af fect of sedimentation on str omatolite reef gr o wth...

IntroductionStromatolites first appear in the Archean rock record andare the reef and carbonate platform builders through the majority of Precambrian time, culminating with theintroduction of thrombolite reefs and stromatolite-thrombolite hybrids in Neoproterozoic rocks(Grotzinger, 1989; 2000; Grotzinger and James, 2000;Riding, 2000). To understand the dynamics of Archeanand Proterozoic carbonate systems requires a betterunderstanding of stromatolite growth dynamics andresponse to environmental conditions. There iscontinuing debate on what aspects of stromatolitemorphology represent the evolution of microorganisms,and what aspects of morphology reflect changingenvironmental conditions (Grotzinger and Rothman,1996; Semikhatov and Raaben 2000; Grotzinger andKnoll, 1999; Riding, 2000). Some attributes of short-termenvironmental variability, such as changes in sea level, may affect stromatolite development (Grotzinger,1989; Narbonne and James, 1996). The variability of mat communities may also help dictate textural andpossibly morphologic responses, such as in thedevelopment of thrombolites (Turner et al., 1997;Grotzinger, 2000). To interpret both environmentalconditions and possible evolutionary trends, therelationships between growth variables and stromatolitemorphology must be deciphered (Grotzinger and Knoll,1999).

Stromatolites are laminated and lithified sedimentarygrowth structures that accrete from a point or limitedsurface of attachment (Grotzinger and Knoll, 1999). The Ediacaran age of the stromatolites in this studyindicates that it is reasonable to assume that microbialmats mediated sediment binding and trapping. As formost microbialites of younger Precambrian age, theformation of microbially mediated laminae andstromatolites is probably a combination of biotic and abiotic processes (Grotzinger and Knoll, 1999).Stromatolitic and thrombolitic textures often occur asproximal components within reefs of Ediacaran age(Grotzinger, 2000; Grotzinger and James, 2000), andmany of the stromatolites referred to in this paper arehybrids that have a combined stromatolite-thrombolitetexture (Kennard and James, 1986; Feldman andMcKenzie, 1998; Grotzinger, 2000). Over geologic time,stromatolites have occurred in a huge variety ofbranching patterns, shapes and sizes (Semikhatov andRaaben, 2000); the stromatolites described here formrelatively straight and rarely branching columns.

Based on analogy to modern carbonate reefs, manyenvironmental factors probably controlled reef growthand nucleation, such as water depth, light intensity,sediment type, sedimentation rates, salinity and nutrientflux (Grotzinger, 1989; James and Bourque, 1994; Woodand Oppenheimer, 2000). Changes in sediment typelikely occur in tandem with changes in other

J. JOHNSON AND J. P. GROTZINGER

SOUTH AFRICAN JOURNAL OF GEOLOGY, 2006, VOLUME 109 PAGE 87-96

87

Affect of Sedimentation on Stromatolite Reef Growth andMorphology, Ediacaran Omkyk Member (Nama Group), Namibia

J. JohnsonDepartment of Earth, Atmospheric and Planetary Sciences, 54-822

Massachusetts Institute of Technology, Cambridge, MA 02139e-mail: [email protected]

J. P. GrotzingerDepartment of Earth, Atmospheric and Planetary Sciences, 54-822

Massachusetts Institute of Technology, Cambridge, MA 02139*Current Address: Division of Geological and Planetary Sciences

California Institute of Technology, Pasadena, CA 91125e-mail: [email protected]

© 2006 March Geological Society of South Africa

ABSTRACTSuperbly preserved reefs of Ediacaran age in Namibia give clues to environmental controls on stromatolite-thrombolite growthmorphology and nucleation. Digital mapping and Measured stratigraphic sections reveal parasequences featuring meter-scalealternation of shallow marine shale and stromatolitic-thrombolitic microbialites associated with other clastic carbonates (grainstones,mudstones). Stromatolite-thrombolite column width, spacing and height vary systematically with the type of sediment beingdeposited, with growth inhibited during shale deposition. Columns are generally wider and more closely spaced during carbonatesediment deposition and narrower and more widely spaced during shale deposition. While stromatolite growth should be sensitiveto sediment type explicitly, we also interpret sediment type as a general proxy for changing environmental conditions (e.g. waterdepth, turbidity) that may directly affect reef growth. A simple rule-based numerical simulation of microbialite growth is formulatedbased on the field interpretations of sedimentological and topographic growth controls. The model cannot explain detailedmorphologic attributes, but can recreate correlations between stromatolite column widths, column spacing and layer bed thicknessas a function of sediment type.

environmental conditions, such as water depth orturbidity, and sediment type may be a proxy for othervariables affecting stromatolite growth. Sediment type isalso the only variable for which evidence is directlypreserved in the rock record. Stromatolites are expectedto be directly sensitive to sedimentation rates and eventsbecause they grow by sediment accretion; too littlesediment and the mat will not be preserved, whereas toomuch sediment may terminate growth because theactive layer of microorganisms becomes buried(Grotzinger and Knoll, 1999).

Recent studies of modern stromatolite growth haveonly considered the interactions between micro-organisms and carbonate sand, not clay or siliciclasticsediment (MacIntyre et al., 2000; Reid et al., 2000). Localdeposition of both clays and siliciclastic sediment wouldlikely be detrimental to the microbial community, due tolight blocking, inhibition of advection and diffusion ofnutrients to the microorganisms, or a lack of earlycementation preventing a hard or stable surface onwhich to develop. Stromatolites usually form in

carbonate sediment because it lithifies early, giving themthe necessary strength to maintain positive topographyand to not get swept away by storms and currents(Grotzinger and Knoll, 1999).

Finally, antecedent topography is likely to be adeterministic control on reef nucleation, again influencedby sediment type and sedimentation rate. Topographichighs will tend to get buried in less sediment, and willtherefore be better locations for stromatolite growth. The importance of antecedent topography has beenshown for coral reefs (Edwards and Brown, 1999; Ferroet al., 1999), and is expected for microbial reefs as well(Grotzinger, 1989; Grotzinger and Knoll, 1999). In general, coarser-grained sediments will tend toaccumulate mostly in lows, where finer-grainedsediments – particularly cohesive clays – will tend to stickto highs as well as accumulating in topographic lows.

The goal of this paper is to evaluate the role ofsediment type and relative sedimentation rate incontrolling the growth and morphology of Ediacaranstromatolite-thrombolite reefs in the northern Namabasin, Namibia. The study involves digital mapping ofreef geometry, reef-sediment relationships, andnumerical simulation of these mapped geometries tobetter understand the controlling variables.

Geologic and Stratigraphic SettingThe reef examined in this study developed within atransgressive systems tract of a carbonate ramp systemdeveloped in a foreland basin of Ediacaran age (Figure 1; Adams et al., 2005; Grotzinger, 2000; Saylor et al., 1995). The reef occurs at the Zebra River farm, inthe Omkyk Member of the Kuibis Subgroup (NamaGroup). An ash bed constrains the Omkyk to besomewhat older than 548.8+/-1 Ma (Grotzinger et al.,1995). The stratigraphic architecture of the OmkykMember at Zebra River was digitally mapped by Adamset al. (2005). The Omkyk Member is subdivisible intotwo sequences: Omkyk Sequence 1 (OS1) consistingprimarily of shelf grainstones, and Omkyk Sequence 2(OS2), which contains grainstones, shales, andthrombolite-stromatolite reefs and biostromes at multiple

SOUTH AFRICAN JOURNAL OF GEOLOGY

AFFECT OF SEDIMENTATION ON EDIACARAN STROMATOLITE REEF GROWTH AND MORPHOLOGY88

Figure 1. Regional map of central and southern Namibia, showing

the location of the field area within the Kuibis Subgroup of the

Nama Group of sedimentary units. After Adams et al. (2005).

Figure 2. Panoramic photograph of patch reefs in unit 1 and unit 2 of OS2, separated by a maximum flooding surface under the talus-

covered slope. The unit 1 reef focused on in this work is in the lower left corner of the photograph, and the adjacent inter-reef sediment

pocket of parasequences is in the lower center, with ledges that correspond to the tops of carbonate beds. Two people are circled for scale.

stratigraphic layers. OS2 in turn is divided into fiveparasequence sets (units 1 to 5). Unit 1 forms thetransgressive system tract and is a backsteppingparaseqence set containing discrete patch reefs whichare progressively onlapped and overlain by fine-grainedsediments, dominated by shale. The reef studied hereoccurs in this lower unit (Adams et al., 2005). Unit 2 –the interval of maximum flooding – contains a smallernumber of reefs, which managed to grow despite themaximum flux of shale at this time.

Local variations in sediment type indicate higherfrequency relative water depth fluctuations wellexpressed by meter-scale parasequences marked byalternating shale and carbonate. Consistent with theincrease in shale deposition, a maximum floodingsurface (representing the greatest water depth) occursabove the studied reef and separates OS2 unit 1 fromunit 2, which in turn contains many well-developedpatch reefs (Figure 2).

These Ediacaran reefs at Zebra River offer excellentexposure of both the internal structure of reefs and thespatial relations between reefs. Adams et al. (2005)digitally mapped the field relations between reefs andexplored quantitative spatial relations between reefs anddepositional patterns within parasequence sets. In thiscomplementary study we focus on internal depositionaland growth structures within a single thrombolite-stromatolite reef. We present quantitative measurementsof reef parameters including stromatolite widths and time-equivalent interfingering clastic carbonates andterrigenous sediments. We show how parasequencesformed of alternating shales and carbonates correlatewith contemporaneous changes in reef growth, and wepresent a numerical simulation of stromatolite growththat suggests ways in which variable sediment type maycause the observed growth changes.

MethodsThe geometric relations between stromatolites andassociated clastic carbonates and terrigenous sedimentswere observed in the context of possible controls onstromatolite nucleation and their subsequent lateral andvertical growth. Measurements were then made ofstromatolite column widths, the widths of sediment fillbetween stromatolite columns, the rotation of fracturedstromatolite pieces, and reef layer thicknesses within asingle stromatolite-thrombolite reef (Figure 3). Widths ofstromatolite columns and sediment fill within the reefwere primarily measured in the field, with someadditional values later estimated from photographs. A stratigraphic section adjacent to the reef was measuredat the centimeter scale in order to establish a referencefor correlation of inter-reef strata into the reef itself. Thesuperb exposures allowed four parasequences ofalternating shale and carbonate to be defined and alsotraced into the reef core.

An attempt was made to correct the measuredthicknesses for the effects of compaction. For thisapproximation, we assumed carbonate sediment

compaction to be negligible due to early lithification. We estimated shale compaction relative to carbonatefrom shale beds that contained carbonate nodules(assumed to lithify early with zero compaction), withchanges in shale bedding thickness compacted aroundconcretions giving an average amount of differentialcompaction. The mean value of compaction, given asstrain (!l/lorig), is 0.63+-0.06 (1", 7 measurements).

Compaction of reef layers was calculated from thefracturing and rotation of stromatolite columns (Figure 3). Because the stromatolites deform rigidly byrotating and breaking, the columns or column piecescan be used as strain markers to estimate verticalshortening of reef layers. In each reef layer, plungeangles of stromatolite fragments were measured in thefield, with some additional angles measured fromphotographs. Plunges were measured systematically atequal vertical spacings, although segment lengths werenot explicitly measured. The vertical rotated length ld ofeach segment is related to the unrotated length lu andplunge angle from horizontal q simply as ld=lu sin(#).The average vertical strain e in the layer is then given as$=%(1-ld)/n for n measurements, assuming lu =1. Whilethis method of estimating the strain in each reef layer isonly approximate, the measured strains are relativelysmall, and lateral variability in reef compaction appearedto be larger than the error that is caused by the strainmeasurement methodology. Uncompacted reef andinter-reef parasequence thicknesses were calculatedusing the respective reef layer and shale compactionestimates.



89

Figure 3. Centre of unit 1 study reef, showing labeled reef

parasequences that correspond to alternating carbonate and shale

deposition of the inter-reef parasequences. Reef layers suggest

variable compaction. Circled staff is 2 m in length. Layers are

numbered corresponding to the parasequences and labeled “r” for

reef, “s” reef growth equivalent to shale deposition, and “c” for

growth equivalent to carbonate deposition. Note the fractured and

rotated segments of stromatolite columns and sediment fill, used to

constrain compaction of reef layers.

Field observations of reef growth relationshipsField observations suggest that shales tend to inhibitstromatolite nucleation and growth. Stromatolites werenever observed to nucleate directly on clay-richsediment. However, even thin carbonate beds allowstromatolite growth (Figure 4). The general progressionof morphology and texture begins with nucleation on ahorizontal surface formed by the tops of carbonategrainstone beds, and to a lesser degree carbonatemudstone beds. Crinkly, subhorizontal laminationsdevelop on the carbonate sediment, and local highsgradually grow in amplitude until a distinct initial domeformed of crinkly laminations can be seen, with distinctsediment pockets adjacent to the dome. Domes thendevelop into well-formed columns with distinct inter-column sediment fills. Tracing of bridging stromatolitelaminae through these fills indicates that synoptic reliefof the columns was typically small (cm scale).

Shale appears to be more effective than carbonatesediment at smothering stromatolite columns, because athin layer of shale is observed over the top of manyindividual columns and entire reef bodies. Shale alsocompacts more than carbonate sediment which couldcause an observational bias that thin shale layers caneffectively smother columns. In some places it appearsthat shale deposition effectively smothered the tops oflower stromatolites but not higher ones, suggesting thatstromatolite growth could at least tolerate theenvironmental condition of high shale flux (Figure 4).Some reef growth must have occurred during shaledeposition because upper and lower bounds of reefalgrowth increments correspond to shale deposition in theadjacent inter-reef depressions. During shale deposition,reef growth in some places reverted to more of a crinklylamination fabric with individual stromatolite columnsbecoming less clear and distinct.

Lateral progradation of reef margins occurred mostoften when reefal facies were able to downlap againstcarbonate beds; lateral expansion of reefs directly on

shales was not observed. These episodes resulted in aninterfingering pattern of growth in cross section. Lateralgrowth seems to occur by lateral extension of crinklylaminae, which then grew in amplitude to becomestromatolites.

Quantitative CorrelationsFigure 5 shows the stratigraphic column measuredadjacent to the reef. The stratigraphy showsparasequences that alternate between dominantlycarbonate and dominantly shale layers. We categorizedbeds as shale, carbonate mudstone, carbonategrainstone, or crinkly laminate (which encompasses allmicrobially-laminated beds including stromatolites).Parasequences 2 through 5 correspond to the mainphase of reef growth exposed in outcrop. Parasequence6 shale is continuous over the top of the reef outcropmarking the termination of reef growth; the reef has anoverall narrowing (backstepping) trend fromparasequences 2 through 5, consistent with its earlierinterpretation as part of a transgressive system tract(Adams et al., 2005). No column widths or spacings arereported from parasequence 3 shale deposition because



Figure 4. Small stromatolite-thrombolite columns amongst

alternating carbonate and shale layers. These are not part of the

studied reef, but are in a laterally-equivalent stratigraphic interval.

Columns nucleate on carbonate grainstone or mudstone layers, and

are commonly covered by shale. Visible staff length ~60 cm.

Figure 5. (a) Measured stratigraphic column thicknesses adjacent

to the reef (Figure 2), with numbered parasequences of shale and

carbonate deposition. Individual beds were categorized as shale,

carbonate mudstone, carbonate grainstone, or microbially

laminated carbonate, which included crinkly laminites and small

stromatolite heads (e.g. Figure 4). These thicknesses are compared

to measured reef thicknesses. (b) The same stratigraphic section

in which shale has been decompacted, compared to uncompacted

reef layer thicknesses. The decompacted inter-reef parasequences

sum to be ~2.5 m thicker than the decompacted reef

parasequences.

the reef at this level is composed only of crinklylaminations without distinct stromatolite columns.

When uncompacted, the thickness of shale depositedis significantly larger than the carbonate sedimentthickness for each inter-reef parasequence. Within eachparasequence that is traced from reef to inter-reef, thecalculated total reef thickness is less than the calculatedinter-reef sediment thickness. Figure 6a shows ratios ofuncompacted reef thickness to inter-reef sedimentthickness for each parasequence. With the exception ofparasequence 4, the thickness of reef facies wassignificantly higher during inter-reef carbonatedeposition than during inter-reef shale deposition.During shale deposition, the thickness of reef faciestends to be significantly less than the thickness of shaledeposited. Therefore, the reef tends to make positiverelief during carbonate deposition and to lose relief

during shale deposition. It is not possible to constrainthe absolute rates of growth or deposition. Figure 6bshows a positive correlation between the ratio of reef to inter-reef sediment thickness and the fraction of carbonate sediment deposited during eachparasequence.

During each parasequence, the internal structure ofthe reef changes as the composition of inter-reefsediment varies. Figure 6c shows the relationshipbetween stromatolite column width and inter-column fillwidth for reef growth during each inter-reefparasequence. Reef growth during shale and carbonatedeposition forms distinct populations, although thevariability in widths is large for most of the layers.Stromatolite columns are wider than the sediment fillduring carbonate deposition and narrower than the fill during shale deposition. There is also a separate,



91

Figure 6 (a) Ratios of uncompacted reef thickness to uncompacted sediment thickness, for shale and carbonate deposition during each

parasequence. Thickness ratios >1 indicate reef growth thicker than sediment deposition, while values <1 indicate that sediment deposition

was probably thicker than equivalent reef growth. The overall trend is decreasing thickness ratios higher up in the reef. Error bars 1".

(b) There is a positive correlation between the reef to sediment thickness ratio and carbonate sediment fraction. Carbonate sediment fraction

is the ratio of carbonate sediment thickness to the entire decompacted sediment thickness for a given parasequence. Numbers refer to the

inter-reef parasequence, “s” for shale deposition, “c” for carbonate deposition. Error bars 1". (c) Stromatolite columns grown during shale

deposition are narrower and more widely spaced than columns grown during carbonate deposition. Parasequence 3 shale-equivalent reef

growth does not have discernable columns, so is not plotted. In both populations, widths of both sediment and fill increase upwards in the

reef. Error bars 1". (d) The ratio of stromatolite column width to fill width tends to increase with carbonate sedimentation. Error bars 1".

longer-term trend of increasing column widths from thebottom to the top of the reef. Finally, Figure 6d shows apositive correlation between the column-to-fill widthratio of each reef parasequence and the fractionalthickness of carbonate beds in each corresponding inter-reef parasequence.

Numerical simulation of reef growthTo explore ways in which the observed reef growthrelations could be explained by sedimentation and localtopography alone, a simple 2-dimensional geometricmodel was implemented numerically. The model rulesare based on the qualitative growth controls inferredfrom field observations: 1) stromatolite growthpreferentially occurs on broad topographic highs,

because topographic lows preferentially accumulatesediment, and 2) stromatolite growth only occurs oncarbonate sediment, not on shale. The utility of thismodel is its ability to explore whether the suggestedgrowth rules alone can explain the observed growthrelations (e.g. Drummond and Dugan, 1999). The modeldoes not try to recreate in a realistic way the physicalprocesses of either sedimentation or stromatolite growth(Grotzinger and Rothman, 1996; Batchelor et al., 2000).

Each model timestep has a period of sedimentdeposition followed by a period of reef growth.“Stromatolite” facies represent growth and cementationof the stromatolites. In addition, two kinds of sedimentcan be deposited in the model: “carbonate” and “shale”.Sedimentation occurs in the model by horizontally fillinglows in the topography. Sediment can be set to fill ineither local minima or absolute minima in thetopography (Figure 7). These two sedimentation rulesare crude representations of differing sedimentationregimes, such as where sediment settles out ofsuspension and then relaxes into local topographic lows,versus sedimentation by traction flows that hug the bedand favor deposition in absolute lows. As will be shown,whether sediment fills local or absolute minima makeslittle difference in the correlations between variables inthis unscaled model. Sediment area (in two dimensions)is conserved, so the narrower the sediment pockets thethicker the deposits. User inputs related to sedimentdeposition are the initial topography and initial substratematerial (shale, carbonate or stromatolite) specified ateach x domain location, and the sediment type(carbonate or shale) and average sediment thickness(savg) to be deposited over the model domain pertimestep.

Lateral and vertical growth rules for the reef facieswere formulated. The incorporation of sediment into thereef is not explicitly modeled; the binding and trappingof sediment that leads to stromatolite growth is assumedto occur on length scales smaller than the x domainspacing. When stromatolite facies are directly adjacent tocarbonate sediment then lateral reef growth occurs as 1 horizontal grid spacing per timestep over thosecarbonates. In contrast, lateral growth does not occurwhen stromatolites are adjacent to shale. The inabilityfor lateral stromatolite growth to occur directly on shalesis the only explicit difference in the model betweencarbonates and shales.

Vertical reef growth occurs at each reef location as anexplicit combination of a specified mean growth rate, arandom component, and a dependence on local relief:

g (x) = gavg + an (x) + b [t (x-1) + t (x+1) - 2t (x)] (1)

Sediment type deposited in a given timestep does notexplicitly affect the vertical growth rate. Parameter g (x)is the vertical growth rate at location x, gavg is a user-specified mean growth rate, n is a gaussian noise term,t (x) is the elevation at horizontal location x, and a andb are scaling factors to vary the relative importance of



Figure 7. Cartoon of the geometric stromatolite model.

The illustrations of local and absolute minima sedimentation

are cartoons, while the inset closeup of numerical model growth

showing seven timesteps (tn to tn+6) is a small part of a model run

with mean growth rate gavg=0.1 vertical units, a=1, b=2.

Reef growth does not occur laterally because the adjacent sediment

is shale. Reef growth from timestep to timestep follows (1); note

the significant random component of growth rates, but that local

lows a single x spacing wide tend to grow faster than local highs,

as highlighted by two growth arrows from timestep tn+5 to tn+6. The

sedimentation rate (averaged over the entire model domain width)

is savg=0.05 vertical units; the sediment thickness seen in the inset

figure in each timestep is larger than this value because sediment

is confined to relatively narrow pockets (the local or absolute

topographic lows).

the noise and local relief terms. Provided that b ispositive, local lows in the reef tend to grow faster thanlocal highs (Figure 7). The rationale for this local reliefdependence is that at length scales smaller than thehorizontal grid spacing, stromatolites grow by trappingand binding sediment and sediment will be trapped in

local lows, not local highs. As described below, the localrelief effect is important to the model dynamics. The sedimentation that allows the stromatolites to growby binding and trapping is not explicitly modeled,although carbonate sediment becomes incorporated intothe reef by adjacent growth. Erosion is not allowed in



93

Figure 8. Simulated stromatolite geometries. The left-hand column shows sedimentation following the local minima rule, while the right-

hand column is absolute minimum sedimentation. Only a subset of each simulation is shown. Shale is red, carbonate sediment is blue, and

stromatolite is green. From top to bottom, the ratio of carbonate to shale sediment is infinite (100% carbonate), 2:1, 1:1, 1:2, 1:10, and 0

(all shale). Model parameters (1): gavg=0.1, a=1, b=2; sedimentation rate savg=0.05. Symbols next to each plot correspond to Figure 9.

this simple model (and we acknowledge that this doesoccur in nature), so g(x) is always forced to be >=0.

Figure 8 shows model runs with different ratios ofcarbonate sediment and shale, using both local andabsolute minimum sedimentation rules. The initialtopography is a random uncorrelated surface. Thevertical growth rate (gavg) is set to be 0.1, and thesedimentation rate savg=0.05, which ensures that moststromatolite columns during shale deposition do not getsmothered by sediment. Stromatolite width andsediment fill widths were calculated from thesesimulations by calculating the distribution of contiguoussections of sediment or stromatolite for each timesteponce growth had approximately equilibrated withsedimentation. Figure 9a shows average stromatolite andsediment fill widths for the models shown in Figure 8.

The numerical model produces results that areconsistent with the two main field observations: the ratioof stromatolite width to sediment fill width increases asthe proportion of carbonate sediment increases, and reefbed thicknesses are larger during carbonate depositionthan shale deposition. The distribution of widths is verylarge, but systematic trends are observed as a function ofsediment type, similar to the field data (Figure 9a;compare to Figure 6c). Likewise, the column to fill widthratio varies systematically with carbonate sedimentfraction in numerical simulations (Figure 9b; compare toFigure 6d). Figure 9c shows that the mean growth rateattained in the model depends on sediment type as well.These results again show a similar trend to the field dataon reef bed thickness (Figure 6b).

Discussion First-order trends in observed stromatolite growthpatterns can be recreated using a simple model ofsedimentation and stromatolite growth. The numericalmodel results can be understood in terms of geometryand the vertical growth rule in the model (1). Duringcarbonate deposition, the stromatolite facies width canadjust by lateral growth, and so inter-column fill widthstend to narrow until the sediment thickness deposited inthe inter-column trough balances the mean growth rate.During carbonate deposition the reef growth rate istherefore controlled largely by the mean vertical growthrate (gavg). The mean reef growth rate during purecarbonate deposition is actually higher than gavg,because of carbonate sediment that becomes explicitlyincorporated into the stromatolite. Column width canadjust during carbonate deposition, and so thestromatolite configuration becomes relatively insensitiveto initial topography. The column to fill width ratio oftwo for pure carbonate deposition (Figure 9b), whichcorresponds well to the field data (Figure 6c), wasadjusted in the model by setting the mean verticalgrowth rate (gavg) to be twice the sedimentation rate(savg), leading to stromatolite columns twice as wide, onaverage, as the sediment fill.

Local vertical growth does not explicitly depend onsediment type, but the local relief term in (1) causes an



Figure 9 (a) Mean stromatolite versus sediment width for model

simulations, unitless. Error bars are 1" for the distribution of

widths. A broad distribution of widths is found for each simulation,

but the mean values show a trend with sediment type. Solid

symbols are model simulations where sediment fills local minima,

while open symbols are for sediment filling absolute minima. “s”

represents shale sediment, “c” is carbonate sediment. Symbols

represent model runs shown in Figure 8 (b) Simulated stromatolite

column to fill width ratio versus the fraction of carbonate sediment.

Error bars 1". (c) Simulated growth rate depends on the fraction

of carbonate sediment. The local mean growth rate parameter was

set to 0.1, while the sediment rate was 0.05. Sediment type

influences vertical growth rates indirectly in the simulations

through the local topography factor. Mean reef growth, plotted

here, is the average thickness of both stromatolite and sediment in

the last timestep.

implicit dependence on sediment type (Figure 9c).Because sediment adjacent to the stromatolite margins istopographically lower (otherwise the stromatolite wouldbe smothered), the local relief term in (1) retards verticalgrowth at the stromatolite edge. This edge retardationcauses the convex morphology that is ubiquitous for theupper surface of stromatolites in general, and which isseen most clearly in the lower right plot of Figure 8.During shale sedimentation, reef growth rates thereforetend to be explicitly tied to the rate of shale deposition(Figure 9c). During shale deposition, column locationsand maximum widths are set early on since lateralgrowth does not occur, and the column to fill ratiodepends strongly on the rate of sediment deposition aslong as the local vertical growth rate calculated in eachtime step depends on local topography (b>0).

Simple geometric models like this one can be usefulfor generating insight into complex patterns observed inthe field. However, this numerical model does notrecreate all observations. Unlike the simulation results,field observations suggest that shale deposition oftencorrelates with a change in reef structure from well-defined columns back to crinkly, stratiform laminations.In the model, shale deposition restricts growth to onlyvertical columns. The branching trend in Figure 8 is arelatively poor match to the columnar stromatolitesobserved in our field area, although the modelbranching is perhaps reasonable for stromatolites ingeneral. Additionally, an overall trend in the studied reefnot addressed by the model is the increase in columnand fills widths and bed thicknesses from the lower tothe higher layers of the reef (Figures 5, 6a, 6c). Theincrease in widths could reflect an overall trend ofincreasing accommodation space, perhaps generated byan increase in water depth, superimposed on theparasequence scale variations in accommodation space.

Overall, the fundamental limitation of this work isthat sediment type cannot be isolated from othervariables that could also cause the observed trends instromatolite growth. Specifically, we cannot separate theeffects of accommodation driven changes insedimentation (e.g. water depth increase leads to shaledeposition, which modifies stromatolite geometry) fromflux driven variations in accommodation (e.g. climatechange diminishes carbonate deposition, shaleaccumulates as background sediment, which modifiesstromatolite geometry) The best way to isolate theeffects of these variables may be to combine morerigorous process-based numerical models of stromatolitegrowth with careful field observations of more regional,dip-dependent stratal relationships. The importance of sediment type on stromatolite growth could be further analyzed through studying the petrographicmicrostructure of stromatolite laminations to show howmuch shale is incorporated into the columns and howlaminations change with sediment type.

Many other significant aspects of macroscopicstromatolite morphology were not systematicallyexplored in this study. Field observations suggest that

synoptic relief of stromatolite columns also vary withsediment type and perhaps accommodation history.Patterns of stromatolite growth, such as changes fromthe center to the margins of a reef, were not studied,although observations indicate that columns tend to benarrower at the reef margins, and wider at the centers ofreefs. Finally, additional stromatolite reefs in otherlocations should be studied, to see if the observationswe present from one stromatolite reef are universal.

ConclusionsIn the realm of biological interactions stromatolitesrepresent a simple system of accretion, and localenvironmental conditions likely correlate to patterns ofstromatolite growth in predictable and diagnostic ways.We quantify, for a well exposed stromatolite-thrombolitereef, how changes in the thickness of reef beds andstromatolite column width and spacing correlate withsediment type. A simple numerical illustrationdemonstrates that the basic field relations can beexplained by the dependence of stromatolite growth onsediment type, with lateral growth inhibited by thedeposition of shale-rich sediment. Stromatolites may beuseful for interpreting specific environmental conditionsif diagnostic aspects of their growth sensitivities can beisolated.

AcknowledgementsThanks to David Fike, Abdullah Al-Habsy, Rashid Al-Hinai, Issa Al-Mazroui, Hamad Al-Shuaily, Aus Al-Tawil,and Erwin Adams for help and good times in the field.Thanks to Marianne and Rob Field for enthusiasticallyallowing work on Zebra River farm. Logistical supportwas provided by the Geological Survey of Namibia. Thiswork was supported by National Science FoundationGrants EAR-9904298 and EAR-0001018, and PetroleumDevelopment Oman. We appreciate thoughtful reviewsby K.H. Hofman and N.J. Beukes.

ReferencesAdams, E.W., Grotzinger, J.P., Watters, W.A., Schroeder, S., McCormick, D.S.

and Al-Siyabi, H.A. (2005). Digital characterization of thrombolite-stromatolite reef distribution in a carbonate ramp system (terminalProterozoic, Nama Group, Namibia). American Association of PetroleumGeologists Bulletin, 89, 1293-1318.

Batchelor, M.T., Burne, R.V., Henry, B.I. and Watt, S.D. (2000). DeterministicKPZ model for stromatolite laminae. Physica A, 282, 123-136.

Drummond, C.N. and Dugan, P.J. (1999). Self-organizing models of shallow-water carbonate accumulation. Journal of Sedimentary Research, 69, 939-946.

Edwards, D.J. and Brown, R.J. (1999). Understanding the influence ofPrecambrian crystalline basement on Upper Devonian carbonates incentral Alberta from a geophysical perspective. Bulletin of CanadianPetroleum Geology, 47, 412-438.

Feldmann, M. and McKenzie, J.A. (1998). Stromatolite-thromboliteassociations in a modern environment, Lee Stocking Island, Bahamas.Palaios, 13, 201-212.

Ferro, C.E., Droxler, A.W., Anderson, J.B. and Mucciarone, D. (1999). LateQuaternary shift of mixed siliciclastic-carbonate environments induced byglacial eustatic sea-level fluctuations in Belize, In: P. M. Harris, A.H. Sallerand J.A. Simo (Editors), Advances in carbonate sequence stratigraphy;application to reservoirs, outcrops and models, Special Publication -Society for Sedimentary Geology, 63, 385-411.



95

Grotzinger, J.P. (1989). Facies and evolution of precambrian carbonatedepositional systems: emergence of the modern platform archetype. In: P.D. Crevello, J.L. Wilson and J.F. Sarg (Editors). Controls on CarbonatePlatform and Basin Development. Society of Economic Paleontologists andMineralogists (SEPM) Special Publication, 79-106.

Grotzinger, J.P., Bowring, S.A., Saylor, B.Z. and Kaufman, A.J., (1995).Biostratigraphic and Geochronological Constraints on Early AnimalEvolution. Science, 270, 598-604.

Grotzinger, J.P. and Rothman, D.H. (1996). An abiotic model for stromatolitemorphogenesis. Nature, 383, 423-425.

Grotzinger, J.P. and Knoll, A.H. (1999). Stromatolites in Precambriancarbonates: Evolutionary mileposts or environmental dipsticks? AnnualReview of Earth and Planetary Sciences, 27, 313-358.

Grotzinger, J.P. (2000). Facies and paleoenvironmental setting ofThrombolite-Stromatolite Reefs, Terminal Proterozoic Nama Group (ca. 550-543 Ma), central and southern Namibia. Communications of theGeological Survey of Namibia, 12, 221-233.

Grotzinger, J.P. and James, N.P. (2000). Precambrian carbonates; evolution ofunderstanding. In: J.P. Grotzinger and N.P. James (Editors). CarbonateSedimentation and Diagenesis in the Evolving Precambrian World, (SpecialPublication - Society for Sedimentary Geology, 67, 3-20.

James, N.P. and Bourque, P. (1994). Reefs and Mounds, In: R.G. Walker andN.P. James (Editors), Facies models: response to sea level change,Geological Association of Canada, St Johns, Newfoundland, Canada, 323-347.

Kennard, J. M. and N. P. James (1986). Thrombolites and Stromatolites: TwoDistinct Types of Microbial Structures. Palaios, 1, 492-503.

Macintyre, I.G., Prufert-Bebout, L. and Reid, R.P. (2000). The role ofendolithic cyanobacteria in the formation of lithified laminae in Bahamianstromatolites. Sedimentology, 47, 915-921.

Narbonne, G.M. and James, N. P. (1996). Mesoproterozoic deep-water reefsfrom Borden peninsula, Arctic Canada. Sedimentology, 43, 827-848.

Reid, R.P., Visscher, P.T., Decho, A.W., Stolz, J.F., Bebout, B.M., Dupraz, C.,Macintyre, L.G., Paerl, H.W., Pinckney, J.L., Prufert-Bebout, L., Steppe, T. F.and DesMarais, D.J. (2000). The role of microbes in accretion, laminationand early lithification of modern marine stromatolites. Nature, 406, 989-992.

Riding, R. (2000). Microbial carbonates: the geological record of calcifiedbacterial-algal mats and biofilms. Sedimentology, 47, 179-214.

Saylor, B.Z., Grotzinger, J.P. and Germs, G.J.B. (1995). Sequence stratigraphyand sedimentology of the Neoproterozoic Kuibis and SchwarzrandSubgroups (Nama Group), southwestern Namibia. Precambrian Research,73, 153-171.

Semikhatov, M.A. and Raaben, M.E. (2000). Proterozoic stromatolitetaxonomy and biostratigraphy, In: R. Riding and S. M. Awramik (Editors),Microbial Sediments, Springer-Verlag, Heidelberg, Germany, 295-306.

Turner, E.C., James, N.P. and Narbonne, G. M. (1997). Growth dynamics ofNeoproterozoic calcimicrobial reefs, Mackenzie Mountains, NorthwestCanada. Journal of Sedimentary Research, 67, 437-450.

Wood, R. and Oppenheimer, C. (2000). Spur and groove morphology from aLate Devonian reef. Sedimentary Geology, 133, 185-193.

Editorial handling: J. Gutzmer



af fect of sedimentation on str omatolite reef gr o wth...

Documents