Sequence stratigraphy has been defined as ‘the study of rocks within a framework wherein the vertical succession of rocks is subdivided into genetically related units bounded by surfaces, including unconformities and their correlative conformities’ (Mitchum et al. 1977a). Sequence stratigraphy is first and foremost a method that guides observations in the stratigraphic record across an array of depositional settings, stratal attributes and datasets, explicitly recognizing that the stratigraphic record is composed of both rocks and surfaces in various forms. These observations are then summarized in models that generalize details to facilitate prediction away from data control points. For completeness, sometimes the models are interpreted in terms of mechanisms (e.g. eustasy, climate, etc.) that may help explain observations and enhance prediction (Bohacs 1998). Few publications focus on the method to make these observations (e.g. Mitchum et al. 1977b; Van Wagoner et al. 1990; Emery & Myers 1996; Abreu et al. 2011; Miller et al. 2013) compared with a majority of papers exploring aspects of various models and mechanisms (e.g. Catuneanu et al. 2011, and references therein). Sequence stratigraphic mechanisms, nomenclature or inferred implications are lively sources of debate and academic study. The depositional sequence boundary in particular has been vigorously discussed (Posamentier et al. 1992; Christie-Blick & Driscoll 1995; Embry 2009; Bhattacharya 2010, 2011; Catuneanu et al. 2011; Holbrook & Bhattacharya 2012; Blum et al. 2013; Miller et al. 2013), but the most important starting point has to be a common observational framework. With such a framework based on criteria directly observable from outcrop, core, well-log and seismic data, independent of causal mechanisms or the duration or magnitude of events, alternative models, implications and possibly even terminology can be discussed without requiring a priori adherence to one ‘school’ or another.

The building blocks of sequence stratigraphy are observations of facies, facies association, vertical stacking, stratal geometries and stratal terminations, supplemented with tracking of shoreline position and trajectory through a stratigraphic succession. Strata are preserved if there is a sediment supply and accommodation or space for the sediments to accumulate and be preserved (Jervey 1988). In shallow water siliciclastic margins, the stacking patterns of these building blocks are fairly simple in concept but can be challenging in application. The accommodation succession method of sequence stratigraphy assumes that these building blocks form in response to varying rates of coastal accommodation increase and decrease (δA) relative to the rate of sediment flux (δS) (Neal & Abreu 2009). The method can be applied to any base-level-controlled deposition (sensu Barrell 1917), most commonly shallow water siliciclastic environments. In carbonate environments, not discussed here, the method can be used with caution and recognition of the complexities of carbonate sediment production and distribution rates relative to changes in rate of accommodation creation (Pomar 2001). In deep marine settings, where accommodation is more a function of bathymetric gradient, depositional patterns and fluid dynamics (e.g. Gerber et al. 2008), stratigraphers may infer coastal accommodation successions through changing sediment flux or biostratigraphic correlation (Neal et al. 1995).

In this paper we will demonstrate in more detail the method and techniques to produce an observation-based accommodation succession (hereafter abbreviated δA/δS) interpretation. Three examples will be used here to demonstrate the technique. The first is from a flume tank experiment with known sediment input, base-level change and subsidence. The second is a high-resolution seismic dataset from a Pleistocene delta in the Gulf of Mexico in which sea-level history is known but sediment input is not. The final example is from the Cretaceous Ferron Sandstone where none of the boundary conditions are known. These datasets are important to (1) demonstrate how the δA/δS method produces sequence stratigraphic interpretations without needing to know sea-level and (2) demonstrate how generalized models can be used to facilitate prediction away from data control points. As higher resolution subsurface data with broader areal coverage become available, the importance of a rigorous and consistent interpretation method becomes ever greater. The future of sequence stratigraphy depends on improved communication and rigour in interpretations. Already, the complexity of sequence stratigraphy terminology rooted in sea-level interpretation (Catuneanu et al. 2011) creates a cumbersome jargon with interpretation placed before objective, repeatable recognition criteria. Because increasing coastal accommodation is associated with lowstand, highstand and transgressive systems tracts evidenced by coastal onlap (Jervey 1988), and regression or transgression of the shoreline and shelf margin is strongly controlled by sediment supply (Carvajal et al. 2009), a method to incorporate both variables with a descriptive terminology is needed (Neal & Abreu 2009; Abreu et al. 2014). Observations must be clearly separated from interpretations (of generalized models or mechanisms of formation) to improve sequence stratigraphy application and sharpen research focus. This observational approach, with proper documentation, also allows for more effective communication and independent validation or subsequent enhancement of stratigraphic frameworks over time, across datasets and across scales of data resolution.

Basics of the method

Stratigraphic observations should be as objective as possible, independent of a generalized model or assumptions about mechanism of formation. A good method is one that guides the interpreter to make accurate and diagnostic observations. With accurate observations that are unencumbered by a priori sea-level implications, impartial discussion of interpretation choices can occur. It has been argued that sequence stratigraphic observation and interpretation are inseparable and cannot be objectively tested (Miall & Miall 2001) or that all interpretations are allowable if the terminology is applied correctly (Catuneanu et al. 2011). We agree that many sequence stratigraphic observations have inherent interpretation implications; however, it is critically important to understand what inherent assumptions are embedded in the sequence stratigraphic method and which ones are model or explanation choices (e.g. eustasy driven). An extensive debate on the topic of ‘model-driven’ versus ‘data-driven’ interpretation is captured on the SEPM Strata website (http://www.sepmstrata.org/TerminologyList.aspx).

The depositional sequence (sensu Mitchum et al. 1977a) is the largest stratigraphic unit that can be considered internally relatively conformable and genetically related, bracketed by depositional sequence boundaries. At the scale of a depositional sequence, Neal & Abreu (2009) observed the geometry of stratigraphic successions to show components that are both vertical (aggradation–degradation) and lateral (progradation–regression and retrogradation–transgression) simultaneously at different magnitudes in an ever-changing succession that follows a repeated and predictable pattern if viewed at a sufficiently large regional scale. A complete depositional sequence (Fig. 1) can be described using three types of stratal termination, three types of vertical or lateral shoreline trajectories and three key bounding surfaces. Stratal terminations occur where sedimentary layers end onto an underlying surface (baselap) or under an overlying surface (toplap) (Mitchum et al. 1977b). The ability to differentiate significance of stratal terminations can tell a great deal about the accommodation or sediment supply history of a margin.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Simplified line drawing illustrating key stratal terminations, stacking patterns and key surfaces of the accommodation succession (δA/δS) method of sequence stratigraphy. Stacking patterns: R, regressive; PA, progradation–aggradation; APD, aggradation–progradation–degradation. Key surfaces: MTS, maximum transgressive surface; MRS, maximum regressive surface; SB, sequence boundary.

Stratal terminations indicate missing or condensed time in the rock record and represent the evolving depositional substrate. It is important to distinguish coastal onlap (horizontal coastal strata terminating onto an inclined exposure surface) from subaqueous relative onlap or downlap (inclined strata onto a less dipping surface) in that coastal onlap represents the updip extent of strata within a depositional sequence. The aggradation of horizontal clinoform topset beds requires that coastal onlap occurs updip (Fig. 1) even if progradation continues downdip owing to excess sediment supply relative to accommodation creation rate. Downlap onto horizontal clinoform topset beds is also an indicator of increased accommodation, discussed further below. Top-discordant stratal terminations, toplap and truncation, occur with sedimentary bypass and erosion (Mitchum et al. 1977b). Truncation can occur as a result of fluvial incision on a regional exposure surface or angular erosion of horizontally deposited beds that are tectonically deformed. Stratal terminations point to key surfaces and key surfaces mark changes in the observed vertical or lateral stratal stacking.

Stratal stacking patterns used in the δA/δS method discussed below are progradation to aggradation (PA), aggradation to progradation to degradation (APD) and retrogradation (R). These stacking patterns represent a linkage of contemporaneous depositional systems, defined as ‘systems tract’ by Brown & Fisher (1977), and are analogous to lowstand, highstand and transgressive systems tracts of Van Wagoner et al. (1988) without the sea-level-linked connotations of Helland-Hansen & Gjelberg (1994) (Abreu et al. 2014). This method combines the effects of sediment supply and accommodation, with each factor having characteristic stacking pattern signals. Sufficient regional perspective is required to distinguish δA/δS stratal stacking patterns and key surfaces from local autogenic surfaces that often result from local sediment supply variation and depositional geometries. Key surfaces defined with stratal terminations and changes in stacking pattern are the depositional sequence boundary, maximum regressive surface and maximum transgressive surface (Abreu et al. 2014). Abrupt facies dislocations that are potentially important for petroleum exploration and production commonly occur across key bounding surfaces (e.g. Van Wagoner et al. 1990; Bhattacharya 2011).

The δA/δS sequence stratigraphic method can be summarized in five steps: (1) observe facies, facies association vertical and lateral stacking trends, and lateral stratal terminations; (2) use stratal terminal patterns and stratal geometries to delineate sequence stratigraphic surfaces; (3) use surfaces and stratal geometries together with stacking patterns to identify systems tracts and (4) use surfaces and systems tracts to define depositional sequences; (5) observe stacking trends of depositional sequences to define sequence sets and composite sequences.

The challenge of sequence stratigraphy is to take a vast array of possible observations and differentiate regionally significant stratal units and surfaces from locally developed ones. Locally, the ratio of sediment supply to accommodation can vary and stratigraphic architecture might diverge from a generalized or representative pattern of the entire margin, well documented in the Holocene (e.g. Anderson et al. 2013). The definition of a depositional sequence (Mitchum et al. 1977a) considers data resolution (‘relatively conformable’ implies that data resolve the presence or absence of significant gaps) but not time, sea-level or regional extent. Fortunately, stratal terminations and vertical stacking patterns of strata can record changes in the balance and vector of δA/δS on a margin scale to make identification of key surfaces easier.

Simple stratal geometric relations have significant implications. For example, observed shoreface parasequence stacking was summarized by Van Wagoner et al. (1988, fig. 1) in three simple patterns: progradation, aggradation and retrogradation. As illustrated there, if each vertical subdivision (parasequence) is assumed to be of equal time duration, the changing factor responsible for shoreline movement is purely sediment flux relative to a constant rate of accommodation increase (e.g. Vail et al. 1977). This is obviously an idealized pattern, but similar geometric constructs are common in the literature or output from computer modelling runs to produce patterns that could occur in nature under certain input conditions (e.g. Helland-Hansen & Martinsen 1996; Paola 2000; Burgess & Prince 2015). In nature, the signals of accommodation and sediment flux can have local variation owing to autogenic processes but follow regional trends occurring in predictable successions and in hierarchal stacks (Neal & Abreu 2009). The δA/δS method focuses on regional allocyclic observations; ones that record significant changes in the ratio and vector of rate of coastal accommodation increase or decrease (δA) and rate of sediment flux (δS). The three types of vertical or lateral shoreline trajectories that change across key bounding surfaces are as follows: (1) progradation to aggradation stacking (PA), bound by sequence boundary below and maximum regressive surface above; δA/δS <1 and increasing; (2) retrogradation stacking (R), with maximum regressive surface below and maximum transgressive surface above; δA/δS >1; (3) aggradation to progradation to degradation stacking (APD), bound by maximum transgressive surface below and sequence boundary above; δA/δS <1 and decreasing.

It should be noted that the δA/δS ratio is never stable at unity as lateral movement of the shoreline is always occurring as tracked by the offlap break. Offlap break refers to the convex part of a depositional clinoform representing the shelf edge, delta front or shoreface basinward of the subparallel topset that represents coastal-plain deposition at the shoreline (Fig. 1). Use of maximum regressive surface (MRS) and maximum transgressive surface (MTS) terminology is intended to clearly communicate the nature of these surfaces as recorders of the most basinward (uneroded) and landward positions of shoreline, respectively, within a sequence (modified from Embry & Johannessen 1992; Embry 2009; Abreu et al. 2014).

The method tracks the δA/δS ratio as it is preserved in the rocks. To do so, the core assumptions of the δA/δS method are that (1) coastal accommodation varies over time at a changing rate and is quasi-periodic over multiple scales of observation and (2) sediment fills local accommodation to base level and any local surplus is transported to the offlap break where excess accommodation exists. These assumptions simplify interpretation of the complexity in fluvial or coastal-plain environments and put greater emphasis on shoreline trajectory, which is the shoreline path viewed along a cross-sectional depositional-dip section (Helland-Hansen & Gjelberg 1994). These concepts place focus on vertical stacking, stratal geometries and stratal terminations that preserve the record of continually varying rates of coastal accommodation and sediment flux represented by facies and facies associations. Knowledge of sea-level or time duration is not required.

To illustrate the δA/δS interpretation method, two high-resolution datasets were selected. The first is a depositional dip line of section through an Experimental EarthScape (XES02) flume tank run made at St Anthony Falls Laboratory, University of Minnesota (Kim et al. 2006; Martin et al. 2009) and the second is a high-resolution seismic traverse constructed along depositional dip (Sydow & Roberts 1994; Roberts et al. 2004) from the Pleistocene–Holocene northern Gulf of Mexico shelf. Each of these datasets has the advantage of a high resolution; clearly preserved stratal architecture and high-resolution age control that tie to actual external forcing mechanisms of base-level changes. Although it is useful for reference to have such good sea-level control, the observations and interpretation method do not require such information. The final example, from the Upper Cretaceous Ferron Formation outcrop, demonstrates application of the δA/δS interpretation method and sequence boundary definition in an environment without known sea-level history.

Experimental stratigraphy example

The XES02 experiment (Fig. 2) was designed and executed with control of subsidence rate and gradient, sediment and water input, basin ‘sea-level’ change (both magnitude and duration) and real-time 3D imaging of basin fill (Martin et al. 2009). The details of the experiment have been described by Kim et al. (2006). The sediment mixture consisted of 63% 120 mm (very fine) silica sand, 27% bimodal anthracite coal (75% 190 mm and 25% 490 mm), and 10% kaolinite clay. The silica sand acts as the coarse sediment fraction because of its relatively high specific gravity (sg = 2.65), and the coal, although of larger grain size, is lighter (sg = 1.3) and behaves as the finer-grained sediment fraction. Although the results obtained look very similar to a regional stratigraphic dip section, it must be recognized that the stratigraphic section image produced represents a length of less than 6 m long by 1 m tall and was built during a run of sedimentary fill into a flume tank for over 310 h with 10–15 cm basin water-level oscillations. Despite its small size, this experiment is not simply an illustration, but rather is a scaled version of hydrodynamic and depositional processes giving a unique perspective into depositional responses to controlled external forcing (Paola et al. 2001; Kim et al. 2006).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Dip oriented stratigraphic section taken from Experimental EarthScape XES02 run and the conditions of base-level change that produced it (modified from Martin et al. 2009), overlain with sequence set stacking trends.

δA/δS method applied to experimental data

The XES02 section (Fig. 2) starts with a Stage 1 shelf-building phase made from constant sediment influx and steady subsidence rate overprinted with a slow cycle of basin ‘sea-level’ fall and rise, followed by a rapid cycle, and then by an equilibrium period (Kim et al. 2006; Martin et al. 2009). Next, a long-term fall and rise cycle, Stage 2, was superimposed with high-frequency cycles of base level to examine sequence-set stacking. To illustrate the δA/δS method, we will focus more closely on the geometries resulting from the rapid cycle of Stage 1 and the falling limb of Stage 2 and interpret the model output following a flowchart of steps (Fig. 3).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Flowchart of interpretation steps in the δA/δS method.

In the centre of the uninterpreted section (Fig. 4a) we observe the facies and facies associations of two tongues of darker material (anthracite coal particles) representing a suspended fine material fraction and recording shoreline transgressions that bracket a prograding wedge of lighter-coloured sandy material that represents proximal fluvial and deltaic to bypass subaqueous gravity flow deposits. First, we mark stratal terminations and the position of the offlap break (Fig. 4b). When marking stratal terminations, we note if the termination is of overlying strata onto a underlying surface (downlap or onlap) or underlying strata against an overlying surface (truncation or toplap). Starting at the lower dark layer (Fig. 4a, I10–B5), dipping strata above terminate against this layer at an angle so it is marked with downlap arrows (Fig. 4b). At the upper end of these dipping layers (Fig. 4a, B4–F6), truncation under an erosional surface is observed. Here too we can see the offlap break stacking pattern emerge. Starting at Figure 4a, B4, the offlap break appears to stack upward and basinward to D5 (subsidence has tipped the whole section basinward; Kim et al. 2006) and then basinward to downward from D5 to F6. Thus we describe this stacking pattern as aggradation–progradation–degradation (APD) and interpret that δA/δS was <1 and decreasing to negative through this interval.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Expanded view of XES02 section (location shown in Fig. 2) for illustration of the δA/δS method: (a) uninterpreted with reference grid; (b) interpreted with stratal terminations and offlap breaks; (c) overlain with offlap break stacking patterns; (d) interpreted with key bounding surfaces based on stratal terminations and stacking patterns.

The basinward end of the stratal package with APD stacking shows toplap under a surface that also has widespread onlap terminations (Fig. 4a, G6–I9) and a basinward shifted unit of coarser grained sediment with increased bypass (Fig. 4a, H8–J10). These characteristics define a depositional sequence boundary (Fig. 4d). Onlap terminations occur subaqueously throughout the shelf-building prograding packages as delta avulsion occurs, but the surface that is correlative to the sequence boundary is significantly more widespread than any other onlap surfaces within the prograding package (Martin et al. 2009). Such widespread onlap surfaces have good preservational potential, even in high-frequency base-level falls superimposed over a long-term fall. These recognition criteria make it easier to differentiate local autogenic surfaces from regional allogenic ones that are key to stratigraphic frameworks.

The next major sediment package above the sequence boundary is the onlapping and downlapping bypass unit (Fig. 4b, H8–J10) that is itself downlapped by a unit of renewed progradation, toplap terminations and clinoform topset preservation (Fig. 4b, G6–I7). Topset preservation is evidence for coastal accommodation increase, as is offlap break aggradation. Offlap break basinward movement with topset aggradation can either be accelerating (aggradation to progradation; AP) or decelerating (progradation to aggradation; PA). In this interval the lateral movement of offlap breaks is decelerating and the stacking pattern is PA, which indicates δA/δS <1 and increasing (Fig. 4c, G6–I7). Further, the top of this package represents the most basinward regression (MRS; Fig. 4d) between our two reference transgressions.

Overlying the MRS is a stratal package with a landward-thickening wedge of coastal deposition, landward-stepping of the offlap break, and internal units that backstep and lap down (Fig. 4a, I7–A2) onto the MRS or the previous prograding unit. All of these characteristics indicate retrogradation (or R) stacking that culminates at the MTS (owing to δA/δS >1; Fig. 4c).

With this set of observations and initial interpretations, we can complete the steps of the δA/δS method to define systems tracts and a depositional sequence. The MTS (Fig. 4c) is overlain by a package of strata that stacks aggradationally, then increasingly progradationally and eventually degradationally into the basin (APD stacking; Fig. 4c). The top of this first package is defined by a surface with erosional truncation beneath it on the shelf and widespread marine onlap, basinward shift of facies and increased bypass to the basin atop the correlative continuation of this surface into the basin. These are the original defining criteria for a depositional sequence boundary (SB; Mitchum et al. 1977a; note that there is no mention of sea-level or time duration involved). Overlying the sequence boundary, regression continues and the stacking pattern of offlap breaks is progradation to aggradation, culminating at the MRS. The MRS marks the onset of transgression and retrogradational stratal stacking that culminates in the next MTS.

Above the upper MTS, the basal interval of the next regressive package laps down onto the MTS (Fig. 4a, 2A–2B) and its uppermost interval has toplap and truncation terminations of topsets (Fig. 4a, 1A–1B). Offlap break stacking is progradation to degradation. An onlap surface is observable at the basinward end of the degradational stacking (Fig. 4a, 2C–5G), indicating the next SB. Above this level, much of the expected succession is missing owing to erosion by the overlying composite incised valley that is formed during the long-term base-level fall and overprinting high-frequency falls and rises (Martin et al. 2009). Through this long-term fall, the dominant stratal termination is topset truncation but two key surfaces can be observed within the underlying obliquely prograding unit: (1) a downlap surface from F3 to H5; (2) an onlap surface from H4 to J7 (Fig. 4a). In the XES02 experiment, downlap surfaces associated with maximum transgressive surfaces are muted during the falling limb of the composite base-level cycle Stage 2 as are offlap breaks owing to erosional truncation but the widespread marine onlap surfaces are persistent and mappable (Martin et al. 2009; see also discussion by Van Wagoner et al. 1990).

Considering the full XES02 section, the Stage 2 long-term cycle with superimposed high-frequency cycles shows sequence-set stacking (Fig. 2). A sequence set was defined by Mitchum & Van Wagoner (1991) as a set of sequences arranged in a distinctive progradational, aggradational or retrogradational stacking pattern that can be aggregated into a succession of genetically related sequences, known as a composite sequence, in which the single sequences stack into lowstand, transgressive and highstand sequence sets. Mitchum & Van Wagoner (1991) proposed that component depositional sequences of a given sequence set have different character. Neal & Abreu (2009) refined this approach to be more descriptive and less tied to sea-level by showing how sequence sets stack with similar trajectories as systems tracts in an APD–PA–R succession. We have seen in the XES02 section that widespread marine onlap surfaces merged with the fluvial erosion surfaces are SBs. Identification of depositional sequences (SB to SB) that stack in a degradational manner define a degradational sequence set. In the XES02 example, input parameters are known as is the amount of time represented by strata and by key surfaces (Martin et al. 2009). Although base-level cycles are incompletely recorded by the strata, the observed sequence trajectory is APD for the equilibrium period to src2, PA for src3 to src4, and R for src5 and src6 (Fig. 2). The asymmetry of sequence preservation for the XES02 section is a function of the limitations of the experimental design. In regional seismic sections over passive margins (e.g. Pelotas Basin; Neal & Abreu 2009), the PA sequence set is best developed because it has the best preservational potential.

Application to high-resolution seismic data, Pleistocene Lagniappe Delta

Although the experimental data from the XES02 run, with its well-calibrated input and output, are very instructive, application of the method must survive limitations of real world data. The seismic line selected for our next example is from a grid of data with core calibration acquired by the Gulf of Mexico Shelf–Slope Research Consortium (GOMSSRC) in the late 1980s to 1990s (Fig. 5a). Details on the history and analysis of these data have been reported by Sydow & Roberts (1994) and Roberts et al. (2004). The dataset covers a Pleistocene incised valley (Fig. 5b) that fed the shelf-edge Lagniappe Delta at the last Glacial Maximum lowstand of sea-level (Sydow & Roberts 1994; Roberts et al. 2004). The seismic data source used triple-plate boomer (500 J per plate) with tandem-mounted 4 cu. in. sleeve guns to produce high-resolution single-channel data with 200 ms or more penetration and vertical resolution of <1 m (Sydow & Roberts 1994). A core hole, MP 303, was acquired in 1988 to provide age control and analyse depositional facies tied to seismic facies. Radiometric ¹⁴C age dating from MP 303 gives a precise (±100 year) calibration to δ¹⁸O-derived eustasy proxy curves (Marine Isotope Stage, or MIS, Imbrie et al. 1984). Despite multiple azimuths of data collected across the buried Lagniappe Delta incised valley, no single line existed that clearly intersected both core hole MP 303 and the terminus of the incised valley fill in a well-imaged depositional-dip direction. As such, two lines were selected with an 18 km along-strike jump tie (Fig. 5b). This is a robust selection, as these lines tie at the sea floor and the incised valley can be correlated and mapped throughout the rest of the grid. The orientation of the selected line segments shows an oblique strike view of the incised valley through MP 303 that gives a clear relationship of the cored fluvial valley fill facies and the complexity of the underlying delta lobe deposition and internal stratal surfaces. The jump tie projection captures the terminus of the valley as it shallows, broadens and becomes relatively conformable (Sydow & Roberts 1994) then makes a transition into the overlying transgression.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

(a) Location map for high-resolution seismic grid over the US Gulf of Mexico shelf Pleistocene Lagniappe Delta incised valley (IVF); (b) isochrons of incised valley facies with the selected seismic line to illustrate the δA/δS method (modified from Roberts et al. 2004).

The most prominent feature in the MP 303 core and on the seismic line was identified by Sydow & Roberts (1994) as the MIS 2 lowstand fluvial incised valley cutting from 115 ms at sp 225 down to 140 ms at sp 606 (Figs 6 and 7). This surface is a sharp contact between pre-MIS 2 deltaic or bay facies below and fluvial facies of MIS 1–2 above. In seismic data, stratal terminations are observed throughout the section, but the regionally significant surfaces become more apparent through application of the δA/δS method. Preserved remnants of strata with offlap breaks below the MIS 2 erosional surface show a progradation to degradation stacking (Fig. 6). Above the erosional surface basinward of sp 615 offlap breaks are observed to stack in a progradational to aggradational pattern to sp 637, where the maximum regression is observed. Above this point a landward-thickening wedge of parallel continuous seismic facies drapes the internally complex seismic facies of the incised valley fill. In the MP 303 core this seismic facies transition is marked by a deepening surface that separates fluvial from marine shelf deposits. Candidate sequence boundary and maximum regressive surfaces can be interpreted based on these observations.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

Regional high-resolution seismic line over the IVF (location shown in Fig. 5) in strike view through the MP 303 core hole and dip view projected c. 18 km along-strike, overlain with interpreted stratal terminations, offlap breaks, stacking patterns and key surfaces (modified from Sydow & Roberts 1994). MP 303 core gamma ray (left) and grain size (right) crosses the MIS 2 incised valley between Sydow & Roberts (1994) Units 2 and 3.1

• Download figure
• Open in new tab
• Download powerpoint

Fig. 7.

Expanded strike view of the IVF (location shown in Fig. 6), (a) uninterpreted and (b) interpreted to illustrate regionally significant stratal terminations (surfaces that mark significant changes in coastal accommodation) that differ from local autogenic ones (surfaces with relatively conformable facies juxtaposition).

At greater detail, stratal terminations are present along multiple surfaces in the section, so it is important to separate surfaces that are local in extent (and most probably autogenic) from surfaces that are regional (mostly allogenic; Fig. 7). Sydow & Roberts (1994) mapped several distributive delta lobes within their seismic data that have surfaces of internal downlap and toplap that most probably record delta-lobe progradation following lobe switching. In MP 303 core, there is little doubt that the sequence boundary should be placed at the base of the fluvial Unit 2 of Sydow & Roberts (1994) but correlation of that surface away from the core presents some options, based on the lateral relation with Unit 3.1 that occurs just below Unit 2 in MP 303. Unit 3.1 is 3 m thick and fine grained with current-ripple beds and a mixed-salinity faunal assemblage, interpreted as an inter-distributary bay that overlies progradational delta units (3.2–3.4 of Sydow & Roberts 1994). Unit 3.1 tied to seismic data has a sharp contact of internal downlap or onlap at its base and truncation at its top (Fig. 7; sp 245.5, 97 ms and sp 244, 93 ms). Sydow & Roberts (1994) interpreted their sequence boundary directly above a thin unit of shingled seismic facies to the west (left) of the well (Fig. 7; sp 237, 87–91 ms to sp 244, 88–82 ms). Roberts et al. (2004) went even higher with their sequence boundary, which raises the question: ‘Why would there be a jump in accommodation such that the shingled unit laps down onto the Unit 3.1 truncation surface?’ The δA/δS method recognizes these strata terminations and change in vertical succession as a change from APD to PA stacking and places the sequence boundary below the shingled unit at the truncation surface above Unit 3.1.

More evidence of increasing accommodation following sequence boundary formation comes from the terminal end of the incised valley (Fig. 8). Stratal terminations observable between sp 609 and 616, 105–115 ms (truncation and onlap; Fig. 8b) point to a surface of fluvial erosion cutting into underlying prograding delta facies, correlated to the incised valley fill penetrated in MP 303 (Sydow & Roberts 1994). As in Figure 7, there are many stratal terminations in the seismic section so it is difficult to distinguish the regional (allogenic) from the local (autogenic) ones based on termination character alone. Here the δA/δS method helps by placing terminations in the context of stratal stacking patterns to differentiate local from regional surfaces. Basinward of sp 614, 110 ms, delta-lobe stacking is progradational with an increasingly aggradational vector. Although numerous surfaces of truncation and toplap are seen within the full seismic section, only one surface demarcates a change in stratal stacking pattern (from APD to PA, sequence boundary; Figs 6 and 8b).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 8.

Expanded dip view of the IVF terminus (location shown in Fig. 6), (a) uninterpreted and (b) interpreted to illustrate the transition to PA stacking above the sequence boundary and the differences between local autogenic and regional allogenic terminations. A regional-scale illustration (c), shows aggradation of the offlap break in the full context of the incised valley fill updip.

In the δA/δS method, a depositional sequence boundary is characterized by an abrupt basinward shift in depositional facies, a change from APD to PA stacking, and basinward shift in coastal onlap. The rise of the offlap break and aggradation of delta-lobe topsets are the physical stratigraphic evidence of coastal accommodation. In this example, fluvial seismic facies are interpreted to aggrade and backfill the incised valley in a PA pattern (Fig. 8b). Thus, we interpret the depositional sequence boundary at the base of the incised-valley fill (as did Sydow & Roberts 1994). PA stacking records positive δA/δS <1 and increasing that overlies the basinward shift or degradational stacking at the end of the underlying depositional sequence. Regionally extensive coastal onlap is the key diagnostic stratal termination, especially in basinward reaches. Onlap, however, is not unique to the sequence boundary and it is commonly obscured within the incised valley fill, complicating the separation of fluvial terrace deposits (formed during base-level fall) from incised-valley fill (associated with base-level rise). Figure 8c summarizes the physical stratigraphic observations from the Lagniappe Delta high-resolution seismic dataset that define the depositional sequence boundary: truncation beneath a regionally extensive surface that separates APD from PA stacked strata above. Above this PA stacking a transgression is recorded above the fluvial Unit 2, tying back to the MP 303 core, which marks the correlative MRS.

Application to outcrop strata, Upper Cretaceous Ferron Formation

Notwithstanding very fine details in the chronostratigraphy of the sequence boundary, the power of the δA/δS approach can be seen from application to outcrop sections with less constrained input parameters. When detailed and accurate observations of facies, facies associations, vertical stacking, stratal geometry and stratal terminations are made, interpretation choices are clear and can be objectively discussed. For example, the stratigraphic framework of the Cretaceous Ferron Delta in Utah by Zhu et al. (2012) carefully documents the physical stratigraphic relations across outcrop sections covering a 30 km depositional dip traverse (Fig. 9a). This framework is sufficiently detailed and observational that stacking pattern trends can be recognized and alternatives can be proposed as discussed by Zhu et al. (2012). In their framework, depositional sequence boundaries were placed above parasequence (PS) sets 17, 15, 12, 8 and 4, with possible sequence boundaries above and below PS set 11d–11a (Fig. 9a). Zhu et al. chose to recognize a forced regression systems tract (FSST) (sensu Posamentier et al. 1992), which places a sequence boundary below the FSSB with a lowstand systems tract (LST) overlying, based on a model for onset of sequence boundary formation with the initial drop in base level. An alternative interpretation, from the δA/δS approach, would recognize the continuum of aggradation–progradation–degradation stacking for PS set 12 and PS 11e with a basinward shift and renewed coastal onlap occurring for PS 11d–11a and again for PS 10c–10a (Fig. 9b). PS 11e appears to degradationally stack relative to PS 12a but a definitive surface of basinward shift in coastal onlap is absent (Zhu et al. 2012). Such a surface is present between PS 11e and PS 11d, where upper shoreface deposits of PS 11d onlap offshore transition mudstones of PS 11e. Such a non-Walther's Law relationship requires basinward shift of coastal onlap and objective placement of a sequence boundary.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 9.

Dip section stratigraphy (a) of the Upper Cretaceous Ferron Formation Notom Delta from Zhu et al. (2012) with (b) tracking of offlap break movement through numbered parasequences illustrating stacking patterns that combine with stratal terminations to define key surfaces of a δA/δS method interpretation. In a hierarchal stack (c), APD and PA sequence sets are identifiable with the composite sequence boundary above PS set 11.

The importance of the alternative interpretation is that the key bounding surfaces are defined by objective recognition criteria (stratal terminations of onlap above and truncation below) as opposed to an interpretation of when relative fall in sea-level began (FSST). Practically, the alternative interpretation highlights that PS 10c may become disconnected from the earlier deposits of PS set 11 and set up a possible hydrocarbon stratigraphic trap. Characteristics of a depositional sequence are influenced by its position within a sequence set (Mitchum & Van Wagoner 1991). Using the δA/δS method for the Ferron Formation shows an APD sequence set from PS 18 to PS 11 with a PA sequence set from PS 10 and above (Fig. 9c). In shelf-slope basins, the surface overlying degradational stacking at the margin is commonly, but not uniquely, the location of the best-developed deep marine fan reservoirs (e.g. Gong et al. 2015). It was shown in experimental data too that the surface capping degradational stacking had the best-developed basinal bypass sand deposition (Martin et al. 2009). On the shelf, recognition of sequence set stacking can reveal overlooked stratigraphic trapping potential and predict potential fan reservoir presence in the basin.

Conclusion

The δA/δS method of sequence stratigraphy is a return to an observation-based approach to constructing frameworks based on stratal terminations, key surfaces and stacking patterns of stratal units that define systems tracts and depositional sequences. Analogously, at a larger scale within a hierarchal framework, observations of stacking trends of depositional sequences define sequence sets and composite sequences. Sequence stratigraphy is a powerful tool to construct an ordered stratigraphic framework that leads to testable predictions away from control points. When a robust framework is in place, mechanisms that control the character of that framework can be hypothesized. Mechanisms (e.g. eustasy, climate, etc.) and time duration (order by any definition) are not parts of sequence stratigraphy interpretations by the δA/δS method but the method produces frameworks that can be used to more reliably interpret causal mechanisms. The future of sequence stratigraphy may be to use higher resolution subsurface imaging to produce robust frameworks that lead to a better understanding of causal mechanisms. A rigorous method is needed to generate the most accurate predictions possible.

Sequence boundary formation is a good example of a problem that can be better addressed with a rigorous method. Using observational criteria to decipher stratigraphic complexity and the δA/δS method, repeatable interpretation of high-resolution data in clastic depositional margins shows the sequence boundary to separate degradational from progradation–aggradational stacking and widespread marine onlap. Research showing that the incised valley fill portion of the sequence boundary is a chronostratigraphically complex surface does not invalidate its use as a significant surface and a base for predictions from frameworks using it. Indeed, the complexity of a sequence boundary makes it an ideal focus for future research, using a rigorous method to interpret its location and the variety of processes that occur during its formation and how they vary along it.