|
|
 |
|||||||||||||||||
|
|||||||||||||||||
 |
|||||||||||||||||
| JOURNAL HOME | HELP | CONTACT PUBLISHER | SUBSCRIBE | ARCHIVE | SEARCH | TABLE OF CONTENTS |
SEDIMENTARY GEOLOGY-SILICLASTICS |
,11 CeREES (Centre for Research into Earth Energy Systems), Department of Earth Sciences, Durham University, Science Laboratories, South Road, Durham, DH1 3LE, UK
ABSTRACT
The conversion of biogenic silica (opal-A) to opal-CT (cristobalite and tridymite) in biosiliceous sediment causes increased rates of water expulsion because of the reduction in sediment porosity and dehydration of the amorphous opal-A phase. This release of water occurs over large tracts of sedimentary basins during sediment burial within discrete, diagenetic, reaction zones. Analysis of two-dimensional and three-dimensional seismic data sets from basins in the Northern Hemisphere provides geophysical evidence for a variety of fluid conduits and roughly circular erosional depressions at the contemporaneous seabed. We interpret these features as indicative of water expulsion and focused fluid flow emanating from opal-A to opal-CT reaction zones at burial depths within the range 200–800 m.
The rate at which water is expelled depends upon the degree of porosity reduction and the weight fraction of bound water at the reaction zone as well as the rate of advance of the reaction zone. Where the reaction is actively taking place within homogeneous biosiliceous sediment, the rate of water expulsion is independent of the reaction rate. This is because water is released across the entire reaction zone; therefore, slow reaction rates are compensated for by expulsion of water across wider reaction zones. We calculate the rate and volume of water expulsion for the Faeroe-Shetland Basin, where the sediment immediately below the reaction zone contains, on average, 
30% opal-CT by weight. The estimated volumetric rate of water expulsion per unit surface area at the present day is 6 m3 My–1 per square meter, which is greater than the vertical flux of water at the same depth from compaction of the deeper basin fill. The average volumetric rate of water expulsion is 120 km3 My–1 across the whole basin. Biogenic silica is particularly rich in Neogene successions in high latitude and equatorial regions, and where silica reaction zones are identified, they should be factored into sediment compaction and fluid-flow histories.
Key Words: reaction zone • opal-A • opal-CT • fluid flow • diagenesis
Fluid flow is a complex and poorly understood aspect of sedimentary basin evolution (e.g., Bitzer et al., 2001). It controls the transport of heat and solutes, the rate and extent of chemical reactions, the rheological behavior of rocks, and the accumulation of oil and gas. In sedimentary basins, it is driven by the compaction of sediment and sedimentary rocks, topography, and hydrocarbon buoyancy.
In clastic successions, the effect of compaction, due to both mechanical and chemical processes, is generally considered to be a gradual change in sediment properties during burial, with compaction rates being well constrained by sampling from scientific ocean drilling sites (e.g., Giles, 1997). However, in successions that contain sediment rich in biogenic silica, the basin fill passes through two silica-phase transitions during burial, from opal-A to opal-CT and then to quartz, and the reaction zones can cover substantial areas of sedimentary basins (up to 105 km2). Within the shallower transition zone, the compaction process is significantly accelerated as silica undergoes a phase change from opal-A to opal-CT. The reaction causes rapid compaction as a result of the dissolution and dehydration of silica and its reprecipitation in a denser and more compact form (Tada, 1991). Consequently, one would expect water to be expelled at higher rates than would be the case for mechanical compaction of non-biosiliceous lithologies. The release of water is well supported by the study of outcrops of pure biogenic oozes (diatomites) that have undergone conversion to porcelanite and chert. These outcrops show evidence for water expulsion and fluid flow (Eichhubl and Behl, 1998; Eichhubl and Boles, 1998; 2000) in the form of multiple fracture populations, brecciation, and cementation within fractures and faults.
Two-dimensional and three-dimensional seismic reflection data have provided a new tool for the study of diagenetic processes within the upper crust. These data have proved particularly useful for the study of silica reactions during early burial because they cause a significant change in acoustic impedance and are often very well imaged. Interpretation of the data is leading to better understanding of the role that silica phase changes play in the evolution of sedimentary basin fill. For instance, the opal-A to opal-CT transformation may cause soft sediment deformation and clastic injectites (Davies et al., 2006), slope failures (Davies and Clark, 2006), and differential compaction folds (Davies, 2005). The underlying process responsible for the development of these phenomena is the release of water during the advance of the opal-A to opal-CT reaction zone (Fig. 1). In this paper, we use seismic data from high-latitude regions in the Northern Hemisphere to analyze the evidence for focused fluid flow at kilometer scales that has resulted from opal-A to opal-CT transformation. We consider the factors that control the rate of water expulsion and outline a method for estimating it, which is then applied to the Faeroe-Shetland Basin, northwest of the United Kingdom.
|
The crystalline structure of opal-A and opal-CT, as well as the terms themselves, were first properly defined by Jones and Segnit (1971), who subdivided these natural hydrous silicas into three, well-defined structural groups—opal-C (well-ordered 
-cristobalite), opal-CT (disordered -cristobalite, -tridymite), and opal-A (highly disordered, near amorphous). From work on the Monterey Formation in California, Bramlette (1946) had already concluded that these different forms of silica are related to each other and occur due to diagenesis during sediment burial. He also showed that diatomaceous rocks were common at the top of the formation, whereas porcelaneous and cherty rocks were progressively more common toward the base of the formation.
Later research demonstrates that the diagenetic change in silica from opal-A to opal-CT occurs by a dissolution-reprecipitation reaction (e.g., Wise et al., 1972) and because both phases of silica are metastable states. Temperature increases the rate of the reaction (Williams and Crerar, 1985), and sediment age is a significant factor, with the reaction occurring at lower temperatures in older successions (Hein et al., 1978; Tada, 1991). After increased burial, opal-CT is then converted to quartz. The reaction zones are usually a few tens of meters thick but can reach thicknesses of 300 m (Isaacs, 1981; Keller and Isaacs, 1985; Nobes et al., 1992).
Host-sediment surface area, sediment composition, and pore-water chemistry also influence reaction rates (Isaacs, 1981; Hinman, 1990; Tada, 1991; Nobes et al., 1992). The presence of clay and organic matter can inhibit the conversion process (Kastner et al., 1977; Isaacs, 1982), whereas calcium carbonate can increase the rate of opal-CT nucleation (Keene, 1975; Kastner et al., 1977). In the presence of dissolved organic matter, the reaction rate for opal-A to opal-CT conversion is reduced (Hinman, 1990). Pure diatomite converts earlier than less pure sediment (Isaacs, 1981; Behl and Garrison, 1994); therefore, the temperature required for conversion in different host lithologies of similar age is variable (Hein et al., 1978; Bohrmann et al., 1994). In very young sediments, conversion takes place at temperatures of 40 °C to 55 °C, whereas in sediments of age ca. 50 Ma, conversion takes place at temperatures of 10 °C to 20 °C (Tada, 1991).
Where the temperature gradient in a basin is vertical, opal-A in diatomaceous mud of a particular composition begins to convert to opal-CT at a specific burial depth known as the depth of conversion (DOC). In heterogeneous successions containing beds with differing concentrations of opal-A, conversion to opal-CT takes place at different depths, and therefore the reaction zone is complex. Because the conversion causes an increase in sediment density and compressional-wave velocity, the reaction zone is commonly imaged on seismic data as a moderate- to high-amplitude reflection with the same polarity as the seabed (Fig. 1).
During conversion of opal-A to opal-CT, 1.5–15.3 wt % of structurally bound water is released (Jones and Renaut, 2004). Sediment rich in opal-A maintains anomalously high porosity during burial because the frustules of silica provide a rigid framework that is resistant to mechanical compaction processes. It is therefore possible to have a high porosity of 75% at 500 m burial depth (Volpi et al., 2003). Porosity is reduced after conversion to values of 45% (Tada, 1991) because pore spaces are either filled by microcrystalline silica or because they are destroyed as a result of the dissolution and reprecipitation processes (Fig. 1). The reduction in sediment porosity can be sharp or gradual (Chaika and Williams, 2001) and necessarily involves expulsion of pore water (Isaacs, 1981).
The seismic examples that show evidence for water expulsion at silica reaction zones are taken from two- and three-dimensional seismic data sets that were acquired in the Faeroe-Shetland Basin and in the North Sakhalin Basin, offshore of Sakhalin Island. We examine data from within the first 800 m of burial, where vertical seismic resolution is 10–30 m. The opal-A to opal-CT reaction zones are identified on the basis of (a) sampling of hydrocarbon and scientific boreholes, (b) clear cross-cutting relationships with stratal reflections, and (c) an increase in acoustic impedance. Other details of the typical geophysical character are provided by Davies et al. (1999), Davies and Cartwright (2002), Volpi et al. (2003), Davies (2005), Davies et al. (2006), and Davies and Clark (2006). Unfortunately, very limited data are available from commercial boreholes in both regions; what sparse information that does exist for the Faeroe Shetland Basin is summarized later in this paper.
Isolated Giant Depressions
The opal-A to opal-CT transition in the Faeroe-Shetland Basin (Fig. 2A) is hosted within Oligocene to Miocene strata and has been drilled by well 214/4-1 (Davies and Cartwright, 2002). There are a number of isolated, roughly circular, buried depressions 200–400 m above the reaction zone at the level of an Early Pliocene unconformity (Fig. 2B). They are 100–800 m wide and 100 m deep, and are filled with high-amplitude, continuous seismic reflections (Fig. 2C). Some reflection events have been truncated around the margins of the depressions (marked T in Fig. 2C). The depressions are linked to the opal-A to opal-CT reaction zone via discontinuous or chaotic seismic reflections. Each zone of discontinuous seismic reflections has a roughly columnar form. Similar features have been described within the same three-dimensional seismic data set by Davies and Clark (2006).
|
|
|
Several phenomena described in the preceding section need to be accounted for: the buried depressions and kilometer-scale columnar zones of chaotic seismic reflections; faults linking the reaction zone to the contemporaneous seabed, where they are marked by trails of amplitude anomalies; and downward-tapering collapse structures. In the case of the Faeroe-Shetland Basin, we should also account for the fact that the depressions are buried by sediment of Early Pliocene to Holocene age, with much less evidence for fluid venting within this succession or on the present seabed. We start with the seismic expression of fluid flow at the seabed and then consider the conduits that fed these seabed features.
Seabed Evidence for Fluid Venting
The depressions in the Early Pliocene unconformity surface in the Faeroe-Shetland Basin (Fig. 2) are roughly circular and were probably erosional, based upon the identification of truncated seismic reflections around their margins. Their shape rules out erosion by deepwater contourite bottom currents, as described elsewhere in this basin (Smallwood, 2004). We interpret them to be large pockmarks, formed as a result of fluid seepage out of the seabed, where soft sediment is entrained at the sediment-water interface (Hovland and Judd, 1988). Similar, but even larger, fluid-escape craters have been recognized previously using three-dimensional seismic databases (e.g., Cole et al., 2000).
The small, roughly circular, high-amplitude anomalies at the Early Pliocene unconformity (Fig. 3) are more problematic because distinctive internal structure is not resolved. The polarity of the reflections indicates that the circular features are infilled by sediment of a lower acoustic impedance than the surrounding fine-grained sediment. They track faults, and the faults have no obvious dominant orientation, indicative of a nontectonic origin. We propose they formed due to volumetric reduction during the diagenesis, and that they were pathways for fluid migration. The change in seismic amplitude is consistent with a change in lithology, which could be the result of: (a) diagenetic change in the sediment at this level, (b) the development of a biological community at a seep, or (c) the formation of small sediment volcanoes.
The depressions interpreted as large pockmarks and the tracks of amplitude anomalies occur at an unconformity overlain by sediment of Early Pliocene to Holocene age (Fig. 2B–C). There is only very sparse evidence for fluid vents on the present-day seabed and pipe-like conduits within the Early Pliocene to Holocene succession. We interpret this evidence to indicate that water expulsion at the opal-A to opal-CT reaction zone was greatly enhanced during the Early Pliocene. Davies and Cartwright (2002) observed that there is a tendency for the opal-A to opal-CT reaction zone to be parallel to the Early Pliocene unconformity, rather than to the present-day seabed. They suggested that the reaction zone underwent a pronounced phase of advance during the Early Pliocene and the advance subsequently stopped, and that, therefore, vents occur on the Early Pliocene unconformity.
Fluid Conduits
The amplitude anomalies that track faults (Fig. 3C) suggest that fluid flow occurred up the faults in the Faeroe-Shetland basin, from the level of the opal-A to opal-CT reaction zone to the contemporaneous seabed. Since some of the anomalies occur as isolated, discrete features, fluid flow was probably focused within pipe-like conduits, at least immediately below the seabed.
We also find evidence for vertical fluid flow without there being any clear faults. This is not uncommon: for example, vertical pipes only a few tens of meters wide and up to a kilometer in vertical extent have been documented by Loseth et al. (2001). Water flows can entrain sediment, causing volume loss at the level of the remobilized strata, resulting in localized collapse of the overburden. The funnel-shaped collapse features identified in the Sakhalin Basin (Fig. 4) are similar to those described by Davies (2003), who considered them to have formed by this mechanism following overpressure buildup and seal breach.
The funnel-shaped collapse features in the Sakhalin Basin, without clear faults, are consistent with overpressure generation at the reaction zone. When pore pressure becomes greater than the sum of the minimum principal stress and the tensile strength of the sediment, hydrofracturing occurs (Engelder, 1993). Fluid flow along the faults, or the newly established conduit systems, would have occurred until fluid pressure dropped to hydrostatic pressure or until fractures closed or filled with diagenetic cements.
WATER EXPULSION AT THE REACTION ZONE
When opal-A converts to opal-CT, the reduction in sediment porosity can be used to give a rough estimate of the volume of water expelled. The higher the opal-A content of a sediment, the greater is the porosity reduction on conversion (Compton, 1991). Scientific boreholes and outcrop sampling provide an empirical database for porosity reduction across opal-A to opal-CT reaction zones, and two example data sets are shown in Figure 5. These data reveal a step in the porosity-versus-depth trend. Porosity reduction can be large, with siliceous sediment porosity dropping by as much as 30% from values as high as 75% when opal-A transforms to opal-CT (Tada, 1991).
|
1 density of opal-A–bearing sediment, 2 density of opal-CT–bearing sediment, W density of free water, 
1 porosity of opal-A–bearing sediment, and 2 porosity of opal-CT–bearing sediment. Other symbols used in the following derivation are:
First, we derive an expression for the reduction factor, R. For a unit volume of opal-A–bearing sediment, the mass of solid grains including bound water is 1 – 1 w, and the mass of the detrital component, other than opal-A, is (1 – W1)(1 – 1 w). Similarly, the mass of the detrital component in a unit volume of opal-CT–bearing sediment is (1 – W2)(2 – 2 w). We assume that the detrital constituents are unaffected when opal-A transforms to opal-CT. Then R equals the ratio between the masses of the detrital component in unit volumes of opal-A–bearing and opal-CT–bearing sediment:
|
| (1) |
The mass of water present in a unit volume of opal-A–bearing sediment as both free water in the pore space and bound water in opal-A grains is 1 w + B1W1(1 – 1 w). In this paper, we assume that any bound water present in the other solid grains is unaffected by the transformation of opal-A to opal-CT. Similarly, the mass of water present in a unit volume of opal-CT–bearing sediment as both free water in the pore space and bound water in opal-CT grains is w + B2W2(2 – 2 w) Given that a unit volume of opal-A–bearing sediment is reduced by the factor R on transformation, the volume of water expelled per unit volume of opal-A–bearing sediment is
|
| (2) |
To determine the flux of water expelled (i.e., the volume of water expelled per unit time per unit surface area of the reaction zone), VE must be multiplied by the rate of advance of the opal-A to opal-CT reaction zone through the opal-A–bearing sediment. In idealized circumstances, where the lithology of the opal-A–bearing sediment is uniform and both burial rate and geothermal gradient are constant, the rate of advance would equal the burial rate.
Application to the Faeroe-Shetland Basin
In the Faeroe-Shetland Basin, the opal-A to opal-CT reaction zone covers an area of 20,000 km2 (Davies and Cartwright, 2002). Well 214/4-1 was drilled through the reaction zone which is hosted in Oligocene strata at the well location, 500 m below seafloor, but no core was taken, and very little wireline logging was carried out through the shallow section in the well. However, some measurements of the weight fraction of opal-CT were made on cuttings that were deposited on the seafloor and collected by an ROV (remotely operated vehicle) (Fig. 6). The weight fractions of opal-CT in these cuttings were calculated using X-ray diffraction techniques and a reference intensity method, where known mixtures of minerals were analyzed. The heights of the X-ray diffraction peaks for the retrieved samples were measured and multiplied by a factor derived from measurements of the standard known mixtures. The average weight fraction of opal-CT in the eight deepest samples of cuttings, all taken from below the reaction zone, was 30% (Davies, 2005).
|
density of silica-bearing sediment, S grain density of silica, D grain density of the detrital component, porosity of silica-bearing sediment, S porosity of pure silica sediment with 100% weight fraction of silica, D porosity of mudstone with 100% weight fraction of the same lithology as the detrital component in the silica-bearing sediment,
The density of water-saturated, silica-bearing sediment is related to its porosity by
|
| (3) |
For a unit volume of the solid grains of the silica-bearing sediment, the mass of silica present is SVS, and the mass of the detrital component is S (1–VS). Hence, the weight fraction of silica present as opal-A in the dry sediment is
|
| (4) |
Rearranging, the volume fraction of silica present in the solid grains of the silica-bearing sediment is
|
| (5) |
We estimate the porosity by assuming it equals the average porosity of beds composed either of pure silica or of the detrital component, with the overall weight fraction of silica being WS. This assumption is reasonable for laminated sediments in which individual laminae are predominantly composed of either silica or detritus. We note that it may be less accurate for homogeneous sediments. The volume of solid grains of silica in a unit volume of silica-bearing sediment is (1 – )VS, and the volume of detrital solid grains is (1 – )(1 – V S). These expressions must be divided by the factors 1 – S and 1 – D, respectively, to obtain the volumes occupied by each constituent together with its associated pore space, and these volumes must sum to unity:
|
| (6) |
Rearranging, the porosity of the silica-bearing sediment is
|
| (7) |
We now use Equations (5), (7), and (3) to estimate the porosity and density for both the opal-CT–bearing sediment encountered in well 214/4-1 and the opal-A–bearing sediment from which it was derived, using the measured weight fraction of opal-CT from the cuttings in the opal-CT zone and published values for the other parameters. We then use Equations (1) and (2) to estimate the amount of water expelled on conversion of opal-A to opal-CT.
We assume porosities of 75% for pure opal-A diatomite and 45% for pure opal-CT porcelanite (Tada, 1991), and 32% for a mudstone comprising 100% weight fraction of lithology corresponding to the detrital component, which is the porosity on the Baldwin-Butler curve for normally compacted muds at 500-m depth (Baldwin and Butler, 1985). During conversion of opal-A to opal-CT, 1.5–15.3 wt % of structurally bound water is released (Jones and Renaut, 2004). The specific gravity of opal-A is 2.01–2.16 (Deer et al. 1992); thus, we assume a mid-range density of 2.09 g cm–3 for the solid grains of opal-A and a mid-range, bound-water fraction of 8%. We assign the grain density of 2.3 g cm–3 for cristobalite (Carmichael, 1984) to the solid grains of opal-CT, assuming that the bound water is negligible, a grain density of 2.65 g cm–3 to the detrital solid grains, which are siliciclastics, and density of 1.03 g cm–3 to the pore water. Assuming that the mass of silica was conserved on transformation of opal-A to opal-CT, the measured weight fraction of 30% opal-CT in the opal-CT zone corresponds to an estimated weight fraction (including bound water) of 32.6% opal-A in the opal-A–bearing sediment before conversion.
Substitution of the above values into Equations (5), (7), and (3) yields the following estimates: for the opal-CT–bearing sediment, the porosity 2 is 37%, and the density 2 is 1.98 g cm–3; and for the opal-A–bearing sediment before conversion, the porosity 1 was 59%, and the density 1 was 1.61 g cm–3. Further substitution of these porosity and density values into Equations (1) and (2) with B1 = 0.08, B2 = 0, W1 = 0.326, and W2 = 0.3 gives the volume of water expelled on conversion per unit volume of opal-A–bearing sediment that has undergone transformation as 39%.
At the well location, the shallowest cuttings sampled came from 59 m below seabed and were of Early Pliocene age. Accordingly, we can only make a very rough estimate for the burial rate through the Quaternary as 15 m My–1. This estimate gives an average volumetric rate of water expulsion per unit surface area of the reaction zone of 6 m3 My–1 per square meter, and the average volumetric rate of water expulsion due to conversion of opal-A to opal-CT is 120 km3 My–1 across the whole basin.
For comparison, we may consider the average vertical flux of water at comparable depths due to compaction of an underlying, mud-rich basin fill during burial. The calculation is done by integrating the rate of pore volume loss during ongoing burial. The Baldwin-Butler curve for normally compacted muds predicts a porosity of 32% at 500-m depth (Baldwin and Butler, 1985), and we assume an irreducible mudstone porosity of 7% (e.g., Hunt et al. 1998). For sediment below the reaction zone, the net burial rate is obtained by offsetting the compaction that takes place as opal-A converts to opal-CT against the burial rate for sediment above the reaction zone. In this example, where the net burial rate below the reaction zone is 9 m My–1, the integrated rate of pore volume loss below 500-m depth is 2.5 m3 My–1 per square meter. This estimate of the vertical flux of water due to compaction of underlying sediment is a maximum estimate because expulsion of water from compaction of deeper sediments does not necessarily keep pace with burial rate (e.g., when undercompaction is occurring), the lithology of deeper sediment is not all mudstone, and much of the fluid expelled from depocenters by compaction may take place laterally toward the flanks of basins (Moss et al., 2003). Therefore, we conclude that the average rate of water expulsion due to conversion of opal-A to opal-CT in the middle of the basin is more than double the vertical flux of water due to compaction of all the underlying sediment.
At the top of the opal-A to opal-CT reaction zone in well 214/4-1, there is a sample where the estimated percentage of opal-CT is 76%. This percentage estimation is consistent with the well-site description of the sample, as "chert" with "concoidal fracture." Based on a reduction in the rate of penetration of the drill bit, this silica-rich bed has an estimated thickness of 25 m (marked by gray band in Fig. 6). When the silica reaction zone was passing through this bed, the rate of water expulsion per square meter would have been 9 m3 My–1 per square meter locally for the same burial rate of 15 m My–1. This highly siliceous unit can be mapped on seismic data around the well, where it is seen to dip below the present position of the reaction zone (Fig. 6). There are other seismic reflections that have anomalously high amplitudes relative to the typical seismic character. Some extend above the level of the reaction zone and may represent particularly rich diatomaceous beds that are undergoing transformation at a slightly lower temperature and hence shallower DOC.
Significance of Reaction Zone Thickness
Previous research shows that the variability in host-rock composition is a primary control on the width of the reaction zone (Kastner et al., 1977; Isaacs, 1982; Keller and Isaacs, 1985). In pure diatomites with little heterogeneity, we can expect fast reaction rates and simple, narrow reaction zones. In less pure siliceous rocks that are heterogeneous, we expect slower reaction rates and more complex, wider reaction zones. We briefly consider whether reaction zone width affects the rate of fluid production.
A reaction zone can be considered as a curve that defines the number of moles opal-A being converted per unit time against depth (Fig. 7A). For lithology that is uniform before the onset of conversion, the rate of conversion increases smoothly to a maximum and then decreases to zero again, where all the available opal-A has been converted to opal-CT. A schematic example of a narrow reaction zone, where pore fluid and structurally bound fluid are expelled within a few meters of the DOC, is marked W1 in Figure 7B. A wider reaction zone is marked W2. Both reaction zones cause the same porosity reduction. For the same rate of advance of the reaction zone, the rate of water production per unit volume would be relatively high where the reaction boundary is narrow and relatively low where the reaction boundary is wide. However, whether the zone is sharp or transitional, the total rate of water production is the same in both cases because fluid is produced across the entire reaction zone. Therefore, the rate of fluid production across a wide reaction zone with slow reaction rates should be the same as the rate of fluid production across a narrow reaction zone where the reaction rate is faster. Obviously, this rule breaks down when the reaction rate becomes so low that the whole reaction zone is not contained within a bed of initially uniform lithology.
|
Fluid Escape Features
The funnel-shaped collapse features identified in the Sakhalin Basin (Fig. 4) are similar in form to the larger collapse structures, 1 km wide, that have been documented for mud volcano systems (Davies and Stewart, 2005; Stewart and Davies, 2006), and funnel-shaped geometries also occur in diatremes (e.g., Nemeth et al., 2001). In mud volcano systems, complex linked fracture and intrusive dike systems have been evoked as pathways for significant volumes of fluid containing entrained sediment to pass through great thicknesses of overburden stratigraphy. Collapse occurs because of differential compaction of the conduit system following volume loss from the source beds and by erosion of the wall rock.
Vertical, fluid-escape conduits have been demonstrated in physical model experiments (Nichols et al., 1994). Based upon the seismic examples illustrated in Figures 2–4 and experimental analogues, we conclude that the Sakhalin collapse structures and the columnar chaotic seismic zones in the Faeroe-Shetland Basin resulted from focused fluid flow, perhaps with entrained sediment, through fractures and faults from the reaction zone to the seabed.
In the siliceous Monterey Formation, the water expelled by the opal-A to opal-CT transition has generated increased pore pressure which caused hydraulic fracturing (Eichhubl and Boles, 1998). The increase in pore pressure also makes fault slip on new or preexisting faults easier. Fluid flow within the Monterey Formation occurred up new and preexisting faults, probably along pipes, rather than being dispersed along the fault strike (Eichhubl and Boles, 2000).
Overpressure
To develop such large-scale fluid conduits and pockmarks as those found in the Faeroe-Shetland Basin requires sealing lithologies to be breached as a result of fracture development and propagation. Initiation of fractures requires the pore pressure to be higher than the sum of the minimal principal stress and the tensile strength of the sediment (Engelder, 1993). Overpressure can be generated by several different mechanisms (Osborne and Swarbrick, 1997), of which the diagenesis of silica is only one. For example, disequilibrium compaction of fine-grained lithologies such as mudstones is a common mechanism of overpressure generation (Osborne and Swarbrick, 1997). However, disequilibrium compaction acting alone does not cause seal breach because it does not reduce effective stresses: it only retards the normal increase in effective stresses with increasing burial depth. Effective stresses can be reduced by fluid expansion mechanisms such as gas generation or oil cracking to gas. We can, however, be confident that the conduits and pockmarks identified in the Faeroe-Shetland Basin data set are not the result of gas generation because the main pulse of fluid appears to have been generated during the Early Pliocene just prior to unconformity development. There are no reasonable grounds for supposing that gas generation preferentially took place within such a short time period. Furthermore, there is no seismic or borehole evidence for the presence of gas hydrates; thus, gas hydrate dissociation is also an unlikely mechanism. Therefore, our preferred interpretation is that the release of water as opal-A converted to opal-CT caused overpressure development, as proposed for fractures in the Monterey Formation (Eichhubl and Boles, 1998), by transfer of the lithostatic load to the fluid.
To generate overpressure, the rate of fluid production needs to exceed the rate of fluid seepage. Since we have good constraints on the magnitude of the porosity reduction, the two main uncertainties that should be considered are the rate of water release and the permeability of the overburden.
Water Release Rates
The rate of advance of the reaction zone is a major control on the rate of water release. Given that the boundary is primarily controlled by temperature and time, for uniform sediment composition and a constant burial rate in a thermally stable tectonic regime, one could assume that the rate of advance of the reaction zone equals the burial rate. Thus, the rate of advance would be a few tens of meters per million years in passive margin settings such as the Faeroe-Shetland Basin.
Seismic data generally reveal only one opal-A to opal-CT reaction boundary reflection. However, the advance of an idealized wide or narrow reaction boundary does not take into account the likelihood that in an interbedded succession with variations in detrital content there would be multiple depths of conversion (Eichhubl and Boles, 1998) because the concentrations of opal-A in the sediment affect reaction rates (Isaacs, 1982). In reality, the rate of conversion of opal-A could vary in interbedded successions. For example, in the Monterey Formation the heterogeneity results in a conversion zone that is up to 300 m thick (Isaacs, 1981; Keller and Isaacs, 1985). In the Bering Sea, the diagenetic boundary takes the form of a large-amplitude BSR (bottom simulating reflector) on seimic-reflection profiles that correspond to a basin-wide velocity discontinuity between overlying lower velocity sediment rich in opal-A and underlying higher velocity sediment rich in opal-CT (Cooper et al., 1987). This seismic reflection merely represents the onset of the diagenetic transition and the first occurrence of sediments containing opal-CT, not the complete conversion of opal-A to opal-CT. Opal-A–bearing sediments persist below the seismic expression of the diagenetic boundary, depending on the detrital content of the sediment. Hence, the curves presented in Figure 7A–B are considerably simplified.
Permeability
No measurements of permeability are available in the sediment above the opal-A to opal-CT reaction zones offshore Sakhalin Island or in the Faeroe-Shetland Basin. However, we can estimate to an order of magnitude the maximum possible permeability that would still permit the buildup and maintenance of high overpressure. From the well information given by Davies and Cartwright (2002), it is reasonable to infer that the burial rate in the Faeroe-Shetland Basin could have been as high as 100 m My–1 during the Early Pliocene. Consequently, the rate of water release might have been as high as 40 m3 My–1 per square meter of the reaction zone. For significant overpressure to develop, an overpressure gradient of 104 Pa m–1 would have been necessary. Given that the viscosity of water is 10–3 Pa s, the permeability required to maintain this overpressure gradient is 100 nD. This permeability value is much less than that of pure diatomite, but is plausible for sediment that contains a substantial clay fraction (Bryant, 2003).
Timing of Fluid Escape
In all the examples of opal-A to opal-CT transitions studied, we have found very little evidence for fluid flow onto the present-day seabed, but instead we find ample evidence for fluid flow onto a discrete, paleoseabed surface that extends across large sections of the sedimentary basin. The seismic reflection method does not detect evidence for pervasive fluid flow that occurs due to capillary leakage, and will also not detect fractures and faults that are below seismic resolution. Therefore, the lack of seismic evidence for fluid flow does not mean that it is not occurring, but only that it is not deforming and disrupting the sediment sufficiently to create seismically observable features such as conduits and pockmarks. The large-scale conduits and pockmarks imaged by seismic data measure tens of meters in width and tens to hundreds of meters in height.
There are several potential explanations for abundant evidence of fluid flow at a specific paleoseabed surface, some of which were considered by Davies and Cartwright (2002), who suggested that the opal-A to opal-CT reaction zone in the Faeroe-Shetland Basin is fossilized. However, the present depth of the reaction zone is between 500–800 m below the seafloor, whereas in the Early Pliocene the strata that now host the reaction zone were only 200–400 m below the seafloor. If the reaction zone has been inactive from the Early Pliocene to the present day, the implication would be that the geothermal gradient was much higher in the Early Pliocene. There is some evidence to suggest that the geothermal gradient was higher in the middle Miocene, but not specifically in the Early Pliocene (Scotchman et al., 2006).
In this paper, we give a revised interpretation that is consistent with the seismic evidence and the outcrop observations reported by others. We suggest that opal-A to opal-CT conversion began at the base of the biosiliceous formation in the Faeroe-Shetland Basin during late Miocene to Early Pliocene times. Pore pressure built up during the Early Pliocene because of an increased burial rate and the low permeability of the biosiliceous mud in the overburden. Fluid release rates would have varied because of lithological variations as well as variations in burial rate. For example, the chert bed penetrated in well 214/4-1 contains a weight fraction of opal-CT that is four times greater than the more clay-rich sediment immediately beneath it, and would have started to convert earlier because of its lower clay content. Thus, the water release rate may have increased significantly when this bed was undergoing conversion. The conduits leading up to the seafloor and the pockmarks found on the Early Pliocene unconformity surface developed just prior to and during the hiatus that gave rise to the unconformity. During subsequent burial, there is some limited evidence for continued fluid flow, but the features are not on the scale of those that terminate at the unconformity.
An interesting observation made by Davies and Cartwright (2002) is that the opal-A to opal-CT reaction zone in the Faeroe-Shetland Basin tends to be more closely parallel to the Early Pliocene unconformity than to the present-day seabed. A parallel relationship with unconformity surfaces has also been identified in other sedimentary basins (Meadows and Davies, 2007). This observation contributed to the suggestion by Davies and Cartwright (2002) that the reaction zone is fossilized. If the reaction zone is active at the present day, then its position may be expected to approximate an isotherm, albeit with some deviations due to the variations in sediment composition. It is quite possible that isotherms in the Oligocene-Miocene section are more closely parallel to the Early Pliocene unconformity than the seabed because of lateral fluid flow. Andrews-Speed et al. (1984) observed regional variations in heat flow in the western North Sea Basin, which they attributed to large-scale, lateral fluid flow. The Paleocene sands in the central North Sea Basin are known to be a conduit for fluid flow from the compacting Mesozoic and Paleozoic sediments in the Central Graben, with fluid escaping to the seabed in the Outer Moray Firth (Moss et al., 2003). In the Faeroe-Shetland Basin, water released at the opal-A to opal-CT reaction zone may move laterally until it reaches a conduit where it can pass upward to the Early Pliocene unconformity. Above that horizon, it may again move preferentially in a lateral direction before escaping to the seabed on the flanks of the basin. The available evidence on geothermal gradients in the Faeroe-Shetland Basin is far too limited to evaluate this hypothesis at present.
The conduits appear to emanate from the present position of the opal-A to opal-CT transition, rather than from greater depths. Regional seismic data from offshore Sakhalin Island and the Faeroe-Shetland Basin show that this appearance cannot be due to the reaction zone reaching a particular compositional boundary, because it cross cuts stratigraphy and only locally follows stratal boundaries. We infer, instead, that evidence for deeper fluid conduit systems has been overprinted as a result of advance of the reaction zone with the associated dissolution and reprecipitation.
Associated Geologic Phenomena
A number of physical geologic processes have been linked to the water expulsion across opal-A to opal-CT diagenetic reaction zones. Eichhubl and Boles (1998) inferred that the release of water causes increased pore pressure and hydrofracturing of the transforming sediment. Davies et al. (2006) described such a process in which fluidized sediment has been emplaced into strata overlying the reaction zone in the form of giant clastic intrusions (injectites) spanning depth intervals from 50 m to more than 300 m. There are also two documented examples of submarine, continental slope failures where the detachment of the failed section occurred close to the level of the opal-A to opal-CT reaction zone. In both cases, it was inferred that the trigger for the failures was the water released by the advance of an opal-A to opal-CT reaction zone (Volpi et al., 2003; Davies and Clark, 2006) that caused shear failure.
The reaction zones can also develop a highly variable, vertical relief with cylindrical promontories, termed "cells," separated by troughs where the reaction zone is deeper (Davies and Cartwright, 2007). The cells have widths up to 2.7 km and elevations up to 200 m above the remainder of the reaction zone. These patterns cover large tracts of sedimentary basins, with cells growing and perpetuating as the reaction zone passes upward through biosiliceous sediment during ongoing burial (Fig. 8). The origin of the extraordinary cell-like pattern is uncertain, but the release of fluid during the conversion may be part of the driving mechanism (Davies and Cartwright, 2007).
|
We provide the first, strong, seismic reflection evidence to support the occurrence of focused fluid flow above opal-A to opal-CT reaction zones. This evidence indicates that pore pressure increased to a point at which natural hydrofractures formed, with fluid moving to the seabed through conduits. Alternatively, fluids moved up new or preexisting faults. Evidence for focused fluid flow only occurs in some regions of the sedimentary basins studied. Gradual, pervasive, fluid seepage, with no associated overpressure buildup and no focusing of fluid, is also likely to occur, but is not detected on seismic reflection data.
The reaction zones cover large tracts of sedimentary basins; therefore, their advance provides a mechanism for increased rates of basin-wide expulsion of fluid from a discrete level. In a uniform lithology, the volumetric rate of water release per unit area of reaction zone depends only on the porosity drop at the boundary, the weight fraction of bound water, and the rate of advance of the reaction zone. In silica-rich successions, especially where favorable lithological variations allow multiple reaction zones to develop, the rate of water release can be several times greater than the total fluid flux from compaction of deeper sediment. Given these characteristics, silica reaction zones should be factored into models of basin-fill evolution and their potential for overpressure generation should also be considered.
ACKNOWLEDGMENTS
We are grateful to Schlumberger for the use of IESX software and BP and PGS Exploration for permission to publish seismic data. TGS-NOPEC kindly provided permission to publish the seismic line in Figure 4. Discussions with Joe Cartwright and Torben Seidel are appreciated. We thank Associate Editor Peter Eichhubl for his guidance and comments. We are very grateful to Dave Scholl, Andy Aplin, Knut Bjørlykke, and two anonymous reviewers for their detailed and thought-provoking comments.
FOOTNOTES
richard.davies{at}durham.ac.uk 
REFERENCES CITED
Andrews-Speed, C.P., Oxburgh, E.R., and Cooper, B.A., 1984, Temperatures and depth-dependent heat flow in western North Sea: American Association of Petroleum Geologists Bulletin, v. 68 p. 1764-1781.[Abstract][GeoRef]
Baldwin, B., and Butler, C.O., 1985, Compaction curves: American Association of Petroleum Geologists Bulletin, v. 69 p. 622-626.[Abstract][GeoRef]
Behl, R.J., and Garrison, R.E., 1994, The origin of chert in the Monterey Formation of California (USA): in Iijima, A., ed., Siliceous, phosphatic and glauconitic sediments of the Tertiary and Mesozoic: Proceedings of the 29th International Geological Congress, Part C, p. 101-132.
Bitzer, K., Trave, A., and Carmona, J.M., 2001, Fluid flow processes at basin scale: Acta Geologica Hispanica, v. 36 p. 1-20.[GeoRef]
Bohrmann, G., Abelmann, A.M.S., Gersonde, R., Hubberten, H., and Kuhn, G., 1994, Pure siliceous ooze, a diagenetic environment for early chert formation: Geology, v. 22 p. 207-210 doi: 10.1130/0091-7613(1994)022<0207:PSOADE>2.3.CO;2.
Bramlette, M.N., 1946, The Monterey Formation of California and the origin of its siliceous rocks: U.S. Geological Survey Professional Paper 212. 57 p.
Bryant, W.R., 2003, Permeability of clays, silty-clays and clayey-silts: in Scott, E.D., et al., eds., Siltstones, mudstones and shales: Depositional processes and characteristics: Tulsa, Society for Sedimentary Geology (SEPM), p. 76-84.
Carmichael, R.S., 1984, Handbook of physical properties of rocks: Vol. III Boca Raton, CRC Press. 340 p.
Chaika, C., and Williams, L.A., 2001, Density and mineralogy variations as a function of porosity in Miocene Monterey Formation oil and gas reserves in California: American Association of Petroleum Geologists Bulletin, v. 85 p. 149-167.
Cole, D., Stewart, S.A., and Cartwright, J.A., 2000, Giant irregular pockmark craters in the Palaeogene of the Outer Moray Firth Basin, UK North Sea: Marine and Petroleum Geology, v. 17 p. 563-577 doi: 10.1016/S0264-8172(00)00013-1.[CrossRef][ISI][GeoRef]
Compton, J.S., 1991, Porosity reduction and burial history of siliceous rocks from the Monterey and Sisquoc Formation, Point Pedernales area, California: Geological Society of America Bulletin, v. 103 p. 625-636 doi: 10.1130/0016-7606(1991)103<0625:PRABHO>2.3.CO;2.
Cooper, A.K., Scholl, D.W., and Marlow, M.S., 1987, Structural framework, sedimentary sequences, and hydrocarbon potential of the Aleutian and Bowers Basins, Bering Sea: in Scholl, D.W., Grantz, A., and Vedder, J.G., eds., Geology and resource potential of the continental margin of western North America and adjacent ocean basins, Beaufort Sea to Baja California: Houston, Texas, Circum-Pacific Council for Energy and Mineral Resources, Earth Science Seriesv. 6 p. 473-502.
Davies, R.J., 2003, Kilometer-scale fluidization structures formed during early burial of a deepwater slope channel on the Niger Delta: Geology, v. 31 p. 949-952 doi: 10.1130/G19835.1.
Davies, R.J., 2005, Differential compaction in sedimentary basins due to silica diagenesis: A case study: Geological Society of America Bulletin, v. 117 p. 1146-1155 doi: 10.1130/B25769.1.
Davies, R.J., and Cartwright, J., 2002, A fossilized opal-A to opal C/T transformation on the Northeast Atlantic Margin: Support for a significantly elevated palaeogeothermal gradient during the Neogene?: Basin Research, v. 14 p. 467-486 doi: 10.1046/j.1365-2117.2002.00184.x.[CrossRef][ISI][GeoRef]
Davies, R.J., and Cartwright, J.A., 2007, Kilometer-scale chemical reaction boundary patterns and deformation in sedimentary rocks: Earth and Planetary Science Letters, v. 262 p. 125-137.
Davies, R.J., and Clark, I.R., 2006, Continental slope failure primed and triggered by silica and its diagenesis: Basin Research, v. 18 p. 339-350.[ISI][GeoRef]
Davies, R.J., and Stewart, S.A., 2005, Emplacement of giant mud volcanoes in the South Caspian Basin: Three-dimensional seismic reflection imaging of root zones: London, Journal of the Geological Society, v. 162 p. 1-4.
Davies, R.J., Cartwright, J.A., and Rana, J., 1999, Giant hummocks in deep-water marine sediments: Evidence for large-scale differential compaction and density inversion during early burial: Geology, v. 27 p. 907-910 doi: 10.1130/0091-7613(1999)027<0907:GHIDWM>2.3.CO;2.
Davies, R.J., Huuse, M., Hirst, P., Cartwright, J.A., and Yang, Y., 2006, Giant clastic intrusions primed by silica diagenesis: Geology, v. 34 p. 917-920 doi: 10.1130/G22937A.1.
Deer, W.A., Howie, R.A., and Zussman, J., 1992, An introduction to the rock forming minerals: Harlow, Longman. 696 p.
Eichhubl, P., and Behl, R.J., 1998, Diagenesis, deformation, and fluid flow in the Miocene Monterey Formation: in Eichhubl, P., ed., Diagenesis, deformation, and fluid flow in the Miocene Monterey Formation of Coastal California: Los Angeles, Pacific Section, Society for Sedimentary Geology (SEPM) Special Publication 83, p. 5-13.
Eichhubl, P., and Boles, J.R., 1998, Vein formation in relation to burial diagenesis in the Miocene Monterey Formation, Arroyo Burro Beach, Santa Barbara, California: in Eichhubl, P., ed., Diagenesis, deformation, and fluid flow in the Miocene Monterey Formation of coastal California: Los Angeles, Pacific Section, Society for Sedimentary Geology (SEPM) Special Publication 83, p. 15-36.
Eichhubl, P., and Boles, J.R., 2000, Focused fluid flow along faults in the Monterey Formation, coastal California: Geological Society of America Bulletin, v. 112 p. 1667-1679 doi: 10.1130/0016-7606(2000)112<1667:FFFAFI>2.0.CO;2.
Engelder, T., 1993, Stress regimes in the lithosphere: Princeton, Princeton University Press. 457 p.
Giles, M.R., 1997, Diagenesis: A quantitative perspective: Implications for basin modelling and rock property prediction: Dordrecht, Kluwer Academic Publishers. 544 p.
Hein, J.R., Scholl, D.W., Barron, J.A., Jones, M.G., and Miller, J., 1978, Diagenesis of late Cenozoic diatomaceous deposits and formation of the bottom-simulating reflector in the southern Bering Sea: Sedimentology, v. 25 p. 155-181 doi: 10.1111/j.1365-3091.1978.tb00307.x.[CrossRef][ISI][GeoRef]
Hinman, N.W., 1990, Chemical factors influencing the rates and sequences of silica phase transitions: Effects of organic constituents: Geochimica et Cosmochimica Acta, v. 54 p. 1563-1574 doi: 10.1016/0016-7037(90)90391-W.[CrossRef][ISI][GeoRef]
Hovland, M., and Judd, A.G., 1988, Seabed pockmarks and seepages: Impact on geology: Biology and the Marine Environment: London, Graham & Trotman. 293 p.
Hunt, J.M., Whelan, J.K., Buxton Eglinton, L., and Cathles, L.M., III, 1998, Relation of shale porosities, gas generation, and compaction to deep overpressures in the US Gulf Coast: in Law, B.E., Ulmishek, G.F., and Slavin, V.I., eds., Abnormal pressures in hydrocarbon environments: Tulsa, American Association of Petroleum Geologists Memoir 70, p. 87-104.
Isaacs, C.M., 1981, Porosity reduction during diagenesis of the Monterey Formation, Santa Barbara coastal area, California: in Garrison, R.E., and Douglas, R.G., eds., The Monterey Formation and related siliceous rocks of California: Los Angeles, Society for Sedimentary Geology (SEPM) Pacific Section, p. 257-271.
Isaacs, C.M., 1982, Influence of rock composition on kinetics of silica phase changes in the Monterey Formation, Santa Barbara area, California: Geology, v. 10 p. 304-308 doi: 10.1130/0091-7613(1982)10<304: IORCOK>2.0.CO;2.[Abstract][CrossRef][ISI][GeoRef]
Jones, B., and Renaut, R.W., 2004, Water content of Opal-A: Implications for the origin of laminae in geyserite and sinter: Journal of Sedimentary Research, v. 74 p. 117-128.
Jones, J.B., and Segnit, E.R., 1971, The nature of opal. I. Nomenclature and constituent phases: Journal of the Geological Society of Australia, v. 18 p. 56-68.
Kastner, M., Keene, J.B., and Gieskes, J.M., 1977, Diagenesis of siliceous oozes–I. Chemical controls on the rate of opal-A to opal-CT transformation—An experimental study: Geochimica et Cosmochimica Acta, v. 41 p. 1041-1059 doi: 10.1016/0016-7037(77)90099-0.[CrossRef][ISI][GeoRef]
Keene, J.B., 1975, Cherts and porcellanites from the North Pacific, DSDP Leg 32: in Initial reports of the Deep Sea Drilling Report 32: Washington, D.C., U.S. Government Printing Office, p. 425-507.
Keller, M.A., and Isaacs, C.M., 1985, An evaluation of temperature scales for silica diagenesis in diatomaceous sequences including a new approach based on the Miocene Monterey Formation, California: Geo-Marine Letters, v. 5 p. 31-35 doi: 10.1007/BF02629794.[Medline]
Loseth, H., Wensaas, L., Arntsen, B., Hanken, N., Basire, C., and Graue, K., 2001, 1000 m long gas blow-out pipes: Amsterdam, 63rd European Association of Geoscientists and Engineers Conference and Exhibition, Extended Abstracts, p. 524.
Meadows, D., and Davies, R.J., 2007, Morphological development of basin-scale silica diagenetic fronts revealed with two-dimensional seismic reflection data: Offshore Sakhalin, Russian Far East: Journal of the Geological Society [London], v. 164 p. 1193-1206.
Moss, B., Barson, D., Rakhit, K., Dennis, H., and Swarbrick, R., 2003, Formation pore pressures and formation waters: in Evans, D., et al., eds., The millennium atlas: Petroleum geology of the central and northern North Sea: London, Geological Society, p. 317-329.
Nemeth, K., Martin, U., and Harangi, S., 2001, Miocene phreatomagmatic volcanism at Tihany (Pannonian Basin, Hungary): Journal of Volcanology and Geothermal Research, v. 111 p. 111-135 doi: 10.1016/S0377-0273(01)00223-2.[CrossRef][ISI][GeoRef]
Nichols, R.J., Sparks, R.S.J., and Wilson, C.J.N., 1994, Experimental studies of the fluidization of layered sediments and the formation of fluid escape structures: Sedimentology, v. 41 p. 233-253 doi: 10.1111/j.1365-3091.1994.tb01403.x.[CrossRef][ISI][GeoRef]
Nobes, D.C., Langseth, M.G., Kuramoto, S., Holler, P., and Hirata, N., 1992, Comparison and correlation of physical-property results from Japan Sea Basin and rise sites, legs 127 and 128, 1987: Proceedings of the Ocean Drilling Program, Scientific Results, v. 127, 128 p. 1275-1296.
Osborne, Mark.J., and Swarbrick, R.E., 1997, Mechanisms for generating overpressure in sedimentary basins: A reevaluation: American Association of Petroleum Geologists Bulletin, v. 81 p. 1023-1041 doi: 10.1306/1D9BBA5D-172D-11D7-8645000102C1865D.[Abstract][GeoRef]
Scotchman, I.C., Carr, A.D., and Parnell, J., 2006, Hydrocarbon generation modelling in a multiple rifted and volcanic basin: A case study in the Foinaven Sub-basin, Faeroe-Shetland Basin, UK Atlantic Margin: Scottish Journal of Geology, v. 42 p. 1-19.[ISI][GeoRef]
Shipboard Scientific Party, 1994, Site 903: in Mountain, G.S., Miller, K.G., Blum, P., et al., eds., Proceedings of the Ocean Drilling Program, initial reports, 150: College Station, Texas, Ocean Drilling Program, p. 129-205.
Smallwood, J.R., 2004, Tertiary inversion in the Faroe-Shetland Channel and the development of major erosional scarps: in Davies, R.J., et al., eds., 3D Seismic Technology: Application to the exploration of sedimentary basins: London, Geological Society, Memoir 29, p. 187-198.
