Abstract

An ongoing study utilizes outcrop-scale seismic data and lithofacies data from cores and platform borings collected from the east Texas continental shelf to test assumptions and models that relate sedimentary facies patterns and sequence stratigraphy. The current data base consists of nearly 15 000 km of high-resolution seismic data and lithological data from hundreds of sediment cores and platform borings. The Texas shelf is ideally suited for this work because sediment supply during the Pleistocene was high enough to keep pace with the relatively rapid rise and fall in sea level, thus producing sequences that can be imaged on high-resolution seismic records. Furthermore, sediment supply, shelf gradient, and the degree of diapirism and faulting vary across the shelf, so the relative roles of these agents in controlling the overall packaging of facies into systems tracts can be evaluated.

Strong contrasts exist between low-sediment-supply (Trinity/Sabine) and high-sediment-supply (Brazos and Colorado) fluvial systems and their associated systems tracts. The predominant difference is that sediment delivered to the shelf by the Trinity/Sabine fluvial system was, for the most part, deposited within the incised fluvial valleys. Only during lowstands did the Trinity/Sabine system deliver sediment directly to shelf-margin deltas and slope minibasins. Reincision, associated with fifth-order eustatic fluctuations, flushes the valley of sediments deposited during the previous transgression and highstand, thus large quantities of sediment are delivered from this valley system during lowstands. Transgressive shelf sand bodies occur adjacent to the Trinity and Sabine incised valleys and are scattered widely across the shelf. Backstepping parasequences, the product of the episodic nature of glacial eustatic sea-level rise, characterize the incised valley fill. The highstand systems tract is thin to absent in interfluve areas.

The Brazos and Colorado rivers have much larger sediment supplies than the Trinity and Sabine rivers. They have filled their incised valleys with fluvial deposits and abandoned them to occupy several more shallow valleys. The result of this fluvial avulsion has been the sequestering of a significant part of the sediments delivered by these rivers to fluvial valleys on the shelf. Both the Colorado and Brazos rivers constructed large shelf-margin deltas during the lowstand, but these deltas differ in terms of their overall morphology and sediment facies. The ancient Colorado delta is sandy and it directly sourced two slope fans during the maximum lowstand. The Brazos shelf-margin delta consists mainly of mud and there is little evidence of significant bypass during the lowstand. During initial transgression, the Brazos/Colorado shelf-margin deltas backstepped onto the outer shelf. Rapid transgression and associated erosion removed the delta-plain beds. Continued transgression led to decapitation of sandy fluvial and deltaic facies, reworking the sands into widespread shelf sand bodies, and further backstepping of the deltas. During the previous highstand, the Brazos and Colorado rivers constructed fluvial-dominated deltas on the shelf.

The shelf-margin deltas of the area show a complex pattern of progradation and aggradation that varies from one delta to the next. This complexity is due predominantly to the different response of systems with different sediment supplies to fifth-order eustatic fluctuations. Before we understood these autocyclic effects, attempts at relating seismic stratigraphic changes to the oxygen isotope curve, our proxy for sea-level change, were unsuccessful.