Sedimentary Geology 114 (1997) 109-130

,

Contemporary sedimentation in the Cocos (Keeling) Islands, Indian Ocean: interpretation using settling velocity analysis Paul S. Kench * Department of Geography and Oceanography, Australian Defence ForceAcadem), Canberra, ACT 2600. Australia

Received 2 October 1996; accepted 18 July 1997

Abstract Textural parameters of grain-size distributions derived from sieve analysis and component analysis have been used to the exclusion of other methods to examine sedimentary processes in modern reef environments. However, because these methods do not reflect the hydraulic behaviour of bioclastic sediment this study uses settling velocity analysis (accounting for grain size, density, and shape variability), and component analysis of separated settling velocity fractions, to describe and differentiate bioclastic deposits and infer transport pathways in the Cocos (Keeling) Islands, an Indian Ocean atoll. Principal Component and Cluster analysis of settling (251 samples), and component (90 samples) characteristics, discriminated eight settling velocity and five component classes. A derived settling velocity classification, reflecting the hydraulic properties of the sediments, describes a lagoonward gradient from those with fast settling characteristics on the outer reef fiat to those with moderate settling character on sand aprons. The active sediment transport system is constrained between these zones. Component analysis of bulk samples identifies two broad sediment types: (1) those with reef-derived components, found throughout the active transport zone (reef flat to sand aprons); and (2) those characterised by greater proportions of gastropod and Halimeda fragments which are produced and deposited in situ in the lagoon. Component analysis of settling fractions further resolves details of sediment movement in the active transport zone. Sediment is transported from the productive reef fiat to passage entrances which form initial depositional environments. Slower settling material is selectively transported from the entrances, through the passage conduits, to passage exit environments. These form the main depositional areas under mean energy conditions. During moderate- to high-energy events sediment is flushed from the passage to sand apron zones, with slower settling grains deposited in the deep lagoon sinks. Results show that settling velocity analysis and component analysis of settling fractions can resolve the nature of sediment transport patterns in reef environments at a much finer scale than studies in which analysis is based on grain-size distributions. Kevwords: bioclastic sediments; settling velocity analysis; component analysis; carbonate sedimentation

1. Introduction There is a paucity of research examining bioclastic sediment transport processes. Orme (1977) * Present address. Department of Geography and Environmental Studies, The University of Melbourne, Parkville, VIC, 3025 Australia. Tel.: +61 (3) 344-6581; Fax: +61 (3) 344-4972

attributed this in part to the reliance of researchers on sieve-derived textural analysis of sediment. Almost exclusively, sieve-derived textural and component analyses have been used to describe and discriminate spatial patterns o f sedimentary facies in modern reef environments (Ginsburg, 1956; Purdy, 1963; Gabrie and Montaggioni, 1982a,b; Scoffin and Tudhope, 1985; Chevillon, 1990; Wardlaw et al.,

0037-0738/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. PH S 0 0 3 7 - 0 7 3 8 ( 9 7 ) 0 0 0 8 2 - 1

110

P.S. Kench / Sedimentary Geology 114 (1997) 109-130

1992; Smithers et al., 1994). Textural parameters of size-distributions have also been used to determine depositional environments and transport pathways at very broad scales (Flood and Scoffin, 1978; Frith, 1983; Chevillon and Clavier, 1988; Colby and Boardman, 1989; Sagga, 1992). However, Maiklem (1968), Braithwaite (1973), Orme (1977) and Scoffin (1987) noted that these textural parameters may not accurately reflect the hydraulic behaviour of bioclastic deposits as the sieve procedure does not account for the heterogeneous size, shape, density, and component mix of bioclastic deposits. Kench and McLean (1996) also show that the size variation and component mix of grains in settled fractions is markedly different from that of sieve-derived fractions. These characteristics of settled fractions influence the proportion of sample contained within settled fractions and produces markedly different sample population distributions. For these reasons, the use of sieve-derived textural results alone may lead to distorted interpretations of the environment of deposition and transport processes in reef environments (Kench and McLean, 1997). Numerous siliciclastic and pyroclastic sediment studies have used settling velocity distributions (and their descriptive statistics) to discriminate sediment deposits and differentiate depositional environments (Reed et al., 1975; Taira and Scholle, 1979; Oehmig and Wallrabe-Adams, 1993). Bryant (1984) and Lund-Hansen and Oehmig (1992) consider that this method provides a better index of grain behaviour in fluid than sieve-derived grain-size estimates, and is a more useful tool for interpreting sediment transport processes. Braithwaite (1973) also suggested that settling velocity analysis of bulk samples and, in particular, component analysis of separated settling fractions (enabling the transport history of grains to be determined) could greatly aid in the interpretation of bioclastic sedimentary processes. Despite its common use in siliciclastic studies, and potential utility in interpreting carbonate sedimentary processes, the settling technique has not been adopted widely for examination of bioclastic deposits. Braithwaite (1973) suggested this is due to the relative difficulty in separating settled deposits into fractions to examine components. In recent studies Oehmig and Wallrabe-Adams (1993) and Michels (1995) using siliciclastic sedi-

ments, and Kench (1994a) and Kench and McLean (1996) using bioclastic sediments, have introduced techniques whereby deposits can be separated into discrete fractions possessing equivalent settling characteristics. Kench (1994a) and Kench and McLean (1996) also show that settling velocity fractions of bioclastic deposits (analogous to sieve fractions but containing grains with varied shape, size, density) possess a narrow range of hydraulic threshold properties. This settling approach provides an alternative to sieve-based textural analysis for describing carbonate deposits and interpreting sedimentary processes. The information gained is essential for understanding the development and dynamic change experienced by surface morphological features of reef systems. In this study the settling velocity approach is used to analyse bulk bioclastic deposits and interpret a depositional pathway in a reef setting. Surficial samples from the Cocos (Keeling) Islands are examined by this means and component analysis of settling velocity fractions is used to describe and differentiate deposits, to determine the spatial pattern of sediments, and to infer sediment transport and deposition.

1.1. Field location The Cocos (Keeling) Islands are located in the eastem Indian Ocean (Fig. 1). The main South Keeling atoll (96°48-56'E, 12°4-13'S), the focus of this study (hereafter referred to as Cocos, Fig. 1), is comprised of 26 islands situated on a near-continuous annular reef flat (Fig. 1). Twelve shallow passages, with depths up to 1.0-1.5 m below mean sea level (MSL), connect the ocean-side reef flat and lagoon to the south and east of the atoll. Intertidal sand aprons extend laterally between 500 and 1500 m lagoonward of the passages. The lagoon is 102 km 2 in area and can be divided into two broad zones: a shallow southern region (depths up to 3 m below MSL), and a deeper northern section (10-20 m below MSL) separated by a broad area characterised by a network of deep (20 m) blue holes. To the northeast and northwest the lagoon opens to the ocean via two deep (15-20 m below MSL) passages (2-5 km wide, Fig. 1). The atoll can be divided into ten distinct environments based on broad geomor-

P.S. Kench / Sedimentary Geology 114 (1997) 109-130 I][ll

t

l 11 ............

! . . . . . . . . . . . . . . . . . .

96"50"E Horsburgh Island []

Direction I

- - 12"05"S

0 I

COCOS (KEELING)

2km l

Home Island

".~.;.:-;-]: " i';';'~ .

.

.

--12"1

Islands

~Sand ~

~

aprons/flats

lnterisland channels Reef flats

Deep Lagoon ~ Algal covered Acroporarubble , , / s s ,, ~ Massivecorals interspersed with sandy patches ~ Sandylagoonfloor 96"50" I

,

Island

9

12 Shallow Lagoon

~

Inter-tidalsand and mud flats

~

Variablecoral and algal flat

i

Seagrassmeadows Blue hole mosaic

ie

l

Islands !!l

ss'ss" I

Fig. 1. Location of the Cocos (Keeling) Islands and division into physiographic zones, based on the surveys of Williams (1988), Berry (1989) and aerial and satellite photograph interpretation. Modified from Smithers (1994) and supplemented by lield observation. Numerals identify the shallow passages.

P.S. Kench/ Sedimentary Geology 114 (1997) 109-130

112

Table 1 Major biological and physical characteristics of the physiographic zones of the Cocos (Keeling) Islands (Fig. 1) Environment

General surface character

Dominant corals

Marine flora

Benthic fauna

Ocean-side reef fiat

>50% coral-algal substrate; Massive: Porites sp.; patches of rubble and sand. branching: Montipora sp.; 50% coral dead.

Sparse seagrass (Thalassia sp.) and Halimeda sp.

Various molluscs and foraminifera: Amphistegina sp. and Homotrema sp.; sparse holothurians.

Shallow passages

As above with increased abundance of sand and rubble.

As above with Acropora sp.

As above.

Sparse molluscs and bioturbating organisms, no foraminifera in growth position.

Sand aprons

>90% sand and rubble.

Sparse massive Porites sp.

Sparse seagrass, Halimeda sp., brown algae.

Medium-density holothurians and bioturbating organisms (1 m2).

Shallow lagoon: -seagrass

Near-shore areas, fine sand.

Typically absent.

Thalassia sp. and Syringodium sp.; brown algae.

Numerous molluscs, the foraminifer Marginopora sp.; few bioturbators.

-coral and coralline algal flat

Variable cover of coral, sand Massive Porites sp. and algae.

Sparse seagrass Thalassia sp., Caulerpa sp. and Halimeda sp.

Sparse holothurians; medium-density bioturbators; molluscs.

-intertidal mud flats

Mud.

Absent.

Algal mats binding sediment.

Large populations of burrowing crabs.

Blue holes

Muds and corals.

Acropora sp., Echinopora sp., Porites sp.

Algae: Caulerpa sp., Unknown. Halimeda sp., Turbinaria sp.

Deep lagoon: -algal-covered Acropora

Aeroporasp. rubble, covered by red algae,

Dead Acropora sp. and sparse Porites sp.

Red encrusting algae and Padina sp. common.

Barren.

-massive coral

Massive corals with sand patches.

Dead massive corals, sparse Porites sp.

Padina sp. algae and sparse Halimeda sp.

Soft corals, otherwise barren.

-sandy floor

Oscillatory ripple surface.

Occasional Porites sp. and Aeropora sp.

Sparse seagrass.

Sparse holothurians.

Source: Williams (1988), Berry (1989), Smithers (1994), Kench (1994a) and field observation.

phic, sedimentological, and biological characteristics (Fig. 1, Table 1). Cocos is dominated by southeasterly trade winds (mean speed of 20 knots) between March and October which become more variable in the doldrum period (November to February) and are punctuated by infrequent cyclone activity (Falkland, 1994). Cocos experiences a semidiurnal microtidal regime. The mean neap and spring tidal range is 0.5 m and

0.7 m, respectively, and the maximum range is 1.2 m (National Tide Tables). The circulation of the lagoon is tidally driven (Kench, 1994a,b). The tidal wave enters the northeastern passage and penetrates south into the lagoon with the rising tide. This inflow is supplemented by wind- and wave-driven (waves of modal height of 2 m break on the reef) unidirectional ocean-side reef to lagoon flow through the twelve shallow passages

PS. Kench/Sedimenta O' Geology 114 (1997) 109-130

throughout the tidal cycle. The lagoon drains through the northwest passage on the falling tide.

1.2. Previous research The surficial sediments of the Cocos lagoon have previously been described by Smithers (1994), and Smithers et al. (1994), who related the sieve-derived textural characteristics (69 samples) and components (46 samples) of sediments to some of the broad physiographic environments and subsurface stratigraphy. Using sieve-derived textural parameters Smithers (1994) inferred a broad lagoonward transport of sediment, noting that coral dominates the component mix of samples. The present study extends the spatial coverage and number of sediment samples analysed from Cocos, particularly in the intertidal-beaches, reef flats, passages, and sand aprons (zones of greatest hydrodynamic energy), in order to identify sediment dispersal patterns at a finer resolution. 2. Materials and methods

The locations of 252 surficial sediment samples from Cocos (including 15 retrieved by Smithers, 1994) are shown in Fig. 2. Samples were obtained from all physiographic zones of the atoll (Fig. 1). Wet sieve analysis was performed on 60 samples (Fig. 2) following the standard techniques of Folk (1974). Subsequently, the settling velocity of 15 g sub-samples from 252 bulk samples was analysed using a McArthur Rapid Sediment Analyser (RSA). The sedimentation column had a vertical fall of 1.875 m, from release mechanism to balance pan, and a diameter of 0.25 m. The RSA records the settling behaviour of a sample over time, and results are corrected to a water temperature of 20°C. The settling behaviour of a sample is partly described by the chi (X) parameter (May, 1981) in which X = -logz(s/so), where s = settling speed (m s -l) and So = standard settling speed of 1 m s t. This parameter is analogous to the phi scale of particle size with chi values increasing from zero with decreasing settling speed (Table 2). Following the procedure developed by Kench (1994a) and Kench and McLean (1996), 50 g subsamples of 90 deposits are settled through a 6.5 m

t13

settling tube into narrow cores. Cores are manually divided using visual textural, component and particle form changes along the core. Separated fractions (minimum of 7 for each sample containing grains with similar settling properties) are referred to as 'settling fractions' and are then settled through the RSA to identify their mean settling velocity. Component analysis was performed on each settling fraction. Following Weber and Woodhead (1972), Colby and Boardman (1989) and Smithers (1994) the Glagolev-Chayes method (Galehouse, 1971) was employed to determine the component mix of samples. Using a binocular microscope 100 grain counts of loose particles distributed on a microscope slide were made for each settling fraction within the notional sand-size grade giving a total of 700 counts per settled sub-sample. Fourteen component classes were identified including fragments of: corals, coralline algae, alcyonarian spicules, crustaceans, echinoderms, Halimeda sp., Homotrema sp., Marginopora sp., Amphistegina sp., other foraminifera, gastropods, bivalves, unknown molluscs, and indeterminate grains. The percentage weight of components of individual fractions were summed and expressed as a percentage of the total sub-sample. Essentially, component analysis of settling fractions follows that of sieve fractions (Smithers, 1994).

2.1. Statistical treatment: differentiation of deposits To compare and determine the distribution of surficial sediment types throughout Cocos, settling velocity and component data were analysed separately using S-mode Principal Component Analysis (PCA, using Statistical Analysis Systems software). The data matrix for analysis of settling velocity distributions comprised fourteen variables (weight percentage of 0.5X settling intervals between 2.0 and 8.0X, and standard deviation of distributions) for 252 samples. The 14 skeletal component types and 90 samples formed the component matrix. Scree plots (of principal component number vs. cumulative percent of variance explained in the data set) were constructed to establish the number of Principal Components required to describe maximum variance within data sets. Factor patterns (for

P.S. Kench/Sedimentary Geology 114 (1997) 109-130

114

I

I

96" 50"E

96" 55"

INDIAN

Horsburgh dand

A

OCEAN

- - 12"05"S

12"05"--

..~Direction

-i

[]

2km I

Island

::/Reef edge

S,. S~

I

So S•

So .'0 \ O

\x •

•

oS

/

"-

S•

--12"10"

D

O

[]

West Island,,

".,,

• o

..

.

.

.

.

.

.

.

.

.

.

.

South Island

• • []

Sedimentation analysis only (RSA) Sedimentation and compositional analysis Sedimentation, compositional and wet sieve analysis Sedimentation and wet sieve analysis Compositional and wet sieve analysis Wet sieve analysis only

S C

Samples collected by Smithers (1994) Samples for which compositional analysis of fractions is presented 96" 50"

96" 55"

I

I

P.S. Kench/Sedimentao' Geology 114 (1997) 109-130

I%1

•

I

250 m

115

I

•

mD

n".

500 m

t

I)

•

r

C~ Co

• •

t I

•

A °

•°

•

o•

• •

C

N

250 m

•

0

r

•

Sedimentation analysis only (RSA) o Sedimentation and compositional analysis Sedimentation, compositional and wet sieve analysis • Sedimentation and wet sieve analysis • Compositional and wet sieve analysis a Wet sieve analysis only •

•

@

'.

s Samples collected by Smithers (1994) c Samples for which compositional analysis of fractions is presented

Fig. 2. Locations of the surficial sediment samples retrieved from Cocos: (A) presents sample locations from the lagoon and ocean beaches and reef flats; insets are shown in (B), (C) and (D) for the eastern and southern passages.

identified Principal Components) were then analysed using the Ward (1963) hierarchical clustering technique (using Statistical Package for the Social Sciences software) to divide samples into clusters based on the Euclidean distance between samples• Cluster analysis provides an objective method of grouping samples based on the principal components.

To aid in describing and interpreting settling velocity clusters, a settling classification scheme has been developed to provide a physical scale of sediment settling behaviour (Table 2). This classification provides a new vocabulary wherein samples are described by their hydraulic settling velocity characteristics (e.g. fast/slow settling), which notionally

P.S. Kench/Sedimentary Geology 114 (1997) 109-130

116

Table 2 Settling velocity classification. Note the division of the settling scale (chi) according to velocity. The equivalent velocity in cm s - l , and the equivalent sedimentation diameter of aragonite spheres are also presented

through shallow passages 2 and 9 to the east and south of the atoll (Fig. 1). The component totals of fractions were grouped into 0.5X settling velocity classes for easy comparison between samples.

Settling velocity (X)

Settling velocity (cm S- l )

3. Results

1 2

50.00 25.00

3 4 5 6 7 8

12.50 6.25 3.12 1.56 0.78 0.39

Settling classification

Very fast Fast Fast-moderate Moderate Moderate-slow Slow Very slow

Comparable size of aragonite spheres (~) -7.22 -5.24 --3.35 -1.72 -0.50 0.38 1.06 1.65

equates to grains within the pebble to fine sand-size range, respectively (Table 2).

3.1. Mud- and clay-sized particle distribution The mud and clay content of deposits throughout Cocos are presented in Fig. 3. The areas with the highest percentage of fine sediments are the southeastern lagoon (>40%, in the lee of South Island) and sheltered embayments on the western side of the atoll (10-40%, Fig. 3). Deposits elsewhere in the atoll have little mud and clay content, with less than 2% on reef fiats and sand aprons, and a maximum of 4% in shallow and deep lagoon (Fig. 3).

3.2. Sediment settling velocity distributions 2.2. Interpreting transport pathways and energy processes Gradients of settling velocity values are used to infer energy gradients and transport pathways, as the settling parameter reflects the hydraulic settling and threshold properties of the sediment (Kench, 1994a; Kench and McLean, 1996). Variations in components and their abundance within and between samples, and between equivalent settling fractions of different samples (as suggested by Bralthwaite, 1973), allow components of known source (e.g. the foraminifera Amphistegina and Marginopora, and alcyonarian spicules produced on the outer reef flat) to be used as a tracer to determine dispersal directions. Coupled with settling velocity data and the physical appearance of grains, components most easily entrained, transported, and deposited can be identified along with their transport history. For example, unbroken components in faster settling fractions, imply local production and in situ deposition. These grains can be discounted from further analysis of the sediment transport system. Abraded grains imply a transport history and are used to infer transport pathways. This approach begins to separate the relative roles of biological production and transport processes in deposit formation. Samples for this analysis (Fig. 2) are from transects

Eight Principal Components were determined from the settling velocity data matrix. These account for 99.54% of the total variance within the data set. Eight clusters are defined by Cluster Analysis (Table 3). The small standard deviations of sample mean settling velocity values within each cluster, around the representative mean settling velocity (RSVmean, for that cluster, Table 3) show that Principal Component and Cluster Analysis groups samples with narrow settling velocity ranges. Representative settling velocity distributions of clusters are near-symmetrical (Fig. 4), and apart from small secondary modes in those for clusters 2 and 3, there is little bimodality in distributions. Thus, the mean settling velocity is the primary discriminating characteristic of the clusters, describing their hydraulic behaviour. Sediments represented by clusters occupy discrete areas within Cocos (Fig. 5). Fast settling clusters 1 and 5 are found close together on the eastern and southern reef fiats and in passages. Sediments with fast-moderate settling properties (clusters 1 and 4) are found on the ocean-side beaches of South and the eastern Islands and on the lagoon beaches of northern West Island. In contrast, the three moderate settling clusters of sediments are found in disparate locations on ocean-side beaches of West and South Island (cluster 3), on the southwest and eastern lagoon

PS. Kench ~Sedimentary Geology 114 (1997) 109-130

I

I

96"50"E

96°55 •

'-Iorsburgh sland -

-

117

INDIAN

OCEAN

12"05"S

12"05"--

Island

--12"1(:

W

Id Percent mud in deposits >40%

~

20- 39.9% 10-19.9%

I

5-9.9% O- 4.9%

o

2=,

I

Fig. 3. The pattern of silt- and clay-size particle content percentage of surficial samples from the Cocos (Keeling) Islands•

118

P.S. Kench /Sedimentary Geology 114 (1997) 109-130

Table 3 Settling velocity classification of samples into eight clusters, based on Principal Component (PC) and Cluster analyses Sediment settling clusters

Samples

Cluster 1. Fast-moderate settling; 42 samples; Std. Dev. = 0.19 X

1, 19,38,40, 47,49,53,58, 86, 87, 93,101, 105,106, 107, 112, 125, 126, 140, 142, 152, 155, 156, 161,162, 169,180,181,183,185,187, 191,198, 201,203, 204, 206,212, 232, 237, 250

Cluster 2. Fast settling, 28 samples; Std. Dev. = 0.59X

2,7,8,9,10,11, 20, 22, 23, 25,30,37,42,43, 46,50, 51,52,54, 61, 62, 68,82, 89,90,94,135, 153

Cluster 3. Moderate settling; 37 samples; Std. Dev. = 0.27 X

3,15,16, 17, 27, 28,34,56,59,69,74,81,91,92,98, 99, 110,115,128,130,131, 132,134, 137, 148, 165, 174, 186, 192, 193, 194, 195, 196, 197, 205, 209, 244

Cluster 4. Fast-moderate settling; 27 samples; Std. Dev. = 0.31X

4,5,55,95, 97, 100, 102, 113, 122,136, 138, 139,147, 157, 158, 172, 173, 199, 213, 215, 216, 226, 238, 240, 246, 247, 249

Cluster 5. Fast settling, 10 samples; Std. Dev. = 0.18X

6, 29,45, 48,120,121, 127,146,189, 230

Cluster 6. Moderate settling; 67 samples; Std. Dev. = 0.2X

12, 21, 24, 26,36, 41,44, 60, 63,64, 65,88,96, 109, 111, 114, 116,117, 118,119, 123,124, 129,133,141, 143, 144, 145, 149, 150, 151, 154, 159, 160, 163, 164, 166,167, 168, 170,171, 175,176, 177, 178, 179,184, 188, 200, 202, 207, 210, 211,214, 217, 218, 221,222, 223, 224, 225, 229, 231,233, 235,243, 248

Cluster 7. Moderate settling; 34 samples; Std. Dev. = 0.15X

13, 14,18,31,32,35,57, 66, 67,70,71, 72,73, 75,76, 77, 78,79, 80, 83,84, 85, 103, 104,108, 182,190, 208, 219, 220, 227,228, 239, 242

Cluster 8. Moderately slow settling; 6 samples; Std. Dev. = 0.37X

33, 234, 236, 241,245, 251

Standard deviation values refer to the deviation of the mean settling velocity of each sample around the representative mean.

floor (cluster 6) with cluster 7 found in a variety of environments (Fig. 5). Individual samples from clusters 1 to 7 are also found in the deep lagoon. Sediments in cluster 8, with the slowest settling characteristics (6-8X,) are found in more sheltered areas of the lagoon (Fig. 5A).

3.3. Component analysis of bulk samples Five clusters were determined from the data matrix of fourteen variables and ninety observations (samples) which account for 99.05% of the sample variance. Samples contained within the five clusters are shown in Table 4 and the representative component mix of clusters is presented in Table 5. Coral detritus dominates the component mix of Cocos deposits (Table 5) but discrimination of the five clusters is based on secondary constituents. Cluster 1 is differentiated by minor additions of

Alcyonarian spicules, Amphistegina and Halimeda. Clusters 2 and 3 are discriminated by the increased presence of Halimeda (20.63%) and coralline algae (33.4%), respectively. Cluster-4 sediments have the highest percentage of coral of all samples analysed (64.3%) with additions of Halimeda and mollusc fragments (Table 5). Cluster 5 also has significant additions of Halimeda and mollusca and is considered genetically similar to cluster 4. However, it is differentiated by the high percentage of unidentified grains (39.5%). These grains had the slowest settling velocity and equate to a size range of > 3.0~b. The component clusters also occupy discrete locations in Cocos (Fig. 6). Cluster-1 sediments are located on the reef fiat, in passages and on sand aprons and there is a concentration of cluster-2 sediments in the southern passages. Cluster-4 sediments dominate the north and central sections of the lagoon and join the cluster-5 sediments in the southeastern

PS. Kench / Sedimentary Geology 114 (1997) 109-130

A) 40

B)

Cluster 1: Fast-moderate settling

M = 2.45 D = 0.97 S = 1.46

4O

3O

30

o t-"

20

~- 20

O"

"

Cluster 2: Fast settling

50

M = 3.88 O = 0.82

A

119

10

10 I

2

3 4 5 6 7 Settling velocity (chi)

8

Cluster 3: Moderate settling

C) 40

D)

30

i

.

!

,

8

9

Cluster 4: Fast-moderate settling

M = 4.35 D = 0.87

A

3O

I

3 4 5 6 7 Settling velocity (chi)

9

f~

M = 3.63 D = 1.06

20

2o -I cr

LL

~. tO

1

2

3 4 5 6 7 Settling velocity (chi)

8

2

9

.1,81.

Cluster 5: Fast settling

E) 40

~: 20 -1 O"

"

g 10

9

M = 4.22 D = 0.99

f'~ 2o

lO

LL.

2

G) 40

8

Cluster 6: Moderate settling

F) 30

M = 2.97 D = 0.88

30

3 4 5 6 7 Settling velocity (chi)

3 4 5 6 7 Settling velocity (chi)

8

Cluster 7: Moderate settling

A

30

1

9

H) 30

2

3 4 5 6 7 8 Settling velocity (chi)

Cluster 8: Moderate-slow settling

M = 4.86 D 0.82

/x,

g-

M=5.7 D = 0.98 =-

20

.

20 -1 O"

u_ 10

t.L

2

3 4 5 6 7 Settling velocity (chi)

8

9

2

3 4 5 6 7 Settling velocity (chi)

8

9

Fig. 4. Representative settling velocity frequency distributions for clusters derived from the mean value of each 0.5X settling fraction of all samples grouped within a cluster (see Table 3). M = mean, D = dispersion, S = skewness.

P.S. Kench /Sedimentary Geology 114 (1997) 109-130

120

I

I

96"50"E

96"55"

INDIAN OCEAN

Horsburgh sland -12"05"S

A

'*~%.%

12"05"--

"". . . . .

. ~

Direction Island

"'" ...::::iiiiiib:.::i!iii::::.. ~/77~':::::::::::::::::::::::::::::::::::::::: .....

,,=

: : i

= =,

.,~',

.,mull[ Jt¢ir?"" . . . . . . . . .

. :.--Reef edge Home Island

~iiiii!iiiii!il;iiiiiiiililiiii!i!i~

, .....*.*.,.

j

".-:'N':':':':':'. "'~';';';-7

".'.'.'.'.'.'.'.'.'.'.'.'.'.-.'.'.'.'."-'-""""'"""""-""""

"'¥::~ ..... •

--12"10"

)

":i:i:i:i:i:i:i:i:i:i:i:i:i::' ....

JII1IN[1II1111

i

- -

".:.:.:.:.:.:.:.:.'."

~ .......

o,k,o

,

x\

-.%-------.-.-.-.-.-.-.-.--.-.---

West

,\

"

t /

:)'__

liiiiiiiiiiiiii Island

Fast settling A Fast settling B Fast-moderate settling A Fast-moderate settling B

Moderate settling A Moderate settling B Moderate settling C Moderate-slow settling o t

2km I

96"50"

96"55"

I

I

Fig. 5. (A-D) The geographical distribution of settling velocity classes of surficial sediment samples. Diagrams (B), (C), and (D) display the reef fiat and shallow passage patterns. Individual sample locations are shown in Fig. 2.

PS. Kench/Sedimentary Geology 114 (1997) 109-130

121

!:i:i:i:i:i:i:!:i:!:!:!:i;!.!:i:i:i:!:!:!:!:!:

!iiiiiiiiiiiiiiiiili!iiiiiiiiiiii

ii!?iiiiiiiiii?iiii"" , ~ :3!!iiii:iSi:::::""

!:i:!:!:!:!:!:!:!:!:i:!:i:!:!:!:!:!:;:!:!:{:!:;:!

!!ii!ili!i!i!!ii!!!i!i!ii i !i !ii!i!i ":::si:!iiiiiiiiiiiiiiii i!!iii!i!i "::::::i:i:!

Fast setting, mean chi 2.45 (-4.0e) Fast setting, mean chi 2.97 (-3.5~) Fast-moderate setting, mean chi 3.63 (-2.5e) Fast-moderate setting, mean chi 3.88 (-1.80e) Moderate setting, mean chi 4.22 (-1.4e) Moderate setting, mean chi 4.35 (-1.0~) Moderate setting, mean chi 4.86 (-O.5e) Fig. 5 (continued).

section of the lagoon (Fig. 6) which corresponds to the zone of slowest settling sediments.

3.4. Constituent analysis of settling fractions A subsidiary study examined the component mix of settling fractions and the physical character of

grains within fractions. Changes in frequency percent of major constituents and prominent secondary constituents between fractions (within and between samples) along reef fiat to sand apron transects on the eastern and southern sides of the atoll are presented in Figs. 7 and 8 and summarised below. For all but crustacean components the greatest

122

P.S. Kench/ Sedimentary Geology 114 (1997) 109-130

Table 4 Sediment sample componentclassificationinto five clusters based on PC and Cluster analyses Clusters - - Composition analysis

Samples

Cluster 1 - - Coral-dominated(39 samples)

1, 2,3, 4,5,9, 12, 15, 16,17, 20, 22,55, 67, 70,73, 78, 96,97, 109, 120, 136, 176, 177, 180, 185, 189, 190, 192, 193, 194, 195, 198, 203, 210, 227, 243, 246, 260

Cluster 2 - - Coral-Halimeda-rich (18 samples)

13, 14, 28, 31, 32, 33, 95, 101,106, 138, 143, 149, 159, 162, 207, 215,254, 259

Cluster 3 - - Coral-coralline algae (3 samples)

11,94, 221

Cluster 4 - - Coral-molluscan/Halimeda (25 samples)

18, 36, 171,172, 174, 179, 209, 214, 222, 223, 224, 230, 235, 237, 238, 240, 242, 244, 248,249, 250, 255, 256, 257, 261

Cluster 5 - - Coral-molluscan/Halimeda (muds) (5 samples)

234, 252, 253, 254, 258

proportion of fast settling material is in reef flat to passage entrance samples. In particular, the fastest settling coral and coralline algae constituents (1.512.0X) only occur in these locations (Figs. 7 and 8). Between mid-passage and passage exit zones there is a shift from predominantly fast and moderate settling fractions (2.5-3.5X range) to those with moderate to slower settling properties (3.51-5.5X range). Passage exit deposits are dominated by components contained within the slower settling 4.015.0X range. In particular, alcyonarian spicules and Halimeda fragments (east) and unbroken and fragmented Halimeda grains (south) dominate these areas (Figs. 7 and 8). On the southern transect there is a large increase in abundance of crustacean grains (broken and unbroken) in the sand apron (Fig. 8). Halimeda in the 3.01-3.5X settling range occurs along both transects and a large number of grains are unbroken, while in the 4.01-5.5X settling range Halimeda grains are predominantly broken. This constituent is present in greater abundance in the southern transect. Constituents of known (limited) source area (Marginopora, alcyonarian spicules and Amphistegina) exhibit a lagoonward shift from fast to slower settling character and also change in abundance. However, the direction and magnitude of changes differ between the east and south. While Amphistegina components exhibit a lagoonward decline in abundance, trends are more complicated for Marginopora which increases in abundance in both the sand apron (east) and passage exit (south) zones. Alcyonarian spicules do not change in abundance

along the eastern transect although they do become more common in slower settling fractions.

4. Discussion Settling velocity measurements and component analyses of surficial deposits of Cocos identify eight settling velocity and five component sediment types. These characteristics are used to interpret sediment transport processes throughout Cocos and explain the sediment patterns displayed in Figs. 5 and 6.

4.1. The spatial pattern of settling velocity classes: transport and energy interpretation The presence of sediment with the fastest settling properties on the outer reef flats (Fig. 5) suggests that this is the highest-energy depositional zone on Cocos. Similar material also forms large ridges (2 m above MSL) bordering the eastern passages. These ridges consist of pebble-sized material which produces the dominant mode of the RSVmean at 2X (25 cm s -1) with a minor secondary mode of slower settling 4 X (6.3 cm s -1) material (cluster 2, Fig. 4). Two processes control the formation of these deposits and their observed settling velocity distributions. First, coral material produced on the reef flat is transported into the passage and onto the ridges by waves during episodic high-energy storm events. Second, the ridges are porous (due to the large size of clasts of which they are composed) and backwash associated with high-energy waves removes much of the slow settling material. However, slower settling particles are

PS. Kench/Sedimentary Geology 114 (1997) 109-130

123

Table 5 Component cluster statistics Components

Coral Coralline algae Alcyonarian spicules Crustaceans Echinoids

Halimeda Homotrema Marginopora Amphistegina Other foraminifera Gastropods Bivalves

Molluscs Indeterminate

Cluster l, coral-dominated

Cluster 2, Halimeda-rich

57.04 (7. 7) 9.68 (6.06) 5.64 (4.09) 0.79 (1.17) 0.45 (0.46) 4.87 (3.62) 1.92 (2.01 ) 0.78 ( 1.09) 5.16 (3.54) 0.27 (0.55) 1.77 (1.68) 0.35 (0.42) 2.76 (1.25) 8.51 (5.68)

47.81 (6.68) 5.44 (3.44) 3.7 (2.66) 1.0 (0.74) 0.18 (0.19) 20.63 (4.72) 3.54 (3.47) 2.40 (1.38) 4.08 (3.06) 0.61 (0.91) 2.34 (3.09) 0.60 (0.56) 2.22 (1.15) 5.44 (2.18)

Cluster 3. coralline algae

43.47 (5.12) 33.4 (11.39) 0.99 (0.97) 0.28 (0.4) 0.50 (0.47) 0.81 (0.59) 0.25 (0.36) 0.18 (0.15) 0.80 (0.69) 0.45 (0.64) 13.13 (13.8) 0.22 (0.31 ) 1.94 (1.14) 3.53 (2.71)

Cluster 4,

molluscanIHalimeda 61.71 (6. 78) 6.88 (3.72) 1.04 (1.56) 1.62 (0.96) 0.33 (0.34) Z87 (3.98) 0.14 (0.24) 1.12 (1.24) 0.82 (0.99) 0.68 (0.74) 8.44 (4.42) 1.19 (1.22) 2.98 (1.17) 9.47 (2.85)

Cluster 5, molluscan/Halimeda

27.26 (7.86) 3.11 (4.82) 0.14 (0.28) 3.88 (2.66) 0.0 8.57 (5.44) 0.16 (0.33) 2.78 (1,08) 0.02 (0.02) 2.25 ( 1.63) Z 84 (4.38) 1.7 (1.65) 2.67 (1.98) 39.46 (8.97)

The representative percentage of each component within a cluster is presented with the standard deviation of each component within a cluster shown in brackets. Italicised values indicate those components that have a high presence within the cluster and aid in discriminating the cluster.

deposited in ridges during low-energy periods when wave backwash is negligible (field observation). The general pattern of sediment settling characteristics exhibits a gradient from fast settling reef flat sediments to the fast-moderate settling passage and moderate settling sand apron deposits, reflecting a decrease in hydrodynamic energy (Kench, 1994a). This pattern has two implications: first, particles with slow settling properties are selectively transported from the reef flat and passages toward the lagoon; and second, sediment with faster settling characteristics is preferentially deposited on the outer reef flat and passage entrance zones forming a lag. The asymmetric character of settling distributions (skewed to the fast settling end) of sediment from the reef flats and passages (clusters 1, 2 and 5, Figs. 4

and 5 B - D ) suggests that these areas are high-production zones with fast settling, locally produced, grains deposited and slow settling grains (secondary constituents) hydraulically sorted lagoonward. The distribution of sediment settling clusters (Fig. 5 B - D ) also indicates that sediments on the southern reef flat have slower settling properties than those on the east. This relationship implies the presence of lower hydrodynamic energy levels and shear stresses required to mobilise sediments in this zone. This energy difference was identified by Kench (1994a) who found that the wider southern reef flat dissipates wave energy before it reaches the passages. The progressive linear change in settling character of sediments between the outer reef flat and sand aprons contrasts with the more diverse settling pat-

P.S. Kench/Sedimentary Geology 114 (1997) 109-130

124

I

I

96"50"E

96"55"

INDIAN OCEAN

Horsburgh sland 12"05 "S

12"05"M

irection Island

:ym

--12"1C

Wq ld

~

Coral dominated

~

Coral - Halimeda rich

I

Coral- Coralline Algae

Coral -Molluscan/Halimeda Sands Coral - Molluscan/Halimeda Muds 0 I

21an I

96"50"

96"55"

I

I

Fig. 6. The distribution of five component classes of surficial sediments. Individual sample positions are shown in Fig. 2.

P.S. Kench / Sedimentary Geology 114 (1997) 109-130

, . , lOO

so

== 40

0

1

Halimeda ;..~ /

|

J

0

o~

8

~

e

=

4

g ¢3r LL

2 0

6 o ~" ¢D

4

cr

2

IL

0 5

Marginopora

4

o

3

=

2

c

¢T

LL

tern in the lagoon, which is dominated by moderate settling samples, interspersed with patches of faster settling deposits (Fig. 5A). Lagoon sediments are less well sorted (Fig. 4) and most have slow settling tails. These characteristics imply deposition of fine sediment transported from passages and sand aprons, whereas, faster settling material may indicate local production. Sediment with slowest settling properties is concentrated in the southeastern sector of the lagoon (Figs. 3 and 5A). This area, in the lee of prevailing winds, experiences the lowest hydraulic energy in the atoll (Kench, 1994a) and acts as a sink for fine sediment. Beds of seagrass and green and brown algae (Fig. 1) also trap slow settling sediments. Similar biological communities have been identified by Ginsburg and Lowenstam (1958), Swinchatt (1965) and Scoffin (1970) to stabilise sediment and promote the deposition of fine material. Gastropods associated with the seagrass and algae locally contribute to sediments upon mortality. While settling provides an accurate reflection of the hydraulic characteristics of deposits, like the sieve procedure, it cannot discriminate between sediment that is transported and that which is biologically produced and deposited in situ. However, in zones where in situ production controls deposit formation settling velocity provides an indicator of the shear stress required to mobilise sediment (Kench and McLean, 1996). Component analysis of bulk samples and settling fractions was used to further resolve the relative importance of production, transport and deposition throughout Cocos.

1 0 0.8 Crustacean

I

4.2. The distribution of sediment components: transport and biological production In principle, results of component analyses derived from settling fractions are similar to those

~ 0.4

g

125

/11

0.2

U.

0.0

RF

SA1 SA2 Geomorphtc Environment SetUJngV~o=ty (ohl) 1.5-2 BIB 2-2.5

PE

MP

Ill 2~3 B I I 3-3.5

PX

B B S,5-4 ~ 4-,4.5

[] []

4,5-5 5-5.5

~3

Fig. 7. Component analysis of settling velocity fractions of samples in a transect between the reef flat and sand apron through passage 2 (see Fig. 1), eastern atoll. Samples shown in Fig. 2B. RF = reef flat; PE = passage entrance; MP = mid-passage; PX ---- passage exit; SA1, SA2, SA3 = samples in transect from inner to lagoonward edge of sand apron. Addition of constituent frequency percentage values for each geomorphic environment produce the total frequency distribution for deposits.

P.S. Kench /Sedimentary Geology 114 (1997) 109-130

126 80 v

O

80 40 20

14.

0 3O

o>2, °°'a''n°'"e o ~

I

0

~

2o

g 0 8

...

u.

0 6 4

2 0

g It.

0 !

--.

RF

PE

MP

PX

SA1

SA2

SA3

using sieve fractions (Smithers, 1994). Coral is the dominant constituent (>50%) of Cocos sediments and secondary components provide sample differentiation. Ecological zonation and physical transport processes control the observed component mix of sediments throughout Cocos (Fig. 1, Table 1) of which there are three principal types: (1) Coral dominated (cluster 1) containing high proportions of fore-reef and reef fiat derived components, including alcyonarian spicules, coralline algae, and Amphistegina (Table 5). The extension of these deposits through the passages and onto sand aprons (Fig. 6) reinforces the transport pathway between these zones identified from settling velocity patterns (Fig. 5B-D). (2) Southern passage deposits contain the largest proportion (20%) of Halimeda grains. The presence of unbroken grains (indicating a local origin) in faster settling fractions indicates that currents are unable to remove these constituents. (3) Shallow and deep lagoon deposits (clusters 4 and 5) lack reef flat derived secondary constituents and are characterised by an increased abundance of molluscan and Halimeda fragments (Table 5). Many of these particles are unbroken and together with fragile coral skeletons indicate in situ production and deposition. Coral patches are common in parts of the Cocos lagoon (Fig. 1, Table 1) and the sediment they shed, combined with in situ production by other benthic organisms contributes to the faster settling fractions of deposits. Jordan (1973) and Frith (1983) identified similar contributions to lagoonal sediments in Bermuda and the Great Barrier Reef, respectively. Thus, in situ production results in variable settling velocity classes shown in Fig. 5A, with areas devoid of coral patches displaying slower settling properties. There is insufficient hydrodynamic energy to sort sediment within the lagoon (Kench, 1994a). The only evidence of sediment movement is in areas adjacent to the deep passage entrances (sandy lagoon floor, Fig. 1) where oscillatory ripples

Geomo~hic Environment

Settling Velocity (chi) B B 1.5-2 Jmm 2,5.3 [ ] B B 2-2,5 BIB 3-3.5 ~

3.5-4 44.5

[] []

4.5-5 5-5.5

Fig. 8. Component analysis of settling velocity fractions of sampies in a transect between the reef fiat and sand apron through passage 9 (see Fig. 1), southern atoll. Samples shown in Fig. 2B.

RF = reef flat; MR = mid-reef flat; PE = passage entrance; MP = mid-passage; PX = passage exit; SA1, SA2 = samples from front and lagoonward end of sand apron. Addition of constituent frequency percentage values for each geomorphic environment produce the total frequency distribution for deposits.

P.S. Kench / Sedimentar), Geology 114 (1997) 109-130

are found. The sparse bioturbating organisms in the deep lagoon also indicate that biologically induced sediment transport is low in this region. The sharp transition between sediments derived from the reef flat and those of the lagoon, characterised by in situ additions, confirms the conclusions drawn from the settling velocity patterns that active transport is constrained between the reef flat and lagoonward end of sand aprons, including the tongue of sediment progressing northward along the lagoon shoreline of West Island (Fig. 6). The shallow and deep lagoon acts as a long-term sink for sediment with deposition controlled by both autochthonous contributions and episodic transport from sand aprons during high-energy events, adding slow settling material to deposits (clusters 6, 4 and 1, Fig. 4).

4.3. Component analysis of settling fractions in the active transport zone Changes in type and abundance of constituents between and in similar settling fractions in different areas (Figs. 7 and 8) and particle form (fragmented or broken) provide indicators of the nature of sediment dispersal in the active transport zone. Most constituents are found in a wide range of settling fractions in reef flat samples, indicating that this is the dominant production zone. The distribution of components also highlights both the various stages in their life cycles when constituents contribute to the sediment, and the discontinuous mechanical breakdown of components during transport. Results indicate that the passage entrance (east) and mid-reef flat to passage entrance (south) are temporary depositional sites which act as storage areas for outer reef fiat sediments. This association is reflected in the marked decline in abundance of Amphistegina tests in faster settling fractions lagoonward of the passage entrances (Figs. 7 and 8) with Amphistegina behaving as a lag in the passage entrance. Faster settling Marginopora and alcyonarian spicules are also deposited in passage entrances. These areas are longer-term sinks for very fast and fast settling coral and coralline algae constituents which form a lag (Figs. 7 and 8). Lag components are heavily frosted and encrusted with epiphytic and boring organisms indicating a long residence

127

time. Other fast settling particles are found in the rubble ridges bordering passages. Similar deposits are identified by Folk and Robles (1964), McLean and Stoddart (1978) and Scoffin (1993), and reflect transport and deposition during episodic high-energy conditions. Mid-passages form conduits of sediment transfer between the reef flat and passage exit. This role is reflected in shifts in abundance of Amphistegina tests and Halimeda fragments toward slower settling fractions reflecting their selective transport from reef flats and passage entrances. In the southern passage coral components (2.51-3.0X) and Amphistegina tests (3.01-3.5X) decrease in abundance lagoonward, reflecting a continuously decreasing transport potential from a sediment source. In the eastern passage Amphistegina tests and alcyonarian spicules (in the same settling ranges) are deposited in the mid-passage and are not found in the passage exit. The increased abundance of slower settling fractions in passage exits indicate that these are the lowest-energy depositional zones along the transport pathway under fair-weather energy conditions. Grains in all fractions are fragmented and abraded, indicating transport and deposition. Halimeda fragments are most abundant in these areas on both sides of the atoll (Figs. 7 and 8). These slow settling constituents would be entrained and transported further lagoonward during higher-energy (storm) events. The increase in abundance of Halimeda and Marginopora fragments and alcyonarian spicules, in slower settling fractions, across the sand aprons (particularly on the east, Fig. 7) indicates that these form long-term depositional sites. Increases in abundance are attributed to the accumulation of grains without further lagoonward transport. The reappearance of Amphistegina tests in faster settling fractions (3.01-3.5X absent in passage exit) on eastern sand aprons is attributed to high-energy flushing of faster settling grains from passages by-passing the passage exit. In contrast to Amphistegina, Marginopora, and alcyonarian spicules Halimeda is produced in all environments and its presence in all areas in the 2.55.5X settling range indicates its rapid production and ease of breakdown. Increases in abundance of Halimeda across the sand apron can be partly attributed to local production and whole grains are observed in

128

PS. Kench/Sedimentary Geology 114 (1997) 109-130

faster settling fractions. Local production and in situ deposition of sediment is also clearly demonstrated by the marked increase in crustacean constituents on the eastern sand apron, but particularly on the southern sand apron (Figs. 7 and 8). These additions lead to an increase in the mean settling velocity of samples across the sand apron. The inferred reef flat to sand apron transport pathway is clearly demonstrated by the distribution of the foraminifera Marginopora (Fig. 7). Whole Marginopora plates are produced on the reef flat and are present in the 3.51-4.0X settling range. These constituents have slow settling rates, despite their large physical size, which suggests that they are transported large distances once entrained (Maiklem, 1968; Braithwaite, 1973). The absence of Marginopora in slower settling fractions of higher-energy zones, and the increase in abundance of fragments in slower settling fractions on the sand aprons, confirms this notion as grains with large size are preferentially transported lagoonward. Komar and Wang (1984) and Li and Komar (1992) identify a similar preferential transport of larger-sized, lower-density grains in the formation of heavy mineral deposits and attribute this to an inverse relationship between grain size and density. Such a relationship is not as simple in bioclastic sediments in which the bulk density of constituents varies due to skeletal architecture and can change as components are broken. It is possible to have some constituents with both larger grain size and bulk density than smaller components. These results support the use of the settling velocity approach to interpret transport pathways as settling fractions account for the size, shape density differences between components and grains. The settling procedure avoids misinterpretation of the hydraulic behaviour of such grains that can result from sieve-based textural analysis (Kench and McLean, 1996). 5. Conclusions Settling velocity analysis of bulk deposits and settling velocity fractions describes and differentiates bioclastic deposits using a hydraulic index of grain behaviour. Results from analysis of Cocos sediments identifies eight settling and five component classes. Coral detritus dominates the component mix

of samples which are differentiated by secondary components. The pattern of settling characteristics implies that active sediment transport on Cocos is constrained between the high-energy and productive reef fiats and lower-energy sand aprons. Lagoon sediments have a variable settling pattern and component analysis indicates an autochthonous supply of material. However, like sieve analysis, settling analysis of bulk samples is unable to differentiate physically deposited and biologically produced grains for bulk deposits, to resolve details of the sediment transport system. As suggested by Braithwaite (1973) this study examines the component mix of settling fractions and this allows the nature of the sediment transport system to be resolved at a finer scale than broad reef fiat to sand apron interpretations identified from bulk settling and sieve analyses. This analysis is shown to have several advantages. First, particles (of varied size, shape, and density) within fractions possess a narrow range of hydraulic properties from which to infer depositional energy; second, inspection of skeletal components, and degree of degradation of grains helps to differentiate those which have been transported or locally produced; and third, changes in the abundance of components between settling fractions help determine transport pathways. Results show that on Cocos sediment is transported from the reef fiat, with initial long-term deposition of faster settling lag material on the southern inner reef fiat and in eastern passage entrances. Passage entrances are holding zones while passages act as conduits for lagoonward transport. Passage exits are the main deposition zones during fair-weather energy conditions. During moderate-energy events sediment by-passes these areas to be deposited on the sand aprons which form medium-term sediment sinks. Locally produced material increases the apparent settling velocity properties of these sediments. Finally, during high-energy (storm) episodes sediment with slow settling properties is flushed from the sand aprons to the deep lagoon. The study shows that the settling approach is a viable alternative to sieve analysis and that component analysis of settling velocity fractions can resolve sediment transport and sedimentation processes within carbonate environments. Use of the settling param-

P.S. Kench/Sedimentary Geology 114 (1997) 109-130 e t e r as a d e s c r i p t o r o f b i o c l a s t i c s e d i m e n t s , c o u p l e d w i t h m e a s u r e m e n t s o f current energy, p r o m i s e s to further r e s o l v e p h y s i c a l s e d i m e n t a t i o n p r o c e s s e s in r e e f e n v i r o n m e n t s . T h i s a p p r o a c h w o u l d a l l o w deposits to be d e s c r i b e d r e l a t i v e to currents o f r e m o v a l (i.e. the p o r t i o n o f s e d i m e n t able to be transported f r o m a deposit) and differential rates o f p r o d u c t i o n to be detected. D u n h a m (1962), O r m e (1977) and Scoffin (1987) h a v e all n o t e d that currents o f rem o v a l p r o v i d e s a better a p p r o a c h to u n d e r s t a n d i n g the d e v e l o p m e n t o f b i o c l a s t i c deposits than currents o f delivery.

Acknowledgements T h i s r e s e a r c h was s u p p o r t e d by the A u s t r a l i a n N a t i o n a l G r e e n h o u s e A d v i s o r y C o m m i t t e e . T h e author w o u l d like to thank the C o c o s ( K e e l i n g ) Islands A d m i n i s t r a t i o n for their support, Scott S m i t h e r s for s u p p l y i n g additional s a m p l e s , Prof. R o g e r M c L e a n for c o n s t r u c t i v e c o m m e n t t h r o u g h o u t the study, rev i e w e r s Dr. H. Roberts, Dr. C. B r a i t h w a i t e and Dr. B. S e l l w o o d w h o s e c o m m e n t s i m p r o v e d the manuscript, and C h a n d r a J a y a s u r i y a for p r o d u c t i o n o f figures.

References Berry, P.E, 1989. Background habitat notes and overall conclusions and recommendations. In: Berry, EE (Ed.), Survey of the Marine Fauna of Cocos (Keeling) Islands, Indian Ocean. Report to ANPWS by the Western Australian Museum, Perth. Braithwaite, C.J.R., 1973. Settling behaviour related to sieve analysis of skeletal sands. Sedimentology 20, 251-262. Bryant, E., 1984. Sample site variation in settling velocity distribution subpopulations using curve dissection analysis. Sedimentology 33, 767-775. Chevillon, C., 1990. Sedimentology of the Great Northern Lagoon of New Caledonia: description of depositional environments using Principal Component Analysis. Proc. Int. Soc. Reef Stud. Meeting, Noumea, pp. 165-172. Chevillon, C, Clavier, J., 1988. Sedimentological structure of the Northern Lagoon of New Caledonia. Proc. 6th Int. Coral Reef Symp. 3, 425-430. Colby, N.D., Boardman, M.R., 1989. Depositional evolution of a windward high energy lagoon, Grahams Harbour, San Salvador, Bahamas. J. Sediment. Petrol. 59, 819-834. Dunham, R.J., 1962. Classification of carbonate rocks according to depositional texture. Am. Assoc. Pet. Geol. Mem. 1, 108121. Falkland, T., 1994. Climate, hydrology and water resources of the Cocos (Keeling) Islands. Atoll Res. Bull. 400, 52 pp.

129

Flood, EG., Scoffin, T.E, 1978. Reefal sediments of the northern Great Barrier Reef. Philos. Trans. R. Soc. London A 291, 55-71. Folk, R.L., 1974. Petrology of Sedimentary Rocks. Hemphill, Austin, Texas. Folk, R.L., Robles, R., 1964. Carbonate sands of Isla Perez, Alacran Reef Complex, Yucatan. J. Geol. 72, 255-292. Frith, C.A., 1983. Some aspects of lagoon sedimentation and circulation at One Tree Reef, southern Great Barrier Reef. BMR J. Aust. Geol., Geophys. 8, 211-221. Gahrie, C., Montaggioni, L., 1982a. Sediments from fringing reefs of Reunion Island, Indian Ocean. Sediment. Geol. 3 l, 281-301. Gahrie, C., Montaggioni, L., 1982b. Sedimentary facies from the modern coral reefs, Jordan Gulf of Aqaba, Red Sea. Coral Reefs l, 115-124. Galehouse, J.S., 1971. Sedimentation analysis. In: Carver, R.E. (Ed.), Procedures in Sedimentary Petrology. Wiley-Interscience, New York, pp. 69-94. Ginsburg, R.N., 1956. Environmental relationships of grain-size and constituent particles in some Florida carbonate sediments. Am. Assoc. Pet. Geol. Bull. 40, 2384-2427. Ginsburg, R.N., Lowenstam, H.A., 1958. The influence of marine bottom communities on the depositional environment of sediments. J. Geol. 66, 310-318. Jordan, C.F., 1973. Carbonate facies and sedimentation of patch reefs off Bermuda. Am. Assoc. Pet. Geol. Bull. 57, 42-54. Kench, ES., 1994a. Atoll Hydrodynamics and Sedimentation: Cocos (Keeling) Islands, Indian Ocean. PhD thesis, Univ. of New South Wales, 344 pp. Kench, ES., 1994b. Hydrodynamic observations of the Cocos (Keeling) Islands lagoon. Atoll Res. Bull. 408, 21 pp. Kench, ES., McLean, R.E, 1996. Hydraulic characteristics of bioclastic deposits: new possibilities for environmental interpretation using settling velocity fractions. Sedimentology 43, 561-570. Kench, ES., McLean, R.E, 1997. A comparison of settling and sieve techniques for the analysis of bioclastic sediments. Sediment. Geol. 109, 111-119. Komar, ED., Wang, C., 1984. Processes of selective grain transport and the formation of placers on beaches. J. Geol. 92, 637-655. Li, M.Z., Komar, ED., 1992. Longshore grain sorting and beach placer formation adjacent to the Colombia River. J. Sediment. Petrol. 62, 429~-41. Lund-Hansen, L.C., Oehmig, R., 1992. Comparing sieve and sedimentation balance analysis of beach, lake and eolian sediments using log-hyperbolic parameters. Mar. Geol. 107, 139147. Maiklem, W.R., 1968. Some hydraulic properties of bioclastic carbonate grains. Sedimentology 10, 101-109. May, J.E, 1981. Chi (X): A proposed standard parameter for settling tube analysis of sediments. J. Sediment. Petrol. 51, 607-610. McLean, R.F., Stoddart, D.R., 1978. Reef island sediments of the northern Great Barrier Reef. Philos. Trans. R. Soc. London A291, 101-117.

130

P.S. Kench / Sedimentary Geology 114 (1997) 109-130

Michels, K.H., 1995. The decomposition of polymodal settling velocity distributions for a comprehensive sedimentological description of sand-sized samples. Sedimentology 42, 31-38. Oehmig, R., Wallrabe-Adams, H., 1993. Hydrodynamic properties and grain-size characteristics of volcaniclastic deposits on the Mid-Atlantic Ridge north of Iceland (Kolbeinsey Ridge). J. Sediment. Petrol. 63, 140-151. Orme, G.R., 1977. Aspects of sedimentation in the coral reef environment. In: Jones, O.A., Endean, R. (Eds.), Biology and Geology of Coral Reefs, 4. Geology 2, 129-182. Purdy, E.G., 1963. Recent calcium carbonate facies of the Great Bahama Bank, 2. Sedimentary facies. J. Geol. 71,472-497. Reed, W.E., Le Fever, R., Moir, G.J., 1975. Depositional environment interpretation from settling velocity (Psi) distributions. Geol. Soc. Am. Bull. 86, 1321-1328. Sagga, A.M.S., 1992. The use of textural parameters of sand in studying the characteristics of coastal sediments south of Jeddah, Saudi Arabia. Mar. Geol. 104, 179-186. Scoffin, T.P., 1970. The trapping and binding of subtidal carbonate sediments by marine vegetation in Bimini Lagoon, Bahamas. J. Sediment. Petrol. 40, 249-273. Scoffin, T.P., 1987. An Introduction to Carbonate Sediments and Rocks. Blackie, Glasgow, 274 pp. Scoffin, T.P., 1993. The geological effect of hurricanes on coral reefs and the interpretation of storm deposits. Coral Reefs 12, 203-221.

Scoffin, T.E, Tudhope, A.W., 1985. Sedimentary environments of the central region of the Great Barrier Reef of Australia. Coral Reefs 4, 81-93. Smithers, S.G., 1994. Sediment facies of the Cocos (Keeling) Islands lagoon. Atoll Res. Bull. 407, 34 pp. Smithers, S.G., Woodroffe, C.D., McLean, R.F., Wallensky, E.P., 1994. Lagoonal sedimentation in the Cocos (Keeling) Islands, Indian Ocean. Proc. 7th Int. Coral Reef Symp. 1,273-288. Swinchatt, J.P., 1965. Significance of constituent composition, texture and skeletal breakdown in some recent carbonate sediments. J. Sediment. Petrol. 35, 633-648. Taira, A., Scholle, P.A., 1979. Discrimination of depositional environments using settling tube data. J. Sediment. Petrol. 49, 787-800. Ward, J., 1963. Hierarchical grouping to optimise an objective function. J. Am. Stat. Assoc. 58, 236-244. Wardlaw, B.R., Thomas, W.H., Martin, W.E., 1992. Sediment facies of Enewetak atoll lagoon. U.S. Geol. Surv. Prof. Pap. 1513-B. Weber, J.N., Woodhead, P.M.J., 1972. Carbonate lagoon and beach sediments of Tarawa atoll, Gilberts Islands. Atoll Res. Bull. 157, 29 pp. Williams, D.G., 1988. Report to the environmental resource advisor Cocos (Keeling) Islands, to the Minister for the Arts, Sport, Tourism, Environment and Territories. Unpublished report to the Australian Government.