2.a. Ammonite biochronology and correlation

The Jurassic System is divided into a sequence of 11 globally applicable stages that are subdivided, without gaps or overlap, into regional sequences of ammonite-correlated ‘zones’ (Ogg & Hinnov, Reference Ogg, Hinnov, Gradstein, Ogg, Schmitz and Ogg2012). Since ammonite zones completely fill each stage, and are now usually explicitly defined with a basal stratotype, they can also be considered to be chronostratigraphical units (Callomon, Reference Callomon, Michelsen and Zeiss1985, Reference Callomon and Le Bas1995; Page, Reference Page1995, Reference Page, Ineson and Surlyk2003, in press). The great majority of zones do not conform in any way to a classical biozone since they do not correspond to the range of any ammonite species or assemblage.

In the present work, we refer neutrally to the ammonite-correlated stratigraphical units as ‘Standard’ Zones and Subzones (sensu Callomon, Reference Callomon, Michelsen and Zeiss1985), rather than using an epithet to describe their character (i.e. ‘chronozone’ and ‘subchronozone’ as in Page, Reference Page2010a,b,Reference Page, Lord and Davisc). However, as chronozones, these units can be explicitly correlated using proxies other than ammonites, such as local lithological changes, other faunal or floral elements and isotopic ‘events’ (Jenkyns et al. Reference Jenkyns, Jones, Gröcke, Hesselbo and Parkinson2002).

Throughout the Jurassic System, infra-subzonal units that are often referred to as biohorizons provide a very high-resolution biochronology. Biohorizons typically correspond to the stratigraphic range of a specific indicator species. Normally, the bases of ammonite Standard Zones are explicitly, or effectively, defined by the bases of specific biohorizons, and hence correlated, as potential timelines, by the stratigraphically lowest occurrences of the indicator species.

The initial biohorizonal framework for the Lower Sinemurian Stage in the UK of Page (Reference Page1992) led to the schemes applied by Page (Reference Page2002, Reference Page, Lord and Davis2010b) to the Devon and Dorset coastal sections. The scheme for the Hettangian Stage of the Devon/Dorset coast (Page, Reference Page, Lord and Davis2010b,Reference Page, Lord and Davisc) was based largely on the version for West Somerset by Page (Reference Page2005), but recent re-sampling of the latter area has revealed even greater biostratigraphical detail, which is used here. This new scheme provides a sequence of 55 biohorizons for the Hettangian, as opposed to the 27 used by Page (Reference Page, Lord and Davis2010b,Reference Page, Lord and Davisc).

Zonal and subzonal boundaries throughout the Hettangian have been reviewed, based on new sampling by K.N.P. on coastal sections (Fig. 1a) in East Devon (Lyme Regis), West Somerset (Doniford, St Audries Bay, East Quantocks Head to Kilve) and South Glamorgan, South Wales (St Mary's Well Bay, Lavernock). The biohorizon boundaries, where constrained to within a few centimetres stratigraphically, are indicated in Table 1. Within the Lower Sinemurian of West Somerset, the placement of biohorizon boundaries are as previously published (e.g. Bloos & Page, Reference Bloos, Page, Hall and Smith2000b; Page Reference Page, Lord and Davis2010b,Reference Page, Lord and Davisc).

Table 1. Ammonite biohorizon limits on the West Somerset (St Audries Bay and Quantock's Head), Devon/Dorset (Lyme Regis) and South Wales (Lavernock) coast sections

Height in metres of biohorizon boundaries above base of the Blue Lias Formation. The value in bold italics for the base Angulata Zone at Lyme Regis was estimated using the correlation with the St Audries Bay section (see Fig. 8, Section 5.a.). Pos. – Position; St A. – St Audries Bay section, Somerset; Qu. H. – Quantock's Head section, Somerset; CWS – Composite West Somerset coastal sections using common splice level at 56.90 m (below this level the heights for Quantock's Head differ from the CWS heights); Lav. – St. Mary's Well Bay section near Lavernock, South Wales; L.R. – Pinhay Bay to Devonshire Head section, Lyme Regis, Dorset.

2.b. The base of the Jurassic

The Global Stratigraphic Section and Point (GSSP) for the base of the Jurassic has been defined in the Kuhjoch section in the Austrian Calcareous Alps (Hillebrandt, Krystyn & Kuerschner, Reference Hillebrandt, Krystyn and Kuerschner2007; Hillebrandt & Krystyn, Reference Hillebrandt and Krystyn2009). Subsequently Page (Reference Page, Lord and Davis2010b,Reference Page, Lord and Davisc) revised the zonal framework for the base of the Hettangian Stage in the UK by inserting the former authors’ Tilmanni Zone below the Planorbis Zone (which was previously considered to be the lowest zone of the Jurassic System).

Page (Reference Page, Lord and Davis2010b,Reference Page, Lord and Davisc) concluded that only the top of the Tilmanni Zone in the UK had yielded ammonites, specifically Psiloceras erugatum (Phillips) as illustrated by Bloos & Page (Reference Bloos, Page, Hall and Smith2000a). Nevertheless, in the UK the base of the Zone and hence the base of the Jurassic System can still be approximated using the most positive part of a positive δ¹³Corg interval, as demonstrated by Clémence et al. (Reference Clémence, Bartolini, Gardin, Paris, Beaumont and Page2010). At St Audries Bay on the West Somerset coast (Fig. 1), the base of the Hettangian corresponds to a level less than 1.5 m above the base of the Blue Lias Formation (Clémence et al. Reference Clémence, Bartolini, Gardin, Paris, Beaumont and Page2010). At St Mary's Well Bay near Lavernock, South Wales (Fig. 1), the δ¹³Ccarb curve of Korte et al. (Reference Korte, Hesselbo, Jenkyns, Rickaby and Spotl2009), based on measurements of Liostrea sp. samples, suggests that the base of the Jurassic System lies about 1.9 m above the top of the Langport Member (Penarth Group, Upper Triassic). The δ¹³Ccarb curve produced from bulk sediment samples from the coast at Lyme Regis on the Devon/Dorset border by Korte et al. (Reference Korte, Hesselbo, Jenkyns, Rickaby and Spotl2009) is, however, of limited use for correlation because it does not show the same signature as elsewhere. Nevertheless, it is likely that the base of the Jurassic System in the area also lies in the lowest part of the Blue Lias Formation.

2.c. The base of the Planorbis Zone

The base of the Planorbis Zone, as correlated by the first appearance of Neophyllites at the base of the Hn3 imitans Biohorizon (Page, Reference Page2010a), corresponds to the base of bed 9 of Whittaker & Green (Reference Whittaker and Green1983) on the West Somerset coast (Page & Bloos, Reference Page and Bloos1995; Bloos & Page Reference Bloos, Page, Hall and Smith2000a). In South Glamorgan, in St Mary's Well Bay, Lavernock, the base of the Planorbis Zone is about 12 cm below the base of bed 30 of Waters & Lawrence (Reference Waters and Lawrence1987), i.e. the level indicated by Hodges (Reference Hodges1994) within his bed 38. At Lyme Regis, this level correlates with the base of bed H25 of Lang (Reference Lang1924); see Page (Reference Page2002).

2.d. Southam Quarry near Long Itchington

At Southam Quarry, Long Itchington, Warwickshire, below bed 10 of Clements et al. (Reference Clements1975, Reference Clements1977) ammonites are rare (Radley, Reference Radley2008) and the lowest part of the Blue Lias Formation is inferred to belong to the Liasicus Zone (i.e. the Tilmanni and Planorbis zones are missing). This interpretation follows from the location of Long Itchington in relation to the reconstructed age of onlap of the Blue Lias Formation onto the London-Brabant Platform by Donovan, Horton & Ivimey-Cook (Reference Donovan, Horton and Ivimey-Cook1979). Furthermore, it is consistent with the records of Old, Sumbler & Ambrose (1984, p. 34), who cite Laqueoceras and Waehneroceras ‘close above the Langport Member’ at Southam Quarry. Laqueoceras certainly indicates the lower part of the Laqueus Subzone Hn18 Biohorizon of Page (Reference Page, Lord and Davis2010b,Reference Page, Lord and Davisc). However, specimens of Waehneroceras from Southam Quarry include species from the upper part of the Portlocki Subzone, including abundant examples of W. portlocki (Wright), in limestone nodules and rare W. cf. shroederi Lange in limestone beds. These observations suggest that at least biohorizons Hn16 and Hn17 of Page (Reference Page, Lord and Davis2010b,Reference Page, Lord and Davisc) are also present.

3.a. Measurements of magnetic susceptibility in the field

Volume magnetic susceptibility (vol. MS) logging of sections used a Bartington Instruments MS2 meter combined with an F-probe in direct contact with fresh faces of rock at right angles to the bedding. Instrument drift, related to temperature changes, was removed by taking measurements in the air more than 1 m from the cliff face between each rock-surface measurement. At the levels of limestone nodules, MS was measured in the limestone rather than light marl in cases where limestone constituted more than half of that stratigraphic level (Weedon et al. Reference Weedon, Jenkyns, Coe and Hesselbo1999). As far as possible, calcite ‘beef’ veins and lenses (Richardson, Reference Richardson1923; Marshall, Reference Marshall1982), most prevalent at the base of laminated shale beds at Lyme Regis, were avoided during MS measurement.

Measurements of weight-specific MS (wt MS) using a Bartington Instruments MS2B sensor for samples from Lyme Regis and the West Somerset coast show the expected strong, linear correlation with vol. MS (Pearson's r = +0.967, N = 74, P < 0.001, Fig. 2a). This relationship and the similarity of both the vol. MS and wt MS logs to %CaCO₃ in Figure 2b confirm that the field measurements are neither significantly affected by changes in magnetic properties due to weathering, nor by incomplete rock sensing due to imperfect surface contact and/or undetected voids.

Figure 2. (a) Scatter plots of field measurements of vol. MS, and laboratory measurements of wt MS and %CaCO₃. The dashed lines indicate the 95% confidence intervals of the regressions. N – number of samples; r – Pearson's r (degree of correlation); ○ – sample from Lyme Regis; + – sample from West Somerset coast. The stratigraphic locations of the samples are indicated in Figures 4 and 5. (b) Profiles of vol. MS and %CaCO₃ measurements at Lyme Regis in beds 13 to 23 (bed numbers from Lang, Reference Lang1924). + – measurement of vol. MS, wt MS or %CaCO₃; ○ – estimated %CaCO₃ based on regression between vol. MS and %CaCO₃ (Fig. 2a); see text Section 3.b. (c) Profiles of measurements of δ¹⁸O and δ¹³C and %TOC at Lyme Regis in beds 13 to 23. + – bulk sample δ¹⁸O and δ¹³C from Weedon (Reference Weedon1987a); □ – bulk sample δ¹⁸O and δ¹³C from Paul, Allison & Brett (Reference Paul, Allison and Brett2008); ○ – Gryphaea sp. sample δ¹⁸O and δ¹³C from Weedon (Reference Weedon1987a). Key to rock-types in Figure 3.

The lithological columns shown in Figures 3–6, with captions indicating the references used in bed numbering, illustrate large variations in the thickness of the marls and shales and highly variable spacing and overall proportions of the limestone beds. In the logs, the laminated black shales are shown as the most recessed lithology on the profiles as a convention to aid their identification rather than as an indication of the lowest resistance to weathering. The fixed spacing of the MS measurements at each section was designed to resolve the smallest variations in bulk composition. The measurements were obtained at 2 cm spacing at Lyme Regis; 3 cm at Southam Quarry, Long Itchington; and 4 cm spacing on both the West Somerset coast and at St Mary's Well Bay, Lavernock. In Southam Quarry, the faces were measured for MS in 1994 when the section in the southwest between the A423 and the A426 was fully accessible down to the disconformity with the Upper Triassic Langport Member. The vol. MS log for Lyme Regis was reported previously (Weedon et al. Reference Weedon, Jenkyns, Coe and Hesselbo1999) but Figure 4, primarily of the Hettangian portion, shows revisions to the zonal and subzonal boundaries.

Figure 3. Lithostratigraphic log, vol. MS measurements and ammonite biostratigraphy at St Mary's Well Bay, Lavernock, Wales. Homogeneous limestone nodules are indicated by black ellipses. Laminated limestone nodules are indicated by white ellipses. At the levels where vol. MS was measured within limestone nodules rather than light marl, the nodules are indicated on the right of the lithostratigraphic log. Bed numbers from Waters & Lawrence (Reference Waters and Lawrence1987).

Figure 4. Lithostratigraphic log, vol. MS and %CaCO₃ measurements and ammonite biostratigraphy to the west of Lyme Regis, Devon. The location of the base of the Angulata Zone is discussed in Section 5.a. For key to lithologies see Figure 3. The concentration of samples at the level of bed 37 relates to figure 3.41 of Weedon (Reference Weedon1987a). ○ – sample measured for %CaCO₃ plotted with reference to lower horizontal axis; B – level with abundant calcite beef; Lport Mbr – Langport Member; Angul. – Angulata; Buc. – Bucklandi; Extra. – Extranodosa; Ext. – Extranodosa. Bed numbers from Lang (Reference Lang1924). Lang (Reference Lang1924) and Hesselbo & Jenkyns (Reference Hesselbo, Jenkyns and Taylor1995) provide the names of limestone beds on the Devon/Dorset coast.

The log for the West Somerset coast (Fig. 5a, b; Palmer, Reference Palmer1972; Whittaker & Green, Reference Whittaker and Green1983) is composite. The section at St Audries Bay was measured from the base of the Blue Lias Formation to the top of bed 101 of Whittaker & Green (Reference Whittaker and Green1983), overlapping the MS log at Quantock's Head by about 3.5 m. The Quantock's Head to Kilve section was measured for MS from bed 97 to bed 163. The splice level of the composite record in bed 96, at 56.90 m above the base of the Blue Lias Formation, represents the base of the Quantock's Head MS log. Note that in Table 1 the biohorizon levels of the Quantock's Head section are listed both relative to the base of the log (i.e. at the splice level) and relative to the base of the Formation according to the heights of the St Audries Bay section within the composite column.

Figure 5. (a) Lithostratigraphic log, vol. MS and %CaCO₃ measurements and ammonite biostratigraphy at St Audries Bay, Somerset. For key to lithologies see Figure 3. Black circles indicate samples measured for %CaCO₃ plotted with reference to lower horizontal axis. ○ – sample measured for %CaCO₃ plotted with reference to lower horizontal axis; B – level with abundant calcite beef. Bed numbering from Whittaker & Green (1983).

Figure 5. (b) As for (a) but a composite section from St Audries Bay and Quantock's Head. The join (splice) of the St Audries Bay section (St A.) with the Quantocks Head section (Q. H.) is indicated within bed 96 (Section 3.a). GSSP – Global Stratigraphic Section and Point.

3.b. Origin of the variations in vol. MS

It was reported previously for the Blue Lias Formation that MS is inversely correlated with %CaCO₃ (Hounslow, Reference Hounslow1985; Weedon et al. Reference Weedon, Jenkyns, Coe and Hesselbo1999; Deconinck et al. Reference Deconinck, Hesselbo, Debuisser, Averbuch, Baudin and Bessa2003; Ruhl et al. Reference Ruhl, Deenen, Abels, Bonis, Krijgsman and Kürschner2010). This relationship is supported by current measurements of wt MS using a Bartington Instruments MS2B cavity sensor and vol. MS (Fig. 2a). The right-hand linear regression line of Figure 2a allows estimation of carbonate contents from the field measurement of vol. MS (using: %CaCO₃ = 99.06 – (13.79×vol. MS), r = −0.892, N = 74, P < 0.001). Calcium carbonate measurements have been shown as open circles on top of the MS logs in Figures 4 and 5 by scaling %CaCO₃ (bottom scale) relative to the vol. MS (top scale) according to this regression. Despite the good agreement between the estimated and measured %CaCO₃ shown on the right of Figure 2b, the uncertainty (95% confidence interval) in individual estimates of %CaCO₃ from vol. MS using the regression is about ±20.0%, ranging between ±19.6 and ±20.4%, depending on the difference between the individual vol. MS and the mean vol. MS (Williams, Reference Williams1984, p. 313).

The average MS values for these mudrocks are fairly low, thereby contrasting with younger British Jurassic cyclic mudrocks such as the Kimmeridge Clay Formation (Weedon et al. Reference Weedon, Jenkyns, Coe and Hesselbo1999). It was shown for the Blue Lias Formation (Hounslow, Reference Hounslow1985; Bixler, Elmore & Engel, Reference Bixler, Elmore and Engel1998; Deconinck et al. Reference Deconinck, Hesselbo, Debuisser, Averbuch, Baudin and Bessa2003; Hounslow, Posen & Warrington, Reference Hounslow, Posen and Warrington2004) that the stratigraphic variations in MS are consistent with variable dilution of paramagnetic clays by non-ferroan, and thus diamagnetic, calcite (wt MS = +0.84×10⁻⁸ SI; Bleil & Petersen, Reference Bleil, Petersen and Angenheister1987). Note that despite the presence of locally rather large percentages of pyrite (maximum 12% in the laminated shales; Weedon, Reference Weedon1987a) in a pure, unweathered, state this mineral is also diamagnetic (wt MS = −0.48×10⁻⁸ SI; Collinson, Reference Collinson1983) and consequently contributes very little to the whole-rock MS signal. Hence, stratigraphic logs of MS in the Blue Lias Formation provide an accurate picture of the occurrence of the limestones (Figs 2a, b, 3).

The marls and shales of the Liasicus Zone at all four localities have consistently higher average MS than the underlying Tilmanni and Planorbis and overlying Angulata and Bucklandi zones (Figs 3–6; Weedon et al. Reference Weedon, Jenkyns, Coe and Hesselbo1999; Deconinck et al. Reference Deconinck, Hesselbo, Debuisser, Averbuch, Baudin and Bessa2003). This observation suggests that there were long-term (zonal-scale) variations in the composition of the paramagnetic (mainly clay mineral) components (Deconinck et al. Reference Deconinck, Hesselbo, Debuisser, Averbuch, Baudin and Bessa2003). The relatively high average MS of the lower part of the section in Southam Quarry (Fig. 6) is consistent with the interpretation (Section 2.d) that the lowermost strata that lie disconformably on the Langport Member belong to the Liasicus Zone and not to the Planorbis Zone.

Figure 6. Lithostratigraphic log, vol. MS measurements and ammonite biostratigraphy at Southam Quarry, near Southam and Long Itchington, Warwickshire. For key to lithologies see Figure 3. L. M. – Langport Member. Bed numbers from Clements et al. (Reference Clements1975).

At about the 31 m level in the mid-Liasicus Zone (upper Portlocki Subzone) of the St Audries Bay section (Fig. 5a), the average vol. MS is unusually high with values extending outside the range of the %CaCO₃ versus vol. MS regression of Figure 2a. Lower stratigraphic resolution sampling by Deconinck et al. (Reference Deconinck, Hesselbo, Debuisser, Averbuch, Baudin and Bessa2003) indicates that, at this level on the West Somerset coast, the clay minerals switch to higher kaolinite/illite ratios, but surprisingly, not to iron-rich clays. The unusually high MS might indicate an association of increased kaolinite with a source of different paramagnetic components and/or small amounts of ferromagnetic components such as magnetite-like minerals that, unusually for the Blue Lias Formation, survived dissolution by H₂S during diagenesis (Deconinck et al. Reference Deconinck, Hesselbo, Debuisser, Averbuch, Baudin and Bessa2003). A similar interval of unusually high values of vol. MS, of around 8.0×10⁻⁸ SI, is found at about 8.2 m in Southam Quarry in the Liasicus Zone (Fig. 6). This interval of high average MS in the marls and shales is located at the top of the Portlocki Subzone, Liasicus Zone at both St Audries and Lyme Regis and may provide a correlatable change in paramagnetic mineralogy (Figs 4, 5a).

4.a. Marginal facies and ‘near-shore’ Blue Lias

The Blue Lias was deposited at the same times as palaeo-coastline deposits that include the Sutton Stone and Southerndown Beds of South Glamorgan (Trueman, Reference Trueman1920, Reference Trueman1922; Hallam, Reference Hallam1960; Wobber, Reference Wobber1965; Fletcher, Reference Fletcher1988; Sheppard, Reference Sheppard2006) and the Brockley Down Stone of the Radstock Plateau near Bristol (D. L. Loughman, unpub. Ph.D. thesis, Univ. Birmingham, 1982; Donovan & Kellaway, Reference Donovan and Kellaway1984). These so-called marginal facies, formed at and close to palaeo-coastlines, lie unconformably on Carboniferous Limestone and comprise conglomeratic limestones with grainstone textures including Carboniferous Limestone and chert lithoclasts, derived bioclasts, ooids and bioclasts (Wobber, Reference Wobber1965, Reference Wobber1966; Fletcher, Reference Fletcher1988).

On the coast near Southerndown, Glamorgan, 31 km west-northwest of the exposure at Lavernock (Fig. 1a), the unconformity surface is exposed and consists of multiple wave-cut platforms overlain by associated successive breccia and conglomeratic and grainstone limestones of the Sutton Stone (Fletcher, Reference Fletcher1988; Sheppard, Reference Sheppard2006). Both the unconformity surface and Carboniferous Limestone lithoclasts in the overlying Sutton Stone have Trypanites borings and are encrusted by oysters and colonial corals (Johnson & McKerrow, Reference Johnson and McKerrow1995; Simms, Little & Rosen, Reference Simms, Little and Rosen2002; Sheppard, Reference Sheppard2006). Some lithoclasts in the Sutton Stone are imbricated (Wobber, Reference Wobber1963) and elongate bioclasts have been used as palaeocurrent indicators that indicate long-shore drift and complex current movements influenced by the local islands of Carboniferous rocks (Wobber, Reference Wobber1966). The matrix-supported and rounded boulder- and cobble-sized lithoclasts in the Sutton Stone are succeeded stratigraphically by pebble-grade lithoclasts, micrites and thin shales of the Southerndown Beds, which grade laterally into the near-shore facies of Blue Lias (Trueman, Reference Trueman1922; Wobber, Reference Wobber1965; Wilson et al. Reference Wilson, Davies, Fletcher and Smith1990; Sheppard, Reference Sheppard2006).

Although the platform erosion probably started in latest Triassic time, ammonites have been recovered in stratigraphic order from the Sutton Stone and Southerndown Beds from the Planorbis, Liasicus and Angulata zones (Hodges, Reference Hodges1986; Sheppard, Reference Sheppard2006). These observations are not consistent with nearly instantaneous deposition of the Sutton Stone by a single hurricane (Ager, Reference Ager1986), but rather with the latest Triassic and Early Jurassic erosion of the wave-cut platforms and deposition of the marginal facies under storm influence over millions of years during episodic rises in relative sea level (Trueman, Reference Trueman1922; Fletcher, Reference Fletcher1988; Sheppard, Reference Sheppard2006).

Just 6 km south of the palaeo-coastline of Southerndown at Nash Point, the Blue Lias of the Bucklandi Zone represents an atypical limestone-dominated facies often full of bioclast fragments in both the marls and limestones (Hallam, Reference Hallam1960; Weedon, Reference Weedon1987a; Trueman, Reference Trueman1930). These strata representing Units B and C of the Porthkerry Formation have much higher limestone proportions than found in more ‘typical’ Blue Lias (an average of 60 to 85% limestone by thickness in sections tens of metres thick; Waters & Lawrence, Reference Waters and Lawrence1987; Wilson et al. Reference Wilson, Davies, Fletcher and Smith1990; Warrington & Ivimey-Cook, Reference Warrington, Ivimey-Cook and Taylor1995). Units B and C consist of metre-scale bedding units formed of closely spaced nodular limestones and 'anastomosing mudstone beds’ that probably result from pressure dissolution (Waters & Lawrence, Reference Waters and Lawrence1987). Local decimetre-scale mudstone beds are present, permitting correlation of limestone-shale bed groups, but organic-carbon contents are low and neither laminated shales nor laminated limestones are present (Sheppard, Houghton & Swan, Reference Sheppard, Houghton and Swan2006). Although much of the bedding was formed purely during diagenesis, the presence of hummocky cross-stratification at Nash Point proves strong storm influences (Sheppard, Houghton & Swan, Reference Sheppard, Houghton and Swan2006).

Thirty kilometres from the palaeo-coastline of South Glamorgan, typical ‘offshore’ Blue Lias facies is present at Lavernock in the Tilmanni to Liasicus zones as logged here (Fig. 3). In the lower part of the Angulata Zone, Unit A of the Porthkerry Formation consists of about 55% by thickness of normal offshore Blue Lias with nodular limestones and marls. However, in the same area near Cardiff, the upper part of the Angulata Zone, ‘Unit B’ of the Porthkerry Formation of Waters & Lawrence (Reference Waters and Lawrence1987), is of a similar facies to the Bucklandi Zone at Nash Point. The Porthkerry Formation is overlain gradationally by uppermost Angulata Zone and Bucklandi Zone oolitic ‘marginal facies’ (Waters & Lawrence, Reference Waters and Lawrence1987). Thus, Units B and C of the Porthkerry Formation near Cardiff and Nash Point represent a ‘near-shore’ facies of the Blue Lias, which has been envisaged as a storm-derived aragonitic lime mud deposited on a marine shelf (Sheppard, Houghton & Swan, Reference Sheppard, Houghton and Swan2006).

4.b. Rock-type compositions

The different rock-types of the ‘typical’ or ‘offshore’ Blue Lias consist mostly of various proportions of bioclasts, calcite microspar, clay minerals, organic matter, pyrite and minor quartz silt (Hallam, Reference Hallam1960; Weedon, Reference Weedon1986, Reference Weedon1987a). Whole-rock compositions quoted here are based on the samples from both Lyme Regis and the Somerset coast (Weedon, Reference Weedon1987a). Numbers of samples by rock-type and locality are indicated in the caption for Figure 1c and sample locations are indicated in Figures 4 and 5. The limestones (total organic carbon, TOC, mean = 0.54%±0.09 (±95% confidence interval), range = 0.14 to 1.64%, N = 48; CaCO₃, mean = 79.3%±2.2, range = 51.8 to 88.7%, N = 48) form beds with contacts ranging from planar to nodular. Limestone nodules are typically entirely enclosed by light marl but can be ‘welded’ onto limestone beds. On the Devon/Dorset and West Somerset coasts, when seen in plan on the foreshore, limestone nodules are often found to be linked laterally at some levels. The limestones generally have a homogeneous bioclastic wackestone to bioclastic carbonate mudstone fabric reflecting the thoroughly bioturbated nature of the precursor light marls. The light marls (TOC mean = 1.42%±0.26, range = 0.38 to 4.41%, N = 40; CaCO₃, mean = 43.8%±2.7, range = 25.7 to 69.9%, N = 40) are friable and homogeneous with planar contacts against dark marls and laminated shale beds.

Normally the dark marls (TOC mean = 2.35%±0.41, range = 0.51 to 6.51%, N = 39; CaCO₃, mean = 40.1%±3.7, range = 19.9 to 61.0%, N = 39) are homogeneous like the light marls and friable, but can become fissile on weathering. However, Figure 7a shows a laminated dark marl fabric, not visible in the field, consisting of abundant microspar lenses in places forming nearly continuous laminae that alternate with other laminae comprising clay and organic matter (Weedon, Reference Weedon1987a). The figure shows a Chondrites burrow with a homogeneous light marl filling that penetrates the dark marl fabric. Burrow mottles at the contacts of light and dark marl beds are common at Lyme Regis as both light marl burrow-fills within dark marl and vice versa (Hallam, Reference Hallam1964).

Figure 7. (a) Photomicrograph of dark marl thin-section (plane polarized light, width of view 3 mm). Sample BL120 and fig. 2.2M of Weedon (Reference Weedon1987a), bed 32a, Rotiforme Subzone, Bucklandi Zone, Lyme Regis. (b) SEM photograph showing coccoliths in situ partly engulfed by euhedral microspar overgrowth (width of view 14 μm). Dark marl sample BL120 and fig. 2.4G of Weedon (Reference Weedon1987a), bed 32a, Bucklandi Zone, Lyme Regis. (c) Scanning electron microscope (SEM) photograph showing an aggregate of coccoliths surrounded by a matrix of organic matter and clay (width of view 56 μm). Dark marl sample BL120 and fig. 2.4F of Weedon (Reference Weedon1987a), bed 32a, Bucklandi Zone, Lyme Regis. (d) SEM photograph showing an aggregate of coccoliths next to Schizosphaerella punctulata (centre) both surrounded by a matrix of organic matter, clay and calcite microspar (width of view 53 μm). Laminated shale sample BL117, and fig. 2.4H of Weedon (Reference Weedon1987a), topmost laminated shale of bed 32, Bucklandi Zone, Lyme Regis. (e) SEM photograph showing a corroded coccolith in limestone (width of view ~ 8 μm). Limestone sample BL103 and fig. 2.4A of Weedon (Reference Weedon1987a), bed 29, Conybeari Subzone, Bucklandi Zone, Lyme Regis.

The laminated shales (TOC mean = 5.46%±0.77, range = 1.53 to 12.80%, N = 51; CaCO₃, mean = 35.5%±3.4, range = 11.6 to 68.4%, N = 51), usually with sharp basal bedding contacts against dark marl or light marl, produce a brown streak when scratched and weather into fissile, sometimes paper thin, laminae. Sub-microscopically they have scattered calcite microspar lenses dispersed within wavy sub-millimetre laminae of clay and organic matter (Weedon, Reference Weedon1987a). Laminated limestone (TOC mean = 1.42%±0.67, range = 0.90 to 3.00%, N = 7; CaCO₃, mean = 77.2%±6.2, range = 62.4 to 87.2%, N = 9) forms beds that have planar contacts with laminated shale beds or nodules that are enclosed entirely by laminated shale (Weedon, Reference Weedon1987a; Arzani, Reference Arzani2004).

Hallam (Reference Hallam1987) and Arzani (Reference Arzani2006) doubted the presence of rock-forming quantities of nannofossils as a possible precursor to the microspar. However, calcareous nannofossils, both as true coccoliths as well as Schizosphaerella, are well documented from throughout the Tilmanni to Bucklandi Zone interval at Lyme Regis and on the West Somerset coast (Hamilton, Reference Hamilton and Lord1982; Bown, Reference Bown1987; Bown & Cooper, Reference Bown, Cooper and Bown1998; van de Schootbrugge et al. Reference van de Schootbrugge, Tremolada, Rosenthal, Bailey, Feist-Burkhardt, Brinkhuis, Pross, Kent and Falkowski2007; Clémence et al. Reference Clémence, Bartolini, Gardin, Paris, Beaumont and Page2010). Indeed, Weedon (Reference Weedon1986, Reference Weedon1987a,Reference Weedonb) argued that the microspar lenses in the dark marls and laminated shales represent neomorphosed zooplankton faecal pellets containing calcareous nannofossils. Very similar fabrics to Figure 7a are known in the Kimmeridge Clay Formation of Dorset (Kimmeridgian to Tithonian; fig. 2d, f, g of Pearson, Marshall & Kemp, Reference Pearson, Marshall and Kemp2004), which has abundant nannofossils (e.g. Young & Bown, Reference Young and Bown1991; Lees, Bown & Young, Reference Lees, Bown and Young2006). Aggregates of nannofossils surrounded by clay minerals and organic matter in the Blue Lias Formation (Fig. 7b, c, d; fig. 3 of Weedon, Reference Weedon1986; fig. 6B, C of Arzani, Reference Arzani2006) have similar dimensions to the microspar lenses, supporting their interpretation as neomorphosed zooplankton faecal pellets (Pearson, Marshall & Kemp, Reference Pearson, Marshall and Kemp2004).

Weedon (Reference Weedon1986, Reference Weedon1987a) argued that nannofossil preservation was best in the more organic-carbon-rich laminated shales and dark marls, as confirmed recently by Clémence et al. (Reference Clémence, Bartolini, Gardin, Paris, Beaumont and Page2010). The high organic-carbon contents may have helped the preservation of primary organic coatings on the coccoliths, which in turn prevented calcite dissolution by acidic pore waters during diagenesis (Bukri, Dent Glasser & Smith, Reference Bukri, Dent Glasser and Smith1982). In the less organic-carbon-rich sediments, partial or total microspar aggradation of nannofossils (e.g. Fig. 7b) was inferred to be the normal situation in the Blue Lias (Weedon, Reference Weedon1986, Reference Weedon1987a,Reference Weedonb). The homogeneity of the light marls and most of the dark marls was explained as resulting from burrowers mixing the organic-matter- and clay-rich laminae with faecal pellets. In the limestones and light marls, however, very rare isolated elliptical structures, representing partially corroded or overgrown coccoliths provide the only in situ evidence for calcareous nannofossil precursors to the microspar (Fig. 7e; fig. 2.4B of Weedon, Reference Weedon1987a).

If most carbonate mud in the ‘offshore’ Blue Lias Formation was derived from very shallow coastal waters rather than from nannofossil material, different localities would be expected to have different rock-type compositions. However, the average %CaCO₃ and %TOC of the limestone, light marl, dark marl and laminated shales are statistically indistinguishable between the Devon/Dorset coast and the West Somerset coast sections (i.e. overlapping 95% confidence intervals in Fig. 1c). Therefore, considering the large difference in zonal and average bed thicknesses between these locations, the lithological composition of the offshore Blue Lias facies was not influenced by either the local net accumulation rates or by local differences in water depth. The consistency of facies between localities and across the range of accumulation rates represented is consistent with a primarily hemipelagic, rather than shoreline-influenced origin for the carbonate mud (Weedon, Reference Weedon1986).

The majority of the laminated limestone beds were formed by the coalescence of laminated limestone nodules (Weedon, Reference Weedon1987a; Arzani, Reference Arzani2004). However, at Lyme Regis, bed H30 categorized as laminated limestone does not have internal lamination but is homogeneous and without macrofossils. As for laminated limestone beds H32 and H36, it is underlain and overlain by laminated shale, has sharp planar top and bottom contacts, and has exceptionally low, slightly negative MS values. The lowest measured vol. MS of −0.13×10⁻⁸ SI implies, by extrapolating the regression of Figure 2a (Section 3.b), 97.3 %CaCO₃ and hence these limestones represent cementation of a primary sediment almost devoid of organic matter and clay. Partly owing to the sharp, planar base, Hesselbo & Jenkyns (Reference Hesselbo, Jenkyns and Taylor1995) inferred a dilute turbidity-current origin for bed H30 (which they named ‘Intruder’). However, in the absence of direct evidence for basal erosion, grading or cross-lamination, alternative explanations for the deposition of almost pure carbonate mud should be considered.

4.c. Lateral and stratigraphic variations in limestone proportions

Unlike the marls and laminated shale beds, average limestone bed thicknesses are consistently close to 10–15 cm at all the logged localities and in every zone of the Hettangian (Fig. 1b). Hallam (Reference Hallam1964, Reference Hallam1986) regarded the consistency in limestone bed thickness of the Blue Lias, independent of locality and zonal thickness, as indicative of a diagenetic rather than a primary control on limestone formation. Although average limestone bed thicknesses remain nearly constant, the proportion of limestone beds by thickness per ammonite zone varies substantially from place to place and stratigraphically (Fig. 1b). Areas with thinner biozones (lower net accumulation rate or ‘condensed sections’) have higher average proportions of limestone (Fig. 1b; fig. 2 of Page, Reference Page1995). For example, all four ammonite zones of the Hettangian of Lyme Regis have a far higher proportion of limestone than those on the West Somerset coast (Fig. 1b).

At Lyme Regis, representing an area of low net accumulation rate, all the zones of the Hettangian are of similar thickness and the average bed thicknesses for all rock-types are also similar in the different zones (Fig. 1b). By contrast, on the West Somerset coast the oldest zones (Tilmanni and Planorbis) are much thinner than the succeeding (Liasicus and Angulata) zones; significantly, the large change in zonal thickness there is mirrored by changes in average bed thickness of the light and dark marls and the laminated shales (Fig. 1b).

The association of changes in average non-limestone bed thickness with changing zonal thickness is interpreted here as indicating a large change in net accumulation rates at the end of the Planorbis Zone times or start of the Liasicus Zone for sections on the West Somerset coast. In this study, only the base of the Liasicus Zone was logged at Lavernock (Fig. 3). However, the much greater thickness of the Liasicus Zone interval at Lavernock compared to the Tilmanni and Planorbis zones (Waters & Lawrence, Reference Waters and Lawrence1987; Warrington & Ivimey-Cook, Reference Warrington, Ivimey-Cook and Taylor1995) indicates a similar change in net accumulation rate to that of the West Somerset coast.

The Liasicus Zone is consistently associated with lower proportions of limestone than preceding and succeeding Hettangian zones in different localities (Fig. 1b; Hallam, Reference Hallam1960; Palmer, Reference Palmer1972). This difference in character has led to member-scale subdivisions of the Blue Lias Formation with locally applied names, i.e. the Lavernock Shales (Richardson, Reference Richardson1905; Waters & Lawrence, Reference Waters and Lawrence1987); the St Audries Shales (Palmer, Reference Palmer1972) and the Saltford Shales (Donovan, Reference Donovan1956; Ambrose, Reference Ambrose2001).

In the lower Liasicus Zone, higher velocities on downhole sonic logs related to relatively higher clay content, and higher gamma-ray counts related both to thorium and potassium in the marls and shales and to uranium within organic matter, allow ready correlation of this portion of the Blue Lias across Britain (Whittaker, Holliday & Penn, Reference Whittaker, Holliday and Penn1985). The base of this mudrock-dominated interval is diachronous because, in the English Midlands, there is biostratigraphical evidence that it occurs within the Planorbis Zone (Old, Sumbler & Ambrose Reference Old, Sumbler and Ambrose1987; Ambrose Reference Ambrose2001), whereas in West Somerset and Lavernock the base is close to the base of the Liasicus Zone (Figs 3, 5).

Mapping of the oldest strata of the Blue Lias Formation on the London-Brabant Platform and the Radstock Plateau near Bristol shows that the Liasicus Zone in the Hettangian and Semicostatum Zone of the Sinemurian were times of accelerated onlap (Donovan, Horton & Ivimey-Cook, Reference Donovan, Horton and Ivimey-Cook1979; Donovan & Kellaway, Reference Donovan and Kellaway1984). An inferred sea-level rise during these zones was suggested by Hallam (Reference Hallam1981) and broadly supported by Hesselbo & Jenkyns (Reference Hesselbo, Jenkyns, Hardenbol, Thierry, Farley, Jacquin, de Graciansky and Vail1998) and Hesselbo (Reference Hesselbo2008).

At Lyme Regis, in the Liasicus Zone, the zonal thickness and the average thicknesses of the various rock-types are similar in the overlying and underlying zones, but there is nevertheless a lower proportion of limestone (Fig. 1b). Therefore, rising sea levels were apparently associated with reduced probability of limestone formation. The near-constancy of net accumulation rate at Lyme Regis, as indicated by the near-uniform average bed thicknesses of the marl and laminated shale, presumably resulted from an underlying tectonic block maintaining a relatively level sea floor within turbulent-water depths (due to storm winnowing) despite sea-level rise (Sellwood & Jenkyns, Reference Sellwood and Jenkyns1975). Apparently, rising sea levels during the Liasicus Zone times at Lyme Regis led to slight increases in water depth and less probability of limestone formation, but the increase in accommodation space was too small to allow substantially increased net sedimentation rates.

On the West Somerset coast and at Lavernock, increased accommodation space and higher sedimentation rates during the Liasicus Zone account for locally greater thicknesses of mudrocks (i.e. greater zonal thicknesses and greater average thickness of marl and laminated shale beds) compared to the Tilmanni and Planorbis zones (Figs 1b, 3, 5a). In these localities, the rapid increase in net accumulation rates apparently resulted from faster, probably fault-related, subsidence that was coincident with the sea-level rise (i.e. the increased subsidence rates exaggerated the effect of sea-level rise on local water depths). Fault-related activity is documented for the Hettangian in the Wessex Basin (Jenkyns & Senior, Reference Jenkyns and Senior1991) and South Glamorgan in South Wales (Wilson et al. Reference Wilson, Davies, Fletcher and Smith1990) and fault movement occurred at some point during the Jurassic Period associated with the Bristol Channel area (Waters & Lawrence, Reference Waters and Lawrence1987; Bixler, Elmore & Engel, Reference Bixler, Elmore and Engel1998). Furthermore, fossil methane seeps dating from the Bucklandi Zone in the Sinemurian might be related to fault movement on the West Somerset coast (Allison, Hesselbo & Brett, Reference Allison, Hesselbo and Brett2008; Price, Vowles-Sheridan & Anderson, Reference Price, Vowles-Sheridan and Anderson2008).

At Southam Quarry near Long Itchington, the majority of the Liasicus Zone strata consist of laminated shale (Figs 1b, 6). This observation might be thought to indicate that greater water depth due to sea-level rise in the Liasicus Zone was associated with the conditions required for the preservation of lamination (especially bottom-water dysoxia and anoxia, cf. Hallam & Bradshaw, Reference Hallam and Bradshaw1979). However, in both West Somerset and Lyme Regis, although there is a reduction in the proportions of limestone in the Liasicus Zone, unlike Long Itchington the proportions of laminated shale are lower than in the succeeding Angulata Zone (Figs 1b, 4, 5a, b; Weedon, Reference Weedon1987a). Hence, it is the formation of a lower proportion of limestones, not the development of laminated shales, which can be consistently related to rising sea level/greater water depth.

5.a. Shaw plot

In Table 1 there are 63 levels in the Hettangian and Lower Sinemurian, out of 130 available, where the same biohorizon level (i.e. biohorizon top or biohorizon base) has been located to within a few centimetres both at Lyme Regis and on the West Somerset coast. This stratigraphic refinement allowed construction of the Shaw plot in Figure 8 that illustrates the correlation of these two sites. At Lyme Regis, the base of the Angulata Zone lies somewhere between the lowest level with recorded Schlotheimia at 11.06 m (base of bed H84) and the highest recorded level with Waehneroceras at 9.66 m (within bed H71), as indicated by short horizontal lines in Figure 8. Using the line of correlation between St Audries and Lyme Regis, its location has been inferred to be close to 10.54 m (within bed H77, Fig. 4, Table 1) as indicated using the long grey horizontal arrow in Figure 8. Alternatively, the line of correlation from Quantock's Head rather than St Audries Bay, with fewer biohorizon constraints, would imply that the base of the Angulata Zone lies close to 10.13 m at Lyme Regis (within bed H73).

Figure 8. Shaw plot constructed using the joint locations of biohorizon bases and biohorizon tops at Lyme Regis and the West Somerset coast (St Audries Bay and Quantock's Head sections) as listed in Table 1. The location of the base of the Angulata Zone at Lyme Regis has been estimated using the West Somerset data via the line of correlation data (long grey arrows, Section 5.a). The ratios of the thicknesses of subzones in the West Somerset composite section to subzones in the Lyme Regis section are indicated at the bottom of the plot. Vertical or horizontal arrows with bed numbers indicate the levels at which breaks in slope of the line of correlation imply relative condensation or intra-formational hiatuses. + – St Audries Bay section versus Lyme Regis section; × – Quantock's Head section versus Lyme Regis section. Tilm. – Tilmanni; Planorb. – Planorbis; Pl. – Planorbis; Jo. – Johnsoni; Laq. – Laqueus; Extran. – Extranodosa; Depr. – Depressa; Conyb. – Conybeari; Roti. – Rotiforme.

Subzones on the West Somerset coast sections are thicker than those in the corresponding section at Lyme Regis, as indicated in figures at the bottom of Figure 8. Treating the biohorizon picks as time lines, the amount of strata within each interval results from the combination of the sedimentation rates and the effects of intra-formational hiatuses, i.e. the net accumulation rate. Shallower slopes on the line of correlation indicate higher net accumulation rates on the West Somerset coast relative to the Devon/Dorset coast.

In contrast to the Lyme Regis section, on the West Somerset coast the much thicker Liasicus and Angulata zones compared to the Tilmanni and Planorbis zones are mirrored by the increased average thicknesses of the corresponding marl and shale beds (Fig. 1b; Section 4.c). Hence, the major decrease in the average slope of the line of correlation in Figure 8 at the base of the Liasicus Zone is interpreted as due to substantially increased net accumulation rates in West Somerset. In detail, the breaks in slope (nearly horizontal or nearly vertical segments) of the line of correlation are interpreted as due to hiatuses or intervals of exceptionally low accumulation rate, and have been labelled with the corresponding bed numbers. These breaks in slope of the line of correlation are discussed in Section 5.c. The breaks in slope are defined by multiple points, significantly reducing the likelihood that they simply result from misplacing the tie-levels owing to collection failure for key ammonites.

5.b. Minor erosion

Several lines of field evidence for minor sea-floor erosion were used by Weedon (Reference Weedon1986) to argue for storm-induced bottom-water turbulence. Protrusive Diplocraterion burrow mottles, which are found only within the limestones and light marls, have been described at all four localities studied here as well as at Saltford, Avon (Fig. 9c; Sellwood, Reference Sellwood, Crimes and Harper1970; Donovan & Kellaway, Reference Donovan and Kellaway1984; Weedon, Reference Weedon1987a; Mogadam & Paul, Reference Moghadam and Paul2000; Barras & Twitchett, Reference Barras and Twitchett2007). Tiering analysis at Lyme Regis (Mogadam & Paul, Reference Moghadam and Paul2000) showed that Diplocraterion cross-cuts all other trace fossils except Chondrites, which is well known to have been produced by late-stage deposit feeders within anoxic sediments (Bromley & Ekdale, Reference Bromley and Ekdale1984). The observations of Mogadam & Paul (Reference Moghadam and Paul2000) suggest re-excavation of the vertical U-shaped burrow following rapid centimetre-scale sediment removal (Seilacher, Reference Seilacher2007).

Figure 9. (a) Photograph of polished limestone section showing bored and encrusted limestone intraclast with a hardground-like surface (diagrammatic description below) collected in situ from within bed 25 (Top Copper) below Devonshire Head, Lyme Regis. (ai) Photographic negative image of acetate peel showing Liostrea encrusting the intraclast surface and overlain by bioclastic packstone. Two Liostrea individuals (left and right) encrust a third individual, demonstrating at least two generations of encrustation. Short arrows indicate the Talpina ramosa Von Hagenow borings into the intraclast surface and within the encrusting Liostrea. (aii) As for (ai) but a different section of the intraclast surface (see diagram for location). (b) Limestone intraclasts on the surface of a fallen block of limestone, Monmouth Beach, Lyme Regis. The lens cap is 5 cm in diameter. (c) Protrusive Diplocraterion forming dark marl burrow-fill within limestone bed 19 (Specketty), Seven Rock Point, Lyme Regis. (d) Horizon of isolated dark marl burrow fills indicating the former presence of a bed of dark marl that has been removed locally by sea-floor erosion, bed 19, Seven Rock Point, Lyme Regis.

At Lyme Regis, dark marl beds that are only a few centimetres thick are locally missing for several metres laterally (Weedon, Reference Weedon1986). Within limestone bed 23 (‘Mongrel’), locally a thin dark marl bed is replaced laterally by centimetre-scale, metre-wide scours filled with bioclastic packstone. The former presence of a thin bed of dark marl is in some places recorded by a layer of isolated dark marl burrow mottles within what otherwise appears to be a single limestone bed (e.g. in bed 19; Fig. 9d). In West Somerset, this phenomenon is occasionally observed in thick light marl beds but more typically an opposite arrangement occurs with isolated bands of light burrow-fills within dark marls representing relics of thin light marl beds. Paul, Allison & Brett (Reference Paul, Allison and Brett2008) also noted that occasional ‘exotic’ sediment fills within uncrushed ammonites at Lyme Regis could indicate material subsequently removed by erosion.

Radley (Reference Radley2008) argued that at Southam Quarry, Long Itchington, erosion within laminated shale was linked to distal storm influences below normal storm wave-base. He described siltstone scour fills within Liasicus Zone laminated shales as well as highly elongated limestone nodules containing imbricated ammonites, which were interpreted as gutter casts. The scour fills are associated with a trace fossil indicative of a crustacean escape structure and therefore very rapid deposition (Radley, Reference Radley2008; O'Brien, Braddy & Radley, Reference O'Brien, Braddy and Radley2009). Similar presumed scour fills, full of ammonites and shell fragments, within light marl beds are also known in West Somerset, albeit mainly in the Planorbis Subzone (e.g. bed 24).

A scour fill at Lyme Regis in the top of a limestone bed, probably from the Rotiforme Subzone, Bucklandi Zone, consists of a shallow depression around 50 cm across filled with coarse shelly debris and abundant small articulated echinoids (?Diademopsis) and common articulated ophiuroids (Page collection, Bristol City Museum and Art Gallery). These echinoderms retained their articulation owing to rapid burial, presumably as waning storm-currents deposited previously suspended sediment. A second level with common, but more scattered Diademopsis in laminated mudstone bed 17 (Planorbis Zone and Subzone) near Watchet, West Somerset, may indicate a similar event, although there is no equivalent concentration of shelly debris (K. N. P. pers. obs. 2015).

Found commonly throughout the sections on the Devon/Dorset and West Somerset coast are centimetre-scale horizons with: (a) concentrated bivalve fragments; (b) lenses of echinoid debris; (c) abundant large ammonites on the surface or base of limestone beds; and/or (d) scattered large ammonites and nautiloids with encrusting oysters and/or crinoid debris on the tops of limestone beds. Some of these features have been previously recorded by Paul, Allison & Brett (Reference Paul, Allison and Brett2008) and they indicate either periods of increased sea-floor turbulence that led to winnowing of fines, fragmentation of shells and the concentration of bioclasts or periods of very slow, or halted deposition. In the latter case, epifauna such as oysters and crinoids had time to attach to the ‘benthic islands’ created by large shells. Such observations can coincide with additional evidence for significant intervals of missing strata.

5.c. Major hiatuses

The following sub-sections summarize the Shaw plot and field evidence for significant hiatuses at many levels within the Hettangian and Lower Sinemurian offshore facies of the Blue Lias of southern Britain. Note that in some cases these major hiatuses can be potentially linked to sea-level rises and in others to sea-level falls.

7.a. Lyme Regis

Figures 2b and 10a show that, at Lyme Regis, oxygen isotopes in the centres of the limestone beds are as high as −1.5‰ VPDB (Vienna Pee-Dee Belemnite) with carbon-isotope values close to 0.0‰ VPDB. These values are not much lighter than the isotopes of unaltered benthic bivalve (Gryphaea) calcite (Weedon, Reference Weedon1987a). Hence, carbon and oxygen isotopes of the Blue Lias, combined with the near-absence of evidence for compaction of macrofossils, have long been considered consistent with early cementation (Campos & Hallam, Reference Campos and Hallam1979; Gluyas, Reference Gluyas1984; Weedon, Reference Weedon1987a; Arzani, Reference Arzani2006; Paul, Allison & Brett, Reference Paul, Allison and Brett2008). The decreases of oxygen-isotope values towards the edges of limestone beds, where calcium carbonate contents are lower, are consistent with cementation occurring progressively during the early stages of compaction (Gluyas, Reference Gluyas1984; Weedon, Reference Weedon1987a; Arzani, Reference Arzani2006; Paul, Allison & Brett, Reference Paul, Allison and Brett2008). Disseminated framboidal and euhedral pyrite found throughout the limestones and the non-ferroan nature of the calcite microspar indicate at least some carbonate generation during sulphate reduction (Gluyas, Reference Gluyas1984).

Figure 10. (a) Compilation of oxygen- and carbon-isotope measurements for Lyme Regis from the Angulata and Bucklandi zones. Note several data points from limestone beds (black squares) near δ¹⁸O = −4.5 and δ¹³C = −1.5‰ VPDB are obscured by the data for marl or laminated shale (unfilled triangles). Sources of measurements for whole-rock: Campos & Hallam (Reference Campos and Hallam1979), Gluyas (Reference Gluyas1984), Weedon (Reference Weedon1987a), Arzani (Reference Arzani2004, Reference Arzani2006), Paul, Allison & Brett (Reference Paul, Allison and Brett2008); for Gryphaea sp.: Weedon (Reference Weedon1987a). (b) Compilation of oxygen- and carbon-isotope measurements for West Somerset coast (Kilve) for the Bucklandi Zone. Sources of measurements for whole rock: Allison, Hesselbo & Brett (Reference Allison, Hesselbo and Brett2008), Price, Vowles-Sheridan & Anderson (Reference Price, Vowles-Sheridan and Anderson2008). Measurements plotted exclude atypical rock-types associated with the Bucklandi Zone mud mounds. (c) Compilation of oxygen- and carbon-isotope measurements for West Somerset coast (St Audries and Doniford) for the Tilmanni and Planorbis zones. Sources of measurements for whole rock: Arzani (Reference Arzani2004, Reference Arzani2006), Clémence et al. (Reference Clémence, Bartolini, Gardin, Paris, Beaumont and Page2010 as listed in supplementary information of Paris et al. Reference Paris, Beaumont, Bartolini, Clémence and Gardin2010); for Liostrea sp.: van Schootbrugge et al. (Reference van de Schootbrugge, Tremolada, Rosenthal, Bailey, Feist-Burkhardt, Brinkhuis, Pross, Kent and Falkowski2007, their ‘unaltered’ samples only).

Raiswell (Reference Raiswell1988) postulated that cementation of the Blue Lias limestone beds and nodules was associated with anaerobic methane oxidation by sulphate within the sulphate reduction zone. Both normal sulphate reduction and anaerobic methane oxidation generate carbonate that neutralizes the acidity produced during the precipitation of iron monosulphide and pyrite oxidation (Raiswell, Reference Raiswell1988; Bottrell & Raiswell, Reference Bottrell and Raiswell1989). Since there is a restricted depth interval involved, anaerobic methane oxidation was used by Raiswell (Reference Raiswell1988) to explain Hallam's (Reference Hallam1964) observations of the limited range of limestone bed thickness. In particular, Raiswell's model requires that shallow cementation (< 1 m burial) occurred during a pause in sedimentation.

Bottrell & Raiswell (Reference Bottrell and Raiswell1989) further demonstrated that pyrite formation occurred over a longer period in the limestones than in the light marl and dark marls. To explain the sulphur-isotope composition of the pyrite in the limestones they invoked incorporation of sulphur derived from the bioturbational oxidation of previously formed pyrite. Such a process implies limestone cementation close to the zone of burrowing and hence not greatly below the sediment–water interface. δ¹³C of carbonate from sulphate reduction or aerobic methane oxidation is typically less than −10.0‰ VPDB. Bottrell & Raiswell (Reference Bottrell and Raiswell1989) explained the relatively heavy carbon-isotope values of around 0.0‰ VPDB in the limestones as indicating that the bulk of the carbonate was derived from dissolution of aragonite and high-Mg bioclasts. Later authors concurred with this suggested source of carbonate (Wright, Cherns & Hodges, Reference Wright, Cherns and Hodges2003; Arzani, Reference Arzani2004, Reference Arzani2006; Paul, Allison & Brett, Reference Paul, Allison and Brett2008). Some limestone beds at Lyme Regis contain centimetre-scale cracks partly filled with bioclastic sediment, indicating that cementation and fissuring occurred close enough to the sediment–water interface to allow access to unconsolidated material, presumably following a phase of sea-floor erosion (Hesselbo & Jenkyns, Reference Hesselbo, Jenkyns and Taylor1995).

Plots of isotope ratios in the Blue Lias Formation show measurements lying near a trend-line that slopes from near the origin down towards more negative oxygen- and carbon-isotope ratios. Trends in isotope composition from heavier oxygen isotopes at the centres of nodules to lighter oxygen isotopes at the edges (Fig. 10; Weedon, Reference Weedon1987a; Arzani, Reference Arzani2006) apparently confirm the supposition that oxygen isotopes became lighter through time, most likely due to increasing pore-water temperatures. However, the carbon isotopes in limestone nodules can increase rather than decrease towards the outside of nodules and consequently do not lie along the main trend of measurements at Lyme Regis (Fig. 10a) and West Somerset (Fig. 10c).

Assuming that oxygen-isotope values only decreased during the diagenesis of the Blue Lias Formation, an explanation is needed for why the limestone beds can have average carbonate with δ¹⁸O = −4.5 and δ¹³C = −1.5‰ VPDB while limestone nodules can have average carbonate with the same δ¹⁸O, but δ¹³C = −3.0‰ VPDB (Fig. 10a). The key to this problem is to regard the main trend in the isotopes in the limestone beds as indicating a mixing line between the composition of early and late carbonate rather than as a time series.

It is now recognized that early meniscus calcite cement associated with carbonate nodule formation could have provided a supporting framework that allowed the bulk of cementation to occur much later during compaction (Curtis et al. Reference Curtis, Cope, Plant and Macquaker2000; Raiswell & Fisher, Reference Raiswell and Fisher2000). This model helps explain why it is so rare to find shallow limestone nodules forming in modern organic-rich marine sediments (Raiswell & Fisher, Reference Raiswell and Fisher2000). In the Blue Lias Formation, the very early cement in the limestones with oxygen and carbon isotopes relatively close to 0.0‰ VPDB was apparently sufficient to prevent the compaction of large macrofossils. At a later phase of diagenesis, once compaction had started in earnest, the remaining pore spaces within the limestone beds could then be partly filled with carbonate that had much lighter oxygen and, typically also lighter carbon isotopes. This model means that, if cemented relatively quickly, limestone nodules can show progressive changes in isotopes that, unlike the limestone beds, record a snapshot of the history of changing pore-water isotopic composition in concentric zones (including pore-water carbon isotopes becoming heavier while oxygen isotopes became lighter). Although Arzani (Reference Arzani2006) demonstrated that microspar growth was displacive, the cementation cannot have continued much beyond the end of occlusion of the pore spaces because the limestone beds and nodules have oxygen isotopes not lighter than about −6.0‰.

At Lyme Regis, the oxygen isotopes of the calcite microspar in the light and dark marls and the black laminated shale (−3.5 to −6.0‰ VPDB) are much lighter on average than the limestone beds (Fig. 10a). Hence, the marls and shales appear to have been cemented later than the limestone beds. However, one sample of light marl from mid-way through bed 15c (BL307, table 3B of Weedon, Reference Weedon1987a; Figs 2b, 10a) with %CaCO₃ = 38.2 and %TOC = 0.61 has an oxygen-isotope value that is far heavier (δ¹⁸O = −1.99‰ VPDB) than the other marl and shale samples. This sample is interpreted to represent light marl that was cemented by early meniscus cement, as though destined to become the centre of a limestone bed or nodule, but the amount of cement was apparently insufficient to provide enough framework support to prevent compaction during burial.

Ferroan and non-ferroan calcite beef occurring near the bases of some laminated shales at Lyme Regis with δ¹⁸O less than −6‰ (Campos & Hallam, Reference Campos and Hallam1979), not shown in Figure 10a, probably formed as a result of over-pressuring during deep burial (Marshall, Reference Marshall1982).

7.b. West Somerset coast

Unaltered carbonate of Liostrea from the Tilmanni Zone of the West Somerset coast have generally much heavier carbon isotopes (+1.6 to +3.9‰, Fig. 10c; van de Schootbrugge et al. Reference van de Schootbrugge, Tremolada, Rosenthal, Bailey, Feist-Burkhardt, Brinkhuis, Pross, Kent and Falkowski2007) than those found in Gryphaea from the Angulata and Bucklandi zones at Lyme Regis (+0.5 to +2.5‰ VPDB, Figs 2b, 10b; Weedon, Reference Weedon1987a). Korte et al. (Reference Korte, Hesselbo, Jenkyns, Rickaby and Spotl2009) measured δ¹³C of between +1.5 and +5.0‰ VPDB for Liostrea in the Tilmanni Zone of Lavernock (not shown) with progressively lighter average values in the upper part of the Tilmanni, Planorbis and lower Liasicus zones. The different ranges of values in the Tilmanni and the Angulata–Bucklandi zones have been related to secular trends in carbon isotopes within the Hettangian rather than diagenetic factors (Korte et al. Reference Korte, Hesselbo, Jenkyns, Rickaby and Spotl2009). However, all three groups of measurements of unaltered benthic bivalve calcite indicate δ¹⁸O between about 0.0 and −2.0‰ VPDB.

The limestones on the West Somerset coast with δ¹⁸O = −4.0 to −12.5‰ VPDB have far lighter oxygen isotopes than those at Lyme Regis (compare Fig. 10a with Figs 10b and 10c). The isotopically lightest values from a limestone bed (Fig. 10b) are derived from beds associated with mud volcanoes developed in the Bucklandi Zone strata near Kilve (Cornford, Reference Cornford2003; Allison, Hesselbo & Brett, Reference Allison, Hesselbo and Brett2008; Price, Vowles-Sheridan & Anderson, Reference Price, Vowles-Sheridan and Anderson2008). Isotope values from material described as comprising the mud volcano mounds themselves (i.e. breccia, ‘tufa’, crust, etc) have been excluded from Figure 10b.

Oxygen isotopes from limestone beds from the Tilmanni and Planorbis zones on the West Somerset coast are not only lighter than the limestone beds at Lyme Regis, but also generally lighter, rather than heavier, than their associated marls and black laminated shales (Fig. 10c; Arzani, Reference Arzani2004). The much lighter oxygen isotopes suggest that the limestones were cemented later than the marls and shales. Nevertheless, large macrofossils within the limestone beds of the West Somerset coast are preserved uncrushed in the same manner as observed at Lyme Regis. Accordingly, cementation started prior to compaction. These apparently conflicting interpretations can be reconciled if it is hypothesized that the West Somerset coast limestone beds had an early framework cement with oxygen isotopes close to 0.0‰ VPDB that was joined by carbonate formed much later that had much lighter oxygen-isotope values. Thus, the marls and laminated shales became fully compacted and fully cemented at a time when there remained framework-supported pores within the limestone beds still available for later cementation by calcite with very light oxygen-isotope values.

Campos & Hallam (Reference Campos and Hallam1979) explained their observations of much heavier average oxygen isotopes in the limestones of the Lower Jurassic of Devon/Dorset compared to the Lower Jurassic of the Yorkshire coast in terms of net sedimentation rate. Their explanation is followed here with respect to the comparison of the West Somerset with the Devon/Dorset coast, assuming that the total time for cementation was approximately the same in both cases. In both areas, cementation appears to have started very early, but the limestones at Lyme Regis were buried more slowly and so their final carbonate was probably precipitated at relatively low temperatures (overall oxygen isotopes heavier than the marls and shales). The more rapidly buried limestone beds on the West Somerset coast apparently ceased full cementation at much greater depths and thus were formed from pore waters with higher temperatures (overall oxygen isotopes lighter than the marls and shales).

A prolonged cementation history on the West Somerset coast is indicated by very negative δ¹⁸O in vein calcite (−9 to −12‰ VPDB, not shown in Fig. 10), probably associated with Late Jurassic to Early Cretaceous extensional and Cenozoic compressional faulting (Bixler, Elmore & Engel, Reference Bixler, Elmore and Engel1998).