Stratigraphic aliasing and the transient nature of deep-water depositional sequences: Revisiting the Mississippi Fan

Sequence stratigraphy remains the foremost methodology used to describe the stratigraphic record and to interpret the controls on deposition; yet, it relies on long-standing assumptions that few studies have sought to validate. Here, we present results from testing hypotheses related to the deep-water depositional sequence model by revisiting the seismic-based type locality: the Mississippi Fan, in the central Gulf of Mexico. By independently testing the relationship between cycles of relative sea-level change and those of sedimentation, we demonstrate that >98% of Neogene–Quaternary deep-water sequences do not accumulate in a manner prescribed by long-held sequence stratigraphic conventions. Instead, over the past 5.5 Myr, sequences show a temporal mismatch in frequency, phase, and amplitude with cycles of relative sea-level change, a concept we refer to as stratigraphic aliasing. Divergences are attributed to variable rates of sedimentation, which were responsible for creating cycle frequencies that were both higher and lower than those of relative sea-level change, and that resulted in two modes of Mississippi Fan accumulation: lower average sedimentation rates in older sediments, and an opposite trend in younger successions. The latter mode occurred ~2.2 Myr after the onset of North American glaciation, a period marked by significant continental drainage reorganization and salt-tectonic deformation. Based on our conclusions, we recommend that the future of sequence stratigraphy be rooted firmly in assessing the reproducibility of preexisting spatiotemporal predictions and in the rigorous use of multiple working hypotheses. INTRODUCTION For over 40 years, depositional systems have been thought to be controlled by the interaction of eustasy and subsidence (i.e., “relative sea level” of Jervey, 1988), with variability in sediment supply providing a secondary effect (see Vail et al., 1977; Posamentier et al., 1988; Van Wagoner et al., 1990). According to this view, relative sea-level change drives reciprocal sedimentation, whereby condensed sections and nearshore deposits accumulate during rises and high stands, and erosional surfaces and deepwater accumulations deposit during falls and low stands (Figs. 1A and 1B). While active debate regarding both controlling mechanisms and their stratigraphic responses has been ongoing since the late 1980s (see Haq et al., 1987; Christie-Blick et al., 1988), the community remains divided on the level of applicability of sequence stratigraphy (see Madof et al., 2016; Burgess, 2016), on its status as a workflow as opposed to a paradigm (see Burgess and Prince, 2016; Catuneanu and Zecchin, 2016), and on its ability to be treated as a set of testable hypotheses (this study). The Mississippi Fan, in the central Gulf of Mexico, serves as the type locality for the seismic-based model of the deep-water depositional sequence. In addition to Weimer (1990), other workers identified Mississippi Fan lobes from downlapping reflections at the base of gullwingshaped deposits (Garrison et al., 1982) and related accumulations to falls in relative sea level (Bouma et al., 1985). While these concepts have been applied globally to deep-water systems, few studies have sought to validate the initial findings. Here, we revisited the Mississippi Fan in an attempt to independently reproduce the relationship between relative sea-level change and sedimentation, and to test whether the latter is controlled by changes in the former. By testing long-standing conventions, we offer new insights into the time-varying interactions between purported drivers and accumulation, and into the effect of climatic and tectonic signals on modulating inherently nonstationary deposition. MISSISSIPPI FAN, GULF OF MEXICO The Neogene–Quaternary Mississippi Fan in the central Gulf of Mexico is one of the most well-studied submarine fans on Earth. Because of the abundance of subsurface data, it has been used to understand deep-water stratigraphic archi tecture and cyclicity, as well as relative sea-level change and sedimentation. Architecture and Cyclicity The Mississippi Fan is a submarine accumulation >4 km in thickness situated outboard of the Sigsbee Escarpment. Based on downlapping reflections (i.e., geometric evidence for sequence boundaries), Weimer (1990) interpreted 17 laterally compensatory deep-water sequences (Fig. 2A) and inferred each to be composed of channelized (coarse-grained) deposits flanked by levee-overbank (fine-grained) deposits. By incorporating age control, Weimer and Dixon (1994) showed that sequence duration generally decreased from 0.66 Myr (5.5–4.84 Ma; sequence number 1) to 0.023 Myr (0.23–0.0 Ma; sequence number 17) and increased in lateral extent and overall thickness through time (Figs. 2B and 2C). In addition to the duration of sequences, Weimer and Dixon (1994) interpreted two major hiatuses, from 3.0 to 1.9 Ma and from 0.45 to 0.086 Ma. While Weimer and Dixon (1994) may have incorrectly assumed that seismic reflections were laterally continuous synchronous surfaces, alternative interpretations have not yet been proposed. Sea Level and Sedimentation We calculated relative sea-level change and sedimentation rates independently for the last CITATION: Madof, A.S., et al., 2019, Stratigraphic aliasing and the transient nature of deep-water depositional sequences: Revisiting the Mississippi Fan: Geology, v. 47, p. 545–549, https:// doi .org /10 .1130 /G46159.1 Manuscript received 26 December 2018 Revised manuscript received 19 February 2019 Manuscript accepted 25 March 2019 https://doi.org/10.1130/G46159.1 © 2019 The Authors. Gold Open Access: This paper is published under the terms of the CC-BY license. Published online 16 April 2019 Downloaded from https://pubs.geoscienceworld.org/gsa/geology/article-pdf/4707771/545.pdf by guest on 31 October 2019 546 www.gsapubs.org | Volume 47 | Number 6 | GEOLOGY | Geological Society of America 5.5 Myr of growth of the Mississippi Fan (Figs. 3A and 3B). The former was constructed by adding the eustatic record of Miller et al. (2005, their fig. 3, created from the benthic foraminiferal δ18O record) to the back-stripped subsidence rate of the northern Gulf of Mexico of Diegel et al. (1995, their fig. A-3). While specific details regarding the construction of the sub sidence curve remain unclear, it serves as the most robust (if not the only) example for the area. Sedimentation rates for each Mississippi Fan sequence were calculated by dividing maximum preserved thickness (from Fig. 2A) by duration (from Fig. 2B); rates span two orders of magnitude, with older sequences displaying lower average accumulation rates (i.e., 104 cm/Myr), and younger ones showing an opposite trend (i.e., 106 cm/Myr). Based on our analysis, the bulk of long-term relative sea-level change in the Mississippi Fan was caused by subsidence related to salt withdrawal, rather than by eustatic change. Fi g u re 1. S ch e m a t i c diagram of reciprocal sedimentation and deep-water depositional sequence model. A: Two-dimensional transect along continental margin showing shelfal accumulations developing during relative sea-level high stands and rises, and slope-tobasinal sedimentation occurring during low stands and falls. B: Each deep-water sequence develops during one cycle of relative sea-level change (modified from Posamentier and Kolla, 2003), during which rate of sedimentation decreases. Sedimentation-rate curve was created by dividing facies thickness (from hypothetical well log) by duration (from relative sea-level curve) and normalizing to maximum rate. Abbreviations are as follows: RSL—relative sea level; SR—sedimentation rate; HS—high stand; LS—low stand. A


INTRODUCTION
For over 40 years, depositional systems have been thought to be controlled by the interaction of eustasy and subsidence (i.e., "relative sea level" of Jervey, 1988), with variability in sediment supply providing a secondary effect (see Vail et al., 1977;Posamentier et al., 1988;Van Wagoner et al., 1990).According to this view, relative sea-level change drives reciprocal sedimentation, whereby condensed sections and nearshore deposits accumulate during rises and high stands, and erosional surfaces and deepwater accumulations deposit during falls and low stands (Figs. 1A and 1B).While active debate regarding both controlling mechanisms and their stratigraphic responses has been ongoing since the late 1980s (see Haq et al., 1987;Christie-Blick et al., 1988), the community remains divided on the level of applicability of sequence stratigraphy (see Madof et al., 2016;Burgess, 2016), on its status as a workflow as opposed to a paradigm (see Burgess and Prince, 2016;Catuneanu and Zecchin, 2016), and on its ability to be treated as a set of testable hypotheses (this study).
The Mississippi Fan, in the central Gulf of Mexico, serves as the type locality for the seismic-based model of the deep-water depositional sequence.In addition to Weimer (1990), other workers identified Mississippi Fan lobes from downlapping reflections at the base of gullwingshaped deposits (Garrison et al., 1982) and related accumulations to falls in relative sea level (Bouma et al., 1985).While these concepts have been applied globally to deep-water systems, few studies have sought to validate the initial findings.Here, we revisited the Mississippi Fan in an attempt to independently reproduce the relationship between relative sea-level change and sedimentation, and to test whether the latter is controlled by changes in the former.By testing long-standing conventions, we offer new insights into the time-varying interactions between purported drivers and accumulation, and into the effect of climatic and tectonic signals on modulating inherently nonstationary deposition.

MISSISSIPPI FAN, GULF OF MEXICO
The Neogene-Quaternary Mississippi Fan in the central Gulf of Mexico is one of the most well-studied submarine fans on Earth.Because of the abundance of subsurface data, it has been used to understand deep-water stratigraphic archi tecture and cyclicity, as well as relative sea-level change and sedimentation.

Architecture and Cyclicity
The Mississippi Fan is a submarine accumulation >4 km in thickness situated outboard of the Sigsbee Escarpment.Based on downlapping reflections (i.e., geometric evidence for sequence boundaries), Weimer (1990) interpreted 17 laterally compensatory deep-water sequences (Fig. 2A) and inferred each to be composed of channelized (coarse-grained) deposits flanked by levee-overbank (fine-grained) deposits.By incorporating age control, Weimer and Dixon (1994) showed that sequence duration generally decreased from 0.66 Myr (5.5-4.84Ma; sequence number 1) to 0.023 Myr (0.23-0.0 Ma; sequence number 17) and increased in lateral extent and overall thickness through time (Figs.2B and 2C).In addition to the duration of sequences, Weimer and Dixon (1994) interpreted two major hiatuses, from 3.0 to 1.9 Ma and from 0.45 to 0.086 Ma.While Weimer and Dixon (1994) may have incorrectly assumed that seismic reflections were laterally continuous synchronous surfaces, alternative interpretations have not yet been proposed.

Sea Level and Sedimentation
We calculated relative sea-level change and sedimentation rates independently for the last CITATION: Madof, A.S., et al., 2019, Stratigraphic aliasing and the transient nature of deep-water depositional sequences: Revisiting the Mississippi Fan: Geology, v. 47, p. 545-549, https:// doi .org/10 .1130/G46159.1 5.5 Myr of growth of the Mississippi Fan (Figs. 3A and 3B).The former was constructed by adding the eustatic record of Miller et al. (2005, their fig. 3, created from the benthic foraminiferal δ 18 O record) to the back-stripped subsidence rate of the northern Gulf of Mexico of Diegel et al. (1995, their fig. A-3).While specific details regarding the construction of the sub sidence curve remain unclear, it serves as the most robust (if not the only) example for the area.Sedimentation rates for each Mississippi Fan sequence were calculated by dividing maximum preserved thickness (from Fig. 2A) by duration (from Fig. 2B); rates span two orders of magnitude, with older sequences displaying lower average accumulation rates (i.e., 10 4 cm/Myr), and younger ones showing an opposite trend (i.e., 10 6 cm/Myr).Based on our analysis, the bulk of long-term relative sea-level change in the Mississippi Fan was caused by subsidence related to salt withdrawal, rather than by eustatic change.

ASSESSING DEVIATIONS FROM CONVENTIONS
To test the effect of accommodation on Mississippi Fan deposition, conditions both conforming to, and deviating from, conventional sequence stratigraphic models must be evaluated.Although no framework currently exists to address the latter, we considered mismatches between relative sea-level change and sedimentation, and the incompleteness of the stratigraphic record.

Stratigraphic Aliasing
Sequence stratigraphic models for deep-water settings assume that cycles of relative sea-level change and sedimentation have equal frequencies, zero phase shift, and moderate amplitudes (Figs.4A-4D).As a result, each deep-water sequence is required to correlate to a single lowstand interval, with the greatest amount of sediment delivered during the maximum rate of relative sea-level fall (Weimer, 1990;Posamentier and Kolla, 2003).To address alternative scenarios, we introduce a conceptual framework termed stratigraphic aliasing (see Fig. DR1 in the GSA Data Repository 1 ), which makes use of varied waveforms to describe input parameters and to facilitate multiple working hypotheses.Although aliasing generally refers to the misidentification of signal frequencies due to sampling biases, our use of the term "stratigraphic alias-ing" is aimed at objectively comparing (and in the time domain) the cycle frequency, phase, and amplitude of relative sea-level change to those of sedimentation.For utility, we define frequency as the number of sedimentary cycles deposited within one cycle of relative sea-level change; frequencies can be equal to, less than, or greater than one and correspond to equivalent, longer, or shorter sedimentation periods.Phase relates to the timing of sediment delivery with respect to changes in relative sea level and can be used to describe lags and leads in accumulation.We describe deposition occurring at phase shifts of 0° (falls; conventional model expectations), 90° (lowstands), 180° (rises), and 270° (highstands).Amplitude is the rate of accumulation during one cycle of relative sea-level change; its variance describes increased and decreased flux.

"Sadler Effect"
In addition to stratigraphic aliasing, we considered the preservation of Mississippi Fan sequences.Because sequences (i.e., strata bounded above and below by sequence boundaries; see Figs. 2A and 2B) are marked by repetitive changes in accumulation rates, they can be classified as cycles.Cycles (i.e., sequences) can be used to assess the completeness of the depositional record by comparing their amplitude (i.e., sedimentation rate) to their frequency (i.e., time span).Sadler (1981) observed that the number and duration of nondepositional events (hiatuses) cause accumulation rates to obey an inverse power-law relationship with respect to the span of measurement.Although this technique (i.e., plotting a function and its reciprocal, which results in a negative correlation) has received some scrutiny (see Anders et al., 1987), it has become a standard practice in quantifying the completeness of the stratigraphic record.The trend, which is observed both in Mississippi Fan sequences and in globally sampled tur bidites (Fig. DR2), manifests as cycles with lower average sedimentation rates (lower amplitude) at longer durations (lower frequency).Higher rates of accumulation are necessarily sustained only over shorter intervals than are lower rates (see Sadler and Jerolmack, 2014;Tipper, 2015;Paola et al., 2018), implying that the "Sadler effect" operates also in the frequency domain.

DISCUSSION
Assessment of the applicability of sequence stratigraphic concepts relies on the number of sequences satisfying the conditions set forth by well-accepted models.According to these assumptions, the 17 preserved deep-water sequences of the Mississippi Fan require an equal number of relative sea-level cycles, with the bulk of sediment delivered during each lowstand interval.However, our analysis shows that only one sequence (comprising <2% of the past 5.5 Myr) coincides with relative sea-level change in the manner prescribed by conventional assumptions (i.e., equal frequencies, zero phase shift, and moderate amplitudes; see Figs. 3A and 3B).
Of the 17 deep-water sequences, older cycles (numbers 1-9; 5.5-0.7 Ma) are characterized by lower frequencies (longer periods) and lower amplitudes (lower accumulation rates), with the opposite pattern observed for younger ones (numbers 11-17; 0.6-0 Ma).In the former, phase (i.e., onset of sediment delivery) is impossible to determine because sequences span multiple relative sea-level cycles.In the latter, because multiple sequences develop between falls and rises, phase is interpreted to be between 0° and 180°.The only deposit that matches the period of coeval relative sea-level change is sequence number 10 (0.6-0.7 Myr); that deposit spans the transition between lower and higher average accumulation rates and has moderate amplitude (i.e., a sedimentation rate of 10 5 cm/Myr).However, phase could not be evaluated because sampling was insufficient to determine timing of sediment delivery.Remarkably, sequence number 10 contains the largest number of deep-water channels of any Mississippi Fan accumulation (see Fig. 2C), even when compared with other Mississippi Fan sequences.The finding that deep-water systems are the most well preserved when conventional sequence stratigraphic conditions are met has significant implications in sediment routing and source-to-sink analyses.
The two modes of deposition for the Mississippi Fan (i.e., lower average sedimentation rates in older sediments, and an opposite trend in younger successions) suggest that stratigraphic aliasing is a more appropriate framework of interpretation than sequence stratigraphy.As such, we find that older Mississippi Fan successions accumulated under markedly different conditions than younger ones.Although younger accumulations were deposited during large-scale climatic and tectonic changes (e.g., North American glaciation [see Aharon, 2003], continental drainage reorganizations [see Bentley et al., 2016], and salt-related deformation [see Tripsanas et al., 2007]), they responded with shorter periodicities.We attribute these short-lived stratigraphic responses not to relative sea-level change, but to the liberation of onshore sediment during glacial advance and retreat, and to the subsequent transport through continental drainages and around regional and local salt-related structures.Based on the climatic and tectonic forces that overprinted the effects of relative sea-level change on the Mississippi Fan, we question the very concept of the deep-water depositional sequence model at its type locality.While these results represent only one location, they are surprising enough to warrant further investigation.We therefore suggest that future research uses similar methods (e.g., stratigraphic aliasing) to rigorously and critically evaluate fundamental sequence stratigraphic assumptions and long-held conventions.

CONCLUSIONS
We used the deep-water Mississippi Fan in the central Gulf of Mexico to test objectively the effect of changes in accommodation on the stratigraphic evolution of 17 seismic sequences, which developed during the past 5.5 Myr.Although this locality was pivotal to the creation of sequence stratigraphic concepts, we find that interactions between purported drivers and sedimentation display temporal mismatches in frequency, phase, and amplitude, a concept we refer to as stratigraphic aliasing.By showing that Mississippi Fan sequences have lower average sedimentation rates in older sediments, and an opposite trend in younger successions, we con-clude that sequence stratigraphic models present an oversimplified and stationary view of deepwater depositional systems.Our results call into question the concept of accommodation-driven reciprocal sedimentation, highlight the necessity of reevaluating type localities, and suggest that stratigraphic aliasing can be a more objective framework than well-accepted sequence stratigraphic conventions.

Figure 3 .
Figure 3.Comparison of cycles of relative sea-level change and sedimentation for Mississippi Fan.Relative sea-level curve was constructed by adding eustasy (Miller et al., 2005) and subsidence (Diegel et al. 1995); sedimentation rates were calculated from Figures 2A and 2B (note change in scale; see text for details).A: Eustatic curve shows two distinct periods and amplitudes: shorter periods (40 kyr) from ca. 2.7 to 1.1 Ma and longer ones (100 kyr) from ca. 1.1 to 0 Ma; lower amplitudes from ca. 5.5 to 3.0 Ma and higher ones from ca. 3.0 to 0 Ma.Note that sequence number 10 is the only accumulation for which period (frequency) appears to match that of relative sea-level change.B: Magnification of last four Mississippi Fan sequences.Sequence numbers 14-16 developed during a fall in relative sea level, whereas sequence number 17 accumulated between a low stand and rise.Marine isotope stages are abbreviated as MIS; timing of meltwater floods was taken from Aharon (2003).

Figure 4 .
Figure 4. Concept of stratigraphic aliasing as it relates to relative sea-level change (blue) and sedimentation (green).Conventional sequence stratigraphic conditions are noted by shaded boxes.Frequency is defined as ratio of periods of relative sea-level change to sedimentation.Phase relates to timing of maximum sedimentation with respect to relative sea level and can occur at any point in cycle.Amplitude describes sedimentation with respect to time and is classified as moderate, low, or high.