Triassic Eustatic Variations Reexamined

Documentation of eustatic variations for the Triassic is limited by the paucity of the preserved marine stratigraphic record, which is confined mostly to the low and middle paleolatitudes of the Tethys Ocean. A revised sea-level curve based on reevaluation of global stratigraphic data shows a clear trend of low seastands for an extended period that spans almost 80 m.y., from the latest Permian to the earliest Jurassic. In the Early and Middle Triassic, the long-term sea levels were similar to or 10–20 m higher than the present-day mean sea level (pdmsl). This trend was reversed in the late Ladinian, marked by a steady rise and culminating in peak sea levels of the Triassic (~50 m above pdmsl) in the late Carnian. The trend reverses again with a decline in the late Norian and the base level remaining close to the pdmsl, and then dipping further in the mid-Rhaetian to ~50 m below pdmsl into the latest Triassic and earliest Jurassic. Superimposed upon this long-term trend is the record of 22 widespread third-order sequence boundaries that have been identified, indicating sealevel falls of mostly minor (<25 m) to medium (25–75 m) amplitude. Only six of these falls are considered major, exceeding the amplitude of 75 m. The long interval of Triassic oceanic withdrawal is likely to have led to general scarcity of preserved marine record and large stratigraphic lacunae. Lacking evidence of continental ice sheets in the Triassic, glacio-eustasy as the driving mechanism for the third-order cyclicity can be ruled out. And even though transfer of water to and from land aquifers to the ocean as a potential cause is plausible for minor (a few tens of meters) sea-level falls, the process seems counter-intuitive for third-order events for much of the Triassic. Triassic paleoenvironmental scenarios demonstrate a close link between eustasy, climates, and biodiversity. INTRODUCTION The Triassic Period encompasses 50.5 m.y., spanning an interval from 251.9 to 201.4 Ma (Ogg et al., 2016). By this time, the megacontinent of Pangaea had already assembled, surrounded by the Panthalassa Ocean that covered >70% of Earth’s surface, and by the mid-Triassic the Pangaean landmass was almost evenly distributed in the two hemispheres around the paleo-equator (see Fig. S1 in the GSA Data Repository1). The interval from latest Permian through the earliest Jurassic, a time span of nearly 80 m.y., represents the longest spell of low seastands of the Phanerozoic. The Triassic is also bracketed by two major biotic extinctions near the Permian-Triassic (P-T) and Triassic-Jurassic boundaries, the one at P-T boundary being the most severe biotic turnover of the Phanerozoic (Raup and Sepkowski, 1982; Hallam and Wignall, 1997; McElwain et al., 1999). The Late Triassic experienced the beginning of the lithospheric swell, ushering the breakup of Pangaea and its eventual split into discrete continents in the later Mesozoic (see Fig. S1 [see footnote 1]). The definite signs of the beginning of Pangaean fragmentation were clearly manifest by the end of the Triassic with the basaltic outpouring of the massive Central Atlantic magmatic province (see, e.g., Marzoli et al., 1999, 2004; Davies et al., 2017). In the past two decades substantial new stratigraphic data from Triassic sections has come to light, and there have been significant refinements in time scales, making a review and revision of the Triassic sealevel variations timely. The documentation for the revised Triassic sea-level curve, though still largely from northwestern and central Europe (western Tethys), now also includes sections further east from other parts of the Tethys, such as the Arabian Platform, Pakistan, India, China, and Australia. From the boreal latitudes documentation includes sections from the Sverdrup Basin, Svalbard, and the Barents Sea. This paper serves to complete a review of the entire Mesozoic as both Cretaceous and Jurassic sea-level variations have already been reappraised (Haq, 2014, 2017). For a background of the paleoenvironmental conditions (oceans and climates) in the Triassic see the GSA data repository (see footnote 1). TRIASSIC TIME SCALE UPDATES A succinct discussion of the methodological advancements and modifications to the Triassic time scale can be found in Preto et al. (2010), and a detailed discussion of Triassic stratigraphy has been presented by Ogg et al. (2014). Since the last update of the Triassic third-order sea-level variations (Haq and Al-Qahtani, 2005) that was calibrated to an earlier version of the time scale, there have been several refinements to the Triassic chronostratigraphy. The latest version of the time scale (Ogg et al., 2016) modifies the boundaries of Triassic standard stages (ages) by anywhere between <1 m.y. to almost 6 m.y. Like earlier versions, the new time scale is mainly based on biostratigraphy, anchored by selected radiometric dates, with some intervals refined by astronomical and cyclostratigraphical fine-tuning, and others aided by magnetostratigraphy. Conodonts and ammonoids constitute the mainstay of the Triassic biostratigraphic correlations. Special problems concerning wider correlations using these fossil groups in the Triassic include taxonomic standardization, rarity of markers, potential diachroniety in conodontsʼ first and last appearance, and provinciality among ammonoids. Since much of the Pangaean landscape was dominated by terrestrial sediments, regional correlations often rely on palynology, ostracods and tetrapods that do not lend themselves to wider correlations with marine records. Ogg et al. (2016) ascribe a Bilal U. Haq, Smithsonian Institution, Washington, D.C., USA, and Sorbonne University, Paris, France composite error of between 0.2 and 0.59 m.y. for the stage boundaries in the Triassic depending on the type of data (see also the GSA data repository for further discussion of time scales [see footnote 1]). LARGE TIME GAPS IN THE RECORD OF THE MIDDLE AND LATE TRIASSIC? Even a cursory look at the most recent update of the Triassic time scale (Ogg et al., 2016) reveals its extreme lopsidedness: while the Early Triassic spans only 5.1 m.y. and the Middle Triassic increases to 9.1 m.y., the Late Triassic jumps to a substantial span of 35.6 m.y. Some unevenness is to be expected, but this extreme asymmetry is also witnessed in the time spans of the stages (ages) and biostratigraphic zones within the stages, as well the lengths of the sequence cycles and corresponding sealevel events that all increase in duration in the Middle to Late Triassic. If the above apparent chronostratigraphic asymmetry is real, then the large differences in the duration of fossil zones imply that evolutionary rates (as measured by appearance of new species/m.y.) were relatively rapid in the Early Triassic (thus the availability of a high-resolution biozonal subdivision), declining somewhat in the Middle Triassic, and slowing down to an extreme thereafter (characterized by a few long-duration biozones), especially in the later Late Triassic. However, the temporal lengths of sequence cycles (based on sedimentary facies shifts) do not have to follow the biotic evolutionary trends, and yet they do. Their long time spans (average of ~5 m.y./cycle in the Middle and Late Triassic) would imply a built-in bias in the record expressed as a lack of preserved marine stratigraphic record. This seems plausible in a scenario where the long-term trend of low seastands for the period means fewer marine records in favor of more terrestrial sedimentary records. This is exacerbated by mostly type-1 sequence boundaries (when the base line withdraws beyond the shelf edges) that may incorporate large erosional time gaps. Large temporal lacunae in the stratigraphic record could explain the potentially specious signal that comes across as slowdown in the biotic evolutionary rates, as well as the dearth of sequence cycles for the interval in question (i.e., Middle and Late Triassic). An oxygenisotopic record of the Triassic derived from Tethyan conodont apatite shows trends that duration = 0.77 m.y.), and expand for the remainder of the Carnian through Rhaetian interval (average ammonoid zonal duration = 2.43 m.y.). Using multiple overlapping criteria (i.e., several fossil groups), these uncertainties can sometimes be narrowed. The long-term sea-level envelope for the Triassic is similar to those shown in Haq et al. (1987, 1988) and Hardenbol et al. (1998). The original long-term curve for the Triassic was based on continental flooding data and this is still the case, because other constraints for this envelope, such as mean age of oceanic crust, are not available since almost all of the Triassic age oceanic crust has since been subducted, with the exception of a limited area of the seafloor on Exmouth Plateau west of Australia (von Rad et al., 1989). Recently van der Meer et al. (2017) have produced independent estimates of the long-term sea level based on Sr-isotope data, which show close similarities to the continental flooding curves and to the long-term Triassic curve presented here, although the interpreted amplitudes differ. The documentation for the short-term (third-order) sea-level events is based on sequence-stratigraphic information pieced together from several available longer duration sections and augmented by some shorter-duration records. In addition to sequence-stratigraphic interpretive criteria that are now well established and do not require repetition, other features that were particularly helpful in stratigraphic interpretations (originally listed in Haq and Schutter, 2008) in the Triassic include forced-regressive facies, organic-rich facies of the condensed sections, transgressive coals, evaporites, exposure-related deposits (including incised valley fills, autochthonous coals, eolian sandstones, and karst), and laterite/bauxite deposits. These features can often aid in the identification of depositional surfaces and system tract bou