Reflooding and repopulation of the Mediterranean Sea after the Messinian Salinity Crisis: Benthic foraminifera assemblages and stable isotopes of Spanish basins

Marine Micropaleontology 176 (2022) 102160 Available online 16 August 2022 0377-8398/© 2022 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/). Reflooding and repopulation of the Mediterranean Sea after the Messinian Salinity Crisis: Benthic foraminifera assemblages and stable isotopes of Spanish basins F. Bulian a , e , * , T.J. Kouwenhoven b , N. Andersen c , W. Krijgsman d , F.J. Sierro a a Dept. de Geología, Univ. de Salamanca, Plaza de los Caídos s/n, 37008 Salamanca, Spain b Faculty of Earth Sciences, Utrecht University, Utrecht, the Netherlands c Leibniz-Laboratory for Radiometric Dating and Isotope Research, Christian-Albrechts-Universit¨ at Kiel, Max-Eyth-Str.11-13, 24118 Kiel, Germany d Palaeomagnetic Laboratory Fort Hoofddijk, Utrecht University, Budapestlaan 17, 3584 CD Utrecht, the Netherlands e Groningen Institute of Archeology, University of Groningen, Poststraat 6, 9712 ER Groningen, the Netherlands ARTICLE INFO Keywords: Miocene-Pliocene boundary Alboran Basin Spanish marginal basins Benthic foraminifera ABSTRACT Benthic foraminiferal, sedimentological, and stable isotope analyses performed on early Zanclean sediments from Alboran Basin ODP Site 976 and southern Spanish land-based sections in the Malaga, Nijar and Sorbas basins have enabled the reconstruction of Mediterranean environmental conditions immediately after the Messinian Salinity Crisis. The presence at the Miocene – Pliocene boundary of dark layers, often enriched in organic matter, suggests that the Zanclean reflooding has created water column stratification, and reduced bottom-water oxygen levels. Considering that such dark layers are recorded at both deep and marginal settings far away from the Gibraltar gateway/s, a Mediterranean-scale water-mass stratification must have occurred. This stratification could be the result of saline Atlantic waters sinking into a less saline Mediterranean Basin still under the influence of the Paratethys. Our early Zanclean benthic δ 18 O data show that the Mediterranean water budget was indeed less negative than at present, explaining the lower salinity of the basin. However, the Atlantic values of the benthic δ 13 C registered in the Alboran basin suggest that bottom-water renewal rates were quite high during the early Zanclean, preventing the reduction of δ 13 C at the seafloor as observed in the Messinian records. Zanclean benthic foraminiferal repopulation sequences show similarities with recovery from low-oxic episodes during sapropel deposition. These observations, paired with the gradual deepening of the basins, suggests that the Zanclean reflooding led to a progressive shift from stressed and unstable environments towards benthic associations typical of efficient circulation and bottom water ventilation. 1. Introduction Since the 1970s there is an ongoing debate regarding the end of the Messinian Salinity Crisis (MSC) of the Mediterranean Sea (Fig. 1A) and the exact dynamics of restoration of marine conditions. Through time three main reflooding scenarios were proposed: (1) an instantaneous inundation of an (almost) desiccated Mediterranean at 5.33 Ma, corresponding to the base of the Zanclean (Hsü, 1972; Blanc, 2002; Loget and Van Den Driessche, 2006; Garcia-Castellanos et al., 2009; García-Alix et al., 2016); (2) high Mediterranean water level during the latest Messinian with rapid restoration of the Atlantic-Mediterranean connection characterised by a minor sea-level rise across the Messinian-Zanclean transition (Loget et al., 2005; Pierre et al., 2006; Roveri and Manzi, 2006; Corn´ ee et al., 2016; Marzocchi et al., 2016; Andreetto et al., 2021b) and (3) a two-stage restoration of the connection with the major reflooding taking place at 5.46 Ma, during the Messinian (Estrada et al., 2011; Bache et al., 2012; P´ erez-Asensio et al., 2013; Bache et al., 2015). The three models differ not only in the dynamics of the reflooding itself, but also in the timing and the magnitude of the base level fluctuations. The Zanclean sediments in the Mediterranean were deposited on top of shallow (Nijar Basin: Bassetti et al., 2006, this work; Malaga Basin: Guerra-Merch´ an et al., 2008, this work; Sorbas Basin, this work; Fig. 1B) and deep, late Messinian Lago-Mare (LM) deposits (Eraclea Minoa: e.g., * Corresponding author at: Dept. de Geología, Univ. de Salamanca, Plaza de los Caídos s/n, 37008 Salamanca, Spain. E-mail addresses: [email protected] (F. Bulian), [email protected] (T.J. Kouwenhoven), [email protected] (N. Andersen), W.Krijgsman@uu. nl (W. Krijgsman), [email protected] (F.J. Sierro). Contents lists available at ScienceDirect Marine Micropaleontology journal homepage: www.elsevier.com/locate/marmicro https://doi.org/10.1016/j.marmicro.2022.102160 Received 9 March 2022; Received in revised form 6 August 2022; Accepted 9 August 2022 Marine Micropaleontology 176 (2022) 102160 2 Brolsma, 1978; Sites 975 and 974: Iaccarino and Bossio, 1999; Iaccarino et al., 1999a). The contact at the Miocene-Pliocene boundary (MPB) is expressed in different ways including conformably above continental facies (Sorbas Basin, e.g., Roveri et al., 2018), and erosional both in deep and shallow basins (e.g., Chelif Basin: Rouchy et al., 2007; Zakynthos Island: Kontakiotis et al., 2016 and Karakitsios et al., 2017; Vera Basin: Caruso et al., 2020; Alboran Basin: Bulian et al., 2021). At some locations conglomerates, breccias or small clasts are intercalated (Site 975: Iaccarino et al., 1999b (Fig. 1E); East Alboran Basin: Garcia-Castellanos et al., 2019; Levant Basin: Madof et al., 2019). In land sections and offshore sites, the boundary is recognized through changes in lithology, carbonate content or stable isotopes of carbonates (Iaccarino et al., 1999b; Pierre et al., 2006; Rouchy et al., 2007). A characteristic lithology identified at the MPB (marking the earliest Pliocene sediments and/or latest Messinian ones) is a cm/dm-thick black layer deposited conformably above the LM deposits in the Piedmont Basin (Fig. 2G; Trenkwalder et al., 2008; Dela Pierre et al., 2016), Northern Apennines (Northern Apennine; Gennari et al., 2008; Grossi Fig. 1. A) Bathymetric map of the Mediterranean region (from https://portal.emodnet-bathymetry.eu/) where the rectangle shows the studied area and the Guadalquivir Basin. B) Bathymetric map of the Alboran Basin and Rifian Corridor. The black dot and rectangles indicate the studied ODP Site and onshore sections (Rio Mendelin section-Malaga basin; Zorreras section-Sorbas Basin and Barranco del Negro section-Nijar Basin) respectively. The hatched areas show the extension of the Messinian gateways (after Martín et al., 2014). Fig. 2. Photographs of the Miocene-Pliocene boundary from some of the mentioned locations. A) Rio Mendelin section (this study); B) Zorreras section (this study); C) Eraclea Minoa section (personal collection); D) ODP Site 974 (Iaccarino et al., 1999b); E) ODP Site 975 (Iaccarino et al., 1999b); F) Barranco del Negro section (this study); G) Moncucco quarry in Piedmont Basin (courtesy of. F. Andreetto) and H) Zakynthos island in Greece (courtesy of K. Agiadi). F. Bulian et al. Marine Micropaleontology 176 (2022) 102160 3 et al., 2008), Northern Italy (Riforgiato et al., 2011), Cyprus (Rouchy et al., 2001; Manzi et al., 2016), Sicily (Brolsma, 1978; unpublished data; Fig. 2C), DSDP Site 376 (Cita et al., 1978), ODP Sites 974 and 975 (Iaccarino et al., 1999b; Fig. 2D, E), Zakynthos (Fig. 2H; K. Agiadi, pers. comm.) and in sections in the Malaga, Sorbas and Nijar Basins included in this study (Fig. 2A, B, F). At some locations (e.g., Malaga Basin) these layers are deposited just above the MPB. This black layer is usually rich in organic matter (Gennari et al., 2008; Trenkwalder et al., 2008; Manzi et al., 2016) and scarce in foraminifera which are usually typical Pliocene species, reworked Miocene to Eocene species, or in some cases both (Cita and Zocchi, 1978; Gennari et al., 2008; Manzi et al., 2016). At some locations microfossils are absent (e.g. Moncucco quarry: Trenkwalder et al., 2008). The palaeoenvironmental significance of such dark layers remains still largely unknown (e.g. Gennari et al., 2008). To test and validate the most probable reflooding scenario and explain the occurrence of dark layers it is necessary to increase our understanding of the palaeoenvironmental conditions present after the MSC and therefore, to analyse the late Messinian - early Zanclean sedimentary record at locations close to the Mediterranean – Atlantic gateway. In this work, we studied three onshore outcrops from Neogene basins in southern Spain which contain well-preserved late Miocene – early Pliocene sedimentary successions: the Rio Mendelin section in the Malaga Basin, the Barranco del Negro section in the Nijar Basin and the Zorreras section in the Sorbas Basin (Fig. 1B). In addition, we reevaluated the early Zanclean sediments retrieved at ODP Site 976 in the West Alboran Basin, located in front of the Gibraltar gateway (Fig. 1B). During the late Miocene, these basins were part of the Mediterranean realm and consequently their sediments could have registered the first influx of Atlantic waters, together with Zanclean water levels and palaeoenvironmental conditions in the western Mediterranean. We performed detailed benthic foraminiferal and stable isotope analyses of these four sites to better understand the dynamics of the reflooding together with the environmental changes that occurred in the Mediterranean after the MSC. 2. Geological setting and material studied 2.1. Neogene Basins of the Betic Cordillera The Malaga Basin (Fig. 1B) is an intermontane E-W oriented basin located in the westernmost sector of the Betic Cordillera (Sanz De Galdeano and Vera, 1992). The basin is drained by the Guadalhorce River and bounded to the north by the Malaga Mountains, to the south by the Mijas Sierra and to the east by the Alboran Basin. Its post-orogenic infilling is composed of upper Tortonian to Quaternary sediments (e.g. Guerra-Merch´ an et al., 2008). The section analysed in this study (Rio Fig. 3. Field photograph of the Miocene-Pliocene boundary in the Malaga Basin (Rio Mendelin section). Fig. 4. Field photograph of the Miocene-Pliocene boundary in the Sorbas Basin (Zorreras section). F. Bulian et al. Marine Micropaleontology 176 (2022) 102160 4 Mendelín section; 36◦45′2.53”N; 4◦25′57.26”W; Fig. 2A, 3, 6A) comprises late Messinian deposits related to the final stage of the MSC (LagoMare unit) and the lower Zanclean. The Río Mendelín section is located in an ancient fluvial valley, incised during the latest Tortonian or during the MSC acme sea level drop, which, based on the Paratethyan affinity of its faunal content, was progressively refilled during the dilution phase at the end of the Messinian (Guerra-Merch´ an et al., 2010; Do Couto et al., 2014). The Pliocene sediments are deposited with an erosional contact at the margins and a gradual conformable contact in the centre of the basin on top of the LM unit. In this study, we analysed 23.5 m of the Rio Mendelin section, including 2 uppermost metres of the LM sequence followed, with a transitional contact, by early Pliocene sediments composed of 25 cm of yellow silty sediments, 75 cm of grey clays overlain by 20.5 m of yellowish clays (Fig. 6A). The Nijar and Sorbas basins are intermontane basins developed above the metamorphic nappes of the southern Betics (Fig. 1; Sanz De Galdeano and Vera, 1992). The sedimentary infilling of the basins occurred during the late Miocene (Tortonian and Messinian), Pliocene and Quaternary (Dabrio et al., 1981; Serrano, 1990; Omodeo Sal´ e et al., 2012). During the Miocene, until the end of MSC Stage 1, the two basins Fig. 5. Field photograph of the Miocene-Pliocene boundary in the Nijar Basin (Barranco del Negro section). Fig. 6. Stratigraphic logs of the three sections studied in Southern Spain. A) Rio Mendelin section; B) Zorreras section, C) Barranco del Negro section. The black dots represent the analysed samples. The colours reflect the colours observed in the field. F. Bulian et al. Marine Micropaleontology 176 (2022) 102160 5 were connected to the Mediterranean, which entered the Nijar Basin from the south and the Sorbas Basin through NW-SE trending corridors north of Nijar (Fortuin and Krijgsman, 2003). In this work, the upper Messinian-lower Zanclean sediments from the Nijar and Sorbas basins have been studied (Figs. 4 and 5). In the Sorbas Basin, lower Pliocene sediments (Gochar Formation) lie on top of the Zorreras Member which mainly consists of palaeosols, fluvio-deltaic reddish silts and sandstones, grey-coloured sandstones and conglomerates intercalated by up to 4 whitish, massive carbonate beds (Manzi and Roveri, 2009; Aufgebauer and McCann, 2011; Roveri et al., 2018; Roveri et al., 2019b) containing euryhaline ostracod specimens of Paratethyan affinity like Cyprideis sp. (e.g., Aufgebauer and McCann, 2011). The Zorreras member is considered as the equivalent of the LM facies (Rouchy and Caruso, 2006). The marine early Pliocene sands show 0.5 m above the boundary a fossiliferous horizon rich in bivalves (Fig. 6B) that can be traced throughout the basin (Mather and Stokes, 2001; Roveri et al., 2019b). Here, we analysed 2 m of the Zorreras section (37◦6′ 9.87′′ N; 2◦6’ 46.78′′ W) composed of 0.5 m of reddish LM deposits, 0.3 m of grey deposits containing carbonate nodules and 1.2 m of Zanclean massive sands including the intercalated bivalve horizon (Fig. 4 and 6B). In the Nijar Basin (Fig. 1B), the upper Messinian is composed of alternations of marly LM facies and conglomerate alluvial beds (Omodeo Sal´ e et al., 2012) of the Feos Formation, which are overlain by the lowermost Zanclean with an unconformity in the marginal parts of the basin and a conformable contact at the centre (Fortuin and Krijgsman, 2003; Aguirre and S´ anchez-Almazo, 2004; Roveri et al., 2019a). This study focuses on the Barranco del Negro section (Fig. 2F, 5, 6C; 37◦ 0′35.02′′ N; 1◦58’ 23.02′′ W) in the north-eastern part of the Nijar Basin. Here, the MPB can be clearly identified, with a sharp contact over uppermost Messinian sediments, a 5 cm-thick black and 5 cm-thick grey layer topped by lower Zanclean massive yellow bioclastic marine sands, characterised at the base by a 1 cm-thick layer with bivalve shell fragments (Fig. 2F, 6C). 2.2. Alboran Basin The Alboran Basin is a transitional area between the semi-enclosed Mediterranean Sea and the Atlantic Ocean (Fig. 1A, B) characterised by vigorous circulation that is strongly related with water exchange at the Gibraltar Strait where the relatively low-salinity Atlantic waters enter the Mediterranean on top of high-salinity Mediterranean water masses. ODP Site 976 (36◦12′18.78” N, 4◦18′45.78” W) is located in the northern sector of the West Alboran Basin (WAB; Western Mediterranean), ~100 km to the east of the Gibraltar Strait on top of a continental crustal horst that formed during earlyto mid-Miocene rifting (Comas et al., 1996). The marine lower Pliocene sediments recovered at this site are mainly composed of homogeneous nannofossil-rich claystone and sandy claystone (Comas et al., 1996) with no visible changes in colour. We analysed 32 samples from the lower Pliocene of core 61. From the underlying core 62, only section 62×-CC was recovered in which three samples were collected. Because of their uncertain stratigraphic position, these have been excluded from the micropaleontological interpretation (Bulian et al., 2021). The position of the MPB was previously estimated with the aid of seismic interpretation and regional biostratigraphy at the base of core 61 (Bulian et al., 2021), corresponding with a visible erosional surface. This erosion has been associated with the Zanclean reflooding, when after the breaching of the Gibraltar Strait, the Atlantic water would have abruptly entered the Mediterranean (GarciaCastellanos et al., 2009; Estrada et al., 2011) and produced a marked incision (Esteras et al., 2000; Blanc, 2002). Alternatively, if a modest drawdown is considered, a hyperpycnal submarine cascading erosion (Roveri et al., 2014b) could be the creation mechanism. The time slice represented by the hiatus has been estimated at 1.67 Ma and includes the majority of the middle-upper Messinian sediments (Bulian et al., 2021). 3. Methodology 3.1. Micropalaeontological analyses For this study, a total of 95 samples (33 from the Rio Mendelin section, 32 from ODP Site 976, 21 from the Barranco del Negro section, 9 from the Zorreras section) taken with variable sampling steps (10–20 cm for the Barranco del Negro and, Zorreras sections and Site 976 and between 10 cm and 1 m for the Rio Mendelin section) were selected for micropalaeontological analyses. The samples were oven dried at 40 ◦C and washed over 63 μ m and 150 μ m sieves. For faunal analysis, aliquots of the 150 μ m fraction of on average 150–200 benthic foraminifera were counted. The counts of benthic foraminifera were then transferred to relative frequencies. Samples yielding <50 specimens were not included in the analyses and interpretations. The benthic foraminiferal content of the Zorreras section has been studied in a semi-quantitative way based on the presence or absence of species considering that very few specimens were present. Similarly, in the three land-based sections the presence of ostracods was recorded, but only the LM marker species (Cyprideis sp.) was identified from the assemblage. The diversity of the benthic foraminiferal assemblages has been estimated through the Shannon index (H, Murray, 1991; Spellerberg and Fedor, 2003), expressed by the formula: H= − K∑ n i=1 piln(pi) where pi is the proportion of the i th species and K a positive constant. For ODP Site 976, of which the Messinian part was deposited at slope depth (Bulian et al., 2021) the sum of shallow-water benthic foraminiferal species, Elphidium spp., Rosalina spp., discorbids and Cibicides lobatulus (Supp. 1) was calculated. Because a relatively high number of displaced specimens indicates downslope transport (Fentimen et al., 2020 and references therein) this sum has been used as an indicator of currents in the basin. 3.2. Statistical analyses Of each data set except from the Zorreras and Barranco del Negro sections (Supp. 2), a hierarchical cluster analyses (Pearson correlation: Past 4.02 software; Hammer et al., 2001; Hammer et al., 2008) was performed on the most abundant (≥3%) variables (taxa) to identify the distributional patterns of benthic foraminiferal species assemblages occurring in similar environments. Species with rare or single occurrences were removed from the data whereas species belonging to the same genus, when considered indicative of analogous environmental conditions, were grouped together. Nodosaria spp. comprises unilocular spp., Nodosaria spp., Pseudonodosaria spp., Lagena spp. and Procerolagena spp., Cancris spp. contains the species C. oblongus and C. auriculus. Uvigerina peregrina includes both Uvigerina peregrina and U. pygmea, and Bulimina striata is the sum of B. striata and B. striata mexicana. In Pullenia quinqueloba, both forms with four and five chambers were included. 3.3. Palaeo – water depth estimates Palaeo-water depth estimates were initially performed using the ratio (P/(P +B))*100 (%P; Gibson, 1989; Van der Zwaan et al., 1990), which, although this reflects general sea-level trends, has disadvantages related to sensitivity to oxygen levels, food availability and preferential dissolution of the planktic fauna (e.g., Sen Gupta and Machain-Castillo, 1993; Jorissen et al., 1995; Kucera, 2007). To improve the reliability of palaeodepth estimations, we excluded from the ratio the infaunal foraminiferal species (e.g., buliminids, bolivinids, uvigerinids, Nonion spp.). This forms the basis for calculation of the regression function (Van der Zwaan et al., 1999; Van Hinsbergen et al., 2005). However, this approach still does not account for dissolution of planktic foraminifera. F. Bulian et al. Marine Micropaleontology 176 (2022) 102160 6 Therefore, to obtain an independent quantification of the palaeobathymetry we estimated the palaeo-water depth using one of several more recently developed equations (e.g. Hohenegger, 2005; Hohenegger et al., 2008; Avnaim-Katav et al., 2016; Milker et al., 2017) based exclusively on the benthic foraminiferal fauna. In this study we apply the transfer function of Hohenegger (2005), since this is considered reliable in deeper environments and has been evaluated by Baldi and Hohenegger (2008) in the Vienna Basin and by P´ erez-Asensio et al. (2012) in the Guadalquivir basin. The equation, here used as modified by Hohenegger et al. (2008) and Baldi and Hohenegger (2008) includes the relative abundances of each species and their depth ranges (Table 1): Paleodepth (m) = ∑k j=1[(lj*nj)/dj]/∑k j=1(nj/dj) where n j is the relative abundance of the n th species, l j the mean species depth, and d j the dispersion. As suggested by Hohenegger (2005), to calculate the mean species depth for shallower marginal basins (e.g., Rio Mendelin section) we used the geometric means, while for deeper basins (ODP Site 976), we used the arithmetic mean in order to avoid underestimation. The foraminifera used in the calculation are all autochthonous species and reported in Table 1. 3.4. Estimation of bottom-water oxygen levels The benthic foraminiferal distribution within the sediment depends on the organic flux and oxygenation at the sea floor (Jorissen et al., 1995; Van der Zwaan et al., 1999). Consequently the microhabitat is used as an indicator of dissolved oxygen (Kaiho, 1994), since species with deep infaunal (>3 cm below the sediment-water interface; BSWI) and intermediate infaunal (>0.7 cm BSWI) microhabitats thrive when oxygen levels are reduced, while epifaunal ones (0–0.7 cm BSWI) prefer better oxygenated bottom waters (e.g. Corliss and Chen, 1988; Jorissen et al., 1995; Schmiedl et al., 2000; Gooday, 2003). These microhabitat preferences, reflected by morphological characteristics (size, wall thickness, shape: e.g., Corliss and Chen, 1988) can be used to define three groups of benthic foraminifera (Table 2) indicating respectively oxic (>1.5 ml/l O 2 ), suboxic (0.3–1.5 ml/l O 2 ) and dysoxic (0.1–0.3 ml/l O 2 ) environments (Kaiho, 1991; Kaiho, 1994; Kaiho, 1999). In addition, within the dysoxic indicators three different groups (A, B, C) have been defined (Kaiho, 1994) where group C includes species that have intermediate characteristics between suboxic and dysoxic markers. Using these indicators, the Benthic Foraminiferal Oxygen Index is calculated (BFOI; Kaiho, 1991; Kaiho, 1994; Kaiho, 1999) obtaining five different conditions of dissolved oxygen: anoxic (−55), dysoxic (−50 to −40), suboxic (−40–0), low oxic (0–50), and high oxic Table 1 The bathymetric ranges, mean living depth and standard deviation of the benthic foraminiferal species used for palaeo – water depth reconstruction. The species used for the Rio Mendelin section are underlined, while the ones employed for ODP Site 976 have an asterisk. Depth ranges are based on previously published work of: 1 Wright (1979), 2 Wright (1978), 3 Lutze and Coulbourn (1984), 4 Pflum et al. (1976), 5 Lutze and Wefer (1980), 6 Haake (1982), 7 Barbieri and Panieri (2004), 8 Van Hinsbergen et al. (2005), 9 Violanti et al. (2011), 10 Corbí (2010), 11 Van Morkhoven et al. (1986), 12 De Stigter et al. (1998), 13 Baggley (2000), 14 Gebhardt (1993), 15 Berggren and Haq (1976), 16 P´ erez-Asensio et al. (2012), 17 Poag and Tresslar (1981), 18 De Stigter et al. (1998), 19 Bandy and Chierici (1966), 20 De Rijk et al. (2000), 21 De Rijk et al. (1999); 22 Sen Gupta and Machain-Castillo (1993), 23 Murray (2006), 24 Bizon and Bizon (1984), 25 Mendes et al. (2012), 26 Ohga and Kitazato (1997), 27 Milker and Schmiedl (2012), 28 Alve (2003), 29 Austin and Evans (2000), 30 Schmiedl et al. (1997), 31 Suokhrie et al. (2021), 32 Russo et al. (2007) and 33 (Hayward, 2004). Species min depth max depth SD Amphicoryna spp.* 9 2860 1425.5 Anomalinoides helicinus* 600 2000 700 Asterigerina planorbis 200 400 1000 Bolivina dilatata 15 3000 1492.5 Bolivina spathulata 30 3547 1758.5 Bulimina aculeata* 5 4000 1997.5 Bulimina elongata 16 200 92 Bulimina mexicana 100 2000 950 Bulimina striata* 100 800 350 Cancris oblongus* 30 500 60 Chilostomella spp.* 700 1900 600 Cibicidoides bradyi* 200 3000 1400 Cibicidoides pseudoungerianus/ungerianus* 50 4000 1975 Cibicidoides kullenbergi* 1000 4000 500 Cibicides lobatulus 20 1300 640 Cibicidoides pachyderma* 30 4000 1985 Cibicides dutemplei* 100 600 250 Dentalina spp.* 30 1200 585 Fursenkoina acuta 0 600 300 Globobulimina spp.* 1000 1500 250 Globobulimina turgida 30 150 60 Globocassidulina subglobosa* 50 4000 1975 Gyroidina altiformis* 30 600 285 Gyroidina soldanii* 100 5000 2450 Karreriella bradyi* 100 3000 1450 Lenticulina spp.* 600 1500 450 Martinotiella communis* 200 3000 1400 Melonis barleeanus* 13 3974 1980.5 Melonis pompilioides* 100 3000 1450 Melonis soldanii* 90 1000 455 Nodosaria spp.* 30 1700 235 Nonion fabum 0 200 100 Oridorsalis stellatus 250 1500 625 Oridorsalis umbonatus* 65 4000 1967.5 Planulina ariminensis* 70 1300 615 Pullenia bulloides 60 4000 1970 Pullenia quinqueloba* 50 2000 975 Sigmoilopsis schlumbergeri* 57 1500 721.5 Sphaeroidina bulloides* 100 2000 950 Stainforthia fusiformis 0 2200 1080 Textularia calva 0 2000 1000 Trifarina bradyi 0 600 300 Uvigerina peregrina* 100 4400 2150 Uvigerina rutila* 200 1000 1400 Table 2 Microhabitat preferences of benthic foraminifera from the Rio Mendelin section and ODP Site 976: epifauna (0–0.7 cm BSWI), infauna (>0.7 cm BSWI) and deep infauna (>3 cm BSWI). Oxic Suboxic (Groups A and B) Dysoxic Cibicidoides brady Amphicoryna spp. Bolivina dilatata Cibicides dutemplei Anomalinoides helicinus Bolivina reticulata Cibicidoides kullenbergi Asterigerina planorbis Bolivina seminuda Cibicides lobatulus Bulimina elongata Bolivina spathulata Cibicidoides pachyderma Bulimina striata Chilostomella spp. Cibicidoides pseudoungerianus Cancris oblongus Fursenkoina acuta Cibicidoides ungerianus Dentalina spp. Globobulimina spp. Sphaeroidina bulloides Epistominella trinacria Globobulimina turgida Globocassidulina subglobosa Stainforthia fusiformis Gyroidina altiformis Gyroidina soldanii Karreriella bradyi Lenticulina spp. Martinotiella communis Melonis barleeanus Melonis pompilioides Melonis soldanii Nodosaria spp. Nonion fabum Oridorsalis stellatus Oridorsalis umbonatus Planulina ariminensis Pullenia bulloides Pullenia quinqueloba Sigmoilopsis schlumbergeri Textularia calva Trifarina bradyi Uvigerina peregrina Uvigerina rutila F. Bulian et al. Marine Micropaleontology 176 (2022) 102160 7 (caption on next page) F. Bulian et al. Marine Micropaleontology 176 (2022) 102160 8 (50–100). The BFOI proved to accurately reproduce long-term dissolved oxygen changes in the Marmara Sea (similar bottom-water conditions as in the Mediterranean) and therefore it can probably be used to evaluate oxygen levels in the Mediterranean too (Kaminski, 2012).The interpretation of the BFOI is prone to bias when very low-diversity assemblages are studied and consequently, this estimate has not been applied to samples with a Shannon index ≤1 (Kaiho, 1994). Moreover, studies on recent benthic foraminifera suggest that quantitative reconstruction of oxygen levels higher than 1 ml/l is probably not feasible (Murray, 2001; Jorissen et al., 2007 and references therein). The BFOI has been calculated following the formula (Kaiho, 1994): BFOI =(O (O+D))*100 where O and D (with O >0) are the numbers of oxic and dysoxic indicator species, respectively. When O =0 and I >0, the equation: BFOI =(( I (I+D))−1)*100 is used instead, where I is the sum of suboxic indicators. The suboxic indicators in group C are excluded from the formula (Kaiho, 1994) although the species are used for the paleoenvironmental interpretations. It is important to mention that oxygen estimates based on any indices can be subject to problems related to, for example, existence of an ecosystem oxygen gradient, coexistence of species with different microhabitat preferences, time-averaged samples or the interplay between export productivity and oxygenation in defining the preferred microhabitat (Jorissen et al., 2007). 3.5. Stable isotope measurements Epifaunal taxa such as Cibicides spp. and Cibicidoides spp. were observed to secrete calcite close to equilibrium with the ambient sea water, displaying minor vital and minimal ontogenetic effects (Theodor et al., 2016a; Theodor et al., 2016b; J¨ ohnck et al., 2021 and references therein), and therefore 2 to 10 specimens (based on the availability) of Cibicidoides pachyderma were picked from ODP Site 976 (core 61) samples for stable isotope analyses. The actual contact between Miocene and Pliocene sediments was not recovered at this site. Nonetheless, our new stable isotope data from three samples of 62×-CC section confirms their Messinian age (see Section 5.4) and consequently attribute, paired with stratigraphic markers (Bulian et al., 2021), an early Pliocene age for core 61. The analysis was only possible until 571.18 m because specimens from the younger part of the sequence were affected by secondary calcite precipitation on the surface and within the aperture, which persisted even after ultrasonic cleaning. For the Rio Mendelin section, other species were chosen given the absence of C. pachyderma. Epistominella trinacria was picked in the basal two samples because the assemblage in the lowermost level is monospecific and in the next level nearly so (88.6%). In the next samples Cibicides dutemplei was picked, while starting from 1 m and going upwards, Cibicidoides ungerianus was chosen. More than one species was picked where possible, to obtain intraspecific correction factors. All measurements were transformed to C. ungerianus. This species has been reported to yield comparable values as C. pachyderma (Kaboth et al., 2017). Based on two paired measurements the E. trinacria values were corrected by adopting a value equal to the average offset from the C. dutemplei values (−0.21 for δ 13 C and 1.46 for δ 18 O values). The C. dutemplei record and the corrected E. trinacria data points were then all adjusted using the average offset established from six paired measurements between C. dutemplei and C. ungerianus (0.33 for δ 13 C and −0.03 for δ 18 O values). Samples from the Barranco del Negro section were not picked for stable isotope analyses because part of the foraminifera are not in situ. The stable oxygen and carbon isotope analyses were performed with a Finnigan MAT 253 mass spectrometer connected to a Kiel IV carbonate preparation device at the Christian-Albrecht University in Kiel (Germany). Sample reaction was induced by individual acid addition (99% H 3 PO 4 at 75 ◦C) under vacuum. The evolved carbon dioxide was analysed eight times for each individual sample. As documented by the performance of international [NBS19: +1.95 ‰ VPDB ( 13 C), −2.20 ‰ VPDB ( 18 O); IAEA-603: +2.46 ‰ VPDB ( 13 C), −2.37 ‰ VPDB ( 18 O)] and laboratory-internal carbonate standards [Hela1: +0.91 ‰ VPDB ( 13 C), +2.48 ‰ VPDB ( 18 O); HB1: -12.10 ‰ VPDB ( 13 C), −18.10 ‰ VPDB ( 18 O); SHK: +1.74 ‰ VPDB ( 13 C), −4.85 ‰ VPDB ( 18 O)], analytical precision of stable isotope analysis is better than ±0.08 ‰ for δ 18 O and better than ±0.05 ‰ for δ 13 C. The obtained values were calibrated relative to Vienna Pee Dee Belemnite (VPDB). 4. Results 4.1. Malaga Basin: Rio Mendelin section 4.1.1. Micropalaeontology and stable isotopes The lowermost 2 m of the Rio Mendelin section (Lago-Mare deposits) are barren of foraminifera and contain high abundances of ostracods, mainly Cyprideis sp. (Fig. 8). At the base of the Pliocene benthic foraminifera appear and remain present throughout the section. Considering the distribution and the good preservation of the specimens, which do not show any signs of transport or alteration, the foraminifera can be considered in situ. Optical microscope pictures of the species are shown in Plates 1 and 2. The first benthic foraminiferal species appearing in the Zanclean is Epistominella trinacria which characterises the lowermost 60 cm of the section (2.1–2.7 m) (Figs. 8, 9). This species disappears almost completely at the base of the grey layer when other species appear: Fursenkoina acuta, Globobulimina turgida, and Stainforthia fusiformis (Fig. 8) which are limited to this grey interval. At the same level, several other species appear which remain present throughout the record (Fig. 9). The most abundant at this level is Nonion fabum (~ 50%). Cluster analysis resulted in two main Clusters 1 and 2 (Fig. 10A and B, top panel). Cluster 2 is composed of two subclusters, Clusters 2.1 and 2.2, the first composed of a total of four subclusters (Fig. 10). These clusters group benthic foraminiferal species based on their distribution, enabling the distinction of three stratigraphic units (Units 1 to 3; Fig. 9). Cluster 1, defining Unit 1, dominates from 2 to 2.5 m in the section and is represented by only one species, E. trinacria. The following 0.5 m (2.5–3 m, in the grey layer), Unit 2, is dominated by F. acuta, G. turgida, B. dilatata and S. fusiformis forming Cluster 2.2. Overlapping with Cluster 2.2 but continuing until ~4 m, is Cluster 2.1.2 A. Highest abundances are reached by N. fabum (~50%), B. striata (~25%) and Lenticulina spp. (up to ~20%). The other species of Cluster 2.1.2 A show relative abundances ranging between ~5 and 12% (Figs. 9 and 10). At ~4 m, Cluster 2.1.2 B becomes dominant (Fig. 10B). Cibicides dutemplei, C. pseudoungerianus and C. ungerianus dominate the assemblage reaching abundances of ~15–16% (Figs. 9 and 10). Apart from Textularia calva (~10%), the remaining species are subordinate, and most do not reach 5% of abundance (Figs. 9 and 10). Cluster, 2.1.2C is composed of Plate 1. Optical microscope photos of the most abundant (>3%) benthic foraminifer species of Rio Mendelin section. 1 Textularia calva; 2 Bolivina seminuda; 3 Bolivina spathulata; 4 Bolivina dilatata; 5 Bolivina reticulata; 6 Stainforthia fusiformis; 7 a, b Trifarina bradyi; 8 a, b Fursenkoina acuta; 9 Bulimina striata var. mexicana; 10 a, b Bulimina striata; 11 Bulimina elongata; 12 Bulimina aculeata; 6; 13 Globobulimina turgida; 14 Nonion fabum; 15 a, b Asterigerina planorbis: a spiral view, b umbilical view; 16 Lenticulina sp. Scale bar =100 μ m. F. Bulian et al. Marine Micropaleontology 176 (2022) 102160 9 (caption on next page) F. Bulian et al. Marine Micropaleontology 176 (2022) 102160 16 Plate 6. Optical microscope photos of some of the most abundant (>3%) benthic foraminifer species identified at Site 976. 1 Planulina ariminensis; 2 Cancris oblongis: a spiral view, b umbilical view; 3 Cibicidoides ungerianus: a spiral view, b umbilical view; 4 Cibicidoides bradyi: a spiral view, b umbilical view; 5 Pullenia bulloides: a apertural view, b lateral view; 6 Pullenia quinqueloba; 7 Melonis soldanii; 8 Melonis barleeanus; 9 Melonis pompilioides; 10 Oridorsalis umbonatus: a spiral view, b umbilical view; 11 Gyroidina altiformis: a spiral view, b umbilical view. Scale bar =100 μ m. F. Bulian et al. Marine Micropaleontology 176 (2022) 102160 17 deposited immediately after the MPB we infer that at the Los Ranchos section no initial deepening after the MPB is recorded. 5.2.1.3. Zorreras section (Sorbas Basin). The Sorbas Basin was characterised by a continental environment during the latest Messinian stage (Roveri et al., 2018 and this work). The first Pliocene sediments deposited are grey sands rich in carbonate nodules and rare shallowwater benthic foraminifera, suggesting a very shallow, at times exposed environment similar to a paleosol. On top of this mostly emerged, continental facies, coastal sandy sediments were eventually deposited because of the rising sea level. The foraminiferal data are only qualitative; however, because of the absence of planktic foraminifera and the presence of shallow-water taxa like N. fabum and Elphidium spp. (Hayward et al., 2001; Murray, 2006; Milker, 2010; Tulbure et al., 2017) the Zanclean palaeodepth probably did not exceed 10–20 m. In most marginal basins, Zanclean sediments are overlying continental and/or shallow, brackish LM facies (Fortuin et al., 2000; GuerraMerch´ an et al., 2010; Caruso et al., 2020; Andreetto et al., 2021a and references therein; Andreetto et al., 2021b). In deeper marginal basins with age control based on planktic foraminifera, where the MPB is continuous and the MPl1 zone is complete, for instance at Eraclea Minoa (Brolsma, 1978; Sgarrella et al., 1997) and sections in Piemonte (Trenkwalder et al., 2008; Violanti et al., 2009) instantaneous refill to upper bathyal depth appears to have occurred immediately after the MSC. This implies a rapid deepening of at least 500 m at the base of the Zanclean. At Cava Serredi (Tuscany: Riforgiato et al., 2011) the refill seems to have been gradual; however, the actual base of the Zanclean may be missing since the MPB is reported to be continuous, but planktic foraminifera of MPl1 and MPl2 (Sphaeroidinellopsis spp., G. margaritae) are absent. For shallower marginal basins, where age control is sub-optimal or absent, the picture is less clear and erosional surfaces are often reported to be associated with the MPB. Despite the erosional surface in the Cuevas del Almanzora section (Vera Basin), the MPl1 zone is reported to be complete. Caruso et al. (2020) reconstruct a water depth of >250 m immediately after the MPB, implying that the erosional surface has removed part of the LM deposits. This may be related to base level variations during the latest Messinian, as suggested by Gargani and Rigollet (2007), Stoica et al. (2016) and Andreetto et al. (2021b). At the Los Ranchos section (Nijar Basin) however, a palaeodepth of 274 m immediately after the MPB (P´ erez-Asensio et al., 2021) is more likely caused by absence of the basal Pliocene since only G. margaritae and no Sphaeroidinellopsis spp. appear, and no benthic repopulation sequence is reported. In the Rio Mendelin section (Malaga Basin; this study) the contact between Messinian and Zanclean is gradual. A benthic repopulation sequence is present and few, but quite large G. margaritae occur above the grey layer. 5.2.2. Deep basin: ODP Site 976 The palaeo-water depth reconstructed for ODP Site 976 is bathyal (≥1000 m). A similar depth has been reported for the early Messinian at the same site (Bulian et al., 2022). Bathyal Zanclean palaeodepths have been reconstructed for the Eraclea Minoa and Capo Rossello sections in Sicily (Sgarrella et al., 1997; Barra et al., 1998: 600–800 m). In general, depth reconstructions for earliest Zanclean DSDP and ODP Sites have been in the order of upperto mesobathyal (1000 m and more). DSDP Leg 42A, Site 371 (South Balearic Basin) was estimated to be 1200–1400 m deep (Wright, 1978). For Leg 107, Sites 652–654 (Tyrrhenian Basin) and Leg 161, Sites 974 (Tyrrhenian Sea) and 975 (Balearic Basin), the depth of MPl1 was also estimated to be bathyal (McKenzie et al., 1990; Sprovieri and Hasegawa, 1990; Iaccarino et al., 1999a). Based on planktic foraminifer biostratigraphy, the MPl1 in Site 975 (Iaccarino et al., 1999a) is complete. This would imply that the Pliocene refill of this – and other - deep basins happened instantaneously on a geological time sale, or alternatively, the basins were not dry at that time (e.g., Marzocchi et al., 2016; Stoica et al., 2016). 5.3. Benthic environments and repopulation Lago-Mare deposits were either barren or containing brackish to fresh/continental faunas, including reworked fossils. Zanclean repopulation by benthic foraminifera shows similarities with repopulation following sapropel deposition. Jorissen (1999 and references therein) discerns general patterns when morphogroups are considered rather than individual species and defines morphogroups generally recognized in repopulation sequences: deep infauna tolerating extreme hypoxia in relatively stable environments (group A, e.g., Chilostomella and Globobulimina spp.); intermediate infauna, less tolerant to hypoxia (group B, elongated biand triserial taxa, e.g., bolivinids, buliminids, Cassidulinoides spp.); so-called ‘phytodetritus species’, small, biconvex and trochospiral taxa tolerating oxic to hypoxic, unstable environments and reproducing fast after seasonal input of fresh phytodetritus (group C, e. g., small Epistominella, Eponides and Gyroidina spp.). Depending on local environmental conditions, either group A or group C will precede group B in a repopulation event. 5.3.1. Marginal basins 5.3.1.1. Rio Mendelin section (Malaga Basin). In the Rio Mendelin section, the benthic foraminiferal assemblage immediately above the LM consists of monospecific Epistominella trinacria (Fig. 8; Plate 2). This species was described in Pliocene sapropels by Verhallen (1991) without further details; its occurrence is explained by analogy with other Epistominella species (mentioned in Group C cf. Jorissen, 1999). In recent environments Epistominella species have been associated with seasonal (spring/early summer episodes) deposition of fresh phytodetritus in oxic to sub-oxic (but not anoxic), otherwise oligotrophic environments (E. exigua: Gooday, 1988; Gooday, 1993; Smart et al., 1994; E. vitrea: Gooday and Hughes, 2002; Platon et al., 2005; Langezaal et al., 2006). In the fossil record Epistominella species were observed in Zanclean repopulation sequences (e.g., Sgarrella et al., 1997 (Sicily); Iaccarino et al., 1999b (Tyrrhenian Sea); Rouchy et al., 2001 (Cyprus); Aguirre et al., 2006 (Nijar Basin, Spain); Cipollari et al., 2013 (Adana Basin, Turkey); Kontakiotis et al., 2016 (Zakynthos, Greece); Caruso et al., 2020 (Vera Basin, Spain)) Their occurrence suggests that these environments were oxic to sub-oxic, oligotrophic but prone to seasonal variations in organic matter input. The abundance of E. trinacria declines in the Rio Mendelin section after the lowermost, monospecific sample and the abundance of Nonion fabum increases to nearly 50% at the top of Unit 1 (Figs. 8, 9). Nonion spp. are versatile species, typical of high organic carbon content, phytodetritus and low oxygen conditions (Fontanier et al., 2002; Diz et al., 2004; Mendes et al., 2004; Mojtahid et al., 2006). Subordinate Stainforthia fusiformis, Fursenkoina acuta and Globobulimina turgida (Group A cf. Jorissen, 1999; Cluster 2.2) suggest a transition from a relatively oxic environment characterised by E. trinacria towards a more differentiated but more hypoxic environment towards Unit 2, the grey layer. At the same time, Bulimina aculeata, B. elongata, and Bolivina dilatata appear, taxa less tolerant to hypoxia (Group B cf. Jorissen, 1999). Towards the top of Unit 2, B. seminuda and Oridorsalis stellatus increase in abundance (cluster 2.1.2A, Fig. 10), suggesting continued repopulation in more oxic and less stable conditions towards the yellowish strata of Unit 3. This repopulation sequence is comparable to the one identified after Quaternary sapropels (S1, S5, S6) (Jorissen, 1999). Unit 3 reflects a normal marine assemblage under relatively stable, oxic conditions: the diversity (H) increases to 3, the BFOI, minimal in the grey layer supporting hypoxia, increases to 80% in Unit 3 (Fig. 10) and the hypoxia-tolerant species of cluster 2 decrease in abundance or disappear. In the Rio Mendelin section, the Cluster 2.1.2 A assemblage prevails in the record until 4 m (Figs. 9 and 10), when Cluster 2.1.2 B F. Bulian et al. Marine Micropaleontology 176 (2022) 102160 18 (C. dutemplei, C. pseudoungerianus, C. ungerianus and Textularia calva) becomes dominant. Although Textularia spp. are tolerant to foodenriched and oxygen-deficient conditions (Naeher et al., 2012), the more common Cibicides/Cibicidoides, epifaunal to shallow infaunal and with low tolerance to oxygen deficiency (Van der Zwaan, 1982; Jorissen et al., 2007), suggest relatively well‑oxygenated waters and moderate productivity up to 6.4 m, in agreement with the BFOI, with values close to high-oxic environments (BFOI =~90%; Fig. 10). Beginning in Unit 3, the benthic foraminiferal assemblages show a clear alternation between Cluster 2.1.2 A and Cluster 2.1.2 B (Fig. 10), hence between a mostly infaunal assemblage (buliminids, bolivinids, N. fabum, Lenticulina spp.) associated with high productivity and reduced oxygenation, and an epifaunal one, characterised by Cibicides/Cibicidoides species, indicating more oxic conditions. We suggest these cycles reflect an alternation of phases of lower versus higher nutrient input and riverine discharge associated with cyclical changes in the freshwater budget (see 5.4). The benthic and planktic δ 18 O (Fig. 10 and Supp. 3) show heaviest values concomitant with Cluster 2.1.2 B peaks, suggesting astronomical control over the cyclicity. In summary, in the Rio Mendelin section monospecific E. trinacria marks the start of benthic repopulation following the incursion of marine waters. The grey clays (Unit 2) more likely reflect water stratification related to further sea-level rise than influx of organic matter. Continued sea-level rise ameliorated circulation and ventilation, eventually supporting a diverse open marine benthic foraminiferal assemblage. Comparable repopulation sequences were described elsewhere in the Mediterranean (see above). 5.3.1.2. Barranco del Negro section (Nijar Basin). The calcarenites above the black and grey layer in the Barranco del Negro section suggest a nearshore, high-energy environment (see Section 5.3.1), but the reconstruction of palaeoenvironment and repopulation by benthic foraminifera is complicated by reworking (see 4.2; Table 3). The number of benthic foraminifera per gram of sediment (BF/g; Fig. 12), which is very low in the black and grey layer, rises steadily suggesting that the number of reworked foraminifera is relatively low, and more indigenous benthic foraminifera start occurring above the black-grey layer. In the first sample above the grey level Chilostomella and Globobulimina species appear which, if in situ, may be part of a repopulation sequence in an initially stable, hypoxic-anoxic environment (Group A of Jorissen, 1999). The benthic foraminiferal species expected to thrive in a deltaic or shoreface palaeoenvironment is C. lobatulus (Hald and Korsun, 1997). Cibicides lobatulus is absent in the sample with Chilostomella and Globobulimina spp. and its abundance then rises (~30%), suggesting that it could be in situ. Not expected to thrive in deltaic or shoreface environments are C. kullenbergi, P. ariminensis, U. rutila, C. pseudoungerianus, C. ungerianus and S. bulloides; this would imply that together with the Miocene planktic foraminifera, these outer neritic to bathyal benthic foraminifera are reworked. Assuming that other species (e.g., S. deperdita, Elphidium spp., P. bulloides, P. quinqueloba, Lenticulina spp., Cassidulina spp.) may be in situ, this benthic foraminifer assemblage is in agreement with shallow, high-energy environments. Several opportunistic taxa feeding on fresh phytodetritus indicate episodes of intermittent, seasonal supply of organic matter. Among these are Valvulineria spp. including V. bradyana (Amorosi et al., 2013; Goineau et al., 2015 and references therein). A maximum abundance of 30% E. trinacia occurs at ±1.5 m and is here apparently not associated with repopulation. Apart from C. refulgens and C. lobatulus, living attached to hard substrates, the taxa assumed to be in situ are epifaunal and shallow infaunal and indicate a generally more oligoto mesotrophic, hypoxic environment towards the top of the sampled section (e.g., Jorissen, 1987; Kaiho, 1999; Murray, 2006). In such an environment, specimens of the ostracod Cyprideis sp., usually regarded as a brackish species (Benson, 1978), are also considered reworked. 5.3.1.3. Zorreras section (Sorbas Basin). In the Zorreras section, where the upper Messinian is continental, the lowermost Pliocene sediments consist of clastic sands with scarce low-diverse benthic foraminifera and no repopulation sequence is found. The first species appearing in the grey layer is monospecific N. fabum, included in the shallow infauna group, which is in the Rio Mendelin section the dominant species immediately following the level with monospecific E. trinacria. Here, the presence of numerous calcareous nodules, together with N. fabum known to tolerate shallow inner shelf environments and eutrophic conditions (Fontanier et al., 2002; Murray, 2006; Duchemin et al., 2008), indicate that the palaeoenvironment was not permanently submerged, but marine water could have intermittently entered the basin through a local geological barrier. While N. fabum dominates the entire section, at 1.1 and 1.2 m above the base of the Pliocene some Elphidium spp. and rare C. lobatulus appear. Together this suggests a gradual ingression of marine waters and a nearshore marine environment which was permanently established after the transgressive pulse that deposited the 5-cm thick bivalve-rich horizon. Fig. 7. Early Pliocene core sections from ODP Site 976 with corresponding core photographs. The black dots represent the analysed samples. F. Bulian et al. Marine Micropaleontology 176 (2022) 102160 19 5.3.2. Deep basin: ODP Site 976 (Alboran Basin) The lowermost Pliocene Unit 1 at ODP Site 976 may not cover the MPB because there is an interval of non-recovery in core 62, right below the first Pliocene sediments (Fig. 7) and a black or grey layer is absent. However, a relatively diverse planktic assemblage is present from the base of the studied interval (core 61×-CC, 573.5 mbsf: Fig. 7), with G. margaritae and Sphaeroidinellopsis spp. appearing at 572.6 m and 572.3 m respectively (Supp. 1). In addition, an interval with high abundances of U. peregrina (Fig. 13) was also identified in other deep sites (from cycle 6 to 10 in Sicily, and from cycle 4 to 12 in the Tyrrhenian Sea). It is referred to as the Uvigerina pygmea-U. peregrina event (Sgarrella et al., 1997; Barra et al., 1998; Iaccarino et al., 1999b) and has Fig. 8. Base of the Rio Mendelin section. From left to right: the photo of the Miocene-Pliocene boundary (red line), the stratigraphic section and the first benthic foraminiferal species that repopulate early Pliocene environments. In addition, a photo of the predominant late Messinian ostracod species Cyprideis sp. is shown. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Fig. 9. Variations in relative abundances of major benthic foraminiferal taxa of the Rio Mendelin section. Colour bands correspond to foraminiferal assemblages characterizing the sample clustering in Fig. 10 which was the basis for the definition of micropaleontological units. F. Bulian et al. Marine Micropaleontology 176 (2022) 102160 20 been used to confirm the early Pliocene age of the sediments. A benthic repopulation sequence comparable to the Rio Mendelin section is not developed at Site 976. Phytodetritus species (epistominellids, G. subglobosa, small Gyroidina spp.) and deep infaunal taxa (e. g., Chilostomella and Globobulimina spp.; Jorissen, 1999) each amount to not >5%. Instead, Unit 1 is characterised by a single peak of B. aculeata (>35%), with C. dutemplei (6%) forming Cluster 1 (Figs. 13 and 14). Bulimina aculeata is one of the most opportunistic taxa in the Mediterranean (De Rijk et al., 2000), feeding on fresh phytodetritus and tolerating suboxic environments (Schnitker, 1993; Gebhardt, 1999). Cibicides dutemplei is generally assumed to tolerate little oxygen deficiency, although Cibicides spp. have been observed in oxygen-deficient environments as well but they may be more sensitive to fresh (undegraded) food particles than to hypoxia (Jorissen et al., 2007). Considering that the abundance peak of B. aculeata co-occurs with stresstolerant species such as M. barleeanum, M. soldanii (Caralp, 1989; Koho et al., 2008), and Lenticulina spp. (Sen Gupta and Machain-Castillo, 1993; Kaiho, 1994) the basal Pliocene bottom-water environment of Site Fig. 10. A) Dendrogram resulting from the hierarchical clustering in the Rio Mendelin section. The clusters have been highlighted with colours. For explanation see text. B) from left to right: Clusters 1, 2.2, 2.1.2 A, 2.1.2 B and 2.1.2C plotted against stratigraphic position; Microhabitat distribution of the most abundant benthic foraminifer species; The BFOI estimation (Kaiho, 1994); Shannon diversity index; % of planktic foraminifera; Benthic δ 13 C and δ 18 O isotopic record (C. ungerianus); The column to the right shows the micropalaeontologically defined stratigraphic units based on the cluster analysis. F. Bulian et al. Marine Micropaleontology 176 (2022) 102160 21 976 was probably characterised by seasonal organic carbon supply and reduced oxygen levels, in agreement with the BFOI values indicative of low-oxic environments (~20, Fig. 14). Unit 2 is still dominated by infaunal, low-oxic taxa. The appearance of the benthic foraminiferal assemblage of Unit 3 (Figs. 13 and 14) and the decline in abundances of taxa present in Units 1 and 2 reflects a change in environmental conditions. Taxa increasing in abundance include G. subglobosa (up to 40%) and later U. peregrina (up to 30%; Cluster 2.2.1; Fig. 14). Although shallow infaunal taxa are still dominating the assemblage and the diversity (H index) does not increase, the BFOI rises from a mean of 40 to a mean of 80 (Fig. 14), suggesting the presence of well oxygenated waters. Dominance of G. subglobosa (in size fractions >150 μ m and 63–150 μ m) has been recorded in the oxygen minimum zone (OMZ) of the Sulu Sea, under high fluxes of organic matter (e.g., Miao and Thunell, 1993), but G. subglobosa is also a phytodetritus feeder in the generally oxic and oligotrophic environments of the Porcupine Abyssal Plain where the spring bloom delivers pulsed supply of fresh organic matter (Gooday, 1993). It is also present in moderately dysaerobic environments characterised by efficient preservation of organic matter (Loub` ere et al., 1988) as well as in early Pliocene recolonization successions (Barra et al., 1998). Uvigerina peregrina is a shallow infaunal species associated with upwelling and labile (fresh) organic material, and generally not associated with low-oxygen conditions (Morigi et al., 2001; Fontanier et al., 2002; Koho et al., 2008; Schmiedl et al., 2010). Other species in this assemblage include S. bulloides and B. striata, indicating upwelling and an elevated organic carbon content (Sen Gupta and Machain-Castillo, 1993; Licari and Mackensen, 2005). Cibicidoides pachyderma and P. ariminensis generally live in well-oxygenated environments (Schmiedl et al., 2000; Schmiedl et al., 2003), although they also tolerate oxygen deficiency and sustained organic matter fluxes (Bernhard and Gupta, 1999). The benthic foraminiferal assemblage reflects better circulation and installation of an upwelling regime. The change towards higher and more stable BFOI (~80–95) suggest a progressive increase in oxygen levels. From 569 mbsf, the Cluster 2.2.2 assemblage returns (Fig. 14 and Supp. 2), with a maximum abundance of M. barleeanus (25%) at the base and with additional presence of C. oblongus and A. helicinus (Cluster 2.1). Anomalinoides helicinus indicates increased productivity and/or preservation of organic matter (Barra et al., 1998) and C. oblongus can also be related to high organic flux and lower oxygen (Murray, 2006). Considering the high BFOI values (~ 50–60), bottom water oxygen levels must have been quite high, even if slightly decreasing from this point onward towards low oxic conditions. Despite differences, repopulation of the early Pliocene bottom-water environments at deep Mediterranean sites has elements in common. The earliest Pliocene is characterised by the presence of phytodetritus feeders at Site 975 and Eraclea Minoa (e.g., Eponides pusillus and Epistominella exigua) and contains shallow infaunal species at Sites 976 and 975 (e.g., Bulimina spp. and Bolivina spp.). All sites are characterised in the early Zanclean by reduced oxygen and high organic flux. After this first interval, all sites are characterised by a gradual amelioration of circulation leading to better oxygenated, and in Eraclea Minoa to fully Fig. 11. Calculated palaeodepths for the Rio Mendelin section. The grey line is the palaeodepth estimated by gradient analysis (Hohenegger, 2005; Hohenegger et al., 2008). The blue line is the palaeodepth estimated by using the %P (Van der Zwaan et al., 1990). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Fig. 12. The Barranco del Negro section with from left to right: the Shannon diversity index, the % of planktic foraminifera, benthic foraminifer per gram and the variations in relative abundances of most abundant benthic foraminiferal taxa are shown. F. Bulian et al. Marine Micropaleontology 176 (2022) 102160 22 open marine conditions. 5.4. Efficiency of the early Pliocene Mediterranean – Atlantic connections In order to better understand Mediterranean-Atlantic connectivity changes at the onset of the Pliocene as well as salinity and temperature differences between the Mediterranean and Atlantic, the newly acquired early Pliocene benthic δ 13 C and δ 18 O isotopic records from Site 976 and the Rio Mendelin section have been compared with ODP Site 982 (Hodell et al., 2001; Drury et al., 2018) and IODP Site U1387 (Hern´ andez-Molina et al., 2013; Van Der Schee et al., 2016; Figs. 16 and 17). Site U1387 is located in the Gulf of Cadiz, on the Atlantic side of the Strait of Gibraltar at 559 m water depth (Hern´ andez-Molina et al., 2013), which is bathed today by Mediterranean Outflow Water (MOW). This water mass is composed mainly of Levantine Intermediate Water and a small component of West Mediterranean Deep Water (WMDW) averaging 13 ◦C and 38.4 psu (Bryden and Stommel, 1984; Bryden et al., 1994; Hernandez-Molina et al., 2014). Site 982 is located in the north Atlantic at a water depth of 1135.3 m (Jansen et al., 1996) and bathed currently by the cold (3–8 ◦C) and less saline (34.95–35.2 ‰) North Atlantic Deep Water (NADW; Ochoa and Bray, 1991; Venz et al., 1999; Hern´ andez-Molina et al., 2016). In the areas of deep-water formation, deep-water δ 13 C is controlled by the δ 13 C of the surface. During photosynthesis, the phytoplankton preferentially extract 13 C-depleted CO 2 , enriching surface waters with the heavy 13 C isotope. In contrast, the degradation of sinking organic matter during deep-water renewal releases 13 C depleted CO 2 leading to a progressive decrease of δ 13 C in deep-water masses. This decrease depends on the quantity of CO 2 added by remineralization of organic matter, and so, on the deep-water residence time and bottom-water ventilation (Laube-Lenfant and Pierre, 1994; Pierre, 1999). The early Pliocene benthic δ 13 C values at Site 976 are initially analogous to the ones registered at the Gulf of Cadiz (Fig. 16; Site 1387) and become increasingly heavier up core when they reach the Atlantic curve (Site 982), which reflects the well ventilated high δ 13 C of the NADW. In contrast, the more depleted Messinian δ 13 C values from Site 976 (average value around −1 per mil) reflect the longer bottom-water residence time of Mediterranean caused by the reduced Mediterranean – Atlantic exchange (Bulian et al., 2022). In the Pliocene, with the reestablishment of an efficient Mediterranean – Atlantic water exchange, Site 976 benthic δ 13 C rises to Atlantic values. Nonetheless, benthic species present at the base of the Pliocene suggest both reduced oxygen levels and high organic carbon content which is not visible from the benthic δ 13 C. This discrepancy can be explained by assuming that the high bottom-water renewal rate and DIC (Dissolved Inorganic Carbon) characterised by high δ 13 C was not large enough to dramatically reduce the δ 13 C at the seafloor, despite high remineralisation rates. Early Pliocene benthic δ 13 C values from the marginal Rio Mendelin section are generally lower than Atlantic and Mediterranean values showing a ~ 2 ‰ offset. This suggests that while the deep Mediterranean was efficiently connected with the Atlantic, this marginal basin could have still been isolated from the Mediterranean. The δ 18 O values for the early Pliocene in the marginal Rio Mendelin section and deep ODP Site 976 are very similar and most probably reflect a comparable temperature and salinity (Fig. 17). At the same time, lighter Mediterranean, and heavier Gulf of Cadiz (Site 1387) values show an offset of almost 2 ‰, which is the opposite situation in comparison with the present one (Fig. 17). As a reference for the latest Holocene data, we used the Cibicidoides pachyderma benthic δ 18 O values recorded in the Alboran Sea (P´ erez-Asensio et al., 2020) and Gulf of Cadiz Site 1387 (Singh et al., 2015). To these data we applied a correction of −0.25 ‰ accounting for the difference in ice volume between the early Pliocene and the late Holocene (Raymo et al., 2018). From this comparison it emerged that the present day δ 18 O difference between IODP site U1387 in the Gulf of Cadiz and ODP site 976 in the Alboran Sea is −0.75 ‰ (Fig. 17, dashed lines). This difference is probably due to the higher δ 18 O of Mediterranean waters due to its negative water budget and higher salinities compared to those of the Atlantic. However, in the early Pliocene we observe an opposite benthic δ 18 O offset between the Gulf of Cadiz and Alboran, which was ~ +1.75 ‰. If the large positive Pliocene offset was due to temperature alone, there would have been an 8 ◦C difference between Mediterranean and Atlantic deep waters on both sides of the Strait (0.23 ‰ =1 ◦C), while this difference today is only 1–3 ◦C (MEDARGroup, 2002). This suggests that at the depth of Site U1387 the thermal influence of the MOW was low during the early Pliocene, which could be explained if the flow of warm Mediterranean waters to the Atlantic was lower than today. A lower flow of heat from the Mediterranean to the Atlantic would have increased the temperature difference between the two basins and consequently the oxygen isotope difference. Another scenario that could justify such Pliocene offset between the benthic δ 18 O in the Alboran Sea and the Gulf of Cadiz is that Mediterranean seawater salinity and δ 18 O Fig. 13. ODP Site 976: Variations in relative abundances of major benthic foraminiferal taxa. Colour bands correspond to foraminiferal assemblages associated with the clusters of the dendrogram (Fig. 14) and Units 1–4. F. Bulian et al. Marine Micropaleontology 176 (2022) 102160 23 were lower than today and/or that the Strait was much wider than today. This would imply that in the earliest Pliocene the Mediterranean water budget would have been less negative than today, or even positive. Lower salinities agree with the latest Messinian scenario, where the Mediterranean water would contain a strong Paratethyan signal which was partly maintained through the early Pliocene as well. A much wider strait in the early Pliocene compared to the present would have exposed the Mediterranean water for a shorter time to a negative water budget, resulting in lower δ 18 O and salinities of Mediterranean water. Even more extreme is the offset between the Mediterranean benthic δ 18 O and Atlantic Site 982 record. This offset is probably due to the difference in temperature between these two water masses. Today, the Mediterranean deep-water temperature is in the order of 13 ◦C, while deep water temperature at the location of site 982 is close to 4–5 ◦C (MEDARGroup, 2002). This different water temperature would result in a δ 18 O offset of 2 ‰, which is still lower than the observed δ 18 O difference in the early Pliocene between site 976 and 982. This could be explained again by the occurrence of lighter δ 18 O waters in the deep Mediterranean probably due to a less negative water budget during the early Pliocene. Fig. 14. ODP Site 976: A) Dendrogram resulting from the hierarchical clustering that divided the benthic foraminifer species in two main species clusters (Cluster 1, Cluster 2) and two subclusters (Clusters 2.1 and 2.2). Cluster 2.2 in turn branches to two subordinate branches as well (Cluster 2.2.1 and 2.2.2). The used clusters have been highlighted with specific colours. B) From left to right: Clusters 1, 2.1, 2.2.1 and 2.2.2 plotted against stratigraphic position; Microhabitat distribution of the most abundant benthic foraminifer species; The BFOI estimation (Kaiho, 1994); Shannon diversity index; % of planktic foraminifers; Benthic δ 13 C and δ 18 O isotopic record (C. pachyderma); The column of the right contains the micropaleontologically defined stratigraphic units based on the cluster analysis. F. Bulian et al. Marine Micropaleontology 176 (2022) 102160 24 6. Conclusions The studied upper Messinian–lower Pliocene sections in the Spanish basins reflect the sedimentological changes linked with the reestablishment of normal marine conditions after the MSC and give important information about the water level across the Miocene-Pliocene boundary. In the marginal sections (Rio Mendelin, Barranco del Negro, Zorreras), the earliest Pliocene sediments show as a dark layer, usually grey or black, often enriched in organic matter. This layer, identified in deep and marginal basins all over the Mediterranean, could imply water column stratification, and reduced bottom-water oxygen levels, which is in agreement with the benthic foraminiferal assemblages. Such conditions could develop in a scenario where the Atlantic inflow reaching the Mediterranean was more saline than the Mediterranean waters still under the influence of the Paratethys, causing Atlantic waters to sink and stratify the water column. The early Pliocene of ODP Site 976 located in the Alboran Basin does not show a dark layer, probably because of its proximity to the Strait of Gibraltar and the inflowing Atlantic waters that could have eroded the basal Pliocene layers. The analyses performed on benthic foraminiferal assemblages enable a more detailed reconstruction of the early Pliocene Mediterranean environments and reflect benthic foraminiferal repopulation of the sea floor. The first benthic faunas that appear are in line with reduced bottom-water oxygen levels and a stratified water column in agreement with the deposition of the dark layers at the MPB. During the early Pliocene, the sea-level rise re-established normal marine conditions in the Mediterranean as can be deduced from the benthic foraminiferal assemblages characterizing this interval. The benthic foraminiferal repopulation identified in the studied basins is comparable with other Mediterranean sections and cores and shows similarities with repopulation following sapropel deposition. The general repopulation trend shows a shift from stressed and unstable environments to benthic assemblages indicating an amelioration of the circulation and bottomwater oxygenation. In the studied marginal Spanish basins, the estimated palaeobathymetry for the early Pliocene was similar ranging between 50 and 150 m. In the Malaga Basin, the earliest Pliocene palaeodepth does not exceed 50 m, and eventually reaches values of 150 m. In the Nijar Basin, the presence of macrofossils and fossil traces suggests high-energy shallow environments just after the MPB, while towards the top of the sections deeper, less high-energy environments were probably established. These observations imply a progressive deepening of the Mediterranean marginal basins as the Mediterranean – Atlantic connectivity was becoming more efficient. At Site 976, the early Pliocene is characterised by a bathyal environment (>1000 m), which is within the range we find today. Although the water column may have been stratified and organic matter accumulating, the high δ 13 C values from Site 976 testify that Fig. 15. Calculated palaeodepths for ODP Site 976. The grey line is the palaeodepth estimated by gradient analysis (Hohenegger, 2005; Hohenegger et al., 2008). The blue line is the palaeodepth estimated by using the %P (Van der Zwaan et al., 1990). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Fig. 16. Comparison between benthic δ 13 C records of Site 976 and the Rio Mendelin section (this study) with Atlantic Ocean ODP Site 982 (Hodell et al., 2001; Drury et al., 2018) and IODP Site 1387 (Van Der Schee et al., 2016). The blue dashed line shows modern Mediterranean benthic δ 18 O values (P´ erez-Asensio et al., 2020). The map shows the locations of the sites. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) F. Bulian et al. Marine Micropaleontology 176 (2022) 102160 25 bottom-water renewal in the deep basins was efficient immediately after the reflooding and that an efficient basin-scale circulation was reestablished. The values are in fact almost 2‰ higher than the ones registered in concomitance with the first gateway restriction at 7.17 Ma. The significant offset between the Alboran Basin and Gulf of Cadiz benthic δ 18 O values for the early Pliocene can be related either to a temperature or salinity difference among the two basins. However, because of the opposite isotopic offset between the benthic δ 18 O in the early Pliocene and late Holocene we believe that the Pliocene water budget was less negative than today, rendering the difference in salinity between the two basins the prevailing factor determining the δ 18 O. A less saline early Pliocene Mediterranean is in line with continuation of the Paratethyan influence present in the latest Messinian, and consequently with stratification of the water column once the Atlantic waters invaded the basin. CRediT authorship contribution statement F. Bulian: Conceptualization, Investigation, Methodology, Data curation, Writing – original draft. T.J. Kouwenhoven: Methodology, Data curation, Supervision, Writing – review & editing. N. Andersen: Formal analysis, Writing - review & editing. W. Krijgsman: Resources, Conceptualization, Supervision, Writing – review & editing. F.J. Sierro: Resources, Conceptualization, Supervision, Writing – review & editing. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments The authors appreciate the work of Jose Ignacio Martin Cruz in sample processing and preparation. Robert Speijer is thanked for his input when classifying benthic foraminiferal species and Federico Andreetto for his advice regarding ostracod identification. We also appreciate the useful information on outcrop location provided by Frits Hilgen. Moreover, all fellow ESRs and supervisors from the SALTGIANT project are thanked for their valuable suggestions and discussions. A part of the samples used in this research were collected through ODP Expedition 161 aboard the Joides Resolution. This research has received funding from the European Union's Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement n◦ 765256 SALTGIANT. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.marmicro.2022.102160. References Aguirre, J., 1998. El Plioceno del SE de la Península Ib´ erica (provincia de Almería). Síntesis estratigr´ afica, sedimentaria, bioestratigr´ afica y paleogeogr´ afica. Rev. Soc. Geol. Esp. 11 (3), 297–316. Aguirre, J., S´ anchez-Almazo, I.M., 2004. The Messinian post-evaporitic deposits of the Gafares area (Almerıa-Nıjar basin, SE Spain). A new view of the “Lago-Mare” facies. Sediment. Geol. 168 (1–2), 71–95. Aguirre, J., P´ erez-Mu˜ noz, A., Sanchez-Almazo, I., 2006. Benthic foraminifer assemblages in the lower Pliocene deposits of the Almeria-Níjar Basin (SE Spain). Rev. Esp. Micropaleontol. 38 (2), 411–428. Alve, E., 2003. A common opportunistic foraminiferal species as an indicator of rapidly changing conditions in a range of environments. Estuar. Coast. Shelf Sci. 57 (3), 501–514. Amorosi, A., Rossi, V., Vella, C., 2013. Stepwise post-glacial transgression in the Rhˆ one Delta area as revealed by high-resolution core data. Palaeogeogr. Palaeoclimatol. Palaeoecol. 374, 314–326. Andreetto, F., Aloisi, G., Raad, F., Heida, H., Flecker, R., Agiadi, K., Lofi, J., Blondel, S., Bulian, F., Camerlenghi, A., 2021a. Freshening of the Mediterranean Salt Giant: controversies and certainties around the terminal (Upper Gypsum and Lago-Mare) phases of the Messinian Salinity Crisis. Earth Sci. Rev. 103577. Andreetto, F., Matsubara, K., Beets, C., Fortuin, A., Flecker, R., Krijgsman, W., 2021b. High Mediterranean water-level during the Lago-Mare phase of the Messinian Salinity Crisis: insights from the Sr isotope records of Spanish marginal basins (SE Spain). Palaeogeogr. Palaeoclimatol. Palaeoecol. 562, 110139. Anthony, E.J., Aagaard, T., 2020. The lower shoreface: Morphodynamics and sediment connectivity with the upper shoreface and beach. Earth Sci. Rev. 103334. Aufgebauer, A., McCann, T., 2011. Messinian to Pliocene transition in the deep part of the Sorbas Basin, SE Spain—A new description of the depositional environment during the Messinian Salinity Crisis.(With 8 figures and 1 table). Neues Jah. Geol. Palaontol. Abh. 259 (2), 177. Austin, W., Evans, J., 2000. NE Atlantic benthic foraminifera: modern distribution patterns and palaeoecological significance. J. Geol. Soc. 157 (3), 679–691. Fig. 17. Comparison between benthic δ 18 O records of Site 976 and the Rio Mendelin section (this study) with Atlantic Ocean ODP Site 982 (Drury et al., 2018) and IODP Site 1387 (Van Der Schee et al., 2016). The blue dashed line shows modern Mediterranean benthic δ 18 O values (P´ erez-Asensio et al., 2020). The orange dashed line refers to the modern Gulf of Cadiz values (Singh et al., 2015). Both are corrected with respect to the Raymo et al. (2018) global curve. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) F. Bulian et al.