Fulthorpe and Austin (1998), Fulthorpe et al. (1999, 2000), Steckler et al. (1999), Metzger et al. (2000), and Delius et al. (2001) all focus on the sequence stratigraphic context of the Leg 174A boreholes and on the correlation of multichannel seismic (MCS) and borehole data. Pekar et al. (2001, in press b) deal with the development of Oligocene sequences best expressed beneath the New Jersey Coastal Plain and preserved only as highly condensed correlatives at slope Site 1073. These strata are geometrically similar to upper Miocene to Pleistocene sequences studied at shelf Sites 1071 and 1072 and provide insights about the manner in which sequences and their intervening boundaries have developed.
The overriding issue addressed by Fulthorpe and colleagues concerns the three-dimensional geometry of Miocene sequences and particularly the extent to which the shallow shelf was subaerially exposed during sea level lowstands. Working with abundant but low-resolution industry data acquired in the 1970s, Fulthorpe and Austin (1998) concluded that Miocene rivers generally did not discharge at clinoform breakpoints (paleoshelf edges) at any point in a sedimentary cycle because evidence for canyon breaching of shelf edges is sparse and ambiguous, in marked contrast with Pleistocene counterparts. A simple backstripping reconstruction of paleoelevations for middle to upper Miocene clinoform breakpoints compared with the eustatic curve of Haq et al. (1987) was used to argue that with the exception of a single mid- to upper Miocene sequence boundary (designated m1c), minimum paleowater depths were 80 to 100 m. Their interpretation depended critically on the amplitude estimates of Haq et al. (1987), estimates that have long been regarded as suspect (Christie-Blick et al., 1990) and that in the case of both the Oligocene and Miocene are now demonstrably too large (Kominz et al., 1998; Kominz and Pekar, 2001). In addition, changes in paleowater depth scale as ~1.48 times eustatic variation, owing to water loading and unloading as sea level rises and falls (apparent sea level change of Pekar et al., in press a). A testable corollary of the comparison with Haq et al. (1987), correcting for this loading effect, is that minimum water depths would have been in excess of 120 m. In a comprehensive two-dimensional backstripping study of the Oligocene to middle Miocene of the New Jersey margin, Steckler et al. (1999) concluded that clinoform rollovers (paleoshelf edges) at sequence boundaries correspond to paleowater depths of ~60-130 m. Only the long-term eustatic estimates of Kominz (1984) were used to obtain this result, permitting minimum paleowater depths to have been several tens of meters shallower.
Working with MCS data acquired aboard the Oceanus in 1995 (Oc270) in preparation for Leg 174A drilling, Fulthorpe et al. (1999) interpreted fluvial channels immediately landward of a number of Miocene clinoform breakpoints. On the strength of the new data, these authors concluded that the entire shallow shelf was subaerially exposed during the development of at least some of these mid- to upper Miocene sequence boundaries and that the marked progradation characterizing this interval resulted from direct fluvial delivery of sediment to the shelf edge during sea level lowstands. This significant change in perspective is consistent with the discovery in the course of Leg 174A of probable estuarine or lagoonal sediments only 3 km landward of the paleoshelf edge for upper Miocene sequence boundary m0.5(s) (Site 1071) (Austin, Christie-Blick, Malone, et al., 1998). However, Oligocene borehole data from the New Jersey Coastal Plain show that complete exposure of the shelf is not necessary to explain the observed progradation. Amplitudes of Oligocene eustatic change are estimated as 10 to 50 m for spans of 1-2 m.y. (Kominz and Pekar, 2001), a range that is comparable to the less well constrained Miocene estimate of at least 20-30 m (Kominz et al., 1998). Yet, with the exception of two lower Oligocene sequence boundaries, for which the paleoshelf edge may have become subaerially exposed, minimum paleowater depths were 20 ± 10 m (Pekar et al., 2001, in press b). Progradation and offlap development began close to eustatic maxima (water depths < 85 ± 25 m) and continued to eustatic minima without the development of onlapping lowstand systems tracts, in spite of rates of eustatic fall that at times greatly exceeded the rate of tectonic subsidence. Although ~65%-80% of the shallow shelf that had been flooded during each eustatic rise became subaerially exposed during the subsequent fall, virtually all of the offlap developed in a marine setting and primarily as a result of sediment bypassing rather than erosion (cf. erosional interpretation of Pleistocene sequence boundary pp3(s) by McCarthy et al., in press). Pekar et al. (2001, in press b) reasoned that this unexpected pattern is due to a trade-off between riverine input of terrigenous sediment and marine dispersal both along and across the flooded portion of the shelf. For comparatively low sediment fluxes, the shallow shelf is inferred to have been drowned during times of eustatic rise and sediment temporarily sequestered in the coastal environment. Sea level lowering was expressed on the shallow shelf mainly by a reduction of water depth, with the efficiency of sediment dispersal increasing as waves began to impinge upon the seafloor. These data suggest that whereas Oligocene to Pleistocene offlap surfaces at the New Jersey margin may have been modified by subaerial exposure and transgressive ravinement, particularly in the Pleistocene, they are not fundamentally due to "forced regression," the currently popular idea for discontinuity development through seaward movement of the shoreline without sediment accumulation (Posamentier and Morris 2000; e.g., Hesselbo and Huggett, 2001; McCarthy et al., in press).
An issue highlighted by the work of Pekar and others concerns the importance of distinguishing between systems tracts as objectively recognizable stratigraphic elements within sequences and the eustatic changes that are in some cases responsible for systems tract development (see also Christie-Blick and Driscoll, 1995). Oligocene through Pleistocene sequences at the New Jersey margin are composed primarily of transgressive and highstand systems tracts. Transgressive tracts overlie sequence boundaries and are characterized by overall marine transgression (although not all of the sediments included are marine). Highstand tracts underlie sequence boundaries and are characterized by progradation with varying degrees of offlap. Lowstand systems tracts, where present, overlie sequence boundaries (with onlap), underlie transgressive deposits, and are recognizable on the basis of evidence for continued progradation and regression of the shoreline. They are also poorly developed to absent at the New Jersey continental shelf and slope. (See Metzger et al., 2000, in part summarized below, for a discussion of exceptions in the Miocene, and Pekar et al., 2001, in press b, for an explanation.) Highstand and lowstand systems tracts are specifically not stratigraphic units accumulating when sea level was either high or low. Placement of sequence boundaries between "lowstand" and transgressive systems tracts (e.g., fig. 9 of Hesselbo and Huggett, 2001; fig. 7 of Savrda et al., 2001a; fig. 3 of McCarthy et al., in press) is inconsistent with the original definitions of these terms and/or with the principle of using offlap-onlap geometry to recognize stratigraphic discontinuities (Christie-Blick and Driscoll, 1995; Pekar et al., in press b; cf. Helland-Hansen and Gjelberg, 1994; Posamentier and Morris, 2000).
Fulthorpe et al. (2000) focused on the morphology and distribution of Miocene slope incisions and showed that many connect with shelf sequence boundaries, as is generally assumed. These authors pointed out, however, that whereas the depth of incision decreases downslope from clinoform breakpoints, it also decreases upslope from clinoform toes. They concluded that the development of canyon systems relates not only to eustasy but also to such factors as the efficiency of downslope sediment transport, sediment supply, grain size, and perhaps slope collapse related to fluid escape.
Metzger et al. (2000) integrated Oc270 MCS data, wireline logs, and core data from Sites 1071 and 1072 on the modern continental shelf. Following the scheme established by Austin, Christie-Blick, Malone, et al. (1998), they identified two sequences in the upper Miocene to Pleistocene interval, bounded by surfaces m0.5(s) (>8.6 Ma below the surface; see discussion of age below), pp4(s) (1.7 to 1.4 Ma; see Wei, Chap. 5, this volume, and discussion below), and pp3(s) (0.46 to 0.25 Ma). In the vicinity of the shelf drill sites, the uppermost surface, pp3(s), merges with a younger surface, pp2(s), and might better be referred to that surface in the area studied by Metzger et al. (see Katz et al., in press). The two sequences defined by the three surfaces are similar, although they represent vastly different spans of time (>8 m.y. and ~1.5 to 1 m.y., respectively). The older sequence (upper Miocene-Pliocene) for the most part deepens upward landward of the rollover in the underlying sequence boundary but is strongly progradational (upward shoaling) seaward of that rollover. The younger sequence (Pleistocene) deepens abruptly at the base, shoals upward, and then deepens again. Seismic geometry indicates that this sequence is also progradational seaward of the rollover in its underlying sequence boundary. Both sequences are clearly associated with higher-order cyclicity and variations in paleowater depth. Contrasts with older Miocene sequences, some of which contain demonstrable lowstand deposits, were ascribed by Metzger et al. (2000) to different long-term eustatic patterns upon which short-term fluctuations are superimposed. In light of the results of Pekar and colleagues (2001, in press b) in their studies of the New Jersey Oligocene and the apparent existence of many more sequence boundaries in the same Miocene to Pleistocene interval in the Bahamas (Shipboard Scientific Party, 1997), this interpretation may or may not be correct. Also unresolved is the rather different sequence stratigraphic interpretation of coeval strata in different portions of the Oc270 MCS data by Fulthorpe et al. (1999, 2000) compared with Metzger et al. (2000). Fulthorpe and colleagues mapped several additional sequence boundaries in the middle to upper Miocene.
Delius et al. (2001) used downhole logging and magnetic susceptibility data from shelf Sites 1071 and 1072, as well as slope Site 1073, to fill in gaps in core recovery, which was poor to nonexistent in sandy intervals on the shelf. Peaks in gamma ray and potassium logs were interpreted to indicate flooding surfaces; sequence boundaries are less easy to interpret from logs because their location depends on stratal geometry and not on any particular lithic character and hence log signature, either above or below a surface. Pronounced porosity variations at Site 1073 may indicate rapid burial of clay-rich sediment by turbidites at times of increased slope progradation.
The Pleistocene sequence between surfaces pp4(s) and pp3(s) is particularly interesting because elements of this interval are well developed at all three sites and they offer the prospect of correlating specific features in logs and magnetic susceptibility data. However, the scheme offered by Delius et al. (2001) is inconsistent with shipboard seismic reflection geometry, biostratigraphy, and magnetostratigraphy (Austin, Christie-Blick, Malone, et al., 1998). Sediments above surface pp4(s) at Site 1073 on the slope are younger than 1.6 Ma (see Wei, Chap. 5, this volume; McCarthy and Gostlin, 2000) and, with the possible exception of a 5-m interval of unstable magnetic inclination immediately above surface pp4(s) (519.8-515 meters below seafloor [mbsf]), are entirely within the Bruhnes Chron (younger than 0.78 Ma) (Austin, Christie-Blick, Malone, et al., 1998). In contrast, the Bruhnes/Matuyama boundary (0.78 Ma) is located immediately below surface pp3(s) at both shelf Sites 1071 and 1072. Therefore, the portion of the intervening sequence on the shelf is for the most part and perhaps entirely older than the portion preserved on the slope; the correlation scheme of Delius et al. (2001, fig. 7) requires the opposite to be true.