Sandstrom et al. (2020) present new elevation and age data for a flight of four marine terraces preserved along the western limb of the Cape Range anticline in western Australia. Their interpretation of these data provides an alternative estimate for the amount of tectonic deformation that has occurred since terrace formation. They conclude that less tectonic uplift has occurred in the region than previously reported and posit that their study provides a template for reducing the uncertainty associated with last interglacial paleoshoreline reconstructions.

We have three principal comments regarding the methodology and data interpretation presented in Sandstrom et al. (2020). First, their method for measuring deformation is neither a measurement of, nor a proxy for, tectonic uplift. Second, their method for calculating deformation rates is internally inconsistent. Third, their interpretation of the upper and lower Tantabiddi notches as two independent time-stratigraphic units together with their age assignment for the upper notch generate warping rates that are untenable. These shortcomings individually and collectively lead to an erroneous conclusion regarding the region's tectonic stability and could lead future paleo–sea-level researchers to underappreciate the tectonic component of land-level changes that are ongoing and well documented in the region. The purpose of this comment is to highlight these issues and their implications for sea-level studies.

In addition to the shortcomings identified in the Sandstrom et al. (2020) methodology for assessing uplift and deducing deformation rates, there are a number of internal inconsistencies in their presentation. Specifically, there are errors and inconsistencies in structural geology and active tectonic nomenclature that confuse their analysis. To establish a framework for our comment, we begin by clarifying this terminology in the following section.

The topographic expression of the Cape Range reflects tectonic uplift consequent to the structural inversion of Mesozoic normal faults in the Carnarvon Basin that began in the late Miocene (e.g., Malcolm et al., 1991). The Cape Range anticline is a fault-cored doubly plunging asymmetric fold with an easterly vergent axial surface that strikes approximately north-south (Malcolm et al., 1991). Fold limbs are defined as dipping approximately perpendicular to the strike of the axial surface of an anticline (e.g., Ramsay, 1967). Therefore, the limbs of the Cape Range anticline dip steeply to the east and shallowly to the west (Figs. 1A and 1B). In detail, the topographic expression of the Cape Range is the sum of the reverse and reverse-oblique motions along the underlying fault and the associated folding across several related geological structures (Fig. 1A). Variability in the amount of displacement along the fault plane, and/or the geometry of the fault, has resulted in curvature of the hinge line of the overlying fold (Fig. 1B).

Sandstrom et al. (2020) define “relative deformation” as the “elevation difference between analogous sea-level indicators on the limbs of the anticline compared to the central apex for each terrace.” They use the terms northern and southern “limbs” of the Cape Range anticline; however, these relate to the northward and southward plunge of the fold axis, not to the eastern and western limbs of the anticline (Fig. 1B). Additionally, the authors refer to the maximum terrace elevations as being at the “apex” or “fold apex,” but this should be referred to as the culmination. Sandstrom et al. (2020) also use the terms displacement, uplift, vertical warping, shoreline elevation changes, and relative deformation interchangeably throughout their text to describe what they have measured. However, these terms are not synonymous. Within this comment, we explicitly use the term warping (so as not to be confused with uplift) to indicate the range in elevation of each individual terrace surface/feature measured by Sandstrom et al. (2020).

The incorrect usage of structural and tectonic terms leads readers to misunderstand what has been measured and how it relates to previous investigations of the Cape Range anticline. Sandstrom et al. (2020) compare their “relative deformation rates” to uplift rates interpreted by others, but they are not comparing like things.

Sandstrom et al. (2020) use differential global positioning system (DGPS) and airborne light detection and ranging (LiDAR) derived elevation data to construct a series of longitudinal and coast perpendicular transects on a flight of four marine terraces along the Cape Range coast. The elevation of each terrace varies along the coast, and they use this variability in elevation to constrain tectonic deformation. Neither what they measure nor how they use the measurements achieves the goal of assessing tectonic uplift; these two flaws are discussed in the following text sections.

Terrace Measurements

The elevation change measured by Sandstrom et al. (2020) is marked as “B” on Figure 2. This measurement is the difference between the maximum elevation of each terrace surface at the culmination and the minimum elevation of the terrace surface to the north and south. This is a measure of the amount of warping that has occurred on an individual terrace surface due to the plunging of the anticline fold axis. As shown on Figure 2, the elevation changes marked “A” and “C” are measures of net tectonic uplift from a mean sea level (MSL) baseline elevation at the time of shoreline formation. Measurements “A” and “C” provide the minimum and maximum uplift since terrace formation. A prior sea-level datum must be known or estimated in order to determine “A” and “C.” These are the measurements of convention for tectonic geomorphological studies (e.g., Carver et al., 1985; Lajoie, 1986; Kelsey and Bockheim, 1994).

There is no interdependence between elevation changes “A,” “B,” or “C” on Figure 2. “B” does not provide an independent indication of uplift, nor does it provide a quantifiable proxy for uplift. By measuring “B,” there is no established datum from which to measure uplift as it pertains to tectonics or sea-level. Rather, each terrace is being measured against itself in a relative reference frame. The value for “B” is herein referred to as warping. We note that the northernmost and southernmost data points from Sandstrom et al. (2020) do not correspond to the northern and southern ends of the anticline. Therefore, the measured warping presented by the authors is only a minimum value derived from measurements along part of the length of the anticline; had the measurements been made farther to the north or south, there would be greater net warping and thus higher warping rates. Nonetheless, these warping rates are not representative of net tectonic uplift.

In every case where the authors use the terms displacement, uplift, vertical warping, shoreline elevation changes, and relative deformation to describe their analyses, they are referring to elevation change “B” in Figure 2. This is a measure that relates to the progressive curvature of the hinge line or axis of the doubly plunging fold. The “relative uplift rates” presented in table 2 of Sandstrom et al. (2020) are not uplift rates pertaining to tectonic geomorphology or sea level and are therefore not comparable to tectonic uplift rates presented by others.

Assessment of Warping Rates

Sandstrom et al. (2020) calculate a warping rate for each terrace by dividing the warping (“B” value in Fig. 2) by the terrace age. This yields four independent warping rates (left panel, Fig. 3). However, the terraces are not warping independently of one another. The warping rate does not change by terrace; the warping rate changes with time (right panel, Fig. 3) and at different positions along the length of the hinge line. Once a terrace is formed, the elevation difference between that terrace and all of the older terraces at any given location is locked. Warping rates are reconstructed over time by removing the amount of warping (elevation difference) and the amount of time (age difference) between successive time-stratigraphic markers (right panel, Fig. 3). This elevation relation is schematically illustrated on Figure 4, which shows an example of progressive tectonic deformation (via progressive warping of successively older terrace surfaces) as observed at the Cape Range; each older terrace exhibits more warping. This reconstruction is illustrated in figure S7 of Sandstrom et al. (2020); however, there is no discussion of that reconstruction in their text, and the implications are not included in the calculation of warping rates.

For example, the warping rate experienced by the 120 ka lower Tantabiddi terrace was also experienced by all of the older terraces during the last 120 ka. The warping was being driven by deformation of the Cape Range anticline, on which the terraces were merely passengers. Therefore, with the exception of the lowest surface, the calculation of rates cannot be derived by simply dividing the amount of warping by the age of the terrace as the authors have done. The terraces provide a means to measure the warping of the anticline, but there is not a warping rate per terrace. Stated simply, the warping is cumulative; all of the warping a younger terrace experiences, by definition, must also have been experienced by the older terraces. Compounding implications of this approach, as it influences morpho-stratigraphic interpretations, are discussed in the following section.

Sandstrom et al. (2020) proposed new ages for the three highest marine terraces (Jurabi, Milyering, and Muiron) based on strontium isotope analysis of samples from each terrace. The new age estimates are significantly older than previous published ages (e.g., van de Graaff et al., 1976; Veeh et al., 1979; Stirling et al., 1998; Clark et al., 2011). No new samples were analyzed from the Tantabiddi shoreline features and no new age estimates are presented in their discussion of terrace chronology or Sr isotope analyses. Despite this, Sandstrom et al. (2020) have included a reinterpretation of the morpho-stratigraphy of the Tantabiddi terrace. They propose two time-stratigraphic units based on elevation differences between two geomorphic notches and the distribution of associated marine sediments. They refer to these two notches as the “upper” and “lower” Tantabiddi notches. Based on an absence of marine limit sediments in proximity to the upper notch, they assert that the upper notch must be related to an older sea-level high stand; the upper notch is cut into the seaward edge of the Jurabi terrace, so they assign the 1.045 Ma Sr age for the Jurabi terrace as a maximum limiting age for the upper Tantabiddi notch. This limiting age is not the age of the upper notch, it only constrains their oldest possible age.

The implications of interpreting the Tantabiddi features as two independent time-stratigraphic units significantly changes the resultant warping rates (when the warping rates are calculated progressively as discussed above). To illustrate how the derived rates change over time, depending on the interpretation, we have recalculated the rates using the data presented on table 2 of Sandstrom et al. (2020). To simplify the table and discussion, we use only the minimum and maximum ages for each terrace to present the scenario that interprets the Tantabiddi notches as two independent units (Table 1). If the Tantabiddi terrace notches are considered to be two independent time-stratigraphic markers with the limiting age assignments of Sandstrom et al. (2020), then the warping rates must have changed by an order of either five or ~100 times compared to the next highest warping rate (gray highlighted text in Table 1); this variability in rates is unjustifiable.

Alternatively, for comparison, we conduct the same analysis considering a single Tantabiddi time-stratigraphic marker of marine isotope stage (MIS) 5e age as previously interpreted by other authors (e.g., Clark et al., 2011, 2012; McPherson et al., 2013; Whitney and Hengesh, 2015a). Again, the warping rate changes with time, however, variations in rates are within the same order of magnitude.

The Sandstrom et al. (2020) interpretation of warping rates for the Tantabiddi shoreline features appears unreasonable. Given the lack of supporting data, we suggest this interpretation should be rejected. The morphological continuity of the MIS 5e shoreline features is remarkable along the coast from Cape Cuvier to the northern tip of the Cape Range (e.g., O’Leary et al., 2008). Minor variations in the elevations of the wave-cut notch and marine limit features are well documented along the coast on both Pleistocene (e.g., Hearty et al., 2007) and active shoreline features (e.g., Whitney and Hengesh, 2015a). The suite of marine limit features identified as the coastal type section include low, high, and storm cut notches, fossil beaches, and reef deposits (Whitney and Hengesh, 2015a); within this section are systematic paired shoreline notches associated with modern and paleo–sea levels. Paired notches are interpreted as forming due to the 1.5 m tidal range creating low and high tide notches with an elevation range of 2–3 m along the active shoreline (Whitney and Hengesh, 2015a) consistent with other global sites (e.g., Fischer, 1980; Wziatek et al., 2011; Hall, 2011).

Minor variations in elevations of marine limit features are illustrated on Figure 5 (modified from figure S4 from Sandstrom et al., 2020). Sandstrom et al. (2020) present the elevations of three marine limit units associated with the Jurabi section. They are: Jurabi terrace, Jurabi beach sands, and Jurabi highest marine. The elevation range between these units within each coast perpendicular transect ranges from less than a meter to greater than 2 m, perhaps greater than 4 m (see cluster of three transects near −22.3 degrees latitude on Fig. 5). Further, their data indicate that there is no stratigraphic concordance between these three units. That is, nine of their 22 transects show the Jurabi highest marine at a higher elevation than the Jurabi terrace, whereas another 10 transects show the Jurabi highest marine at a lower elevation than the Jurabi terrace. For two of their transects, the Jurabi beach unit is at higher elevation than both the Jurabi highest marine and Jurabi terrace, and for one transect all three units essentially occupy the same elevation. Neighboring transects x and y (our labels on Fig. 5) show this discordance with the inversion of stratigraphic position between the Jurabi highest marine and Jurabi terrace. Comparing transects x and z, the total range in elevation of the Jurabi units, as reported, is close to 5 m. Similarly, with the reported elevations for the Tantabiddi sequence, we note that the data points from their shore perpendicular transects near ~22.1 degrees latitude show a greater range in elevation for upper notches measured on two adjacent transects than between the upper and lower notches on a single transect (Fig. 5, unfilled rectangle).

There are four points to the preceding discussion. First, the initial horizontality of terrace formation has a variability of at least 3 m as measured along the active shoreline. This variability could be up to 5 m based on the stratigraphic discordance and inversion observed by Sandstrom et al. (2020) in the Jurabi section. Second, the absence of marine sediments associated with the notch is common as evidenced by the stratigraphic inversion within the Jurabi sequence. Third, the elevation variability of individual units is not considered by Sandstrom et al. (2020). Fourth, this variability is significant given the overall low warping rates and small elevation changes being considered in the analysis. For instance, if the cumulative warping of 6.86 m the authors measure for the Tantabiddi terrace (Table2) considered the variability, then perhaps only 1.86 m of warping has occurred to the Tantabiddi terrace in the past 116 ka, which would yield a warping rate of 0.016 mm yr–1. This warping rate would be within the range of warping rates estimated from the older terraces. Nevertheless, the contribution of uncertainty this variability brings to the rate calculation is not considered.

Further, each symbol shown on Figure 5 is the average of the three highest elevations for each marine-limiting feature per transect. By averaging the maximum elevations, the authors are doing two things: (1) not reporting the actual elevation differences between limiting units within a single transect; and (2) arbitrarily reducing the maximum reported elevations. The former means the reported difference in elevations for each unit, as discussed above, is lower than the actual range. Therefore, the ~5 m elevation range within the Jurabi sequence may not be the full range. The latter point implies that the maximum limiting sea level elevation has been artificially reduced; we suggest that the paleo–sea-level constraint should be based on the highest observed in situ coral, not the average elevation of the three highest corals.

In summary, the range in elevations within the coastal type section for both the Tantabiddi sequence and the Jurabi sequence is at least a few meters. Sandstrom et al. (2020) use the principal of horizontal deposition without considering this range. Where the upper and lower notches are observed on the same transect or adjacent transects, they are within 3 m elevation of each other, except for a single location. This range is similar to the elevation range between high- and low-tide notches reported by others (e.g., Whitney and Hengesh, 2015a) and within the range of elevations reported for multiple short-lived notches formed during a postulated multi-peak MIS 5e high stand (Hearty et al., 2007). At a single transect, the two notches are separated by 6 m, yet compared to the transect adjacent to this location, the difference in elevation between upper notches is locally greater than the difference in elevation between the upper and lower notches (Fig. 5). The Sandstrom et al. (2020) reinterpretation of the Tantabiddi sequence based on elevation differences between two geomorphic notches and the distribution of associated marine sediments is not supported by the data. Furthermore, the warping rates derived from their reinterpretation are untenable. The available data suggest that the two notches can be interpreted as part of a type-section representing a single time stratigraphic sequence observed at varying elevations along the coast. This sequence includes multiple tidal and storm features, fossil beaches, and other shoreline deposits that can be attributed to variability in wave energy consistent with modern analogs (e.g., Whitney and Hengesh, 2015a).

Over the last few decades, the significance of northwestern Australia as a location for sea-level research has grown (e.g., McCulloch and Mortimer, 2008; Stirling et al., 1998). This is anchored on the paradigm that “the Western Australian margin is assumed to have been tectonically stable at a level <0.01 mm/year (1 m/glacial cycle)” (Lambeck et al., 2012, p. 4). Recently, a series of tectonic geomorphological, paleoseismological, and geophysical investigations have revealed active late Quaternary tectonic deformation along an ~2000-km-long section of western Australia's former passive margin (e.g., Clark et al., 2011; Whitney and Hengesh, 2015b; Hengesh and Whitney, 2016) passing through and including the Cape Range anticline (Clark et al., 2012; McPherson et al., 2013; Whitney and Hengesh, 2015a; Whitney et al., 2015). These studies alter the previously held paradigm of this region's tectonic stability and by extension have highlighted the importance of understanding the tectonic contribution to paleoshoreline positions when using marine limit features to assess global paleo–sea levels under the a priori assumption of tectonic stability. Specifically, Whitney and Hengesh (2015a) estimate tectonic uplift rates based on the Last Interglacial marine terrace elevations along the Cape Range of 0.049 ± 0.030 mm yr–1. The uncertainty in this rate incorporates the range of MIS 5e sea level estimates from prior global studies as a frame of reference, but to avoid circular reasoning, excludes elevation data from prior Cape Range studies.

Sandstrom et al. (2020) state that they use age estimations and elevation measurements to determine relative deformation and uplift rates. However, age and elevation must be compared to a datum to achieve this. They have no elevation datum as they measure each terrace feature against itself (assuming initial horizontality with no variability). This approach yields an unconstrained amount of warping for an individual terrace, but provides no independent reference frame, or baseline, to evaluate tectonic uplift of the feature since time of formation. This approach assumes that the entire difference in elevation along the terrace surface is due to post-depositional warping, even though the natural variability in formation elevations are up to 5 m in the data the authors present.

Sandstrom et al. (2020) attempt to reconstruct the vertical displacement of the Cape Range and conclude that “the most recent interglacial shoreline [MIS 5e] has undergone <1.3 m of vertical warping, suggesting minimal deformation since deposition.” The authors repeatedly assert that their results are in contrast to, and significantly lower than, published tectonic uplift rates (e.g., Clark et al., 2012; Whitney and Hengesh, 2015a). However, their rates are not uplift rates and are not comparable to any published rates due to the issues discussed in this comment.

The work of Sandstrom et al. (2020) presents a discussion of the strontium isotopic methods and geochemical analysis used to obtain new ages for the marine terraces on the Cape Range; this is a contribution to the geochronology of the Cape Region for the three older terraces. However, what they measured—the warping related to the double-plunge of the fold—and how they derived their warping rates, are not consistent with typical tectonic geomorphological methods for assessing coastal uplift and are not directly comparable to any of the previously published interpretations of tectonic uplift rates for the region. None of the measurements from Sandstrom et al. (2020) constrains uplift from an independent base-level, and none of the rates presented in Tables 1 and 2 herein, and in Table 2 in Sandstrom et al. (2020), are tectonic uplift rates. We conclude that their analysis does not quantify the net tectonic uplift of the paleoshorelines nor does it constrain tectonic uplift rates for the Cape Region of western Australia.

Science Editor: Brad S. Singer
Associate Editor: Emily Finzel
Gold Open Access: This paper is published under the terms of the CC-BY license.