High-frequency cycles of brachiopod shell beds on subaqueous delta-scale clinoforms (early Pliocene, south-east Spain)

During the early Pliocene, subaqueous delta-scale clinoforms developed in the (cid:1) Aguilas Basin, in a mixed temperate carbonate – siliciclastic system. The facies distribution is consistent with the infralittoral prograding wedge model. Stacking patterns and bounding surfaces indicate that the clinoforms formed during the highstand and falling sea-level stages of a high rank cycle. Twenty-two prograding clinothems were recognized over a distance of ≥ 1 km. Biostratigraphic data indicate a time span shorter than 700 kyr for the whole unit (MPl3 biozone of the Mediterranean Pliocene). Cyclic skeletal concentrations and occasional biostromes of suspension feeders (terebratulid brachiopods, modiolid bivalves and adeoniform bryozoan colonies), slightly evolved glauconite and occasional Glossifungites ichnofacies formed on the clinoforms during high-frequency pulses of relative sea-level rise. During such stages, increased accommodation space in the topsets of the clinoforms caused a strong reduction of terrigenous input into the foresets and bottomsets. This provided favourable conditions for the development of these suspension feeder palaeocommunities. During stillstand stages, however, reduced accommodation space in the topsets eventually resumed progradation in the foresets. There, the abundance of Ditrupa tubes indicates frequent siltation events that extirpated the terebratulid populations and other epifau-nal suspension feeders in the foreset and bottomset subenvironments. The occurrence of shell beds on the clinoforms suggests that this case study represents lower progradation rates than standard examples where shell beds bound the clinobedded units at their base and top only. Importantly, the distributions of biofacies and ichnoassemblage associations contribute signiﬁ-cantly to the understanding of the effects of relative sea-level ﬂuctuations on the evolution of subaqueous delta-scale clinoform systems.


INTRODUCTION
The duration of cycles is the traditional criterion to discriminate the hierarchical order of stratigraphic sequences (Mitchum & Van Wagoner, 1991;Vail et al., 1991). The assignment of sequences to orders, however, can be difficult and arbitrary because the structure of the sedimentary record conforms to a continuum rather than to distinct modes of abundance classes of basin can be designated by the generic rank 'N', and successively lower rank sequences can be designated by ranks 'N-1', 'N-2', etc. (e.g. Massari & Chiocci, 2006). Such hierarchical systems can serve as a template for comparison with other study areas and, if good chronological control is available, these ranks can be evaluated in light of cycle duration to reconcile both approaches (Schlager, 2010). The physical expression of sequences can include the relative extension of unconformities, depth of incision of fluvial valleys, geometric relationships between the building blocks of composite sequences, magnitude of facies shifts, relative scale of clinoforms (Thorne, 1995;Helland-Hansen et al., 2012;Patruno et al., 2015), or the development of onlap, backlap, downlap and toplap shell beds (Kidwell, 1991;Abbott, 1997;Naish & Kamp, 1997;Kondo et al., 1998;Di Celma et al., 2005;Hendy et al., 2006;Zecchin & Catuneanu, 2013). Except for largescale outcrops, however, where the relationships between the rank of sequences and the distribution of shell beds can be directly traced (Beckvar & Kidwell, 1988;Massari & D'Alessandro, 2012; (Fig. 1A), exposures with limited spatial extent hamper the observation of clinoforms. In such cases, the rank of sequences defined by the position and geometric relationship of condensed shell beds with the sequence building blocks can be difficult to elucidate (F€ ursich et al., 1991;Ruffell & Wach, 1998).
This study documents the distribution of cyclically arranged brachiopod shell beds in the Aguilas Basin (south-east Spain) in extensive outcrops of Pliocene sediments. These pavements formed on the distal part of lower rank delta-scale (i.e. tens of metres high) clinoforms (Fig. 1B and C). Importantly, this contrasts with other examples where onlap and backlap shell beds bound lower rank delta-scale clinoforms (e.g. Massari & D'Alessandro, 2012) (Fig. 1A). Determining the scale of clinoforms in the current study area enables the identification of low rank onlap and backlap shell beds. Moreover, the development of condensed shell beds on clinoforms implies lower progradation rates than those where clinoforms of comparable scale/rank lack such hiatal concentrations.
In rare cases, all four types are found to prograde synchronously in the same basin, forming a compound clinoform system (Patruno et al., 2015) (Fig. 2B and C). In physical-accommodation dominated systems (sensu Pomar & Kendall, 2008), the relative progradation rates decrease from subaerial and subaqueous deltas, to shelf prisms and then to continental margin clinoforms (Patruno et al., 2015) (Fig. 2B). This is important because the relative progradation rates can help to reconcile the duration of relative sea-level cycles and their physical expression in the rock record. 1990), whereas the latter have the whole clinoform (topset, foreset and bottomset) submerged (Fig. 2C). Moreover, the rollover in subaerial deltas can be coincident with or very close to the shoreline, whereas in subaqueous deltas the rollover is on average several kilometres away from the shoreline.
In a meta-analytical study of clinoform geometry and scale, Patruno et al. (2015) differentiated two types of subaqueous delta-scale clinoforms: sand-prone and mud-prone. This distinction is relevant because the former display higher foreset gradients and their rollovers are closer to the shoreline than in the latter (Patruno et al., 2015). The terminology of Patruno et al. (2015) focuses on the geometric description of clinoforms. Similar terms related to geometric aspects include 'distally steepened ramp' (Read, 1985) for carbonate environments (see also Pomar, 2001).
Other terms have been proposed following a genetic approach. For example, Hern andez- Molina et al. (2000) introduced the term 'infralittoral prograding wedge' (IPW) for a morpho-sedimentary system characterized by narrow, shore-parallel, sigmoidal-shaped sedimentary bodies that prograde below the wave base in the offshore transitional zone of wave-dominated coasts. Geometrically, this system belongs to the category of sand-prone subaqueous delta-scale clinoforms of Patruno et al. (2015). Such a distinction is helpful because prograding reef platforms also produce subaqueous delta-scale clinoforms (Franseen & Mankiewicz, 1991;Pomar & Ward, 1994;Braga & Mart ın, 1996;Cuevas-Castell et al., 2007;Kleipool et al., 2017) but the genetic factors and resulting facies are quite different from those of an IPW (Pomar & Kendall, 2008).

Mixed carbonate-siliciclastic 'hybrid' deposits
For mixed carbonate-siliciclastic systems, the term 'hybrid' is often used in the literature to Fig. 1. Idealized sketch of the sequences with the location of condensed shell beds and the conceptual framework for a hierarchical classification of ranks based on geometry and scale. (A) Rank N sequence with indication of rank N onlap, backlap, downlap and toplap shell beds (adapted from Zecchin, 2007). (B) Rank N sequence indicating the position of both rank N and rank N-1 shell beds as described in this study. (C) Outcrop example of rank N-1 shell beds on a Pliocene subaqueous delta-scale clinoform (low rank) in the Aguilas Basin (from proximal to distal: Schizoreteporarhodolith debris, Schizoretepora and Terebratula facies).  Patruno et al., 2015). (B) A compound system with subaqueous delta-scale, shelf prism and continental margin clinoforms. The range of relief and progradation rates is indicated in insets (adapted from Patruno et al., 2015). (C) Compound delta-scale clinoform system with either shoreline or subaerial delta-scale, and subaqueous delta-scale clinoforms (adapted from Patruno et al., 2015): m SWWB, mean storm weather wave base; FWWB, fair weather wave base.

Synthem
A synthem is an unconformity-bounded unit (Ruban, 2015). In this study, the term refers to the units bounded by the local high rank unconformities.

'Pristine' preservation
This study uses the term 'pristine' as a shortcut to refer to bioclasts that barely display any signs of biostratinomic alteration. These bioclasts are articulated and complete, with well-preserved ornamental features, and have not been subject to macrobioerosion and/or encrustation by epizoan organisms.

GEOLOGICAL SETTING
The Aguilas Arc (south-east Spain) (Fig. 3A), which belongs to the Inner Zones of the Betic Cordillera, is a tectonic megastructure that extends onshore over a distance of 60 km along a south-west/north-east axis. The megastructure resulted from a north-south or north-west/ south-east rigid-plastic indentation of a crustal block that began in the Early Miocene due to collision of the African and Eurasian plates in the Western Mediterranean and is still active today (Coppier et al., 1989;Griveaud et al., 1990). This arc is delimited to the south-west and north-east by systems of left-lateral and right-lateral strike slip faults (Palomares, Coc on-Terreros and Moreras fault systems) (Coppier et al., 1989;Silva et al., 1993). The internal sector of this arc comprises five small basins that are open to the Mediterranean (Bardaj ı et al., 1999); their opening was probably caused by an important collapse of the southern margin of the arc, associated with transtension (Coppier et al., 1989). These basins probably acted as rias (i.e. estuaries encased in high-relief fluvial valleys) and then as coastal embayments during the early Pliocene (Dabrio et al., 1991;Garc ıa-Ramos et al., 2014). This study focuses on the southwestern sector of the Aguilas Basin ( Fig. 3B and C), located some 5 km south-west of the town of Aguilas, where the succession of Pliocene marine sediments is most complete (Montenat et al., 1978).

STRATIGRAPHIC FRAMEWORK
Pliocene deposits in the current study area can be attributed to three synthems: SP0 (MPl1-MPl2 pro parte biozones), SP1 (MPl3 biozone) and SP2 (possibly MPl4) (Fig. 3D). The focus here is on the prograding succession of synthem SP1. The topmost part of SP1 consists of carbonates of an isolated platform abutting a volcanic ledge, that is an entirely different morpho-sedimentary system and therefore beyond the scope of this study ( Fig. 3C and D). To provide a stratigraphic framework, SP0, SP1 and SP2 are briefly outlined.
The SP0 synthem is represented by glauconyrich condensed deposits. At the 14 m thick El Barcel on section (Fig. 3C), beds are oriented N79°E/8°SE. A sharp transgression over metamorphic rocks of the Palomas Unit (Alpuj arride Complex, Inner Betic Zones) ( Alvarez & Aldaya, 1985) is recorded at the base. Planktonic foraminifera from SP0 indicate the MPl1 and MPl2 biozones of the Mediterranean Pliocene (Montenat et al., 1978;Garc ıa-Ramos et al., 2012). Based on benthic and planktonic foraminifera, Garc ıa-Ramos et al. (2014) proposed a shallowing upward trend evolving up-section to assemblages of shallow-water benthic foraminifera, devoid of planktonic foraminifera. In this section, the top of SP0 is truncated and overlain by Quaternary conglomerates, and the transition from SP0 to SP1 is not exposed. An angular, erosive unconformity is inferred because of the different strike and dip of the two synthems and a conspicuous shift in benthic and planktonic foraminiferal assemblages from shallow-water to a relatively deep-water, offshore environment between the top of SP0 and the base of SP1. This unconformity crops out in a section northeast of Castillo de Terreros (Montenat et al., 1978) (Fig. S1).
Synthem SP1 consists of a succession of clinobedded units that prograded over a distance of about 2 km starting from the hillock of Cabezo Alto across the area of Cañada Brusca and the Cuatro Calas coves (Fig. 3D). In SP1, clinoforms have a strike of N57°E with a variable dip (a few degrees to over 14°SE) along a northwest/south-east transect. The unconformity between SP1 and SP2 eroded part of the upper interval of SP1, which is either missing on the surface or covered by colluvial deposits. The uppermost part of SP1 crops out again, however, in the Cuatro Calas coves sector (Fig. 3C); there, the top of SP1 is also truncated and overlain by Quaternary conglomerates. The SP2 synthem was described and interpreted as a wave-dominated Gilbert-type delta system (Dabrio et al., 1991): SP2 is also truncated at the top by an unconformity and covered by Quaternary marine and terrestrial units, one of which has been dated to the oxygen isotopic stage 5e based on the occurrence of the gastropod Persististrombus latus (Bardaj ı et al., 2001).

Biostratigraphy
The co-occurrence of Globorotalia puncticulata and G. margaritae from the base to the top of the synthem SP1 (Fig. 3D) indicates that it was deposited entirely during the MPl3 planktonic foraminiferal biozone of the Mediterranean Pliocene (4Á52 to 3Á81 Myr) (Iaccarino et al., 2007;Violanti, 2012;Corb ı & Soria, 2016). Because of truncation at the base and the top of the synthem, the exact duration of SP1 is uncertain, but must be <700 kyr.

MATERIAL AND METHODS
Fluvial incision has revealed laterally continuous outcrops oriented subparallel and subperpendicular to the depositional strike that enabled the stratal geometries and stacking patterns to be studied. Clinoforms and clinothems were mapped using outcrop panoramic photomosaics while stratigraphic contacts and facies were checked in the field. Two main sections (Figs 3C, 3D and 4), 44 m thick (Cabezo Alto) and 77 m thick (Cañada Blanca), were logged in detail, for lithology, sedimentary structures, macrofossil composition, biofabrics and ichnoassemblages to evaluate the vertical variation and stacking of facies. These sections were complemented by smaller sections to show details of stratigraphic features. The macrofauna was identified to species level whenever possible, except for most bryozoans, for which only the zooarial morphology was noted. The abundance of macrofaunal taxa was estimated in the field by distinguishing between dominant (26 to 100%), common (11 to 25%) and rare (1 to 10%) categories. This qualitative approach was conducted by visually inspecting each sampling site for 30 min (cf. time-picking of Ceregato et al., 2007). Skeletal concentrations were described using qualitative criteria (Kidwell et al., 1986), while biofabrics follow the semi-quantitative charts of Kidwell & Holland (1991). Macroscopic descriptions of lithofacies were complemented with representative thin sections. Some bulk samples of uncemented sediment were sieved through 500 lm, 125 lm and 63 lm meshes to explore qualitatively the content of benthic and planktonic foraminifera in the 125 lm fraction, to aid in a palaeoenvironmental interpretation and biostratigraphic characterization of the studied synthem. For 26 samples of the Cabezo Alto section, >200 benthic foraminifera were identified and counted. Taxa with >3% proportional abundance are reported.
Magnetic susceptibility, a proxy of terrigenous input (Davies et al., 2013), was measured in the field, with a SM-20 magnetic susceptibility meter (Gf Instruments, Brno-Medl enky, Czechia) for the Cabezo Alto (34 sampling sites) and Cañada Blanca (68 sampling sites) sections. Five to six replicate measurements per sampling site (ca 1 sec measuring time) were taken on flat rock surfaces, and the mean value reported.
Carbonate content was quantified at the Institute of Geography and Regional Research (University of Vienna) with a Scheibler calcimeter for 34 samples in the Cabezo Alto section and 48 samples at the Cañada Blanca section. The procedure specified in ISO 10693:1995 has been followed ( € ONORM L 1084( € ONORM L , 2006. Glauconite maturity has been categorized into four stages, based on the K 2 O content (Amorosi Fig. 4. Stratigraphic logs of the CA (clinothems 1 to 4) and CBL (clinothems 9 to 22) sections. The interpreted systems tracts of the high rank and low rank cycles are included, the latter with abbreviated letter codes; T (transgressive), H (highstand), F (falling stage) and L (lowstand). Numbers indicate the facies types of four facies associations (FA), the latter noted by colour code. The position of benthic foraminifera samples from the CA section is indicated. Skeletal concentrations were simplified to pavements and thick shell beds (dominant taxa indicated by the corresponding icon in the legend). Magnetic susceptibility and carbonate content trends are included. Possible cyclicity indicated by arrows. For a magnified zoom view of the interval including clinothems 9 to 14, see Fig. 12. TST or T, transgressive systems tract; FSST or F, falling stage systems tract; HST or H, highstand systems tract. Amorosi, 2012): (i) nascent (K 2 O = 2 to 4%); (ii) slightly evolved (K 2 O = 4 to 6%); (iii) evolved (K 2 O = 6 to 8%); and (iv) highly evolved (K 2 O > 8%). Glauconite K 2 O content and colour are correlated: nascent to slightly evolved glauconite is light greenyellowish, mature glauconite is dark green. Glauconite composition was examined in one sample to determine its maturity. About 80 to 100 grains were picked from the 125 to 500 lm fraction, embedded in resin and polished on a slide. Glauconite analyses were performed at the Department of Lithospheric Research (University of Vienna) using a Cameca SXFive FE Electron Probe Microanalyzer (EPMA; CAMECA, Gennevilliers Cedex, France) equipped with five wavelength-dispersive and one energy-dispersive spectrometers. Well-characterized homogeneous natural and synthetic minerals were used as standards. All analyses were performed at 15 kV accelerating voltage and 20 nA beam current. Due to the K migration a defocused beam with 5 lm diameter and 10 sec counting time on peak position were used. For matrix corrections, the PAP method (Pouchou & Pichoir, 1991) was applied to all acquired data. The relative error of the laboratory internal standard is below 1%.

DESCRIPTION OF FACIES
Four main facies associations and one facies were recognized in the studied synthem. These are described in detail in Tables 1 to 4. Facies distributions are shown in stratigraphic logs and outcrop photomosaics to highlight vertical and lateral changes. In general, the facies grade into one another along a proximal-distal gradient.

Facies Association 1
The common feature of Facies Association 1 (FA1) is the occurrence of coarse-grained siliciclastics. Three facies are distinguished based on sorting, carbonate matrix and packing of macroinvertebrates.
Coarse-grained friable sandstone -F1.1 This facies was observed in only two clinothems of the Cañada Brusca W area. It consists of friable sandstone composed of well-sorted, coarse, angular grains (mainly of quartz and schists) ( Fig. 5A and B). This sandstone is poorly cemented and pervasively bioturbated, therefore no physical sedimentary structures are preserved. In proximal parts it displays an intensely bioturbated ichnofabric dominated by vertically oriented Macaronichnus and subsidiary Ophiomorpha ( Fig. 5A and D to H). It yields abraded and fragmented microfossils in low abundance, including ostracods (for example, Aurila) and benthic foraminifera, most notably Elphidium crispum and Ammonia inflata (Table 1).
Hybrid rhodolithic sandstone -F1.2 This facies mainly occurs in the Cañada Blanca section and the Cañada Brusca W sectors ( Fig. 5I and J), in more distal positions than F1.1. It consists of poorly sorted coarse sandstone with a carbonate matrix. Granule-sized debris of coralline red algae is characteristic, albeit in varying proportions. The fabric in general is massive; locally, well-defined trace fossils are identifiable (Table 1).

Shell-rich hybrid sandstone -F1.3
This facies occurs in the Cañada Blanca section (clinothem 12). The matrix resembles that of F1.2 but it is distinguished by thick (>1 m), densely packed skeletal concentrations dominated by pectinids (Aequipecten scabrellus). It also displays a complex biofabric, with large gutter casts infilled with pectinids, and overlies an irregular erosive surface.

Interpretation of Facies Association 1
The well-sorted and winnowed texture of the coarse sands, selectively enriched in detrital quartz (Fig. 5B), suggests proximal environments affected by tidal and wave currents above the fair-weather wave base (Blomeier et al., 2013). This interpretation is supported by the lateral facies change, in which beds displaying proximal facies F1.1 evolve distally into poorly-sorted and unwinnowed coarse sandstone facies F1.2 ( Fig. 5I and J). The association of Ophiomorpha and dominant Macaronichnus in similar lithofacies ( Fig. 5D to F) has been interpreted as either foreshore or upper shoreface environments (Fr ebourg et al., 2012;Mayoral et al., 2013;Uchman et al., 2016). This is compatible with the impoverished benthic foraminiferal assemblage, with the poorly preserved shallow shelf species Elphidium crispum and Ammonia inflata (Sgarrella & Moncharmont Zei, 1993;Fiorini & Vaiani, 2001;Rasmussen, 2005). The low species richness, abundance and high taphonomic alteration of these microfossils can be interpreted as an indication of onshore transportation (Davaud & Septfontaine, 1995). The dominance of Flabellipecten bosniaskii in some patches of facies F1.2 (Table 1) is consistent with proximal sandy environments (Aguirre et al., 1996). The lack of physical sedimentary structures is most probably due to thorough bioturbation and/or cryptobioturbation (Pemberton et al., 2008).

Facies Association 2
The main characteristic of Facies Association 2 (FA2) is the fine-grained carbonate-rich matrix (CaCO 3 ca 40 to 80%) (Fig. 4) and the frequent presence of coralline algae, either in the form of complete rhodoliths or rhodolith debris.
Calcarenite -F2.1 In contrast to other facies, this was found only in the upper two clinothems ( Fig. 6A and B). The coarse-grained, well-sorted fabric is similar to F1.1 but is composed of carbonate lithoclasts. Small casts, probably of comminuted aragonitic shells, are visible. This facies is locally crudely stratified and can contain pavements of Flabellipecten and Ostrea (Table 2; Fig. 6). It has variable proportions of rhodolith debris and is pervasively bioturbated, with poorly defined trace fossils, except for intervals with welldefined Thalassinoides (Fig. 6B).

Hybrid rhodolithic floatstone -F2.3
This facies is characteristic of the whole study area. The matrix consists of fine-grained siliciclastic material and micrite in variable proportions (up to 80% carbonate content). The dominant bioclastic material is rhodolith debris, which varies from coarse-grained to gravel size, but complete rhodoliths also occur and one locality exhibits pavements (Fig. 7). It is pervasively bioturbated with variable ichnoassemblages (Table 2); hence only one example of swaley cross-stratification (SCS) has been identified (Fig. 5K). In clinothem 21, however, it displays a crude stratification, forming tabular beds about 30 to 40 cm thick. The most characteristic macroinvertebrates are Clypeaster cf. aegyptiacus (often as complete tests), Spondylus crassicosta (often articulated), Ostrea edulis f. lamellosa and Gigantopecten latissimus (juveniles and adults). In some samples, coralline algae attributable to lithophylloid and melobesioid taxa were identified ( Fig. 7D to F).
Shell-rich hybrid rhodolithic floatstone -F2.4 This is similar to F2.3 but contains densely packed concentrations of pectinids (Aequipecten opercularis) and rhodoliths and locally also Ostrea and Spondylus. This facies usually forms very thick (several metres) beds, often overlying an erosive or irregular surface. Coralline algae are sometimes present as rhodoliths or represented by small proportions of rhodolith debris.

Interpretation of Facies Association 2
The well-sorted, winnowed texture and coarse grain-size of F2.1, together with the dominance of Flabellipecten and Ostrea, points to highenergy proximal environments (Aguirre et al., 1996;Blomeier et al., 2013). The reduced grainsize of siliciclastics in F2.3 and F2.4 points to lower energy levels compared to FA1. The abundance of rhodoliths ( Fig. 7) suggests background low-moderate energy conditions, good oxygenation, low sedimentation rates and low turbidity enabling suitable light penetration; the assemblage of melobesioids and lithophylloids ( Fig. 7D to F) suggests depths in the order of several tens of metres (Aguirre et al., 2012(Aguirre et al., , 2017. Moreover, the characteristic macroinvertebrate species in this facies (Table 2; Fig. 5A) are common in shoreface environments (Malatesta, 1974;Ben Moussa, 1994;Mancosu & Nebelsick, 2017).
Swaley cross stratification ( Fig. 5K) indicates storm deposition events (e.g. Myrow, 2005) in the offshore transitional zone (Dumas & Arnott, 2006). The occurrence of sporadic densely packed lenticular shell beds (Fig. 5), probably the product of 'cut and fill' structures, also points to major storm events . The ichnoassemblage of F2.3 (Ophiomorpha nodosa, Skolithos linearis and Planolites montanus) (Table 2), combined with the features discussed above, is interpreted here to indicate an opportunistic response associated with storms or other high-energy disturbances Gani et al., 2009;Buatois et al., 2015), although individual ichnotaxa can occur under normal marine conditions. In the first scenario, Palaeophycus can be characteristic of the fair-weather assemblage . Finally, the rhodolith-pectinid rudstone infilling Piscichnus traces suggests trapping in burrows by passive filling when coarser particles are entrained during storm traction-transport (Wanless et al., 1988;Zuschin & Stanton, 2002;Yesares-Garc ıa & Aguirre, 2004). The complex biofabric of thick, densely packed skeletal concentrations (facies F2.4) overlying erosive surfaces, together with the observation that they overlie FA1 ( Fig. 5C to E) or F2.1 (Fig. 6), suggests that this facies formed under conditions of sediment bypass or starvation, promoting the amalgamation of event beds during transgressive phases (Kidwell, 1991;Abbott, 1997;Dattilo et al., 2008;. They are therefore interpreted as onlap shell beds in line with conclusions drawn by Kidwell (1991) and  elsewhere. The absence of physical sedimentary structures is interpreted here to be a result of thorough bioturbation. According to Zecchin (2007), this trait is typical of sheltered embayments.

Facies Association 3
The characteristic feature of Facies Association 3 (FA3) is the occurrence of the serpulid polychaete Ditrupa arietina in a fine-grained hybrid matrix. The carbonate content varies between about 30% and 50% (Fig. 4).
Hybrid packstone with Ditrupa and rhodolith debris -F3.1 Facies F3.1 is transitional between Facies Associations 2 and 3. It consists of hybrid finegrained packstone to grainstone with small fragments of rhodolith debris. The main feature is the much smaller proportion of rhodolith debris compared to FA2. Locally it contains Ditrupa, pectinids or fragments of adeoniform zooaria (Schizoretepora sp.). Because of the high carbonate content (ca 50% CaCO 3 ), cementation locally defines beds varying from 30 to 40 cm to over 1 m in thickness. The fabric is massive.
Hybrid packstone with Ditrupa -F3.2 This is the most characteristic facies of FA3. Ditrupa is the dominant macroinvertebrate, often passively infilling pods (Fig. 8A) or forming loosely to densely packed concentrations. The grain-size of terrigenous particles is fine-grained and poorly sorted. Macaronichnus -Teichichnus and other traces are characteristic (Table 3). The bedding is completely disrupted by Similar to model III of Kidwell (1985). The bioclasts of species that occur characteristically in FA1 and FA2 are interpreted as allochthnous.
These can include highly altered or well-preserved shells, as a result of being entrained in cohesive debris flows indicates more sustained conditions of low sedimentation rates in time, with Glossifungites ichnofacies, higher diversity of bioerosion traces and brachiopods, and amalgamation of highdensity gravity flows reworking and resedimenting shells. Increased diversity of suspension feeder bioturbation and the fabric is massive. This facies is rich in benthic (Fig. 9) and planktonic foraminifera.

Interpretation of Facies Association 3
This facies association is dominated by D. arietina, a short-lived, free-living suspension-feeder and opportunist that can attain high densities in fine sands and muddy substrates under high sedimentation rates, high turbidity and unstable conditions (Gr emare et al., 1998;Sanfilippo, 1999;Ceregato et al., 2007;Scarponi et al., 2014). The interpreted opportunistic behaviour agrees with its high dominance (Ceregato et al., 2007), as in Pliocene outcrops in the Aguilas and the neighboring Cope Basins (Martinell et al., 2012). A review on its ecology by Hartley (2014) emphasizes two explanations for the high densities reported in modern environments, both associated with disturbances: (i) disruption of established benthic communities, enabling successful recruitment of high numbers of Ditrupa larvae; and (ii) post-settlement redistribution by storms and concentration in areas of deposition. It is therefore probable that the paucispecific fossil assemblages dominated by Ditrupa concentrations in FA3 are associated with the action of storms or internal waves, either by redeposition, by opportunistic responses to storm-induced siltation producing organic-rich substrates, or both (Ceregato et al., 2007;Hartley, 2014). The abundance of the benthic foraminifera Cassidulina carinata, Bolivina spp., Bulimina aculeata and Globocassidulina subglobosa (Fig. 9) is consistent with the organic enrichment associated with siltation (Jorissen et al., 2007;Abu-Zied et al., 2008;Goineau et al., 2012;P erez-Asensio et al., 2017). Ichnoassemblages support the interpretation of high sedimentation rates and nutrient contents,   although there are variations in a proximal to distal gradient, with increasing ichnodiversity and density of traces towards distal positions. In general, facies F3.2 is dominated by Macaronichnus, which is the product of vagile, detritus-feeding worms (Bromley et al., 2009;Pearson et al., 2013). Such trace fossils are rarely reported from offshore settings (Aguirre et al., 2010;Rodr ıguez-Tovar & Aguirre, 2014;Giannetti et al., 2018) and their producers cope well with high sedimentation rates (Taylor et al., 2003). In distal positions (the base of the Cañada Blanca section), other common ichnotaxa include Teichichnus rectus, attributed to a deposit-feeder in nutrient-rich sediments, which can re-equilibrate to the sediment-water interface (MacEachern et al., 2012a). The intense bioturbation in distal positions (Table 3), however, suggests long colonization windows (Buatois et al., 2015) under background fair-weather conditions, because the effects of siltation and/or gravity flows decrease both in intensity and frequency in these settings. The occurrence of the traces Teichichnus, Diplocraterion and Scalichnus in distal F3.2 (Table 3)

Facies Association 4
The major characteristic of Facies Association 4 (FA4) is the presence of fine-grained hybrid packstone distally and the reduced carbonate content (ca 13 to 40%) (Fig. 4), characteristically dominated either by Costellamussiopecten or Terebratula (Figs 10B and 11). Hybrid packstone with Costellamussiopecten -F4.1 Facies F4.1 consists of hybrid packstones with poorly sorted fine-grained sands to coarse silts (variable proportions of micrite and sparite depending on the locality). The terrigenous particles comprise angulose grains of quartz, schist and abundant mica flakes. Planktonic and benthic foraminifera (Fig. 9) are abundant, the latter including centimetre-sized tests of Pyramidulina raphanistrum and Lenticulina spp. As in the other facies, the fabric is massive and structureless. Macrofossils, most notably Costellamussiopecten cristatum, occur as dispersed, complete, disarticulated valves (Fig. 10B). The density of identifiable traces varies. Outsized, angular floating clasts of metamorphic material from the basement are very rare (Fig. 10C). Some of them are pebble-sized, rounded and bioeroded black dolostones (Fig. 10D).

Paraconglomerate of outsized floating clasts -F4.2
This facies occurs only at the base of the Cabezo Alto section (clinothem 1) (Fig. 10A). The matrix is similar to that of F4.1. It is characterized by a paraconglomerate of outsized, angular floating clasts and loosely packed to dispersed bioclasts (Fig. 10E). The richness of vertebrates and macroinvertebrates is the highest in the whole study area (including fish vertebrae, elasmobranch teeth, crustacean dactyla, wood remains bioeroded by Nototeredo sp., plant detritus and others; Fig. S2). Most shells are disarticulated, consisting of a mixture of pristine, fragmented and bioeroded/encrusted specimens of many species; some of them occur typically in FA1 and FA2 (Table 4; Fig. 10F). This facies is densely bioturbated (mostly indistinct mottling) and the richness of identifiable ichnotaxa is relatively high in comparison to other facies. The density of floating lithoclasts and bioclasts peaks at the base of clinothem 1 and decreases progressively upward ( Fig. 10A and F). In the 125 to 500 lm fraction, yellowish to light-green glauconite grains (often preserved as foraminiferal casts) are frequent.

Terebratula pavements -F4.3
This facies is characteristic of the Cabezo Alto -Cañada Brusca area, where 13 outcrops were identified. They consist of ca 5 to 20 cm thick beds in which brachiopods appear embedded in a fine-grained matrix ( Fig. 11A and B). Two outcrops showed two pavements separated by about 20 cm. These are referred to in this study as 'twin pavements' (Fig. 11C). No erosive or planar surfaces, either at the base or at the top of the skeletal concentration, were observed (Fig. 11A, B and G). No normal grading is visible in the matrix; the sediment is indistinguishable from that underlying and overlying the pavements; the shell orientation varies from random to umbo-down; bioclasts are well-preserved and only a few specimens show minor taphonomic alterations; more than 50% of the specimens are articulated. Some pavements yield small juveniles. The packing of specimens is variable, from dense in the centre to loose towards distal and proximal positions of the pavement (Fig. 11G). In some cases, disrupted biological clumping occurs ( Fig. 11B and F). All of the pavements studied yield yellowish to light-green glauconite grains. Chemical analysis of two glauconite grains from one sample showed a K 2 O content of 4Á4% and 4Á3%. This facies is distributed cyclically, most often alternating with F4.1.
Terebratula biostrome -F4.4 This is one thick bed (>1 m) dominated by loosely to densely packed terebratulids (Fig. 12A). It contains a mixture of well-preserved, articulated specimens (Fig. 12B), sometimes devoid of sediment infill (Fig. 12C), and disarticulated valves (Fig. 12E and F). Many of the latter are fragmented, abraded, heavily bioeroded and encrusted by bryozoans, anomiid bivalves (Fig. 12D), serpulids and craniid brachiopods. The biofabric is variable and complex, with examples of in situ Terebratula clumps, pod concentrations and gutter casts (Fig. 12G). Outcrops of this single interval are recognizable for 850 m parallel to the strike, whilst at Cañada Brusca, a low abundance of additional brachiopod species (Table 4) was observed.

Interpretation of Facies Association 4
This facies association crops out in the more distal positions of the depositional profile, where the fine-grained sediment composition indicates low-energy background conditions. This is supported by the occurrence of C. cristatum, characteristic of F4.1, which is an extinct pectinid with delicate valves, frequently reported from offshore environments (Aguirre et al., 1996;Robba, 1996;Yesares-Garc ıa & Aguirre, 2004;Ceregato et al., 2007). Extant species of the homeomorphic genus Amusium (Waller, 2011) inhabit quiet waters on fine sandy and muddy substrates of the Indo-Pacific region, at depths of 10 to 100 m (Fr eneix et al., 1987;Minchin, 2003). The benthic foraminiferal assemblage ( Fig. 9) is also typical of offshore environments (Rasmussen, 2005). The massive   suggested by the dominance of lined burrows: such lining helps to stabilize burrows (constructed as permanent domiciles) in soft substrates (Bromley, 1996;Buatois & M angano, 2011). The dominance of Domichnia therefore indicates well-oxygenated substrates and stable background conditions (Buatois & M angano, 2011). The local occurrence of Trichichnus isp. at the CA section (clinothem 2) might be related to longer periods of stable conditions and a low food content at the sediment-water interface (Pervesler et al., 2008). An event-bed suite can be interpreted based on the occurrence of some Taenidium and backfilled, unidentified meniscate traces (possibly Scolicia isp.), indicating the activity of deposit feeders. They probably reacted opportunistically to sporadic high-density gravity flows or siltation events associated with storms or other disturbances (de Gibert & Goldring, 2007). The occurrence of outsized floating clasts (Fig. 10) across the depositional profile is interpreted as the product of storm-induced highdensity gravity flows (Postma et al., 1988; Mulder & Alexander, 2001;Talling et al., 2012). This is consistent with source areas dominated by schists and phyllites, such as in the Aguilas Basin. In particular, according to Garc ıa-Garc ıa (2004), the sensitivity of these lithologies to erosion favours the production of high volumes of fine fraction, which in turn enhances the formation of cohesive debris flows. The angularity of these clasts (some of which have weak lithologies) (Fig. 10C) suggests that they bypassed the depositional profile directly from the river or ephemeral stream mouth to reach the distal positions where FA4 was deposited, probably by hyperpycnal flows that transformed into cohesive debris flows. The roundness and the presence of Gastrochaenolites traces on the dolostone clasts suggest that the latter were stored in a delta plain, a beach or a cliff-toe (Uchman et al., 2002;Garc ıa-Garc ıa et al., 2011) and were incorporated into the flows during flash floods. The storage area would have been no further away than a few kilometres (Fig. 13), judging from the distribution and structure of the Palomas Unit ( Alvarez & Aldaya, 1985). A possible alternative explanation for their occurrence is kelp or seaweed rafting as a main transportation means (Bennett et al., 1994;Garden et al., 2011;Frey & Dashtgard, 2012) but cooccurrence of the floating clasts with other allochthonous elements [out of habitat molluscs (Table 4), plant debris, Teredolites isp. and Calpensia bryoliths (Fig. S2)] supports the first hypothesis (MacEachern et al., 2005;Ghinassi, 2007;Moissette et al., 2010;Nalin et al., 2010;. In facies F4.3, the frequent pristine preservation of terebratulids, the occurrence of juveniles and disrupted patchiness ( Fig. 11A and B), together with the absence of diagnostic features for hydraulic reworking (e.g. Roetzel & Pervesler, 2004), suggest that these pavements probably represent obrution deposits of autochthonous palaeocommunities (Brett & Seilacher, 1991;F€ ursich, 1995;Brett et al., 2003). They can be interpreted as mixed assemblages, in part within-habitat time-averaged, and in part census death assemblages (Kidwell, 1998). The occurrence of glauconite in the terebratulid pavements points to conditions of very low terrigenous input (Odin & Fullagar, 1988;Harding et al., 2014). Preservation of glauconite grains (often as well-preserved casts of foraminifera) points to their autochthonous or parautochthonous origin (Amorosi, 1997(Amorosi, , 2012. Therefore, compared with F4.1, the Terebratula pavements (F4.3) indicate conditions of notably reduced sedimentation rates. The taphonomic traits and biofabric of these pavements suggest that the Terebratula palaeocommunities in the study area were extirpated by siltation events (Emig, 1989;Toma sov ych & Kidwell, 2017) because fine terrigenous particles clog the lophophore and smother these animals (He et al., 2007). These siltation events probably represent the onset of the next cycle of F4.1 sedimentation. The interpretation here is, therefore, that the pattern of alternating F4.3 and F4.1 facies represents cyclical changes of decreased and increased sedimentation rates.
The 'Terebratula biostrome' (F4.4) shares the dominance of terebratulids with F4.3 (Fig. 12). The variable taphofacies (Table 4) suggests complex taphonomic pathways of mixed biogenicsedimentological origin (Kidwell et al., 1986). The occurrence of in situ clumps (Fig. 12B) indicates that the terebratulids were autochthonous to the biotope where FA4 was being deposited. The dominance of Terebratula points to high shell productivity [high hard-part input rates sensu Toma sov ych et al. (2006)]. In contrast, pod concentrations, pristine void Fig. 13. Proposed palaeogeographic map of the Aguilas Basin during deposition of synthem SP1 (Zanclean, MPl3 biozone). The exact position of the palaeocoastline in the western sector is tentative due to the paucity of shallow-water outcrops, and uncertainty as to their attribution to SP1. The indented black line indicates the rollover of the infralittoral prograding wedge and is shown with estimation of approximate inferred palaeobathymetry at the time of deposition of the oldest clinothems. The palaeocurrent direction is inferred from outcrops of stacked sandwave fields at the Terreros and La Carolina sectors (see Fig. 3), migrating seaward. specimens (geopetal structures) and gutter casts infilled with highly altered and densely packed terebratulid shells, are all characteristic of episodes of rapid burial and hydraulic reworking in situ (Fig. 12C to F) (Baeza-Carratal a et al., 2014). Hydraulic disturbances are also demonstrated by the occurrence of a few disarticulated valves of Spondylus, a typical element of FA2 (where it is often articulated), which suggests an allochthonous origin. The abundant fragmented, abraded, bioeroded and encrusted shells (assemblage-level alteration) are typical for prolonged exposure at the sediment-water interface, which, together with the loose to dense packing (shelliness), points to reduced sedimentation rates (Kidwell, 1985(Kidwell, , 1989. The rich ichnoassemblage of bioerosion traces (Table 4) is strongly uneven, dominated by Entobia isp. (clionaid sponges), highlighting strong hydrodynamics and background sedimentation rates between <1 g m À2 and ca 7 g m À2 (Carballo et al., 1994). The rare occurrence of echinoid rasping traces Gnathichnus pentax coupled with the absence of Radulichnus (scratch marks by herbivorous gastropods and polyplacophorans) (de Gibert et al., 2007) point to dim light or aphotic conditions (Bromley, 2005). The co-occurrence of the above taphonomically altered brachiopod bioclasts, together with abundant pristine specimens, demonstrates that the Terebratula palaeocommunity was able to recover from multiple episodic disturbances, where background conditions of strong hydrodynamics and low sedimentation rates were environmentally optimal for these brachiopods (Emig & Garc ıa-Carrascosa, 1991;Reolid et al., 2012). The co-occurrence of lined and sharpwalled, unlined burrows (Fig. 12D) suggests that the substrate within F4.4 evolved from a softground to a stiffground, indicating a decrease in sedimentation rates (Taylor et al., 2003;MacEachern et al., 2012b).

Facies 5
Cemented burrowed hybrid packstone (Glossifungites ichnofacies) This facies cannot be attributed to any particular facies association because in some places it adjoins FA4 but elsewhere adjoins FA3. It is treated here as a separate type. The matrix consists of a fine-grained hybrid packstone that is cemented (ca 40% CaCO 3 ) and completely bioturbated (Fig. 8C). In general, traces are poorly defined but some are attributable to non-compacted Thalassinoides burrows, except for one locality, where many well-defined, non-compacted Thalassinoides suevicus occur (Fig. 8D). The material infilling the burrows is similar to the surrounding matrix. In some localities, this facies displays loosely to densely packed skeletal concentrations, but Ditrupa is absent.

Interpretation of Facies 5
This facies, interspersed between either FA4 or FA3, is interpreted as the formation of stiffgrounds during phases of low sedimentation rates, which favoured cementation of the sea floor (Taylor et al., 2003;MacEachern et al., 2012b).

Geometry
The Cabezo-Alto and Cañada Brusca W sectors enable delta-scale clinoforms sensu Patruno et al. (2015) to be identified (Figs 1C, 14A and 14B). For example, the clinoform separating clinothems 5 and 6 extends for about 250 m from the toeset-point to the upper rollover (Fig. 1C). The clinoforms display a sigmoidal profile (sensu Adams & Schlager, 2000) where, in general, FA2 and facies 3.1 occur in the upper rollover, FA3 in the foreset and FA4 from the lower rollover basinward, in the bottomset. FA1 is best observed in the Cañada Brusca W sector where it occurs in the topset (Fig. 14D to F).

Stacking patterns
The mapping of stratigraphic surfaces on photomosaics, the outcrop study of bed surfaces and the facies distribution show that SP1 displays a south-east prograding and offlapping stacking pattern of sigmoidal clinothems. Twenty-two clinothems were identified in SP1 (numbering in Figs 4 and 14). In the Cabezo Alto (CA) sector, clinothems 1 to 6 display a forestepping pattern (progradation plus aggradation), evolving vertically from FA4 at the base to FA2 at the top; the latter is truncated and overlain by Quaternary deposits (Fig. 14A and B). The aggrading pattern is present in clinothems 5 and 6, which display facies 3.1 at the top of the sections; this contrasts with the adjacent clinothem 4, with facies 2.3 at the top (Fig. 14A and B). The contacts between these clinothems in distal positions consist of facies F4.2 (only at the base of the CA section), F4.3 (red lines in Fig. 14A and B), or F5. Starting with clinothem 7, a downstepping pattern is visible, with strong shifts from facies F2.2 and F2.3 to F3.2 ( Fig. 14A and B). The contact between these latter clinothems is erosive in the upper part of the sections (Fig. 14C). The offlapping trend is modulated by clinothem 11, which displays the more distal facies F3.1 at the Cañada Brusca W sector compared with adjacent older clinothems ( Fig. 14D and E). Clinothem 11 is followed by strong shifts with facies F1.1 and F1.2 in clinothems 12 and 14, alternating with facies F2.3 in clinothems 13 and 15 through prominent erosive surfaces (Figs 5 and 14D to F). Clinothems 11 and 14 are very distinctive and make up marker beds recognized in the Cañada Brusca E, Cañada Brusca W and Cañada Blanca sectors. The Cañada Blanca section is characterized, in general, by alternating facies associations FA1 and FA2 and erosive contacts in between (Fig. 4). This pattern holds except at the base of the section which displays facies associations FA3 and FA4 up to clinothem 11, sharply overlain by facies association FA1 through an erosive surface (Figs 4 and 12).

MAGNETIC SUSCEPTIBILIY AND CARBONATE CONTENT
The overall carbonate content ranges from 13 to 80% (Fig. 4). Maximum values are attained in the rhodolith dominated facies of FA2; they decrease proximally and distally from this facies. In the CA section no cyclicity in CaCO 3 is recognizable, whereas in the CBL section the variation appears to roughly coincide with variation trends in magnetic susceptibility. Both the CA and CBL sections display very low magnetic susceptibility (MS) values, increasing only near the clinothem boundaries identified by sedimentologicalpalaeontological criteria. The maximum value in the whole studied area is recorded in the paraconglomerate interval at the base of section CA (Fig. 4).

Depositional model
Clinoforms of SP1 have the diagnostic features of sand-prone subaqueous delta-scale clinoforms (Patruno et al., 2015), in particular: (i) steep foresets (≥7°, up to 14°) (Fig. 1C); (ii) a sigmoidal profile; (iii) development on a narrow shelf (an embayment about 14 km wide) (Fig. 13); and (iv) close proximity to the palaeocoast (indicated by bioeroded dolostone clasts of the Palomas Unit). Furthermore, the Pliocene Molino del Saltador delta occurs 6 km north-east from the base of the CA section. This implies that La Serrata and Los Melenchones (Fig. 3B), adjacent to where this delta developed, were already above sea-level during the early Pliocene (Fig. 13).
In general, the facies distribution in SP1 shows a proximal-distal energy gradient with decreasing grain-size distally, especially basinward, beyond the upper rollover (FA3 and FA4). This grain-size distribution matches systems dominated by physical accommodation in which facies belts reflect the hydraulic competence of the sedimentary particles (Pomar & Kendall, 2008). The geometry of the clinoforms is consistent with a prograding distally steepened ramp (Pomar, 2001;Pomar et al., 2002;Mart ın et al., 2004) or an infralittoral prograding wedge (IPW; Hern andez- Molina et al., 2000;Pomar et al., 2015). In both cases, the rollover zone represents an energy threshold above which episodic highenergy conditions affect the topset and below which overall quiet conditions prevail offshore (seaward of the rollover). According to these genetic models, the upper rollover corresponds to the mean storm-weather wave base (SWWB), fostering sediment bypass at the topset and sediment shedding down on the foreset (Hern andez- Molina et al., 2000;Pomar et al., 2015), the latter in the form of siltation/suspension fall-out and as sediment gravity-flows (Massari & Chiocci, 2006). Immediately offshore of the upper rollover zone, sedimentation rates peak (upper foreset) and gradually decrease distally, both in frequency and intensity (in the lower foreset and bottomset) (Walsh et al., 2004;Mitchell, 2012). In the original example of an IPW from off Cabo de Gata (southern Spain) described by Hern andez- Molina et al. (2000), the rollover lies at about 25 m water depth coincident with the mean SWWB. This bathymetry is compatible with the coralline algal assemblages in the study area (outer topset) (Fig. 7), although, during the early Pliocene, the storm intensities at this latitude were presumably stronger due to warmer sea-surface temperatures (SST) (Emanuel, 2005;Beltran et al., 2011). A conservative depth of 25 to 30 m for the upper rollover zone of the Aguilas subaqueous delta-scale clinoforms is, therefore, proposed. The location of the SP1 IPW in the south-western corner of the basin (Fig. 13) implies that it was probably the area most exposed to easterly storms, as opposed to the laminated silty marls occurring in the north-eastern part of the basin, interpreted as a sheltered bay (Montenat et al., 1978).

Sequence stratigraphy
Two hierarchical sequence ranks were here interpreted for the early Pliocene (late Zanclean, MPl3 biozone) SP1 synthem of the Aguilas Basin. The low rank sequences (LRS) are the basic building blocks of the high rank sequence (HRS). In particular, the systems tracts of the HRS are defined both by the LRS stacking patterns and their bounding surfaces (Zecchin & Catuneanu, 2013). The LRS are represented by the identified outcropping clinothems (1 to 22); older ones were eroded, younger ones truncated or covered by colluviums.

Architecture of the high rank sequence
The interpreted HRS is bounded at the top by an extensive unconformity described by Dabrio et al. (1991). The basal unconformity is inferred at the Cabezo Alto sector based on changes in facies, strike, dip and micropalaeontological assemblages between the top of SP0 and base of SP1. This basal unconformity, however, crops out at the Terreros section (3Á35 km to the southwest) (Fig. S1). Further work is necessary to confirm its presence throughout the study area.
The transgressive systems tract (TST) is interpreted here to crop out at the base of the CA section (only the youngest LRS) (Figs 4 and 10). Since the contact between synthems SP0 and SP1 in the Cabezo Alto does not crop out, the high rank transgressive ravinement surface has not been observed. The highstand systems tract (HST) is interpreted from the forestepping stacked clinothems 1 to 6 ( Fig. 14A and B). These clinothems overlie the paraconglomerate at the base of the CA section (facies 4.2, clinothem 1) (Fig. 10), which is interpreted here as the maximum flooding zone (MFZ) (see below). Evidence for the falling stage systems tract (FSST) is shown by the generally downstepping facies stack of clinothems 7 to 22 (Fig. 14). The high rank lowstand systems tract (LST) has not been identified and is thought to occur in a deeper part in the basin, below the present-day sealevel. The general offlapping stacking pattern of the LRS in synthem SP1 (Fig. 14) indicates an overall regressive trend, typical for subaqueous deltas, which form during relative stillstands (highstands or lowstands) (Hern andez-Molina et al., 2000;Pepe et al., 2014;Patruno et al., 2015), or during falling stages of relative sealevel (RSL) (Hansen, 1999;Massari et al., 1999).

High rank transgressive systems tract and maximum flooding zone
The high rank MFZ is interpreted to correspond to the paraconglomerate interval at the base of the CA section (facies 4.2) (Fig. 10), implying that most of the high rank TST does not crop out. According to Zecchin & Catuneanu (2013), the maximum flooding surface (MFS) may correspond to: "a 'cryptic' conceptual horizon within condensed deposits during the time of maximum transgression, without a clear physical expression". Condensation is interpreted here from the pattern of the dispersing upward packing of lithoclasts and bioclasts (Fig. 10), which can be explained by the R-sediment model of Kidwell (1985) (Fig. 10F). This model argues that clasts are increasingly dispersed upward concomitant with an increase in burial rates or higher sedimentary dilution (Dattilo et al., 2012) at the onset of the HST, when sedimentation rates outpace accommodation space. This interpretation assumes a relatively constant frequency of the high-density gravity flow events that deliver allochthonous clasts to these depths (bottomset). In the rest of the synthem, floating lithoclasts in FA3-FA4 are rare and isolated, as expected from higher burial rates during the high rank HST and FSST (Neal & Abreu, 2009). Furthermore, the paraconglomerate interval is densely bioturbated (Zecchin & Catuneanu, 2013) and yields the deepest assemblage of benthic ( Fig. 9) and planktonic foraminifera (Leckie & Olson, 2003). This includes frequent or common outer shelf taxa, such as Planulina ariminensis and Uvigerina peregrina. The high species richness of macrofossils also implies a longer window of time-averaging. Moreover, the position of this interval at the base of the prograding low rank clinothems reinforces its interpretation as the MFZ. Thus the paraconglomerate interval of clinothem 1 is here interpreted as a high rank backlap shell/clast bed. The paraconglomerate interval also coincides with the strongest magnetic susceptibility in the whole study area (Fig. 4).
The erosive surfaces at the CBL section (Figs 4, 6 and 15) are interpreted here to have formed by scouring associated with high-frequency variations of base level (Massari & D'Alessandro, 2012). However, the 'master RSME' or high rank RSME, which represents the onset of forced regression in the high rank sequence, is not the most prominent interpreted RSME in the study area. This can be explained by gradually stronger shoreface erosion at increasingly lower sea-levels (when the amplitude of the RSL fall of the low rank cycles is enhanced by the falling sea-level trend of the high rank cycle). In the Cabezo Alto sector, the erosive surfaces disappear distally and these distal portions are interpreted here as basal surfaces of forced regression.

Architecture of proximal low rank sequences
In the context of hierarchical sequence stratigraphy, Schlager (2004Schlager ( , 2010 recognized 'Ssequences and P-sequences'. The P-sequences have only TST and HST, while S-sequences also contain FSST and LST: P-sequences are equivalent to the small-scale cycles (metres to decametres in thickness) of Zecchin (2007), where R, T-R or T cycle types were recognized based on the predominant development of transgressive (T) or regressive (R) deposits. In general, the clinothems of SP1 can be interpreted as R and T-R cycles, with variations in the architecture depending on the position in the depositional profile and the systems tracts of the HRS. In the Cañada Blanca sector, the most common motif of LRS associated with the high rank FSST consists of R cycles bounded by erosive surfaces overlain by thick skeletal concentrations and a coarsening upward trend (Figs 4, 6 and 15). Some skeletal concentrations can be attributed to shelly tempestites because the texture and grainsize of the matrix is similar to or coarser than that of the material underlying the erosive surface (Figs 5A, 15A and 15B). Distinguishing low rank onlap shell beds (OSB) in shoreface environments from bedsets, which display tempestite amalgamation unrelated to shoreline shifts, is difficult . This is because high-frequency, low-amplitude RSL fluctuations result in subtle facies variations in shoreface environments (Zecchin, 2007). Onlap shell beds form under low sedimentation rates when transgression creates accommodation space further onshore. Skeletal material then accumulates in the shoreface, producing loose to dense packing due to low sedimentary dilution (Fig. 15F). The resulting biofabric of the OSB thus reflects a complex history of multiple events of biotic (bulldozing organisms) and/or hydraulic reworking, along with differential winnowing by storms and tidal currents (Kidwell, 1991;. In the Cañada Blanca sector, erosive surfaces carved on coarse-grained F1.2 ( Fig. 15C and D) and mantled by thick, shell-rich facies with fine-grained matrix (F2.4) (Fig. 15E) are interpreted to reflect RSL fluctuations (Massari et al., 2002;Cattaneo & Steel, 2003). The abundance of complete rhodoliths in many of these shell beds (F2.4) (Fig. 6C) indicates low sedimentation rates (Aguirre et al., 2017).

Architecture of distal low rank sequences
The interpretation here is that the internal architecture of the clinothems in FA4 consists of low rank TST formed by Terebratula pavements (F4.3) and the overlying hybrid packstone, with dispersed to barren packing (F4.1), represents the low rank HST. These cycles therefore conform to the structure of R cycles. In more proximal positions, the Terebratula pavements are occasionally replaced by the Glossifungites ichnofacies (F5) or shell beds (F3.4). A genetic model of these R cycles is presented below.
During stillstand stages of the low rank RSL, sediment aggraded in the topset until reaching the base level. Accommodation space thus became unavailable at the topset and progradation in the foreset resumed, eventually forming a new clinothem (Rich, 1951;Swift & Thorne, 1991;Pomar & Kendall, 2008;Pomar et al., 2015) (Fig. 16A). Background conditions with frequent siltation events and high-density gravity flows in the foreset (F3.2) are indicated by opportunistic faunal responses, including the dominance of Ditrupa, infaunal benthic foraminifera and ichnoassemblages of vagile deposit feeders. During these stillstand stages, F4.1 was deposited at the lower rollover and bottomset.
During stages of low rank RSL rise (Fig. 16B), progradation in the foreset switched off and clinoforms developed as omission surfaces in the bottomsets and foresets, and often as transgressive lags or low rank OSB in the upper foresets and topsets (Massari et al., 1999) (Fig. 5). This is because the base level rose concomitantly with the RSL, creating accommodation space in proximal settings of the topset. This was accompanied by reduced fluvial gradients and sediment trapping in nearshore environments, while more distal settings (mainly, foreset and bottomset) were left starved (Brett, 1998;Embry, 2009;Dattilo et al., 2012). Compared with other examples of subaqueous delta-scale clinoforms (Pomar et al., 2002), low rank RSL rise stages in the study area did not result in aggrading clinothems. Rather, they were non-accretionary, forming only hiatal skeletal concentrations. This implies lower sedimentation rates of the Aguilas subaqueous deltascale clinoforms compared to those at Migjorn (Pomar et al., 2002). The conditions of low-sedimentation rate fostered: (i) the colonization of the bottomsets-toesets and foresets by palaeocommunities of siltation-sensitive suspension feeders (Brett, 1998), in this case from proximal to distal: Schizoretepora, Gibbomodiola and Terebratula (Fig. 1C); (ii) the formation of authigenic minerals such as glauconite (Kidwell, 1991;Catuneanu, 2006;Amorosi, 2012); (iii) the development of densely packed shell beds in the middle-upper parts of the foresets (Fig. 8B) due to sediment starvation and differential winnowing (R-sediment model of Kidwell,1985); and (iv) the formation of firmgrounds associated with cementation, enabling colonization by callianassid shrimps and development of the Glossifungites ichnofacies (Taylor et al., 2003;MacEachern et al., 2012b) (Fig. 8C and D).
Hypotheses about the formation of the Terebratula pavements Three main hypotheses are considered here to explain the genesis of the Terebratula pavements ( Fig. 17): 1 The allochthonous concentration hypothesis (Kidwell et al., 1986) envisages that the terebratulids were deposited at the lower rollover and bottomsets after being entrained in high-density gravity flows induced by storms or other disturbances (for example, internal waves). Their autochthonous habitat would be located in more proximal environments (for example, the upper rollover and outer topset) (Fig. 17A). This hypothesis is rejected because of: (i) the lack of diagnostic physical sedimentary structures for shelly tempestites (Einsele & Seilacher, 1991;Einsele, 2000;Fl€ ugel, 2004;Roetzel & Pervesler, 2004;Myrow, 2005); and (ii) the absence in nearly all pavements of other taxa that are abundant or dominant in more proximal environments of the depositional profile. An allochthonous concentration would consist of a mixture of taxa entrained and mixed up from different habitats (Leighton & Schneider, 2004). Furthermore, to produce an allochthonous brachiopod-dominated concentration at the bottomsets, there should be high brachiopod productivity in the presumably autochthonous habitat in more proximal environments, which was not observed in the study area.
2 The storm-winnowing model (Dattilo et al., 2008(Dattilo et al., , 2012 considers that the Terebratula pavements are concentrated autochthonous shell lags that result from differential winnowing of the fine-grained sediment during storms (Fig. 17B). This hypothesis is rejected because, to produce such a dominance and abundance of terebratulids, high brachiopod productivity should occur throughout the stratigraphic intervals of FA4, between terebratulid pavements. These, instead, are barren or are characterized by dispersed Costellamussiopecten.
3 The episodic starvation model (Dattilo et al., 2008(Dattilo et al., , 2012 considers that the Terebratula concentrations are the result of biological processes during stages of low sedimentation rates (Fig. 17C). This hypothesis is supported by the disrupted biological patchiness, the presence of juveniles, the dominance of articulated, pristine shells and by the occurrence of glaucony, a typical proxy for condensed deposits. The occurrence of Terebratula clumps (sensu Kidwell et al., 1986) in F4.4 demonstrates that Terebratula is autochthonous to FA4 (Hallam, 1961;Middlemiss, 1962;F€ ursich, 1995).

Magnetic susceptibility and carbonate content
Quartz, calcite and organic compounds yield very weak to negative magnetic susceptibility (MS) values (diamagnetic minerals). In contrast, paramagnetic minerals such as clays (smectite, illite and chlorite); ferromagnesian minerals (biotite, tourmaline, pyroxenes and amphiboles); iron sulphides (pyrite and marcasite) and iron carbonates (siderite and ankerite), yield MS values several orders of magnitude higher than those of diamagnetic minerals, which dominate the signal when present in bulk samples (Davies et al., 2013;Sullivan & Brett, 2013). Magnetic susceptibility in marine sedimentary rocks is usually considered as a proxy for the proportion of iron-rich sediments derived from terrestrial sources (Ellwood et al., 2000;Sullivan & Brett, 2013). High MS values are considered to be attained during regressive stages, when increased erosion delivers proportionally higher amounts of terrestrial iron-rich sediments into the marine basin (Sullivan & Brett, 2013). This argument has been contradicted by the report of distinct MS peaks associated with surfaces of maximum starvation (Ellwood et al., 2011). This can be explained by concentration of paramagnetic particles derived from aeolian sources (Reuter et al., 2013). Likewise, very low to negative MS values, as in the Aguilas Basin, SP1, may be explained by a very low terrigenous input and/or dilution of terrigenous particles in biogenic carbonate (Reuter et al., 2013). When distinct positive MS peaks are the result of increased terrigenous input, MS trends anticorrelate with those of CaCO 3 (Davies et al., 2013;Rothwell & Croudace, 2015). At the CA section, distinct MS peaks are coincident with a relative increase in CaCO 3 content (Fig. 4), suggesting that the MS values in such cases are associated with condensation. These peaks occur at the clinothem boundaries that were interpreted as omission surfaces (Glossifungites ichnofacies) or condensed intervals (paraconglomerate and Terebratula pavements) based on sedimentological and palaeontological features. The weak MS values are also potentially influenced by the slightly evolved glauconite content, which is paramagnetic (Amorosi, 1997).

Progradation rates of the lower rank cycles
Biostratigraphic data constrain the maximum possible duration of the HRS to somewhat less than 700 kyr. The yellowish to light green colour of the glauconite grains that separate the LRS in facies 4.3 suggests that, in terms of maturity, this is a slightly evolved stage. This was confirmed in one sample, where glauconite grains had a K 2 O content of ca 4%. According to Amorosi (2012), the slightly evolved glaucony would indicate sediment-starved periods lasting about 10 4 years, implying that the LRS in the study area can be interpreted as high-frequency cycles in the Milankovitch band. The cyclicity in variation of terrigenous input is also recorded in the patterns of magnetic susceptibility and calcium carbonate content (Fig. 4). These patterns of magnetic susceptibility are similar to those reported by Davies et al. (2013) for the high-frequency cycles of the Llucmajor platform (Miocene, Spain), reinforcing the above interpretation. If this is true, the time-span encompassed by the HRS is considerably less than the 700 kyr suggested by biostratigraphic proxies.

The
Aguilas Basin records subaqueous deltascale clinoforms that prograded during the early Pliocene (MPl3 biozone) in mixed temperate carbonate-siliciclastic environments. The sedimentological and palaeontological features of these clinoforms are compatible with the infralittoral prograding wedge model. The prograding units formed during the highstand and falling stages of a high rank relative sea-level cycle, and the biostratigraphic data indicate that this progradation lasted for less than 700 kyr. The basic building blocks of this sequence are clinothems whose internal architecture generally consists of skeletal concentrations overlain by a stratigraphic interval with a more disperse packing. In distal positions of the depositional profile, the skeletal concentrations consist of terebratulid brachiopod pavements. These pavements are distributed cyclically; they are interpreted here to have formed during high-frequency relative sea-level rise pulses that led to sediment starvation in these distal environments. During stillstand stages, accommodation space eventually became unavailable in the topset of the clinoforms, leading to a resumption in the progradation of the clinoform system, extirpating the brachiopod communities until the next cycle of relative sea-level rise. In other examples of subaqueous deltas, similar brachiopod assemblages bound the clinobedded unit at the base and the top, but did not occur on the clinoforms, as seen in the Aguilas Basin. This implies lower progradation rates of the Aguilas Basin clinoforms, allowing enough time for these benthic communities to develop. The occurrence of slightly evolved glauconite in the Aguilas Basin suggests that these high-frequency cycles fall within the Milankovitch band, probably precession.

ACKNOWLEDGEMENTS
We thank Gregorio Romero S anchez from the Paleontological Heritage Service of the Community of Murcia (Spain), for granting permits to conduct palaeontological fieldwork in the study area; Javier Souto-Derungs, Andrey Ostrovsky and Eric Wolfgring for their help with the SEM images and the identification of Schizoretepora. We thank Alfred Uchman for cross-checking identifications of certain trace fossils and for palaeoenvironmental suggestions about the Macaronichnus-Ophiomorpha ichnofabric. Thanks to Martin Maslo, Christa Hermann, Robert Peticzka and Theodoros Ntaflos for the analysis of carbonate content and glauconite. This work was inspired during fieldwork in the Aguilas Arc with Jes us Soria, Hugo Corb ı, Jordi Martinell, Rosa Dom enech and Cristino Dabrio, all of whom are also thanked for providing literature and for discussions in the field. We thank Guillermo D ıaz-Medina, Johann Hohenegger, Wolfgang Eder, Vlasta Cosovi c, Adam Toma sov ych, Rafał Nawrot and many other colleagues for fruitful discussions and suggestions. Special thanks to Ildefonso Bajo and Enrico Borghi for their input about the taxonomy of Pliocene echinoids; to Julio Aguirre and Juan Carlos Braga for sending literature about rhodoliths and Andr es Guilabert for his help with Digital Terrain Models. We thank Michael Stachowitsch and Alexander Hugh Rice for improving the language. We thank the Editors Peir Pufahl and Christopher R. Fielding for the constructive comments during the review process. The stimulating reviews of Massimo Zecchin, Carlton Brett and an anonymous colleague significantly improved the final version of this manuscript. The authors declare that there is no conflict of interest.