*4.3. Ponta Do Castelo at the Island's Southeast End*

Located approximately 12 km farther east from Prainha (Figure 1b), the study site at Ponta do Castelo occupied the extreme southeast corner of Santa Maria Island. There, a modern CBD was entrained as a berm at mean sea level representing site 3, whereas the Pleistocene (MIS 5e) CBD was lodged above Pliocene strata at a height ~4.3 m amsl, representing site 4 (Figure 6).

**Figure 6.** Modern coastal boulder deposit (CBD) eroded from basalt at Ponta do Castelo (site 3, see map in Figure 1b) and overlying Pleistocene CBD preserved at a level 4.3 m above mean sea level at the same locality (sites 4 and 5).

The high exposure of this site to the main sea currents and the local morphology did not allow the deposition and maintenance of sand deposits. Raw data on clast size in three dimensions collected from the two parallel transects at this locality are available in Tables 3 and 4.

Data points representing individual boulders grouped by transect were plotted on a set of Sneed-Folk triangular diagrams (Figure 4c,d), showing shape variations. Compared to the pair of Sneed-Folk diagrams from Prainha (Figure 4a,b), the Ponta do Castelo set was similar with regard to the slope of points angled to the lower-right corner (Figure 4c,d). The main difference was that the data set from the modern CBD at Ponta do Castelo (Figure 4c) was more diffuse throughout the same subdivisions represented by the modern CBD at Prainha (Figure 4a). A higher number of points registered in the top triangle marked a significant deviation in the Pleistocene (MIS 5e) CBD at Ponta do Castelo (Figure 4d) compared with Prainha (Figure 4b). Overall, there was a greater tendency towards equidimensional clasts in the Pleistocene (MIS 5e) CBD than in the modern CBD at Ponta do Castelo.


**Table 3.** Quantification of boulder size, volume, and estimated weight from the modern coastal boulder deposit at site 3 at Ponta do Castelo, from the southeast end of Santa Maria Island (see map, Figure 1b). The standard density of basalt at 3.0 gm/cm<sup>3</sup> was applied uniformly in order to calculate wave height for each boulder. EWH: estimated wave height (in meters), calculated from equations in Nott [37] and Pepe et al. [38]. See the methods Section 3.2. (hydraulic model).

**Table 4.** Quantification of boulder size, volume, and estimated weight from the Pleistocene (MIS 5e) conglomerate at site 4 at Ponta do Castelo, from the southeast end of Santa Maria Island (see map, Figure 1b). The standard density of basalt at 3.0 gm/cm3 was applied uniformly in order to calculate wave height for each boulder. EWH: estimated wave height (in meters), calculated from equations in Nott [37] and Pepe et al. [38]. See the methods Section 3.2. (hydraulic model).


At Ponta do Castelo, variations in boulder size as a function of maximum and intermediate axis length were plotted using bar graphs (Figure 7), based on the raw data drawn from Tables 3 and 4. The general congruence between the modern and Pleistocene (MIS 5e) CBDs at Ponta do Castelo was strong, especially compared to the marked difference in size variation observed between the modern and Pleistocene (MIS 5e) CBDs at Prainha (Figure 5). Of foremost significance, 50% of measurements for clast size through the long axis (Figure 7a) qualified as boulders based on the criteria of Wentworth [35]. The tendency towards elongated clasts in the modern CBD was shown by the higher frequency in the bin size 6–15 cm for the intermediate axis (Figure 7b), whereas that bin size was void with respect to the long axis (Figure 7a). Compared to the modern CBD, the Pleistocene CBD at Ponta do Castelo exhibited only a small shift of data points into the size interval of 6 to 15 cm, but otherwise was similar. The congruence between modern and Pleistocene CBDs at Ponta do Castelo was even more striking, taking into account clast size measured on the intermediate axes (Figure 7b,d).

**Figure 7.** Bar graphs used to appraise variations in the long and intermediate axes on basalt clasts from modern and Pleistocene CBDs at Ponta do Castelo; (**a**) Long-axis length from clasts in the modern CBD; (**b**) Intermediate-axis length from clasts in the modern CBD; (**c**) Long-axis length from clasts in the Pleistocene CBD; (**d**) Intermediate-axis length from clasts in the Pleistocene CBD.

#### *4.4. Context of the Pleistocene Fauna at Ponta Do Castelo*

The MIS 5e sedimentary sequence at Ponta do Castelo is extremely poor, with only two gastropod species reported: the limpet *Patella aspera* Röding, 1798, and the supra-littoral littorinid *Tectarius striatus* (P.P. King, 1832). Nevertheless, Ponta do Castelo is considered as a geosite of international relevance because of a set of specific conditions that allowed the preservation of a shelf tempestite deposit [21] for which a precise water depth could be estimated. Additionally, it provides a good proxy for island uplift/subsidence reconstructions [15,16].

#### *4.5. Ponta Do Cedro on the Island's East Shore*

Located 3.0 km north of Ponta do Castelo, the third study area treated herein occurred on the east shore of Santa Maria Island (Figure 1b). No modern CBDs occurred at this locality. However, a correlated line of Pleistocene CBDs might be traced laterally at elevations that ranged between

~2.65 m and ~9 m amsl, represented by study sites 5 to 7. Typically, the conglomerate at Ponta do Cedro was as much as ~0.75 m in thickness and laterally continuous (Figure 8, site 6).

**Figure 8.** Pleistocene (MIS 5e) conglomerate 9 m above mean sea level at Ponta do Cedro (site 6).

Raw data on clast size in three dimensions collected from three study sites at this locality are available in Tables 5–7.

**Table 5.** Quantification of boulder size, volume, and estimated weight from the Pleistocene (MIS 5e) conglomerate at site 5 at Ponta do Cedro, on the east shore of Santa Maria Island (see map, Figure 1b). The standard density of basalt at 3.0 gm/cm<sup>3</sup> was applied uniformly in order to calculate wave height for each boulder. EWH: estimated wave height (in meters), calculated from equations in Nott [37] and Pepe et al. [38]. See the methods Section 3.2. (hydraulic model).




**Table 7.** Quantification of boulder size, volume, and estimated weight from the Pleistocene conglomerate at site 7 at Ponta do Cedro, on the east shore of Santa Maria Island (see map, Figure 1b). The standard density of basalt at 3.0 gm/cm3 was applied uniformly in order to calculate wave height for each boulder. EWH: estimated wave height (in meters), calculated from equations in Nott [37] and Pepe et al. [38]. See the methods Section 3.2. (hydraulic model).


Data points representing individual boulders sampled from the conglomerate layer at Ponta do Castelo, correlated at sites 5 to 7, were plotted on a set of Sneed-Folk triangular diagrams (Figure 4e–g).

**Figure 9.** Bar graphs used to appraise variations in the long and intermediate axes on basalt clasts from three sites of correlated Pleistocene CBDs at Ponta do Cedro; (**a**) Long-axis length from clasts at site 5; (**b**) Intermediate-axis length from clasts at site 5; (**c**) Long-axis length from clasts at site 6; (**d**) Intermediate-axis length from clasts at site 6; (**e**) Long-axis length from clasts at site 7; (**f**) Intermediate-axis length from clasts at site 7.

Overall, the distribution of clasts among the three Ponta do Cedro samples was highly consistent, with 68–78% of data points limited to the two central blocks within the triangular plot. The occurrence of data points in the top triangle, as well as the lower-right corner of the plot, were equally rare. Even so, there was a tendency for the data clusters to slope towards the right, indicative of a slight favorability towards elongated shapes. This trend was not nearly as strong as detected in modern CBDs at Prainha (Figure 4a) or Ponta do Castelo (Figure 4c), but entirely consistent with trends in the Pleistocene (MIS 5e) CBDs at Prainha (Figure 4b) or Ponta do Castelo (Figure 4d).

Variations in boulder size as a function of maximum and intermediate axis length were plotted using bar graphs (Figure 9), based on raw data drawn from Tables 5–7. The percentage of boulders in each Pleistocene sample was found to increase from the 20% (the more southern locality at site 5) to 48% (at site 6) up to the maximum value of 56% (the more northern locality at site 7). In each case, there was a marked shift in the abundance of clasts measured on the intermediate axis allocated to the size range between 6 and 15 cm (Figure 9b,d,f). Such a trend reinforced the impression that the Pleistocene (MIS 5e) basalt clasts conformed to a moderately oblong shape. Site 5 (Figure 9a,b) stood out among all the Pleistocene samples as most dominated by clasts defined by cobble size. In contrast, sites 6 and 7 were remarkably similar in their size distributions. Moreover, these samples

were indistinguishable from the Pleistocene CBD sample at Ponta do Castelo, which also included a high concentration of boulders.

#### *4.6. Context of the Pleistocene Fauna at Ponta Do Cedro*

At Ponta do Cedro, preliminary work allowed us to report a total of 13 Last Interglacial mollusk taxa (12 gastropods and 1 bivalve species). As in other MIS 5e outcrops (e.g., Prainha, Vinha Velha, Lagoinhas), the thermophilic element was present through several species of *Conus* and of the Pisaniidae *Gemophos viverratus* (Kiener, 1834) (=*Cantharus variegatus*). Ponta do Cedro was classified as a fossiliferous geosite of regional relevance [15,16], a situation that would change in the near future, as a result of ongoing work.

#### *4.7. Analysis of Calculated Storm-Wave Heights*

A summary of key data was provided (Table 8), pertaining to average boulder size and maximum boulder size from the five Pleistocene (MIS 5e) transects and two modern transects as correlated with weight calculated on the basis of specific gravity for basalt, using the equations derived by Nott [37] and Pepe et al. [38]. These data were applied to estimate the wave heights required to transport boulders from the bedrock source in sea cliffs to their resting place, and variations in the results depended on the equations used.

**Table 8.** Summary data from Tables 1–7, showing maximum boulder size and estimated weight compared to the average values for all boulders (n = 25) from each of transects 1–7 together with calculated values for wave heights estimated as necessary for boulder mobility. Estimated average wave height (EAWH; in meters) and estimated maximum wave height (EMWH; in meters), calculated according to equations derived by Nott [37] and Pepe et al. [38]. (see the methods Section 3.2. (hydraulic model)). Maximum values for each Transect in bold.


First, we have discussed the estimates using Nott's equation. The estimated wave height needed to move the largest boulder encountered in the Pleistocene (MIS 5e) transects (transect 6 at Ponta do Cedro) amounted to 6.1 m. The comparison between the Prainha and Ponta do Castelo modern and the Pleistocene (MIS 5e) CBD gave contrasting results; at Prainha, the average boulder size and weight and estimated average wave height were higher in the modern CBD in comparison with the Pleistocene CBD, whereas at Ponta do Castelo, it was the opposite trend. Moreover, the estimated maximum wave height required to shift boulders was very similar in all sites (ranging from 5.2 to 6.1 m) but Prainha (MIS 5e), where this value was about half (2.7 m). There was also a large difference between the largest boulders from Prainha (MIS 5e) and Ponta do Castelo (modern) CBDs, which were much smaller than those from the other sites (Table 8).

The results using the equation of Pepe et al. [38] were different from those using Nott's equation [37], as the former consistently indicated that the maximum values for both the estimated average wave height (EAWH) and the estimated maximum wave height (EMWH) occurred at the MIS 5e transects (at Ponta do Castelo and Ponta do Cedro, in the case of the EAWH; at Ponta do Castelo, in the case of the EMWH; cf. Table 8).

#### **5. Discussion**

#### *5.1. CBD: Tsunami Versus Storms, and the Use of Flawed Equations*

A recently published field-test of the hydrodynamic equations, based on the Nott-Approach and their derivations [51–54], failed to validate estimations of the wave height from boulder dimensions and concluded that such equations were flawed because they yielded unrealistically large heights [55]. These authors also stated that the Nott-Approach analysis did not differentiate between storm and tsunami waves [55]. Such criticism might be valid, but acceptance in full threatens to nullify any attempt to engage in the quantification of CBDs from the geologic record. The Nott-Approach remains useful and is applied solely to compare the relative storminess imprint left in CBDs from different time-periods (MIS 5e vs. modern, as in our example), but located at the same site. For this particular case and question, it does not matter that equations may be flawed because the errors are uniformly carried through the exercise. What is important is that such equations function to detect consistent differences between the estimated wave heights for the past (MIS 5e) and for the modern situation. Moreover, they also provide the magnitude of possible differences.

CBDs should be portrayed as archives of extreme wave events, and their origin may be due to either storm waves or tsunamis. All CBDs studied in Santa Maria Island have no characteristics of tsunami deposits (e.g., imbrication of the boulders, erosive base (scour-and-fill features), rip-up clasts of the underlying substratum, downward-injected clastic dykes inside the palaeosol, traction figures; [56]), and we have, therefore, interpreted them as the result of storm waves.

The width of Santa Maria Island's insular platform is larger in the N and W (although this is not apparent in the western shores, as a result of the uplift of the island) due to the higher offshore significant wave heights mainly coming from the NW, W, and N [24]; in contrast, at the south and eastern shores, the insular platform is quite narrow [25]. However, the most destructive storms (and, therefore, with higher wave heights) that impacted the Azores in recent times (and we assume a similar pattern in the past) were related to the passage of hurricanes. As these have a counter-clockwise rotation in the northern hemisphere, its passage most strongly affects precisely the southern and eastern shores of the islands, thus providing a possible explanation for the limited occurrence of CBD. Based on the historical record, recent hurricanes that struck the islands in the Azores Archipelago include "Hurricane 15" (1932), Hannah (1959), Debbie (1961), Fran (1973), Gordon (2006 and 2012), Gaston (2016), Ophelia (2017), and Lorenzo (2019). A direct hit on Santa Maria Island at 37◦ N entails a northward shift in the latitude of 10◦ and a longitudinal travel distance of roughly 2000 km. Hurricane frequency in the Azores is regarded as on the rise during the last 50 years [30]. The strong Pliocene sedimentary record on Santa Maria reflects some indication of hurricane activity on the southern coast [57], where the marine shelf is narrowest (Figure 1b). Moreover, the island's geological record holds evidences of Pliocene deep wave interaction with the seafloor, reaching 50 m depth (Ponta do Castelo, [21]), thus suggesting a high likelihood of large-size boulder transport.

During the Last Interglacial, the average values for summer insolation were about 11% higher than those of today [58], the oceanic Polar Front was pushed north-westwards by the Gulf Stream warmer sea surface temperatures (SSTs) [59], and the summer position of the Azores High was forced E of its present location to around 35◦N and 20–25◦W, i.e., between the Azores and Iberia [60]. These oceanographic conditions induced a stronger North Atlantic storm track, shifted northward from its present location, and extended to the east [4]. Altogether, the storminess associated with tropical cyclones is expected to have been higher during the MIS 5e than today [1,3,61], as a result of the North Atlantic Subtropical High eastward shifting from its present location. However, the occurrence of "superstorms" during the Last Interglacial has been recently questioned by authors [62,63]. Therefore, probably, the higher storminess that most authors suggest for the Last Interglacial is related to a higher frequency of hurricanes and not necessarily with higher wave heights.

#### *5.2. Comparison between Model Wave-Height Data and Inferred Modern Wave-Height*

Our study yielded modern CBD estimated averages for wave heights varying from 2.8 to 3.6 m (using Nott's Equation (1)) and from 2.0 to 2.9 (Pepe et al.'s Equation (2)) (Table 8). The former values were higher than the mean values presented by Rusu and Onea [31] for the archipelago, which varied from 2.23 to 2.55 m, a situation that other authors also reported [55], with Nott's Equation consistently yielding wave height values higher than those registered by modern wave buoy. In contrast, the maximum calculated estimate wave heights of 5.5 m (Equation (1)) and 6.4 m (Equation (2)) (Table 8) did not get close to the lowest maximum value of Rusu and Onea [31] of 9.18 m.

The values obtained for the Pleistocene (MIS 5e) estimates were similar using both formulas, with wave height estimates varying from 1.3 to 2.9 m (Equation (1)) and from 1.3 to 3.3 m (Equation (2)); however, the maximum wave height estimates were quite different: 6.1 m (with Equation (1)) and 8.1 m (Equation (2)) (Table 8). These discrepancies highlighted the serious reserves [55], regarding the use of these formulas when solely targeted to the inference of modern (or past) wave heights, based on the dimensions/mass of the boulders.

Considering the general morphology of the sites, the nature of the boulders, and the processes that lead to its shaping, and albeit based on only two sites (Prainha and Ponta do Castelo; cf. Table 8), where it was possible to compare the modern CBD with the MIS 5e CBD, the results showed contrasting site-dependency. At Prainha, modern storminess was estimated to be higher during the Last Interglacial (independent of which equation was used), whereas at Ponta do Castelo, MIS 5e storminess should have been highly similar to the modern one (using Equation (1)) or higher (Equation (2)).

Further analysis is needed from other islands having similar MIS 5e/modern sites situation before sound conclusions might be reached regarding such a restrictive use of formulas, otherwise argued to be flawed [55]. Finally, a set of parallel profiles to the sea (at different heights) should be undertaken with respect to modern CBDs, as well as perpendicular profiles, in order to check for possible boulder size changes according to distance from present-day sea level. This analysis could also help to relate a mean boulder dimension to an average distance to the sea. An in-depth analysis of Pleistocene (MIS 5e) wave-cut surfaces will also enhance our knowledge of coastal retreat, average erosion rates, and wave activity.

## *5.3. Comparison with CBD Studies Elsewhere*

A recent review [64] indicated that although significant studies have dealt with the Quaternary CBDs, only 21 papers studied the Neogene deposits (10 of Miocene and 9 of Pliocene age, plus 2 works dealing with both epochs), mostly due to their low preservation potential [64]. According to our literature survey, this was the first study on Pleistocene CBDs conducted on the Atlantic oceanic islands. Moreover, previous works using the same methodology dealt with the different rocky composition of the CBDs (limestone, rhyolite, and andesite) [17–19]; this paper being also the first to use the same kind of analyses in regard to basalt, which has higher specific gravity than any of the other rocks tested. Therefore, for similar size CBDs, basalt boulders will require a more energetic wave to be moved.

#### **6. Conclusions**

Limited to comparative results from modern and late Pleistocene (MIS 5e) boulder deposits on the southern and eastern shores of Santa Marina Island, the present study permitted the following conclusions:

1) Wave heights estimated on the basis of the largest modern boulders from CBDs on the modern south shore of Santa Maria Island indicated maximum values between 5.2 and 5.5 m (Equation (1)) and between 3.7 and 6.4 m (Equation (2)), which was 2 m higher than the expected average height experienced during winter storms, supporting the conclusions of [55] regarding the advice on the non-use of these formulas because of unrealistic, flawed estimates of wave heights;

2) Regarding storminess, our results were contradictory and did not allow any conclusion about possible higher storminess during the Last Interglacial when compared with the present-day, and recorded in the CBDs of Santa Maria Island;

3) Historical records of storm activity, coupled with the westerlies regime predominant in the Azores area, placed a stronger emphasis on the seasonal effect of winter storms that preferentially influence shore erosion on north-facing coasts. Although the historical record indicates that hurricane activity is less frequent in the islands, its erosional effect against a south- and eastern-facing shores must be considered. From the available data derived from CBDs and coastal geomorphology around Santa Maria Island, it was clear that wave action both today and approximately 125,000 years ago during MIS 5e remained consistently highest against the eastern and south-eastern coasts at Ponta do Cedro and Ponta do Castelo. These localities also correspond to areas where the island's marine shelf is narrowest. Although the historical incidence of hurricanes passing through the Azores Archipelago is statistically low compared to the arrival of annual winter storms, the migration of hurricanes moving in a counter-clockwise rotation across the North Atlantic Ocean conforms to the evidence for excessive erosion rates in those districts beyond the capacity of winter storms.

The abundant fossil record on Santa Marina Island that makes accurate dating possible for the Late Pleistocene (MIS 5e) interglacial epoch is not available elsewhere in the Azores Archipelago, but future research on MIS 5e versus modern CBDs (whenever located in the same site) from other archipelagos is expected to apply the same formulaic techniques to extract information on storminess.

**Author Contributions:** Fieldwork in Santa Maria Island has occurred since 1999 by a large, multidisciplinary team led by S.P.Á. during the workshops "Palaeontology in Atlantic Islands". Field data on CBDs was collected in July 2019 by A.C.R., L.B., and C.S.M., M.E.J. prepared the first draft of this contribution and drafted all figures, except for Figure 1 prepared by C.S.M., S.P.Á. was responsible for working out the mathematics related to storm hydrodynamics, produced all tables, and supplied all ground photos. S.P.Á. and C.S.M. summarized the literature on CBDs in the Azores region. In addition, C.S.M. contributed input on wave regimes and main directions within the study area. Authorship has been limited to those who contributed substantially to the work reported. All authors have read and agreed to the published version of the manuscript.

**Funding:** We thank Direcção Regional da Ciência e Tecnologia (Regional Government of the Azores), FCT (Fundação para a Ciência e a Tecnologia) of the Portuguese Government, and Câmara Municipal de Vila do Porto for financial support. S.P.Á. acknowledges his research contract (IF/00465/2015) funded by the Portuguese Science Foundation (FCT). A.C.R. was supported by a Post-doctoral grant SFRH/BPD/117810/2016 from FCT. L.B. acknowledges her Ph.D. Grant from FCT (SFRH/BD/135918/2018). C.S.M. acknowledges his Ph.D. Grant M3.1.a/F/100/2015 from Fundo Regional para a Ciência e Tecnologia (FRCT). This work was also supported by FEDER funds through the Operational Programme for Competitiveness Factors – COMPETE and by National Funds through FCT under the UID/BIA/50027/2013, POCI-01-0145-FEDER-006821, and under DRCT-M1.1.a/005/Funcionamento-C-/2016 (CIBIO-A) project from FRCT. This work was also supported by FEDER funds (in 85%) and by funds from the Regional Government of the Azores (15%) through Programa Operacional Açores 2020 under the project VRPROTO (ACORES-01-0145-FEDER-000078).

**Acknowledgments:** We acknowledge the field assistance of Manta Maria and Câmara Municipal de Vila do Porto. We are grateful to the organizers and participants of all editions of the international workshops "Palaeontology in Atlantic Islands", who helped with fieldwork (2002–2019). We are grateful to Daniele Casalbore and two anonymous reviewers, whose comments greatly improved this manuscript.

**Conflicts of Interest:** The authors declare no conflict of interest.
