|
|
 |
|||||||||||||||||
|
|||||||||||||||||
 |
|||||||||||||||||
| JOURNAL HOME | HELP | CONTACT PUBLISHER | SUBSCRIBE | ARCHIVE | SEARCH | TABLE OF CONTENTS |
,1
,11 Department of Geological Sciences, Rutgers University, Piscataway, New Jersey 08854, USA
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSION: COMPARISONS OF...CONCLUSIONSREFERENCES CITED |
|---|
0.9
18O increase in the SSQ core hole is correlated to the global earliest Oligocene (Oi1) event using magnetobiostratigraphy; this increase is associated with the Shubuta-Bumpnose contact, an erosional surface, and a biofacies shift in the core hole, providing a first-order correlation between ice growth and a sequence boundary that indicates a sea-level fall. The 18O increase is associated with a eustatic fall of 55 m, indicating that 0.4 of the increase at Oi1 time was due to temperature. Maximum 18O values of Oi1 occur above the sequence boundary, requiring that deposition resumed during the lowest eustatic lowstand. A precursor 18O increase of 0.5 (33.8 Ma, midchron C13r) at SSQ correlates with a 0.5 increase in the deep Pacific Ocean; the lack of evidence for a sea-level change with the precursor suggests that this was primarily a cooling event, not an ice-volume event. Eocene–Oligocene shelf water temperatures of 17–19 °C at SSQ are similar to modern values for 100 m water depth in this region. Our study establishes the relationships among ice volume, 18O, and sequences: a latest Eocene cooling event was followed by an earliest Oligocene ice volume and cooling event that lowered sea level and formed a sequence boundary during the early stages of eustatic fall.
Key Words: Eocene-Oligocene • sea level • climate • ice volume • Alabama • sequence stratigraphy • icehouse • greenhouse
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSION: COMPARISONS OF...CONCLUSIONSREFERENCES CITED |
|---|
18O increase of 1.0–1.5 (Oi1 of Miller et al., 1991) occurred throughout the Atlantic, Pacific, Indian, and Southern Oceans (Shackleton and Kennett, 1975; Savin et al., 1975; Kennett and Shackleton, 1976; Keigwin, 1980; Corliss et al., 1984; Miller et al., 1987; Zachos et al., 2001; Coxall et al., 2005). Most studies agree that the earliest Oligocene marked the beginning of the icehouse Earth, with large (i.e., near modern sized) ice sheets in Antarctica (e.g., Miller et al., 1991; Zachos et al., 1992). However, considerable controversy has surrounded the cause of the 18O increase that culminated in Oi1, ranging from early studies that attribute it entirely to a cooling of deep-water (and hence, high-latitude surface water) temperatures (Shackleton and Kennett, 1975; Savin et al., 1975; Kennett and Shackleton, 1976), to a recent study that attributes the entire 1.5 18O increase observed in the deep Pacific to growth of ice sheets (Tripati et al., 2005). This latter interpretation requires: (1) ice storage that is 1.5 times that of modern ice sheets; (2) the presence of large ice sheets in Antarctica and in the Northern Hemisphere; and (3) a global sealevel (eustatic) fall of 150 m. It is based on the lack of a deep-sea Mg/Ca change associated with the Oi1 18O increase (Lear et al., 2000), implying little or no cooling. However, a dramatic drop in the calcite compensation depth occurred at this transition (Van Andel et al., 1975; Coxall et al., 2005), and may have caused changes in carbonate ion activity that masked cooling in Mg/Ca records (Lear et al., 2004).
There is ample evidence for a major global cooling during the Eocene–Oligocene transition. Paleontological evidence for cooling includes the development of psychrospheric (cold-loving) ostracods (Benson, 1975), a deep-sea benthic foraminiferal turnover (e.g., Miller et al., 1992; Thomas, 1992), the loss of calcareous nannoplankton that thrived in early Paleogene warm, oligotrophic waters (Aubry, 1992), regional evidence for cooling from pollen (e.g., New Jersey; Owens et al., 1988), mammalian turnover (e.g., England; Hooker et al., 2004), and microfossil assemblages (e.g., New Zealand; Nelson and Cook, 2001). Isotopic evidence also indicates global cooling. The 18O increase in the deep Atlantic is typically 1.0 (e.g., Miller and Curry, 1982); the 1.5 18O increase at Ocean Drilling Program (ODP) Site 1218 (deep Pacific; Coxall et al., 2005) implies that there was at least a 2 °C cooling in the deep Pacific. Comparisons of benthic foraminiferal 18O records and a latitudinal profile of planktonic foraminiferal 18O values show a shift in mean values of 0.6 from the late Eocene to the early Oligocene (Keigwin and Corliss, 1986). This global change in 18Oseawater is best explained by ice growth with a consequent sea-level lowering of 50–60 m (using the 18O/sea level calibrations of Fairbanks and Matthews [1978] and Pekar et al. [2002] of 0.11/10 m and 0.1/10 m, respectively). Nevertheless, the precise amount of the 18O increase that is attributable to ice versus temperature remains debatable.
Sea-level studies have established that a major eustatic drop occurred in the earliest Oligocene. Although the Exxon Production Research Company (Exxon) sea-level curve shows no earliest Oligocene change and a dramatic (160+ m) mid-Oligocene fall (Vail et al., 1977; Haq et al., 1987), studies in New Jersey have documented a major earliest Oligocene eustatic fall of 55 m (Pekar et al., 2001; Miller et al., 2005a). Using the sea level/18Oseawater calibrations cited above, this implies that 0.5–0.6 of the deep-sea 18O increase was due to an increase in ice volume and 0.5–1.0 was due to a 2–4 °C deep-water cooling. Sequence stratigraphic and backstripping studies in New Jersey have documented that the mid-Oligocene eustatic lowering was 50–60 m (Pekar et al., 2001; Miller et al., 2005a), far less than the 160 m fall shown by the Exxon curves. The absence of an earliest Oligocene event on the Exxon curve has been a source of discussion and debate for more than 25 yr (Olsson et al., 1980).
Until now, the data sets used to decipher ice-volume changes across the Eocene–Oligocene transition were derived from deep-sea locations largely drilled by the Deep Sea Drilling Project and the ODP. In contrast, sea-level studies of this interval have mostly examined seismic profiles on continental margins (e.g., Vail et al., 1977) or marine sections on land (e.g., Vail et al., 1987; Haq et al., 1987; Loutit et al., 1988; Baum and Vail, 1988; Miller et al., 2005a). It has proven difficult to obtain expanded and reliable 18O records for onshore marine sections because of hiatuses, diagenesis, poorly fossiliferous sections, and other complications due to nearshore influences. Linking deep-sea isotopes and sea-level records requires using magnetobiostratigraphic correlations that have errors of 0.5–1.0 m.y. (e.g., Miller et al., 1990). Miller et al. (1998) provided first-order correlations between Miocene sequences and 18O records at New Jersey continental slope Site 904, directly linking 18O increases and sequence boundaries. Such comparisons provide a prima facie link between ice volume and sequences, but such first-order correlations for the Eocene and Oligocene have been lacking until this study.
St. Stephens Quarry (SSQ) in Alabama has provided one of the global reference sections for the Eocene–Oligocene transition, yet the basic relationships among sequences, sea level, temperature, and biotic events in this section have been controversial. Pioneering sequence stratigraphic studies of the SSQ outcrop were published as part of the Exxon sea-level curve (Baum and Vail, 1988; Loutit et al., 1988; Fig. 1). SSQ outcrop studies have integrated sequence stratigraphy with planktonic foraminiferal biostratigraphy (Mancini and Tew, 1991; Tew, 1992). Keigwin and Corliss (1986) identified a major (1) 18O increase in the lowermost Oligocene in the SSQ outcrop. ARCO Oil and Gas Company drilled a core hole at SSQ in 1987 that spanned the Eocene–Oligocene section and yielded an unambiguous magnetostratigraphy for the core hole reported by Miller et al. (1993). There is general agreement among studies of SSQ, both outcrop and core hole (Fig. 1), except for one critical interpretation: is there an earliest Oligocene sequence boundary and an attendant sea-level fall? One school maintains that the lowermost Oligocene Shubuta-Bumpnose formational contact at SSQ and elsewhere in Alabama and Mississippi is associated with a maximum flooding surface (MFS) of a sequence (Baum and Vail, 1988; Loutit et al., 1988; Mancini and Tew, 1991; Tew, 1992; Jaramillo and Oboh-Ikuenobe, 1999; Echols et al., 2003). A second school maintains that there is a sequence boundary at this level (Dockery, 1982) and associates the contact with a 18O increase (Keigwin and Corliss, 1986) and attendant sea-level fall (Miller et al., 1993). It has been difficult to choose between these hypotheses because of the lack of critical data sets (e.g., log data, benthic foraminiferal biofacies, and 18O). The presence or absence of a sequence boundary at the time of the global earliest Oligocene 18O increase has profound implications for the role of temperature versus ice volume at this time.
|
18O proxy for glacioeustasy. |
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSION: COMPARISONS OF...CONCLUSIONSREFERENCES CITED |
|---|
1–5 ft (0.3–1.5 m), and the core was redescribed for lithology and sequence stratigraphy. Detailed procedures for each type of analysis are provided in the following.
|
|
|
1. The Jackson Stage (approximately upper Eocene) comprises the Moodys Branch Formation, North Twistwood Creek Clay, and Yazoo Formation and its members, the Cocoa Sand, Pachuta Marl, and Shubuta Clay.
2. The Vicksburg Stage (approximately lower Oligocene) comprises the laterally equivalent Forest Hill Sand–Red Bluff–Bumpnose Formations, informal Mint Spring formation, Marianna Limestone, and the Byram Formation, including the Bucatunna Clay Member.
3. The Chickasawhay Stage (approximately mid-Oligocene) comprises the Chickasawhay Formation.
D.T. Dockery (1989, personal commun.) identified the depths of lithologic units in the SSQ core hole as reported in Miller et al. (1993). However, examination of the core hole shows that there are some minor discrepancies in the depths of the core as reported by Dockery. We reinterpreted the depths in the core to standardize measurement; this results in minor shifts (to 0.5 ft [15 cm]) in the depths of the lithologic contacts.
We used lithofacies (% sand fraction, % glauconite, relative abundance of carbonates versus siliciclastics), percent planktonic foraminifera, and benthic foraminiferal faunal data to infer paleoenvironmental changes between and within sequences (Figs. 2 and 3). Samples of 20 cm3 were soaked in a sodium metaphosphate solution to disaggregate the sediment. A few samples required heating in a sodium carbonate solution. Samples were washed through a 63 µm mesh to remove the clay and silt, and the percent sand was computed. The percent glauconite and relative abundances of carbonates versus quartz sand were visually estimated. Glauconite is associated with intervals of low sedimentation rates, usually associated with transgressive systems tracts (TST). Quartz sand increases in highstand systems tracts (HST) in many sequences globally, but here we observe this pattern only in the shallowest sequence (Moodys–North Twistwood Creek). Carbonate in the sand fraction (Fig. 2) is composed of foraminifer and mollusk fragments and shows a less predictable relationship to sequences; it is inversely proportional to the mud fraction (Fig. 2) that tends to increase in the upper HST of two sequences (i.e., HST quartz sands are absent).
Biostratigraphy
Calcareous Nannofossils
Smear slides were prepared from sediment in the interval between 271 ft (82.6 m; in the Moodys Branch Formation) and 73 ft (22.5 m; in the Marianna Formation). An additional sample was taken at 14 ft (4.27 m) from the Chickasawhay Formation. Sample density is high (1–3 ft; 0.31–0.81 m) in the Pachuta Marl to the base of the Marianna Formation interval (178–129 ft; 54.25–39.32 m) and much lower (8 ft; 2.44 m) below the Pachuta and in the bulk of the Marianna Formation. Slides were studied with a Zeiss microscope at 500–1200x magnification. Special attention was given to determine the stratigraphic distribution of the (often scarce) primary and secondary biozonal markers (Martini, 1971; Berggren et al., 1995), while the species inventory helped to determine levels where markers are reworked.
Planktonic Foraminifera
Samples for foraminiferal biostratigraphy were taken from the paleomagnetic samples (Miller et al., 1993) supplemented by additional 20 cm3 samples for a total of 83 samples from 75 to 230 ft (22.86–70.10 m). The >63 µm fraction was dry sieved into three different size fractions, which were studied for their foraminiferal assemblages: 63–125 µm, 125–250 µm, and >250 µm.
Benthic Foraminiferal Biofacies
After washing, the dried samples from SSQ were sieved to obtain the >105 µm fraction and random samples of 300 specimens were picked for quantitative benthic foraminiferal analysis. Benthic foraminifers were identified to species level using the taxonomy of Tjalsma and Lohmann (1983), Jones (1983), Bandy (1949), Enright (1969), Boersma (1984), and Charletta (1980). The data set was normalized to percentages and Q-mode factor analysis was used to document faunal variations among the samples. The factors obtained were rotated using a Varimax Factor rotation using Systat 5.2.1.
Benthic foraminiferal assemblages were quantitatively characterized and used to interpret depositional environments and to establish water-depth fluctuations (Fig. 2). We examined 39 samples and a total of 70 species were identified from 9235 specimens (see GSA Data Repository Table DR11). Four Q-mode Varimax factors were extracted from the percentage data, explaining 80% of the faunal variation (Fig. 2).
Oxygen Isotopic Records
Benthic foraminiferal stable isotope analyses from the SSQ core hole were generated in the Stable Isotope Laboratory in the Department of Geological Sciences at Rutgers University. We selected 1–5 specimens of monogeneric benthic foraminifera (Cibicidoides spp.) from each sample for analysis (Fig. 4; Table DR2 [see footnote 1]); this genus typically comprises <10% of the 300 specimens picked for census studies. Although less common than other taxa, we chose this genus because of its consistent occurrence and the fact that its isotopic calibration is well known (e.g., Katz et al., 2003). Foraminifera were reacted in phosphoric acid at 90 °C for 15 min in an automated peripheral attached to a Micromass Optima mass spectrometer. The 18O and 13C values are reported versus Vienna Peedee belemnite by analyzing NBS-19 or an internal lab standard during each automated run. Typically 8 standards are analyzed along with 32 samples; 1
precision for standards is 0.08 and 0.05 for 18O and 13C, respectively. Sample resolution in the upper part of chronozone (chron) C13r to C13n is 50 cm (40 k.y. or better using average sedimentation rates of 1.2 cm/k.y for this section).
|
18O and 13C records.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSION: COMPARISONS OF...CONCLUSIONSREFERENCES CITED |
|---|
Nonion Biofacies
Factor 2 (15% of the variance explained; Fig. 2) is characterized by the common occurrences of Nonion stavensis, Nonionella spiralis, and Nonionella spissa. It is confined to the North Twistwood Creek Clay and associated with sediments containing the highest percentage of siliciclastic sand. The Nonion biofacies is the shallowest biofacies (50 m water depth), based on the lowest percent planktonic foraminifers (3%) and the coarsest and highest abundances of siliciclastic sediments.
Hanzawaia Biofacies
Factor 1 (37% of the variance explained; Fig. 2) is characterized by the common occurrences of Hanzawaia mauricensis, Cibicidina mississippiensis, and Spiroplectammina alabamaensis. It is characteristic of the Cocoa Sand, the upper Red Bluff Formation, and upper Marianna Formation. The Hanzawaia biofacies represents deeper water than the Nonion biofacies based on lithologic and faunal (5% planktonic foraminifera) criteria. Modern Hanzawaia are typical of shelf environments (generally <100 m; Murray, 1991). In New Jersey, paleoslope modeling was used to estimate water depths of 75 ± 15 m for a middle Eocene Hanzawaia mauricensis biofacies (Olsson and Wise, 1987)), and we infer similar depths for the Hanzawaia biofacies in Alabama.
Siphonina-Cibicidoides Biofacies
Factor 3 (11% of the variance explained; Fig. 2) is characterized by the common occurrences of Cibicidoides cookei and Siphonina eocenica. This biofacies is found in the Pachuta Marl, Mint Spring formation, basal Marianna Formation, and Bumpnose Formation. The Siphonina-Cibicidoides biofacies is indicative of water depths of 100 m, based on paleoslope estimates for a similar biofacies in New Jersey (Olsson and Wise 1987).
Uvigerina Biofacies
Factor 4 (17% of the variance explained; Fig. 2) is characterized by the common occurrences of Uvigerina byramensis and Uvigerina gardinerae. This biofacies is found in the Shubuta Marl and the Bumpnose Formation. The Uvigerina biofacies contains 9% planktonic foraminifera and has an inferred water depth of 125 m. Bulimina jacksonensis, a minor constituent of this biofacies, is typically found in modern outer neritic and deeper environments (van Morkhoven et al., 1986). In the modern oceans, Uvigerina generally occurs in outer neriticbathyal (>100 m) environments and is often associated with low oxygen and/or organic-rich sediments (Miller and Lohmann 1982). Uvigerina is used frequently as a marker for the MFS (e.g., Loutit et al., 1988).
SSQ Lithostratigraphy and Sequence Stratigraphy
There is general agreement about the identification of upper Eocene–Oligocene lithostratigraphic units in Alabama and Mississippi (Baum and Vail, 1988; Loutit et al., 1988; Tew, 1992), but there is disagreement as to their sequence stratigraphy, particularly the significance of stratal surfaces at the top of the Shubuta Marl of the Yazoo Formation and the top of the Glendon Limestone (Dockery, 1982; Baum and Vail, 1988; Tew, 1992; Miller et al., 1993; Jaramillo and Oboh-Ikuenobe, 1999; Echols et al., 2003). We based our identification of sequence on surfaces noted in the core, facies shifts, sharp changes in biofacies, and gamma log changes. The sequences in the SSQ core hole are similar to those in New Jersey: both generally have thin glauconite beds at the base representing the TST, and siliciclastic sediments at the top representing the regressive HST, although high-stand sands are lacking from most of the SSQ sequences. Rapid shifts from shallow-water biofacies to deeper water biofacies are associated with sequence boundaries (Fig. 2). The pattern of deepening across sequence boundaries results from overstepping of facies due to the general absence of lowstand deposits in the coastal plain, as expected from sequence stratigraphic models (Posamentier et al., 1988). One exception to this appears to be the basal Bumpnose (see following). We assign sediments spanning the Eocene–Oligocene transition to seven sequences bracketed by sequence boundaries (Fig. 7): (1) the North Twistwood Creek–Cocoa Sand contact (chron C16n); (2) the mid-Pachuta Marl (mid-C13r–C15r); (3) the Shubuta-Bumpnose contact (latest chron C13r; the earliest Oligocene event); (4) the Mint Spring–Red Bluff contact (C13n-C12r boundary); (5) the Glendon-Byram contact (C12n-C11r); and (6) the Bucatunna-Chickasawhay contact (late C11r; the mid-Oligocene fall). The sequence boundaries at the base of the upper Moodys Branch and the top of the Chickasawhay were not evaluated. Of the six sequence boundaries, four have clear (0.5–1.0 m.y.) hiatuses associated with them (Fig. 3), and short hiatuses (<<0.5 m.y.) may be inferred for the other two (Fig. 4). Three of the sequence boundaries were recognized previously by Exxon (Fig. 1; e.g., Baum and Vail, 1988).
|
30–50 m; Fig. 2). We tentatively place an MFS at 222 ft (67.67 m) at a major gamma peak, with values decreasing (coarsening) upsection above this. The contact of the North Twistwood Creek Formation (177.0 ft; 53.95 m) with the overlying Cocoa Formation is a distinct lithologic change and unconformity (Fig. 7), with an irregular surface associated with a gamma log increase (Fig. 4).
Cocoa–Lower Pachuta Sequence
The Cocoa Member of the Yazoo Formation (172.2–177.0 ft; 52.49–53.95 m) is a sandy micrite and/or chalk in the core hole, with the percent sand never exceeding 50%. The sandsized fraction is almost exclusively carbonate and composed largely of foraminifera. The carbonate in the mud fraction is presumably primarily nannofossils in these relatively deep water (>50 m) deposits, although a contribution by other carbonate mud sources cannot be precluded. The Cocoa Member is only slightly sandier than the Pachuta Member of the Yazoo Formation (159–172.2 ft; 48.46–52.49 m) and there is minimal lithologic difference between the Cocoa and the overlying Pachuta Member in the core hole. The lower part of the Pachuta Member consists of pale yellow (2.5Y8/2) sandy micrite with very little glauconite. It is differentiated from the upper part of the Pachuta Member above 165.3 ft (50.38 m), where glauconite is common. There is an irregular contact at 165.3 ft (50.38 m) within the mid-Pachuta with 0.1 ft (3 cm) of relief separating glauconitic micrite above and sandy micrite below (Fig. 7); the sand fraction below the surface has traces of quartz and little or no glauconite. This contact at 165.3 ft (50.38 m) is associated with a distinct gamma log increase (Fig. 4), and is interpreted as a sequence boundary associated with a major hiatus (see Chronology section). The Cocoa–lower Pachuta sequence is the only sequence that lacks a distinct basal glauconitic interval. The MFS of the Cocoa–lower Pachuta sequence is tentatively placed at 174.5 ft (53.19 m), at a major gamma log peak. The Cocoa–lower Pachuta sequence was deposited in middle neritic environments (Hanzawaia biofacies; 75 m water depth).
Upper Pachuta–Shubuta Sequence
The upper part of the Pachuta Marl of the Yazoo Formation at SSQ consists of light gray (2.5Y7/2) slightly glauconitic to glauconitic sandy micrite, with glauconite generally decreasing upsection from the sequence boundary at 165.3 ft (50.38 m). Maximum water depths occur in a zone of maximum flooding from 163.7 to 157 ft (49.90–47.85 m) in a Uvigerina biofacies spanning the upper Pachuta to lower Shubuta members. The MFS is tentatively placed at 161 ft (49.07 m) in the upper part of the Pachuta, where Uvigerina reaches maximum abundance. The Shubuta Member of the Yazoo Formation (154.1–159 ft; 46.97–48.46 m) consists of a clean, uniform, cream-white (5Y8/1) marly micrite with virtually no sand and silt. A subtle but distinct erosional surface occurs at a lithologic contact at 154.1 ft (46.97 m) (Fig. 7). Above the surface is a glauconitic micrite and/or chalk assigned to the Bumpnose Formation. The Shubuta was deposited in middle-outer neritic environments, with a distinct, abrupt shallowing upsection from a Uvigerina biofacies (120 m water depth) to a Hanzawaia biofacies (75 m water depth) at the top (154.6 ft; 47.12 m). There is a sharp shift in biofacies across the Shubuta-Bumpnose contact from the shallower water Hanzawaia biofacies to the deeper water Siphonina-Uvigerina biofacies.
We follow Miller et al. (1993) in placing a sequence boundary at the Shubuta-Bumpnose contact. In contrast, others (Baum and Vail, 1988; Loutit et al., 1988; Pasley and Hazel, 1990; Mancini and Tew, 1991; Tew, 1992; Jaramillo and Oboh-Ikuenobe, 1999; Echols et al., 2003) have interpreted the Cocoa Sand through the Red Bluff Formation as a single sequence (Fig. 1), the Shubuta Marl and Bumpnose Formation being separated by an MFS. They interpreted the surface separating the two formations as a starvation surface caused by sediments being trapped onshore. However, we show that: (1) a benthic foraminiferal biofacies shift occurs abruptly across the 154.1 ft (46.97 m) contact; (2) the contact is an erosional surface (Fig. 7); and (3) the amount of glauconite increases 2 ft (60 cm) above the contact. The combination of an abrupt biofacies shift and increase in glauconite is typical of other sequence boundaries in Alabama and New Jersey. Thus, we interpret the Shubuta-Bumpnose contact as a sequence-bounding unconformity, not an MFS. This agrees with Dockery's (1982) observation of lowermost Oligocene channels incised into the top of the Shubuta Member in Mississippi that are filled with the Forest Hills Sand, a lateral correlative of the Red Bluff Formation. Part of the controversy in the interpretation of the Shubuta–Red Bluff contact as a sequence boundary probably derives from the fact that the upper Pachuta–Shubuta sequence is a deep-water Uvigerina biofacies, except for the abrupt shallowing at the top. Although the deep-water Uvigerina interval was noted in previous biofacies studies of the SSQ outcrop (Loutit et al., 1988), the abrupt shallowing was missed due to sampling limitations.
Bumpnose–Red Bluff Sequence
The overlying lower Oligocene sequence consists of the Bumpnose and the Red Bluff Formations. The Bumpnose Formation in the SSQ core hole occurs from 140.2 to 154.1 ft (42.73–46.97 m), where it consists of a slightly marly, slightly glauconitic to glauconitic, light greenish-gray (5GY8/1) micrite and/or chalk. Distinctly glauconitic zones occur at 151.3–154.1, 149.0–149.6, and 146.6–147.0 ft (46.12–46.97 m, 45.42–45.60 m, and 44.68–44.81 m) and glauconite generally decreases upsection above this in the lower Bumpnose Formation, beginning at 146 ft (44.50 m). An MFS associated with a peak abundance of Uvigerina could tentatively be placed at 152.1 ft (46.36 m); however, because common glauconite continues upsection to 149.0 ft (45.42 m) in the Siphonina biofacies, we place the MFS at this level, with a TST from 2 ft (60 cm) above the base of the sequence to 149.0 ft (45.42 m) (the lower 2 ft may be a lowstand systems tract [LST], as discussed in the following). A change to the shallower Hanzawaia biofacies above 147.5 ft (44.96 m) indicates the regressive HST. The Red Bluff Formation in the SSQ core hole occurs from 133.0 to 140.2 ft (40.54–42.73 m), where it consists of brown clay with scattered pyrite nodules (15 cm spacing) that is siltier near the base. The formation was deposited in middle neritic (75 m) water depths associated with the Hanzawaia biofacies. This is a typical sequence in that it contains glauconite near the base with carbonate-rich sediments above. The benthic Uvigerina biofacies indicates that the sequence is deepest near its base and progressively shallows upward.
Mint Spring–Marianna–Glendon Sequence
The Mint Spring formation and Marianna and Glendon Formations comprise a thick (133.0–51 ft; 40.54–15.54 m) sequence at SSQ. The Mint Spring formation in the SSQ core hole occurs from 131.2 to 133.0 ft (39.99–40.54 m), where it consists of a glauconitic micrite and/or chalk. The base of the Mint Spring formation (133.0 ft; 40.54 m) is a sequence boundary consisting of a sharp, burrowed lithologic contact with rip-up clasts of the underlying brown clay and white (?kaolinite) clay clast and glauconite and shell concentrations at the base of the formation. There is a consensus that the contact at base of the Mint Spring formation is a sequence boundary (Baum and Vail, 1988; Mancini and Tew, 1991; Tew, 1992; Jaramillo and Oboh-Ikuenobe, 1999; Echols et al., 2003). The Mint Spring formation was deposited in middle-outer neritic environments (100 m water depth) associated with the Siphonina biofacies.
The Marianna Formation in the SSQ core hole (59.5–131.2 ft; 18.14–39.99 m) is a heavily bioturbated, slightly marly, white (5Y8/1) micrite with scattered pyrite and opaque heavy minerals. The formation is generally friable, shows no evidence of cementation across grains, and is best termed a chalk in the diagenetic series oozechalk-limestone, except for a shelly limestone from 66.7 to 67.3 ft (20.33–20.51 m). The base of the Marianna Formation is placed at a change to a glauconitic micrite at 131.2 ft (39.99 m). We interpret this contact as the MFS of the Mint Spring–Marianna sequence; a mixed Uvigerina-Siphonina biofaces is associated with the MFS (125 m water depth). Above this, the Marianna Formation shallows upsection from the Siphonina (100 m water depth) to the Hanzawaia biofacies (75 m water depth).
The interpretation of the Glendon Limestone (51–59.5 ft; 15.54–18.14 m) is problematic. This unit in the SSQ core hole is a limestone consisting of bluish-gray (10B5/1) to pale yellow (2.5Y8/3), heavily indurated, shell-rich micrite to micritic shell hash. Tew (1992) regarded the Glendon Limestone as the highstand deposit of the underlying sequence and placed a sequence boundary at its top. In contrast, Baum and Vail (1988) interpreted it as the lowstand deposits of the overlying sequence. The Glendon Limestone did not disaggregate and benthic foraminifers could not be separated. Thus, we cannot rely on foraminifera to place constraints on water-depth changes and placement of the sequence boundary. A major facies shift occurs at the top of the Glendon Formation in the core hole, associated with a large gamma log increase (Fig. 2), and we follow Tew (1992) in placing a sequence boundary at the top of the Glendon Formation. Integration of magnetobiostratigraphy indicates a significant hiatus (0.8–1.3 m.y.) between the Glendon and Byram Formations (Fig. 3; see Chronology section), further supporting our interpretation.
Byram–Bucatunna Sequence
The Byram–Bucatunna sequence consists of the Byram (undifferentiated) Formation (49–51 ft; 14.94–15.54 m) and Bucatunna Member (22.5–49 ft; 6.86–14.94 m). The basal sequence boundary is an abrupt contact at 51 ft (15.54 m) with the indurated Glendon Limestone below and a poorly sorted shell hash above. There is a very large gamma log increase associated with the contact (Fig. 2). The Byram Formation is a grayish-brown (10YR5/2) shelly silty clay and is more fossiliferous than the overlying Bucatunna Member, which is a grayish-brown micaceous, lignitic silty clay. This is a fine-grained sequence with a minor sand fraction dominated by the Hanzawaia biofacies. Cibicidina and Astigerina subacuta dominate a sample near the top of the Bucatunna Member. We interpret this as indicating shallowing upsection, although an MFS was not identified.
Chickasawhay Sequence
The Chickasawhay sequence (6.0–22.5 ft; 1.83–6.86 m) has a sharp basal sequence boundary with white (10YR8/1) partly indurated silty clayey foraminiferal quartz sand above and a grayish-brown micaceous, lignitic silty clay of the Bucatunna Member below associated with a sharp gamma log decrease (Fig. 2). The base of the Chickasawhay Formation is a major mid-Oligocene disconformity, the global correlation of which was discussed in detail by Miller et al. (1993). The top of the Chickasawhay Formation is heavily weathered between 6 and 10 ft (1.83–3.05 m). The depositional environment of the Chickasawhay Formation and sequence is middle neritic, based on qualitative examination of benthic foraminifera (foraminiferal recovery from the formation was insufficient for quantitative evaluation).
Planktonic Foraminiferal Biostratigraphy
We examined 83 samples at SSQ from 75 to 230 ft (22.86–70.10 m) for planktonic foraminiferal biostratigraphic analysis. Planktonic foraminifera are extremely varied in terms of their abundance, diversity, and preservation. Samples range from full planktonic foraminiferal assemblages consistent with open ocean environments, to almost monospecific assemblages of tenuitellids and Dipsidripella danvillensis. Preservation of planktonic foraminifera appears to be facies dependent, and ranges from extremely well preserved, glassy specimens, to recrystallized (white) specimens.
Planktonic foraminiferal assemblages are dominated by dentoglobigerinids (D. galavisi, D. pseudovenezuelana, D. tripartita) and Turborotalia ampliapertura. The small size fractions (<125 µm) contain common Pseudohastigerina and tenuitellids. Globigerinatheka spp. were absent from all samples, and possibly excluded due to shallow water depths. Therefore, we were unable to constrain the highest occurrence (HO) of G. index at SSQ and the zone E15-E16 boundary.
The Eocene-Oligocene boundary is characterized by the extinction of family Hantkenindae at 33.7 Ma in the global stratotype at Massignano, Italy (Premoli Silva and Jenkins, 1993; Berggren and Pearson, 2005; note that this level was assigned an astronomical age of 33.714 by Jovane et al., 2006). Hantkenina and Cribrohantkenina are extremely rare in the SSQ core hole, and we were not able to confidently place the Eocene-Oligocene boundary using the HO Hantkenina, which occurs in the core hole in the upper Pachuta Member at 163 ft (49.68 m) (see also Miller et al., 1993). Mancini (1979) reported the HO of Hantkenina spp. at the top of the Shubuta Member in outcrop at SSQ (equivalent to 154.1 ft [46.97 m] in the core hole). However, Bybell and Poore (1983) reported that specimens of Hantkenina in the upper Shubuta–Bumpnose are reworked based on analysis of nannofossils included in the foraminiferal tests, and Keller (1985) placed the HO of Hantkenina spp. in the basal Shubuta Member (equivalent to 158 ft (48.16 m) in the core hole, although there are uncertainties in core hole–outcrop correlations). The Eocene-Oligocene boundary is preceded by the extinction of the Turborotalia cerroazulensis group at 33.765 Ma (Berggren and Pearson, 2005). Turborotalia vary in abundance at SSQ; we find the HO of the T. cerroazulensis group at 162.0 ft (49.38 m), associated with the precursor 18O shift. Based on our age-depth plot anchored on the HO of T. cerroazulensis and the base of chronozone C13n, we predict that Hantkenina spp. should range to 157.5 ft (48.01 m) (lower Shubuta member), in agreement with Keller (1985).
Cassigerinella chipolensis (Fig. 6J) is rare. While the lowest occurrence (LO) of this species is not well calibrated to the time scale, the presence of C. chipolensis from 159 ft (48.46 m) at the base of the Shubuta Member suggests placement of the Eocene-Oligocene boundary (zone E16-O1) near this level (Miller et al., 1993).
The extinction of Hantkenina is also associated with the HO of Pseudohastigerina in the >125 µm size fraction (Nocchi et al., 1986). We find the HO of large Pseudohastigerina at 155.4 ft (47.37 m), approximating the top of planktonic foraminiferal biozone E16 (Berggren and Pearson, 2005) and the Eocene-Oligocene boundary.
This discussion highlights the problems in using biostratigraphic markers for very high resolution (100 k.y.) correlations in near-shore sections. Although the Eocene-Oligocene boundary is firmly placed on chron C13n.12 at Massignano stratotype, it is not possible to interpolate the precise position of this chron at SSQ because two sequence boundaries occur within it and biostratigraphic criteria must be used. The Eocene-Oligocene boundary could be placed using biostratigraphic criteria at SSQ at 163 ft (HO of Hantkenina, which is certainly depressed at this location), 162 ft (49.38 m) (HO of T. cerroazulensis group, calibrated as 65 k. y. older than the boundary; Berggren and Pearson, 2005), 159 ft (48.46 m; LO of C. chipolensis, poorly calibrated to time scale), 157.5 ft (48.01 m) (predicted HO of Hantkenina in the core hole and the HO of common, unreworked Hantkenina in the outcrop; Keller, 1985), or 155.5 (47.40 m; HO of large Pseudohastigerina spp.). This difference of 5.5 ft (1.68 m) only represents 140 k.y. based on the sedimentation rate of 12 m/m.y. (Fig. 4). Similarly, the uncertainty in the HO of Hantkenina due to reworking (i.e., reported to an equivalent position of 154.1 ft [46.97 m] in the outcrop versus 163 ft [49.68 m] observed and 157.5 ft [48.01 m] predicted) amounts to only 100–200 k.y. This uncertainly is reflected in the stairstep placement of the Eocene-Oligocene boundary (Fig. 4). We favor placing the boundary in the lower Shubuta Marl (157.5 ft) at the predicted HO of Hantkenina.
Insights into foraminiferal preservation can be gained from SSQ. Planktonic foraminifera preservation tracks sequence boundaries and better preservation is associated with intervals of higher clay content. Within the North Twistwood Creek Clay (177–231.8 ft; 53.95–70.65 m), foraminifera are rare but extremely well preserved (Figs. 6). Foraminifera from the Cocoa–lower Pachuta sequence (177–163.7 ft; 53.95–49.90 m) are recrystallized and appear white under the light microscope. From 163.7 to 150.2 ft (49.90–45.78 m), foraminifera from the Shubuta Marl and lower Bumpnose–Red Bluff sequence are glassy (Fig. 6), except for a thin interval of recrystallization from 156.5 to 154.1 ft (47.70–46.97 m). Preservation deteriorates and foraminifera are recrystallized from 150.2 to 140.2 ft (45.78–42.73 m), with glassy specimens from 140.2 to 131.2 ft (42.73–39.99 m).
Nannofossil Biostratigraphy
The abundance and preservation of coccoliths and the diversity of their assemblages vary considerably in the section. There is a striking contrast between scarce to common, moderately to poorly (micritization) preserved assemblages between 271.1 and 166.8 ft (82.63–51.42 m; Moodys Branch to lower Pachuta) and the abundant, high-diversity, well-preserved assemblages between 163 and 142 ft (49.68–43.28 m; upper Pachuta to Bumpnose). The lower part of the section (i.e., the lower Moodys Branch Formation) is well anchored by the occurrence of a diverse, typical Bartonian (zone NP17) assemblage at 269.5 ft (82.14 m) with Campylosphaera dela (HO in upper zone NP17), Discoaster barbadiensis, D. saipanensis, Ericsonia formosa, Reticulofenestra reticulata, R. umbilicus, and Sphenolithus obtusus (restricted to zone NP17). The interval between 268.8 and 206 ft (81.93–62.79 m; lower Moodys Branch through North Twistwood Creek Clay) comprises an alternation of levels with abundant and well-preserved coccoliths and micritic levels with rare, heavily recrystallized nannofossils. Even where preservation is best, this interval lacks biostratigraphic markers (e.g., Chiasmolithus oamaruensis, Helicosphaera reticulata, and Isthmolithus recurvus). It is assigned to zone NP18 because of the lowest occurrence (LO) of Isthmolithus recurvus at 178 ft (54.25 m; just below the North Twistwood Creek Clay–Cocoa Sand formational contact), which defines the base of zone NP19–20. The Cocoa–lower Pachuta sequence (176–166.8 ft; 53.64–50.84 m) yields very impoverished assemblages because of extremely poor preservation. However, Reticulofenestra reticulata and Ericsonia formosa occur consistently and abundantly through the sequence, with occasional, generally overgrown, rosette-shaped discoasters, including D. barbadiensis (at 168 ft; 51.21 m) and D. saipanensis (at 166.8 ft; 50.84 m). Reticulofenestra reticulata ranges from middle Eocene (upper zone NP16) up to a level within zone NP19–20, and can be used to subdivide that zone. Its common occurrence in the Cocoa–lower Pachuta sequence indicates zone NP19–20, and its HO at the unconformable sequence boundary at 165.3 ft (50.38 m) indicates that the zone is truncated. The upper Pachuta to Bumpnose succession (comprising the upper Pachuta–Shubuta and Bumpnose–Red Bluff partim sequences) up to 142 ft (43.28 m) yields abundant and well-preserved assemblages of zone NP21. In this interval, the fine fraction is mostly composed of coccoliths, with little or no detrital particles (in contrast to the underlying sediments). Preservation and calcareous nannofossil abundance decrease markedly above 144 ft (43.89 m). Zone NP21 extends up to 110.6 ft (33.71 m). The interval between 96 ft and 14 ft (29.26–4.27 m) belongs to zone NP 23.
Chronology
In view of controversies surrounding the sequence stratigraphic framework at St. Stephens Quarry, Miller et al. (1993) conservatively estimated continuous sedimentation between magnetochron boundaries. We reevaluated the SSQ chronology of Miller et al. (1993) using new biostratigraphic data (Figs. 3 and 4) indicating that there are hiatuses at many sequence boundaries. Our age control varies from ±0.1 to ±0.5 m.y. based on integration of magnetobiostratigraphy; individual biostratigraphic datum levels typically have errors on the order of 0.5 m.y., but the identification of magnetochrons and integration with biostratigraphy improves age control to as fine as ±0.1 m.y. Our chronologic control is sufficient to generally constrain sedimentation rates as 13–24 m/m.y. within individual sequences, but is not sufficient to resolve detailed sedimentation rate changes within sequences as they coarsen (shallow) upsection above MFSs. We assume that sedimentation rates were otherwise constant within sequences between these control points (Figs. 3 and 4); although this is certainly not correct, it results in age uncertainties that are relatively minor (<0.5 m.y.)
Integration of the biostratigraphic data with magnetostratigraphy is essential for the temporal interpretation of sections and the determination of the duration of hiatuses at sequence boundaries (Aubry, 1995, 1998). Integration of planktonic biostratigraphy and magnetostratigraphy within a sequence stratigraphic framework allows quantification of hiatuses and identification of unrepresented magnetochrons. Age assignments derived from planktonic foraminifera and the calcareous nannofossils are generally consistent (Figs. 3 and 4), although nannoplankton provide finer resolution for this interval (Aubry, 1995, 1998). The magnetostratigraphic record at SSQ has been previously interpreted to be complete from chron C16r through chron C11n (Miller et al., 1993). This interpretation is not supported by biostratigraphy that reveals three previously undetected hiatuses at the basal upper Moodys Branch–North Twistwood Creek (36.7–?37.5 Ma; Fig. 3), Cocoa–lower Pachuta (35.4–35.9 Ma), and upper Pachuta–Shubuta (33.9–35.0 Ma) sequence boundaries (Fig. 4).
Nannofossil biostratigraphy indicates the concatenation of chronozones C13r and C15r and C16n1 and C16n2 (i.e., chrons C15n and C16n1r are not represented). A magnetozone from 206 to 239.5 ft (62.79–73.0 m) was interpreted as chron C16n (Miller et al., 1993). The chron C16n-C15r reversal is associated with mid-biochron NP19–20 (Berggren et al., 1995); thus, if this normal polarity interval represented chron C16n, it should belong to zone NP19–20. This is not the case; the base of the zone occurs at 178 ft (54.25 m). We reinterpret the interval from 183 to 222.5 ft (55.78 to 67.82 m) as chronozone C16r, which correlates with zone NP18, and place a 0.5 m.y. hiatus at the base of the Cocoa–lower Pachuta sequence (Figs. 3 and 4); two normal points within this interval (206 and 209.5 ft; 62.79–63.86 m) are interpreted as normally overprinted. As a result, the normal from 226 to 239.5 ft (68.88–73.0 m) is interpreted as chron C17, not C16 n1. The thin normal polarity magnetozone between 176 and 178 ft (53.64–54.25 m) was interpreted as chron C15n (Miller et al., 1993), which is associated with the upper part of zone NP19–20. This thin magnetozone is now interpreted as the concatenation of chrons C16n1 and C16n2. Chronozone C13r of Miller et al. (1993) is now interpreted as a concatenation of chrons C13r and C15r, with a hiatus longer than 1 m.y. (33.9–35.0 Ma) associated with the basal upper Pachuta–Shubuta sequence boundary.
Planktonic biostratigraphy support the interpretation of the identification of chronozones C13r through C11n (Fig. 3). The normal magnetozone between 133.8 and 151 ft (40.78–46.02 m; Bumpnose and Red Bluff) is associated with zone NP21 and can be confidently interpreted as chron C13n. The reversed polarity interval between 152.1 ft and 165.3 ft (50.38 m) represents the late part of chron C13r, and the upper surface of the sequence boundary at 165.3 ft (50.38 m) (lower Pachuta–upper Pachuta contact) is ca. 33.9 Ma. The interpretation of the normal polarity magnetozone between 50.8 and 50 ft (15.48–15.24 m) as chron C12n is compatible with the NP23 zonal and the P19 zonal assignment. However, we note the thinness of this interval and its proximity to the Glendon-Byram Formation contact, implying that only part of the chron is recorded. The age-depth diagram (Fig. 3) indicates that there is a >1 m.y. hiatus associated with this disconformity.
Integrating the calcareous biostratigraphy and magnetostratigraphy within the sequence stratigraphic framework, we propose a revised chronology for the stratigraphy in the SSQ core hole (Figs. 3 and 4). Using the age-depth diagrams, we derive ages for the sequences as follows.
1. The upper Moodys Branch–North Twistwood Creek sequence was deposited between 36.7 and 35.9 Ma (Fig. 3; chron C17n partim to C16n2 partim) (the lower Moodys Branch is part of an older sequence).
2. The Cocoa–lower Pachuta sequence was deposited from 35.4 to 35.0 Ma (Figs. 3 and 4). Thehiatusacrossthe165.3 ft(50.38 m)sequence boundary is estimated as 35.0–33.9 Ma.
3. The upper Pachuta–Shubuta sequence spans the Eocene-Oligocene boundary as it is defined at Massignano as the level associated with the HO of Hantkenina spp. (Pomerol and Premoli Silva, 1988; Premoli Silva and Jenkins, 1993). Considerable controversy has attended the placement of this boundary at SSQ because of uncertainties in biostratigraphy, primarily due to reworking (see Planktonic Foraminiferal Biostratigraphy section). The upper Pachuta–Shubuta sequence is entirely in chronozone C13r (Figs. 3 and 4), and we prefer a placement of the Eocene-Oligocene boundary in the middle of the Shubuta at 157.5 ft (48.01 m) immediately above the first occurrence (FO) of Cassigerinella chipolensis at 158.0 ft (48.16 m) (see Planktonic Foraminiferal Biostratigraphy section). Our best estimate for the age of the sequence is ca. 33.9–33.6 Ma (i.e., using a sedimentation rate of 1.0 cm/k.y. between the Turoborotalia cerroazulensis, C. chipolensis, and chron C13n datum levels; Fig. 4).
4. The Bumpnose–Red Bluff sequence was deposited between 33.6 and 33.0 Ma (Figs. 3 and 4), consistent with its correlation to chronozone C13n and uppermost C13r. The hiatus associated with the basal sequence boundary of the Bumpnose–Red Bluff sequence (i.e., the Shubuta-Red Bluff contact) is not discernable, although it appears to be much less than 100 k. y. based on extrapolation of sedimentation rates (e.g., 33.59–33.62 Ma; Fig. 4).
5. The Mint Spring–Marianna–Glendon sequence was deposited between 33.0 and 31.7 Ma (Figs. 3 and 4), consistent with its being rapidly deposited (1.8 cm/k.y.) in chron C12r. The younger age limit is uncertain and is based on the assumption that the HO of Pseudohastigerina spp. is in situ and not reworked as suggested by Miller et al. (1993), but it must be older than ca. 31 Ma (i.e., it is in chron C12r). Chronozone C12r partim is found in the Glendon Limestone (Fig. 2), as is the underlying Marianna Formation (Miller et al., 1993), whereas chronozone C12n-11r is found in the overlying Byram-Bucatunna Formations, suggesting placement of a sequence boundary at the top of the Glendon Limestone (Fig. 3).
6. The Byram–Bucatunna sequence was deposited from 30.3 to 30.5 Ma (chrons C11r–latest C12; Fig. 3).
7. The base of the Chickasawhay Formation is a major disconformity (Fig. 3) that represents the major mid-Oligocene unconformity of Vail et al. (1977). This surface was dated precisely for the first time at SSQ by Miller et al. (1993) as late chron C11r (ca. 30.1 Ma) and correlated with the oxygen isotope increase associated with Oi2 (Miller et al., 1993).
Stable Isotopes
An 0.9 18O increase is associated with the C13r-13n boundary immediately above the HO of Eocene microplankton, and thus can be confidently correlated to the global Oi1 isotopic increase and 18O maximum (Miller et al., 1991) using magnetobiostratigraphic criteria (Fig. 8). The 18O increase begins in the core hole in association with the Shubuta-Bumpnose contact, an erosional surface, and a biofacies shift (i.e., the increase occurs between 155.5 and 152 ft [47.39–46.32 m]; Fig. 4); it culminates with maximum 18O values in the lowermost Bumpnose Formation at 152 ft (46.32 m). This requires that the sequence boundary was eroded very early in the sea-level fall and that deposition resumed during the later stages of the fall. This lag interval represents 50 k.y., using the best estimate of sedimentation rates (1.1 cm/k.y.) for this section. Although water depth appears to increase across the sequence boundary (75 m below to 100 m above), we lack faunal samples from the critical 2 ft (60 cm) immediately above the sequence boundary. This interval lacks glauconite and is interpreted as an LST deposited during a time when sea level was falling globally (Fig. 9); glauconite and planktonic foraminifera increase immediately above this section at 152 ft (46.32 m) at the time of the lowest global sea level (i.e., associated with peak 18O values), marking the turning point and the beginning of the TST (see Fig. 9 and Discussion).
|
|
18O values (i.e., the base of zone Oi1; Miller et al., 1991) occurs in reversely magnetized sediments immediately below the transition to chronozone C13n (Fig. 4). At Pacific Site 1218, the maximum 18O values occur in a section correlated to chron C13n based on correlations from the magnetostratigraphy at nearby Site 1219 using percent carbonate data (Pälike et al., 2005; Lanci et al., 2005). At Site 522, maximum 18O values measured by Zachos et al. (1996) occur in normally magnetized sediments at 133.13 m (Tauxe and Hartl, 1997), although there is a coring gap from this sample to 134.0 m that spans the upper part of the 18O increase and the polarity transition. Although it is clear that the 18O increase associated with Oi1 begins in latest chron C13r, it is not clear when the actual maximum value occurs, although it does appear to be closely associated with the polarity transition.
A precursor 18O increase of 0.5 occurs within the middle of chron C13r (Fig. 4). The precursor 18O increase is associated with the HO of the T. cerroazulensis group at 162.0 ft (49.38 m) and is thus latest Eocene (zone E16). It correlates with a similar event at Pacific Site 1218 (Fig. 8; Coxall et al., 2005). A 0.5 increase is readily resolvable given that precision is 0.08 for 18O and that natural variability is generally 0.1; the similar patterns observed both in the Pacific and at SSQ document that this is a real increase, and that it was geographically widespread.
We have sufficient isotope data across two MFS (161 ft [49.07 m] in the upper Pachuta–Shubuta sequence and 149 ft [45.42 m] in the Bumpnose–Red Bluff sequence) to evaluate their relationships to temperature and ice volume. The 149 ft (45.42 m) surface is associated with decreases in 18O values, which is unexpected because it implies warmer temperatures in the deepest water depths. However, comparison with deep Pacific Site 1218 (Fig. 8) suggests that these 18O decreases may be global and associated with decreases in ice volume and eustatic rises. As noted above, the placement of the MFS in the upper Pachuta&#
