Reappraisal of the sequence boundary in time and space: Case and considerations for an SU (subaerial unconformity) that is not a sediment bypass surface, a time barrier, or an unconformity

The sequence-bounding unconformity bears the key defining traits of being “a surface separating younger from older strata, along which there is evidence of subaerial erosional truncation … or subaerial exposure, with a significant hiatus indicated (Van Wagoner et al., 1988).” This subaerial component of sequence boundaries (subaerial unconformity—SU) is also broadly considered to form as a topographic surface of sediment bypass, carved during relative sea level fall and buried by backfilling during relative sea level rise. Accordingly, the SU is commonly presumed to record an approximate time barrier, which separates older from younger strata along its full length. In this paper we show that regional composite scour (RCS) surfaces that are traditionally mapped as an integral component of the SU were never a single subaerial topographic surface characterized by sediment bypass, are not unconformities, do not record an effective time barrier, and form diachronously at the channel-belt scale over the entire fall to rise of a base-level cycle. These RCS surfaces, and by inference the SU surfaces they comprise, thus do not fully fit key defining characteristics embodied in the conceptual sequence boundary. 

Flume observations and field data show that the RCS is buried by fluvial sediment simultaneously as it is scoured. Accordingly, the RCS is perennially covered with stored sediment during formation, is only exposed as a subaerial topographic surface at the local place and time where it is undergoing active growth, and forms over the duration of local marine drainage during a relative sea-level cycle. This “cut-and-cover” model differs greatly from more established “bypass” models, which assume that the RCS was roughly sediment free and subaerially exposed for long durations of incision during regression and thus preserves a significant depositional hiatus upon later burial. Instead, the RCS may commonly and locally record a hiatus more typical of a facies-bounding diastem without a lacuna significantly greater than that of surfaces within the strata it binds. Fragments of fluvial strata may commonly and sporadically be preserved above the RCS that are older than underlying marine units overrun by this surface. Consequently, the RCS is not an effective time barrier. Lateral planation by migrating and avulsing channels as the RCS expands laterally after nucleation can place younger fluvial strata over much older units, which means that this surface is also composite and highly diachronous laterally at the scale of channel belts. 

The cut-and-cover model has additional implications not captured by the bypass model. First, significant sediments may be stored within fluvial strata above the RCS during regression that are not available for contemporary falling stage and lowstand marine shorelines. This can result in marine sediment starvation, particularly of the sand fraction, and in extreme cases can result in sand autodetachment and an absence of regressive marine reservoir sandstones. Second, cutting of the RCS co-generates a suprafluvial surface above the covering fluvial strata during regression that may be used as a mappable proxy for the conceptual maximum regressive surface (MRS). The MRS may be raised above this surface locally by lowaccommodation aggradation during lowstand normal regression, but in either case preserves an approximate time line where not reworked during later transgression. Third, valley development across the RCS does not exclusively form by landward knickpoint growth, and may include complexly formed and potentially crosscutting buffer valleys. SU valley incision can be divided into four modes, which include denudation, structural, buttress, and buffer valley components, which may work together locally and tend to have variable importance along the shore-to-hinterland profile. 

Although the RCS is not a good rock proxy for the conceptual sequence boundary it remains a very mappable surface which may separate facies of potentially very different origin and reservoir quality. The RCS is also  inseparable from the SU and typically the only terrestrial erosional surface of extent in most short-duration sequences. Its nullification as a sequence boundary would mean abandonment of depositional sequence stratigraphy as a correlation and interpretive tool within these sections. An alternative to abandonment of the SU as a sequence boundary is to loosen the definition of a sequence boundary to ‘a discrete surface of erosional truncation carved landward of contemporary shorelines that is traceable beyond the scale of a single valley or comparable local system, and its correlative surfaces of conformity and/or non-deposition’, and continue its use as before.