DISCUSSION

This study assesses the effects of environmental variables on siliceous sinter deposition in the acidic waters of the Parariki study site. The restriction of different types of sinter morphologies, textures and associated microorganisms with respect to different microfacies shows that local environmental factors are of profound influence. Furthermore, a dynamic interplay between abiotic and biotic factors plays a major role in forming sinter textures. In the following discussion we evaluate the significance of different environmental variables (both abiotic and biotic) on the genesis of the Parariki sinters. Where appropriate, we also draw comparison with previous studies of acid-derived siliceous hot spring deposits.

Subaerial Silica Precipitation and the Role of Microorganisms

Abiotic precipitation of silica was shown previously to occur where silica-rich waters wet subaerial substrates, resulting in increased silica oversaturation due to cooling and evaporation (e.g., Weed 1889; Krauskopf 1956; Walter et al. 1972; Rimstidt and Cole 1983; Hinman and Lindstrom 1996; Renaut et al. 1998; Campbell et al. 2002; Mountain et al. 2003). In the Parariki outflows, monomeric silica polymerises and deposits opal-A as the water cools and evaporates (Teece 2000). However, the acid pH at the site slows silica from precipitating subaqueously by inhibiting monomeric silicic acid from deprotonating, polymerising, and nucleating (cf. Iler 1979; Makrides et al. 1980; Weres et al. 1981; Mountain et al. 2003). This inhibition explains both the monomeric form of dissolved silica in the Parariki outflows and the lack of subaqueous silica deposition. The low propensity for silicic acid to polymerise in acid environments is responsible for the small size (<100-500 nm) of the opal-A spheres that comprise the Parariki sinters (cf. Iler 1979). However, monomeric silica may also deposit directly onto the subaerial substratum. Direct monomeric deposition, concomitant with precipitation of small silica spheres, can result in the formation of the massive vitreous silica texture observed in the cross-sections of all sinters examined in this study (cf. White et al. 1956; Rimstidt and Cole 1983; Handley et al. 2005).

While diatoms take up silicic acid from the water to form their tests, prokaryotes are not known to actively precipitate silica by metabolic activity (Mountain et al. 2003; Konhauser et al. 2004). In this study, it appears that the silicification of microbes and their associated EPS is a passive process. Silicification often does not proceed evenly on the biofilms of Parariki sinters, but is concentrated in places that are more prone to cooling and evaporation, such as the upper portions of microcolonies and spicules (Figure 6) (cf. Jones and Renaut 1997; Lowe and Braunstein 2003). Recent experiments have shown that silica precipitation is affected by its concentration but is independent of microbial growth (Toporski et al. 2002; Yee et al. 2003; Benning et al. 2004). Nevertheless, microbial cell surfaces at the Parariki site seem to act as favourable nucleation sites for silica precipitation and polymerisation. Silica shows an affinity with functional groups on proteins and polysaccharides of cell walls and EPS (e.g., Schultze-Lam et al. 1995; Westall et al. 1995; Konhauser and Ferris 1996; Renaut et al. 1998; Farmer 1999; Asada and Tazaki 2001; Konhauser et al. 2001). However, microbial silicification in acidic conditions may differ from those in near-neutral to alkaline waters. Acidophilic microbes, for example, may act as reactive interfaces that promote silica nucleation and enhance precipitation kinetics (Fortin and Beveridge 1997). Furthermore, the cell walls of some acidophiles are considered to adsorb hydrogen ions to their surfaces, forming a static barrier against proton (H+) influx into cell interiors (Gimmler et al. 1989; Asada and Tazaki 2001). Asada and Tazaki (2001) suggested that highly reactive silica ions could be generated in acidic hot springs when adsorbed hydrogen ions combine with monomeric silicic acid, promoting silica polymerisation if a steady supply of silicic acid is provided.

Microbial silicification at the Parariki site is also important for the textural development of these sinters. Mountain et al. (2003) noted that the small silica sphere sizes occurring in low pH waters can produce a dense silica matrix that accurately preserves microbial matter. In this study, small sphere sizes resulted in the intricate preservation of microbial cell outlines and their EPS (Figure 6, Figure 7, and Figure 12). The precipitation of opal-A spheres on diatom tests and their eventual cementation at the Parariki site shows that these tests can act as sites upon which silica precipitates (cf. Campbell et al. 2004; Jones et al. 2000). Indeed, microbial silicification occurs on all morphotypes at the Parariki site, although species-specific patterns of silicification have been noted elsewhere (e.g., Francis et al. 1978; Westall et al. 1995; Toporski et al. 2002; Lalonde et al. 2005). The tendency of silica to deposit on the margins of diatoms and EPS (Figure 18.2) is likely due to the higher energy encountered at these surfaces (Banfield and Hamers 1997). Atoms on edges have lower coordination and strongly asymmetric bonding configurations (Banfield and Hamers 1997), thereby allowing preferential deposition of silica. The patchy nature of microbial silicification on Parariki sinters (Figure 16.2) may be induced abiotically, or by continuous microbial growth and cell division. The latter would enable microbial populations to survive during constant bathing by silica-rich thermal outflow.

Subaerial Constraints on Sinter Dimensions and Morphogenesis

The formation mechanisms and morphology of Parariki sinters are primarily the products of substrate shape and size and the specific environment. Silica-rich water reaches subaerial substrates in the outflows through water level changes, wave wash and likely through capillary creep. However, these mechanisms are microfacies dependent, and do not occur everywhere at the study site.

In Microfacies 1-3, fluctuating water levels enable thermal outflows to reach subaerial portions of various substrates. Subsequent cooling and evaporation ensue, allowing dissolved silica to become oversaturated and deposit as continuous layers. Wave wash, caused by the pulsating discharge of thermal water, produces a similar effect. However, this mechanism is largely confined to Microfacies 1, where relatively vigorous vent discharge occurs.

Substrate width and shape governs the extent of siliceous covering. On substrates that are relatively wide and higher above water, silica reaches only the margins, producing a ring-like structure. Continuous accumulation of silica between spicules can form a rim (cf. Handley 2004), which, if high enough, blocks further silica deposition inwards. Smaller substrates (e.g., pumice pebbles), in turn, allow silica-rich water to reach surfaces evenly, resulting in sinters that cover the entire original surface. A similar process was suggested to form thrombolites at Lake Clifton, Western Australia (Moore and Burne 1994), and micro-atolls in modern corals (Stoddard and Scoffin 1979). In these two cases, upward growth is constrained by local water level, and subaerial exposure of the upper surfaces restricts growth to the margins (Stoddard and Scoffin 1979; Moore and Burne 1994).

The local energy of the thermal waters surrounding substrates can also affect sinter morphology. In Microfacies 1, where relatively more turbulent conditions prevail, silica-rich water reaches greater subaerial portions of the substrates compared to area- and shape-equivalent deposits in Microfacies 2. Hence, sinters from the former exhibit a greater extent of siliceous coating than those from Microfacies 2 and Microfacies 4, where more quiescent conditions prevail. In places where thermal water availability is low due to an increasingly sandy substratum, only the margins of substrates become silicified. The resulting deposits are observed in parts of Microfacies 2 (Figure 10.4) and, especially, Microfacies 4, (Figure 17), where thermal water moves around increasingly finer grain sizes between adjacent sinter substrate deposits.

Capillary creep and/or diffusion may also play a role in subaerial sinter formation (cf. Henley 1996; Hinman and Lindstrom 1996; Renaut et al. 1998; Campbell et al. 2002; Guidry and Chafetz 2002; Mountain et al. 2003), particularly for the cup-shaped deposits of Microfacies 4. Substrates there are neither exposed to changing water levels, nor are they affected by wave wash. Thermal pore water that is present within the sandy substrate is therefore most likely drawn up towards the surface by capillary rise or diffusion. An upward rise of thermal water would explain the often near-vertical rims of these deposits and their confinement to only the small pumice substrate margins. In this scenario, the margins form convex-upward laminae (Figure 17.4). Small, calcareous stromatolitic structures of similarly low relief (<5 cm in diameter) occur on a sandy shoreline at Lake Clifton, Western Australia (Moore and Burne 1994). Even when exposed by low lake levels during summer, these stromatolites remain saturated by low salinity groundwater seepage through capillary action (Moore and Burne 1994).

In Microfacies 3, silica precipitation will be slowed by the acidic discharge that flows over the flat, parallel-laminated deposits (Figure 15.1). However, periodic exposure above the water and concomitant cooling and evaporation of remaining puddles of silica-rich water would allow for sinter to accumulate. The patchy silicification of these sinters (Figure 15.3) is likely a result of cooling and evaporation of local puddles of thermal water. However, other factors not identified in this study may also play a role in the formation of Microfacies 3 sinter.

Significance of Microbial Biofilms upon Sinter

Biofilms can afford protection from environmental extremes (Hall-Stoodley et al. 2004), which are common in an acid hot spring setting: low pH (McNeill and Hamilton 2003); metal toxicity (Teitzel and Parsek 2003); dehydration and high salinity (Le Magrex-Debar et al. 2000; Sutherland 2001); and UV exposure (Espeland and Wetzel 2001). In addition, associated EPS may prevent cell silicification by providing reactive sites for silica to bind (Lalonde et al. 2005). According to contemporary models (e.g., Stoodley et al. 2002), the formation and development of prokaryotic biofilms requires the transport of microbes to a surface and their initial attachment, followed by microcolony formation. In quiescent waters of low-shear, laminar flow, and with ideal nutrient conditions, microcolonies often resemble pillar, mushroom, or mound-like structures (e.g., Hall-Stoodley and Stoodley 2002; Stoodley et al. 2002; Hall-Stoodley et al. 2004; Purevdorj et al. 2002). These morphologies are formed by clonal division, whereby daughter cells spread outwards and upwards from the attachment surface to form cell clusters (Hall-Stoodley and Stoodley 2002; Stoodley et al. 2002). Microcolonies on the Parariki sinters also exhibit positive relief, with both bacilliform and coccoidal prokaryotic microorganisms developing vertically upright structures. At the Parariki site, lobed coccoidal and bacilli morphotypes also were confined to different microfacies settings, with the former largely restricted to Microfacies 1, and the latter observed elsewhere at the site. Similar differential distributions of microbial morphotypes occur in acid hot springs at Yellowstone (e.g., Brock 1978), Italy (Simmons and Norris 2002), and Montserrat (Burton and Norris 2000). In these studies, lobed coccoidal prokaryotic microorganisms (e.g., Sulfolobus) occur in higher temperature settings (usually >60°C), while bacilli (e.g., Thiobacillus) are found in the relatively cooler waters (>30°C).

Apart from prokaryotes, algae, particularly diatoms and members of the Cyanidiophyceae, are ubiquitous in the relatively cooler waters (52.5°C) at the Parariki site. The predominance of algal mats at the study site is consistent with previous studies of acidic hot springs in New Zealand (Brock and Brock 1971; Brock 1978; Cassie and Cooper 1989; Jones et al. 2000) and elsewhere (e.g., Brock 1973, 1978; Gross 1998; Seckbach 1998; Ferris et al. 2005; Walker et al. 2005). The role of Cyanidiophyceae-dominated biofilms in the formation of laminae is discussed below. Diatom biofilms also are an important constituent of sinters from Microfacies 2-4, where waters are cool enough for diatom survival (typically 45°C; Cassie 1989). Benthic diatoms that are adnate, or closely appressed to the substratum, like those at the Parariki site, tend to be motile (Cohn and Dispari 1994; Cohn and Weitzell 1996). Such diatoms glide up through sediments in a movement that is non-random, following distinctive sets of chemotactic and phototactic responses (Cohn and Dispari 1994; Cohn and Weitzell 1996). EPS secretion by these organisms is also used for their daily migrations across surfaces (Cohn and Weitzell 1996).

Diatoms that attach to siliceous deposits at the Parariki site prefer to inhabit areas of low microrelief, such as pits, cracks, and along small cavities (Figure 13.1-13.3). In studies conducted elsewhere, protective areas were shown to provide refugia for diatoms from grazers (e.g., Dudley and D'Antonio 1991), shield them from abrasion and drag associated with moving water (Luttenton and Rada 1986), and protect them from desiccation (Hostetter and Hoshaw 1970). The effects of grazing activity on diatoms at the Parariki site are unknown. However, the presence of numerous clumps of fractured diatom shells indicates that water is turbulent enough in places to crush diatom tests and transport them (Figure 13.5). Repeated wetting and drying of the siliceous deposits could also cause significant stress to diatoms.

While a preponderance of diatoms has been noted in other acid hot spring settings (Jones et al. 2000), no fungi were observed on Parariki sinters. This observation is in contrast to previous studies of acid hot spring deposits, where fungi are purported to be dominant (Jones et al. 1999, 2000).

Origins of Spiculose Textures

While spicules from the Parariki sinters are similar in their gross morphology to spicular geyserite around spouters (cf. Walter 1976; Braunstein and Lowe 2001; Jones and Renaut 2003; Lowe and Braunstein 2003), there are also distinctive differences between them. Spicular geyserite forms at the air-water interface where water splashes around the inner rims of springs (Walter 1976). While biofilms are present on spicular geyserite and help form porous laminae (Cady and Farmer 1996), the formation of these spicules is largely attributed to abiotic mechanisms, involving splash and spray of silica-rich waters (Walter 1976; Braunstein and Lowe 2001; Jones and Renaut 2003; Lowe and Braunstein 2003). Spicules that form this way tend to be larger at the poolward sides of rims, where greater wave activity and splash occurs, than on the landward sides (Walter 1976; Lowe and Braunstein 2003). At the Parariki site, by contrast, neither splash nor spray was observed, and spicules are progressively larger away from the water. Furthermore, the Parariki spicules are bigger where water is more quiescent, with longer spicules in Microfacies 2 (1 cm in high), than in the more turbulent Microfacies 1 (0.5 cm high) environment. In the latter, spicules occur only on the uppermost portions of the sinters, away from the thermal waters, and within cavities (Figure 5.2). Thus, the spiculose Parariki sinter textures must develop by mechanisms different from those of classic spicular geyserite.

In previous studies, biotic mechanisms for spicular sinter formation have also been proposed (Cassie and Cooper 1989; Campbell et al. 2002) and shown to be facilitated by filamentous microbes at Champagne Pool, New Zealand (Handley et al. 2005). In this study, microspicule initiation and development was observed to involve a dynamic interplay between biofilm growth and silica deposition. Vertically upright colonies of microorganisms (coccoidal and bacilliform) were observed to act as domains of positive relief that commonly became progressively silicified and subsequently recolonised (cf. Handley et al. 2005). The restriction of these microspicules to quiescent sinter portions correlates with the distribution of their macroscopic counterparts.

It was beyond the scope of this study to establish causes governing the development of the Parariki biofilm morphologies. However, biofilm development is a multifactorial process influenced by both environmental factors and genes (Hall-Stoodley et al. 2004). Previous studies have shown how environmental conditions, such as water flow, including turbulence, can potentially affect biofilm development, superseding cell-cell communications as a principal determinant of biofilm morphogenesis (Hall-Stoodley and Stoodley 2002; Purevdorj et al. 2002). The tendency for vertically upright microcolonies to form in less turbulent waters may therefore explain the restriction of the Parariki spicules to more quiescent areas. Nevertheless, environmental factors other than water turbulence, such as nutrient availability and phototrophy, may also play a role (e.g., Doemel and Brock 1974; Brock 1978; Stoodley et al. 2002; Hall-Stoodley et al. 2004). However, for spicule formation to proceed and continue, the supply of silica is likewise important. In Microfacies 4, thermal water is confined to pore spaces within the sandy substrate. Water is not turbulent there and sinters experience only minor contact with thermal water. Therefore, sinter growth rates are slowest in that setting and spicules are generally absent, although vertically upright microcolonies are present. Hence, balances between environmentally induced biofilm morphogenesis and the supply of silica-rich water are most likely the major determinants of spicule formation and growth at the Parariki site (cf. Handley et al. 2005). Fluctuations in the supply of silica would reinforce or dampen spicule growth by the deposition of successive silica laminae (Figure 11.3). Such fluctuations could be achieved by dilution of the thermal water by rain water or diurnal or seasonal variations in temperature.

The occurrence and preservation of spicular textures in fossil sinter deposits may be of importance to palaeoenvironmental analysis. However, care should be applied when interpreting textures. As noted above, spiculose sinters can occur in both quiescent and in turbulent conditions, the latter producing geyserite. Therefore, spiculose textures in ancient hot spring deposits should not be taken as the sole facies indicator of water turbulence but considered in context with other proxies (e.g., oncoids and pisoids; silicified streamers; preserved sinter terraces) (e.g., Walter et al. 1996; Campbell et al. 2001).

Morphogenesis of Ridges, Cavities, Nodules, and Isolated Remnants

Anastomosing ridges, associated cavities, nodules, and pits that occur on surfaces of Microfacies 1 sinter were seen previously on siliceous sinter (Braunstein and Lowe 2001; Lowe and Braunstein 2003) and silica residue (Cook et al. 2000; Rodgers et al. 2002, 2004). For sinter, constructive processes were inferred for their formation. Such constructive processes result when pits retain water between wetting events from geyser eruptions, while capillary action and evaporation draw water along edges, forming rims alongside these pits through concomitant deposition of silica (Lowe and Braunstein 2003, figure 20A). In cross-section, these deposits are typically composed of cavities and pseudo-cross-laminae that mark pit migration (Lowe and Braunstein 2003, figure 20B, C). Micropitted nodular sinters, or "knobs," are also suggested to form from repeated wetting and drying (Braunstein and Lowe 2001, figure 12B).

The formation of ridges, cavities, and nodules on Microfacies 1 sinter at the Parariki site may involve a similar component of construction. However, destructive processes in acid settings should not be discounted. Ridge formation does not necessarily require repeated wetting events by silica-rich water, as similar irregular ridges have been observed to form in sinter buried in humate-rich soil that was exposed to rain water (B.Y. Lynne, personal commun., 2004). Nodules, similar in shape to those in the Parariki sinters of Microfacies 1, also occur on siliceous coverings of pumice on steaming ground at Rotokawa (R. Schinteie, unpublished data).

Anastomosing ridges on silica residue are interpreted to be destructive remnants of surfaces etched by dissolution (Cook et al. 2000; Rodgers et al. 2002, 2004). Alongside these ridges, layers of once-cemented opal-A microspheres become exposed, while depressions between the ridges are lined by irregular clusters of silica spheres at different stages of dissolution (cf. Rodgers et al. 2002, figure 8). Corrosive activity could likewise produce the gnarly texture (Figure 5.4), necking (Figure 8.6, Figure 9.2, 9.6), and the dominance of larger ridges on the outermost sides of isolated remnants (Figure 9.4, 9.6) on Microfacies 1 sinter. Corrosive attack focused around edges, kink sites, or necks would likely occur due to the lower coordination number and hence lower energy encountered in these locations (cf. Banfield and Hamers 1997).

Cross-sections of Parariki sinters, seen in thin section and under SEM, reveal that they are massive in texture, with neither open cavities nor pseudo-cross-laminae present. Therefore, the ridge and cavity morphology is only a surface feature. A later-stage deposition of silica would bury these irregular surface textures.

Steam condensate, made acid by the oxidation of H2S, is often cited as the cause of corrosion of silica residue or sinter surfaces close to vents or steaming ground (e.g., Rodgers et al. 2002; Jones and Renaut 2003; Lynne and Campbell 2004). Complexing of silica by sulphate has been suggested to increase amorphous silica solubilities in aqueous Na2SO4 solutions (Marshall and Chen 1982; Fournier and Marshall 1983). Acid steam condensate may therefore explain the restriction of corrosive-like textures to predominantly Microfacies 1 sinter where copious steam emission occurs. However, further research is warranted.

Changes in Silica Sphere Morphology

The ill-defined silica-spheres (Figure 8.1-8.2) in Microfacies 1 sinter superficially appear as if they were covered by microbially induced, mucus-like substances. However, the pervasive nature of this texture and the lack of association with microorganisms may call for alternative explanations. Similarly ill-defined sphere shapes have been observed on acid-formed sinters elsewhere and attributed to repeated episodes of silica dissolution and redeposition, causing a blurring of particle detail (Rodgers et al. 2004). Silica solubility is greater on convex surfaces than on concave surfaces (Iler 1979). Hence, silica spheres become obscure in form as silica dissolves from the upper convex surfaces and redeposits on the concave surfaces, where solubility is lower. This process forms "interparticle necks." Rodgers et al. (2004) suggested that changes in microchemical conditions were responsible for the episodic nature of silica dissolution and redeposition in acid water derived sinters. Similarly, near fumaroles, indistinct opal-A spheres form closely packed aggregates (Lynne and Campbell 2004). In silica residue, gelatinous, ill-defined spheres (also referred to as "frog spawn" texture) form and are purported to be the result of progressive strengthening of interparticle bonds at the contact between adjacent silica spheres (Rodgers et al. 2002, 2004).

Origins of Sinter Laminae

While laminae in Parariki sinters of all microfacies were visible in hand sample and/or in thin section, they were not observed under SEM, where vertical sections mostly revealed a massive, vitreous sinter body (e.g., Figure 16.3). However, the irregular, gnarled surfaces on Microfacies 1 sinter exhibited layering under SEM (Figure 9.5), while for Microfacies 2 deposits, a succession of silica layers was observed only at the surface (Figure 13.3). Jones and Renault (2004) related indistinctive laminae in other sinters to differences in the water contents of opal-A. Since these differences are texturally featureless, they are not detected by standard SEM methods. However, etching of these surfaces, attributed to acidic steam condensate, reveals alternating laminae caused by differential dissolution of opal-A due to local differences in silica solubility (Jones and Renaut 2003, 2004). Jones and Renaut (2004) interpreted "wet" opal layers to be formed by rapid precipitation of hydrated silica. "Dry" opal, in turn, would form by slow evaporation, resulting in layers of silica with less water. Since the Parariki sinters also form by repeated wetting and drying, a continuous laminated buildup would likewise develop.

Lower portions of sinter from Microfacies 2, by contrast, consist of laminae that are both abiotic and biotic in nature. Green, algal-dominated mats cover a large portion of these sinters close to the air-water interface (Figure 10), and become silicified and incorporated into the sinter (Figure 14). Studies of the taxonomically related alga Cyanidium caldarium, suggest that silicification of its mats result by the continuous proliferation of non-motile algal cells, with parental cells underlying younger cells (Asada and Tazaki 1999). Therefore, older cells will face shortages of light and CO2 for photosynthesis, and O2 for respiration. Asada and Tazaki (1999) suggested that these stresses impair the ability of the algal cells to regulate silica on their walls, so that silica continuously grows on them and progressively fills the interstices between silica crusts of different cells.

The restriction of the green Parariki algal mats to the lower portions of Microfacies 2 sinters (Figure 10 and Figure 11.5) could be temperature controlled, although moisture and nutrient supply may also influence their distributions. Brock (1978) found C. caldarium cells growing in high densities at ~45°C throughout its range (55-20°C). Indeed, the green mats from our study are thickest (1 mm) close to the water level at Microfacies 2 (50-40°C), and gradually disappear closer to the upper portions of the sinters. On the higher substrate areas of Microfacies 2 sinters, where green mats are absent, silica layering occurs only as fine laminae in thin-section and as a succession of silica horizons under SEM (Figure 13.3, red arrows), with no apparent microbial association. Continuous deposition of silica by wetting and drying appears to be the principal method of forming these abiotic laminae in the upper areas.

Stromatolitic Nature of Parariki Sinters

The sinters forming at the Parariki study site are laminated growth structures, morphologically similar in appearance to stromatolites described around other hot springs. However, the criteria used to define stromatolites are not straightforward because the search for a clear and widely accepted definition of them has proven controversial (Ginsburg 1991; Cady et al. 2003). Generally, most authors define stromatolites as layered organosedimentary structures, formed by the trapping, binding, and/or precipitation of sediments as a result of the growth and metabolic activity of microorganisms (Walter 1976; Krumbein 1983). However, the lack of fossilised microbes in many ancient stromatolites, or the potential for stromatolitic laminae to be formed entirely by abiotic means (Grotzinger and Rothman 1996) has caused problems with this genetic definition.

At the Parariki site, sinter laminae and textures such as spicules formed as a result of a combination of abiotic and biotic factors. However, taphonomic constraints can affect the differential preservation of microorganisms. At the Parariki site, ubiquitous microbial mats of prokaryote-sized microorganisms cover sinters from all microfacies. Nevertheless, preservation of these microbes as distinct, silicified laminae is often absent or confined to spicules. By contrast, algal cells, including diatoms, are much better preserved in Parariki sinter laminae. This differential preservation would bias the potential fossil record in assuming that these sinters were once largely colonised by eukaryotes. Therefore, a descriptive and non-genetic definition of stromatolites has been adopted herein, whereby a stromatolite is described as "an attached, laminated, lithified sedimentary growth structure, accretionary away from a point of limited surface of initiation" (Semikhatov et al., 1979).