THE MIDDLE EOCENE CLIMATIC OPTIMUM (MECO) IMPACT ON THE BENTHIC AND PLANKTIC FORAMINIFERAL RESILIENCE FROM A SHALLOW-WATER SEDIMENTARY RECORD

. We present here new quantitative analyses of planktic and benthic foraminifera to assess the impact of the Middle Eocene Climatic Optimum (MECO, ~40 Ma) on these biotic groups studied along a shallow-water succession rich in larger benthic foraminifera (Sealza, Liguria, NW Italy). The MECO is one of the major Eocene global warming events, characterized by ~4–6°C warming, shifts in the global carbon cycle, and rise in atmospheric p CO 2 . The Sealza succession is interpreted as the product of a drowning ramp influenced by tectonic activity and provides an exceptional chance to compare biotic variations in shallow-water assemblages with deep-water communities across the MECO. In the section, the MECO interval is tentatively constrained by stable isotope oxygen data and calcareous plankton biostratigraphy. The marked decline in abundance of the epifaunal benthic Cibicidoides across the lower-middle part of the MECO suggests a decrease in oxygenation at the seafloor. Further evidence of oxygen depletion is the increase in organic matter content (TOC) of the sediment and the presence of infaunal genera Uvigerina and Bolivina . The planktic foraminiferal assemblages record the MECO warming in the upper water column as the mixed-layer warm index genera Acarinina and Morozovelloides markedly increase in abundance. In the post-MECO interval, here poorly exposed, cooler conditions are indicated by the dominance of the cold-water index genus Subbotina . Remarkably, Acarinina decline in abundance in the upper MECO interval and never recover. The MECO perturbance permanently impacted the benthic and planktic communities at Sealza that exceeded the tipping point to move to a new regime, thus proving the fauna to be not resilient, but also not recording any extinctions.


IntroductIon
Marine calcifiers are threatened by the consequences of anthropogenic CO 2 emissions.The ongoing and future global climate changes such as global warming, ocean acidification, eutrophication and anoxia, are all phenomena which weaken their ecological resilience (e.g., Raven et al. 2005).
As highlighted in the most recent IPCC (Intergovernmental Panel on Climate Change 2018) reports, the evaluation of marine ecosystem health needs long-term observation data and studies on modern fauna only provide a limited time span.In contrast, the geological record allows us to evaluate long term responses to global climatic changes.
The Paleogene, being one of the most climatically dynamic periods of Earth's history, offers the opportunity to study the biotic effects.In particular, the early Paleogene is characterized by several short-term warming events (~50-200 kyr), known as hyperthermals (e.g., Zachos et al. 2001); the peak of temperature and pCO 2 were reached during the Early Eocene Climatic Optimum (EECO ~53-49 Ma).A long-term cooling trend (~49 to ~34 Ma) followed the EECO, eventually leading to the establishment of a continental Antarctic icesheet by the early Oligocene (e.g., Coxall et al. 2005).This climatic transition was interrupted by a prominent, though transient global warming event: the Middle Eocene Climatic Optimum (MECO).During the MECO the δ 18 O values of marine carbonatic sediments and of benthic foraminifera shells declined by roughly 1‰ in over ~400 kyr, which is usually interpreted as 4-6 °C of global temperature rise, with a gradual onset and brief peak temperatures at ~40 Ma and followed by a rapid return to pre-event conditions (e.g., Bohaty & Zachos 2003;Jovane et al. 2007;Bohaty et al. 2009;Edgar et al. 2010;Luciani et al. 2010;Spofforth et al. 2010;Savian et al. 2013;Boscolo Galazzo et al. 2014;D'Onofrio et al. 2021).This event is particularly appealing to study as it records temperatures and pCO 2 (Boscolo Galazzo et al. 2014;Cramwinckel et al. 2019;Henehan et al. 2020) that Earth will reach in the next few centuries if anthropogenic emissions do not stop (IPCC, 2018).Several characteristics, such as a duration longer than the Eocene hyperthermals, the absence of a clear trigger mechanism, and the lack of a globally coherent negative δ 13 C excursion in marine carbonates, make the MECO one of the most enigmatic events in the Cenozoic, dubbed a middle Eocene "carbon cycle conundrum" (Sluijs et al. 2013).
Even though the MECO is attracting high attention from the scientific community, studies on the paleoenvironmental changes and paleobiotic repercussions across the MECO are still limited.To improve the understanding of the impact of the MECO on the marine biotic communities and their resilience to global warming, here we present the analysis of the Sealza shallow-water succession outcropping in Liguria (NW Italy) (Fig. 1) that contains a very high diversity of both small benthic and planktic foraminifera, calcareous nannofossils, and larger benthic foraminifera (LBF) (Briguglio et al. 2023), thus making it a perfect study site to detect differential resilience among all taxa.Specifically, a multi-proxy study across the section is presented, including data from foraminiferal quantitative analysis, stable isotopes, and total organic carbon (TOC) to better understand the impact of the MECO on the taxa analyzed and on the environment.The data here presented provide for the first time, an analysis on the resilience of both benthic and planktic foraminifera in a shallow-water setting, consistently within the photic zone where mixed-and autotrophic organisms thrived, whereas most of the known literature focuses on deep-water environments (e.g., Luciani et al. 2010;Edgar et al. 2013;Boscolo Galazzo et al. 2013, 2014;D'Onofrio et al. 2021).

GeoloGIcal settInG
the study area is located close to the border between Italy and France, and the investigated section is exposed along a winding gravel-road starting from the hamlet of Sealza (Ventimiglia municipality, Imperia Province, Liguria, NW Italy) (Fig. 1).

Field sampling
The Sealza sedimentary succession here analyzed, rich in LBF, lies in disconformity on deep-water upper-Cretaceous deposits of the Trucco Formation and is overlain by the deep-water Olivetta San Michele Silty Marl Formation (Briguglio et al. 2023).In this study we selected the interval from 58 m to 124.3 m where the lithology shows alternations between biocalcirudites to biocalcisiltites and marls, thus suggesting potential deepening of depositional setting that may include calcareous plankton.We collected from this interval seventeen samples (SE-21, 23, SE25-SE41), all processed for foraminiferal and calcareous nannofossil analyses.The numerical order of samples is not sequential along the sedimentary sequence due to different sampling phases.In addition, samples SE-34 and SE-40 are missing (Fig. 2).The sample SE-33 was taken at 205 m in the Olivetta San Michele Silty Marl Formation, above an interval of approximately 81 m fully covered by vegetation, where a very little rocky exposure was visible underneath the ruins of an abandoned house.
In addition, 159 powder samples, spaced ~40 cm, were collected, from the 51 m to 124.3 m interval for carbon and oxygen stable isotopes analysis, directly on the outcrop surface using a pressure driller.Before collecting the powder samples, each surface point was carefully cleaned, subsequently initial holes of at least 2 cm were performed.A funnel has been used to collect the powder, which was immediately stored in numbered plastic vials.Additionally, nineteen of the total 159 powder samples collected were analyzed for the amount and type of organic matter by Rock-Eval pyrolysis.

Foraminiferal extraction
Planktic and benthic foraminifera were extracted from the marly-limestones and limestones using two different techniques.The less lithified lithotypes were treated through neo-steramina, a surface-active agent with the following chemical composition: alkyldimethylbenzylammonium chloride diluted at 10%, following the procedure by D' Onofrio & Luciani (2020).The seventeen samples collected for this study were crushed (size 1-2 cm) using a press and immersed in a neo-steramina bath for the time needed to obtain disaggregation (2-3 days).After that, all samples were washed through a 63 μm-mesh sieve and successively dried at 50°C.This procedure was effective for eleven of the total seventeen samples.
and Zeiss) equipped with on-board camera, picked, and isolated in microslides.The taxonomic criteria adopted in this work to identify the benthic fora- In addition, the most significant taxa retrieved in this study have been imaged by Scanning Electron Microscope (SEM, EVO 40 Zeiss), at University of Ferrara.

Quantitative analysis and data treatment
The abundance of foraminifera was counted on a population of ~300 specimens (when possible) in the ≥63 μm size fraction of washed residues, expressed as percentage as well as absolute abundance standardized to 1g of dried sediment (Supplementary data 1).The absolute abundance of benthic and planktic foraminifera is expressed as BFN and PFN respectively.The planktic and benthic ratio (P/B) was calculated as P/(P+B) *100 and expressed as percentages.Additionally, biodiversity parameters such as dominance D' and the Shannon-Weaver (H') index, calculated as proxies for diversity and heterogeneity of the assemblages (e.g., Murray 2006), were determined using the PAST software (Hammer et al. 2001).

Calcareous nannofossils
Nine samples (SE25-SE33) from the Sealza section, the same collected for the foraminiferal analysis, were prepared from unprocessed material as smear slides and examined using a light microscope at ~1250X magnification.Due to the scarcity of calcareous nannofossil content in all studied samples, we applied the semi-quantitative counting method developed by Gardin & Monechi (1998) assessing the presence or absence of index species in an area of ~6 mm 2 (three vertical traverses).This method allows a larger area to be scanned when the calcareous nannofossil content is too low for a statistically sound quantitative count.In addition, a qualitative estimate of the abundance of all nannofossil taxa detected was performed as reported in Rio et al. (1990) and modified following the abundance code outlined below: D (dominant) = taxon with frequencies greater than the 50% of assemblage; A (abundant) = usually more than 5 specimens occurring per field; C (common) = 1-5 specimens per field; F (few) = 1 specimen per 1-10 fields; R (rare) = 1 specimen per more than 10 fields (Supplementary data 2).Taxonomic concepts follow those of Perch-Nielsen (1985) and Nannotax (Young et al. 2017).We adopted the zonations of Fornaciari et al. (2010) and Agnini et al. (2014).

Stable isotopes
A first batch of 86 powder samples (SEP1 to SEP86) were analyzed for carbon and oxygen stable isotopes at the Laboratory of Paleoclimatology and Isotopic Stratigraphy of the Department of Physics and Earth Sciences of the University of Ferrara, using isoFLOW (Elementar©) operating in continuous flow with a PrecisION Isotopic Ratio Mass Spectrometer (IRMS; Elementar©).For each analysis, approximately 150 µg of homogenous powder of sample was weighed and placed into vials, which were flushed with pure He to replace the atmosphere.In each vial the powder reacted with hot viscous water-free orthophosphoric acid for 3 h at 50 °C, triggering the release of CO 2 from the samples.The released CO 2 sample gas was extracted through a needle and transferred to the IRMS for the simultaneous analysis of C and O.The 13 C/ 12 C and 18 O/ 16 O isotopic ratios were expressed with the δ notation (in ‰ units) relative to V-PDB.The in-house MAQ-1 standard was used for the singlepoint calibration and two control standards (IAEA 603 and Carrara Marble) were measured to monitor the quality of the analysis.Analytical uncertainties (1σ) for the isotope analyses were in the order of ± 0.1‰ for both δ 13 C and δ 18 O.
The remaining 73 powdered samples (SEP87-SEP159) were analyzed for δ 13 C and δ 18 O in the stable isotope laboratory of University of Lausanne in Switzerland, with the additional measurements of selected planktic and benthic foraminiferal genera Subbotina and Cibicidoides, respectively (picked from the 11 washed residues).The analyses were done on a Gasbench II coupled to a Finnigan MAT Delta V mass spectrometer via reacting the samples with phosphoric acid at 70 °C (Spötl &Vennemann 2003).The generated CO 2 was introduced with a He-flow into the mass spectrometer.In-house Carrara Marble standards calibrated to NBS-19 were used to normalize the obtained data.The analytical precision for this method was better than ±0.1‰ standard deviation.

Rock-Eval analysis
The amount and type of organic matter of 19 powdered samples were determined by Rock-Eval pyrolysis at the Institute of Earth Sciences of the University of Lausanne, Switzerland, using Rock-Eval 6 following the procedure of Espitalié et al. (1985) and Behar et al. (2001).The total organic car-bon (TOC) content is expressed in weight (wt%), the hydrogen index (HI = S2/TOC × 100) in mg hydrocarbons per gram of TOC and the oxygen index (OI = S3/TOC × 100) in mg CO 2 per gram of TOC.The calibration was performed with the IFP 160000 standard (French Institute of Petroleum), and the precision was better than 0.1%.

Biostratigraphy
The stratigraphic column of the Sealza section with collected samples, isotopic records, and biozones is shown in figure 2. In this study, the biostratigraphy is based on planktic foraminifera and calcareous nannofossils; nevertheless, along the same section, previously published data on LBF are available in Briguglio et al. (2023).These authors identified several Nummulites species that correspond to the Shallow Benthic Zone 17 (SBZ17) (Serra-Kiel et al. 1998), that is in the lower Bartonian (Papazzoni et al. 2017).Specifically, the species N. puschi, N. perforatus, N. biarritzensis, and N. beaumonti are all restricted to SBZ17.The species N. cyrenaicus and N. biedai, also recorded from Sealza, were previously believed to occur only in SBZ18 (Serra-Kiel et al. 1998).However, these species have been observed by Papazzoni & Sirotti, (1995) in their "N.lyelli zone" (=SBZ17) from the Igualada (NE Spain) and Pradipaldo (NE Italy) sections, which allows excluding the occurrence of SBZ18 at Sealza.In our section, the absence of Heterostegina, which first appeared at the base of SBZ18 (Less et al. 2008;Less & Özcan 2012), supports the attribution of the entire larger foraminiferal assemblages to the lower Bartonian (SBZ17) in the studied succession.
As for planktic foraminifera, the occurrence of the species Orbulinoides beckmanni, even though represented by rare specimens, in the interval from 113.5 m to 120.5 m allowed us to identify the Total Range Zone E12 (Wade et al. 2011) that largely corresponds to the MECO interval (Sexton et al. 2006;Bohaty et al. 2009).
Calcareous nannofossils are generally poorly preserved, as expected from the rocks sampled.In the interval from 58 m to 69 m, the assemblages consist mainly of reworked Cretaceous taxa thus preventing detailed biostratigraphic attribution.The presence of very rare specimens of genus Reticulofe-nestra (1-2 specimens in samples SE-25 and SE-27; Supplementary data 2) allows us to ascribe them to the Eocene (e.g., Agnini et al. 2006;Agnini et al. 2014;Fig. 2).At 114 m (SE-29), the occurrence of Sphenolithus furcatolithoides, Cribrocentrum reticulatum, and Reticulofenestra umbilicus allows us to assign the sample to the upper part of Zone CNE14 of Agnini et al. (2014), which correlates to Zone MNP16A of Fornaciari et al. (2010).The absence of Sphenolithus furcatolithoides, the top occurrence of which marks the CNE14-CNE15 boundary, and the presence of a reliable specimen of Dictyococcites bisectus allow the attribution of sample SE-30 to Zone CNE 15 (=MNP16Bb) starting from 115.5 m.The single other productive sample at 205 m (SE-33) confirms, due to the concomitant presence of D. bisectus, few Sphenolithus spiniger and the absence of Sphenolithus obtusus, the attribution of this portion to the Zone CNE15 (Fig. 2; Supplementary data 2).

Benthic and planktic foraminifera
The preservation of benthic and planktic foraminiferal tests is generally poor but adequate to detect the diagnostic morphological features useful to correctly identify the species.The most significant taxa retrieved in this study are reported in Plate 1 and the raw quantitative data are shown in Supplementary data 1.Benthic foraminiferal specimens are generally better preserved than the planktic specimens; a total of 12 benthic foraminiferal genera were recognized (Fig. 3).The most abundant genus is Cibicidoides (Cibicidoides sp., C. lobatulus) that displays a marked abundance decrease of 2.59% from 113.5 m to 118 m (samples SE-21, SE-29 -31), followed by an increase of up to 42.1% in the upper part of the section.Another abundant genus is Anomalinoides (Anomalinoides sp.), which similarly to Cibicidoides, is very abundant in the lower part of the measured section (up to 43.1%) but records a decrease in the same interval, though less pronounced (mean value 14.7%).However, differently from Cibicidoides, the genus Anomalinoides did not recover in the upper part, becoming absent at the top of the section.The genus Heterolepa (Heterolepa sp., H. dutemplei) is very scarce in the lower part of the succession; it increases in abundance around 117 m, markedly decreases in sample SE-30, and finally increases in abundance up to 25.6% in the upper part of the section, from 113.5 m to 122 m.The genus Uvigerina (Uvigerina sp., U. eocaena) exhibits low abundance in the basal and uppermost portions of the section but becomes more abundant from 116 m to 121 m.The genus Bolivina is absent in most of the lower portion of the section, only sporadically occurring in samples SE-21, SE-29-31.This genus displays two abundance peaks of up to 30% at 116 m and at the top of the section.The genera Lagena and Quinqueloculina occur only in samples SE-21 and SE-29-30, constituting a minor component of the assemblages.In turn, Lenticulina (Lenticulina sp., L. rotulata), Melonis affinis, and Dentalina sp.do not show significant changes in abundance, and they are present approximately from the base up to the top of the interval analyzed.
The disaggregated samples contain planktic foraminiferal specimens, except for SE-27-28, and SE-23, which proved to be barren.However, where present, planktic foraminifera are scarce, as expected from a shallow-water succession.The identification of planktic foraminifera is more difficult in the lower part of the section due to the pronounced recrystallization of their tests, while specimens from the washed residues obtained from samples SE21-SE33 are better preserved.A total of 4 genera were identified from the 113.5 m to 120.5 m interval: Acarinina (A. praetopilensis, Acarinina sp.), Morozovelloides (M.bandyi), Subbotina (S. hagni, S. eocaena) and Turborotalia (T.cerroazulensis), together with the species Orbulinoides beckmanni (Fig. 4).The genus Subbotina shows the lowest abundances in the interval from 113.5 m to 122 m and reaches the highest abundance in SE-33 with a partial increase from 115 to 117 m.The genus Acarinina is very scarce in the lower part of the section, but its abundance significantly increases from 113.5 m to 122 m (6.98% on average).The genus Morozovelloides displays a similar trend: it increases from mean values in abundance of 2.90% to 6.38% in the same interval, but it is absent in the lower part of the section.The genus Turborotalia occurs in very low abundances, only in samples SE-29-30.The species Orbulinoides beckmanni is recorded in the samples SE-21 and SE-29-32.

Diversity and statistical data
The values of benthic and planktic foraminifera (BFN and PFN) and the P/B ratio are shown in figures 3 and 4. The P/B ratio varied from 0 to 35%; in particular, lower P/B ratio values occur in the lower and upper parts of the section (except for sample SE-33), while higher values are recorded in samples SE-21 and SE-29-32 (Fig. 4).The benthic foraminiferal Shannon-Weaver (H'b) index (1.23 on average) exhibits slightly high values in the interval from 113.5 m to 122 m, while Dominance D'b (0.35 on average) does not show changes, implying that no one genus dominates over the others (Fig. 3).The Shannon index (H'p) for planktic foraminifera (0.91 on average) exhibits lower values in the lower part of the section and increases from 113.5 m to 122 m as for H'b.Differently, the dominance D'p (0.44 on average) indicates that the genus Subbotina is dominant (Fig. 4).

Calcareous nannofossils
Data from calcareous nannofossils are ambiguous because the section is represented by rocks poorly suitable for calcareous nannofossil analysis and generally dominated by Cretaceous reworking.Pristine assemblages are rare to few, and only three samples (SE-29, SE-30, and SE-33) have a statistically significant content of calcareous nannofossils.Preservation ranges from poor to moderate, with variable degrees of overgrowth and recrystallization in most of the analyzed samples.Raw semi-quantitative data are shown in Supplementary data 2. From the 58 m to 69 m interval, just 1-2 specimens of the genus Reticulofenestra were found (samples SE-25 and SE-27).The pristine assemblages are dominated by the genera Reticulofenestra, Cyclicargolithus, Sphenolithus, and very small Dictyoccocites, while Cribrocentrum reticulatum is scarce to common.The genera Coccolithus and Ericsonia are rare, while the genus Discoaster is virtually absent.

Oxygen and carbon stable isotope trends
Results of bulk stable isotope analyses are shown in figure 2 (red curve -five-point movement average).The δ 18 O bulk values vary between -5.6‰ and -0.4‰.The lower part of the section (58 to 73 m) is characterized by δ 18 O bulk values around -3.4‰.Several minor shifts toward major/minor negative values are recorded in the following interval up to 80.68 m.The interval from 109 m to 112.5 m records high values (between -1‰ and -2‰) with minor oscillations ranging between 0‰ and -2‰.An abrupt negative shift up to -3.7‰ occurs at 112.6 m.Following this shift, δ 18 O bulk values remain between -3.0‰ and -4.4‰ in the 113 m to 122.5 m interval.Following this interval, a sharp increase in δ 18 O bulk is recorded showing values varying between -0.7‰ and -3.0‰.A single negative peak of -5.2‰ occur at the top of the section analyzed.
The δ 13 C bulk values vary between -4.9‰ and 0.2‰ (Fig. 2).The interval from the base of the section to 74 m records roughly constant values around -0.4‰.The δ 13 C bulk curve show a negative trend up to 80.68 m reaching a minimum of -4.9 ‰ at 79.5 m.Above the covered interval of 27.25 m δ 13 C bulk data are similar to those recorded in the lower portion with mean values of -0.4‰.Three negative shifts occur at 109.4 m (-1.2‰), at 120.1 m (-3.6‰) and at 122 m (-1.4‰).
Carbon and oxygen stable isotopes measured on Cibicidoides (blue curve in figure 2), and Subbotina (green curve in figure 2) are constrained to the interval from 113.5 m to 122 m that provides the best foraminiferal preservation.Oxygen isotope values obtained from the Cibicidoides tests vary between -4.2‰ and -2.7‰ and exhibit a negative trend, while the Subbotina values are most constantly comprised from -3.7 ‰ and -3.3 ‰.
The δ 13 C Cibicidoides and δ 13 C Subbotina curves reveal similar trends consisting of a negative shift up to -3.9‰ for Cibicidoides and up to -2.7‰ for Subbotina.These shifts precede the negative peak recorded by the δ 13 C bulk .

Organic matter
The total organic content (TOC) varies from 0.02% to 0.34% in the Sealza section.The TOC values increase in the marly limestone and decrease in the biocalcirudite (Fig. 5).The values of hydrogen index (HI) ranges between 63 mg and 367 mg HC/g TOC (Fig. 5) with the lowest values recorded in sam-ples belong to the biocalcirudites (from 109 to 110 m).The oxygen index values (OI) varies from 73 to 636 mg CO 2 /g TOC (Fig. 5) with maximum values in samples 109.00 m, 109.48 m, 111.08 m, and 120.14 m.

MECO constrains at the Sealza section
The identification of MECO at the Sealza section though stable isotopes is not as straightforward as in most deep-water settings, and this was somehow expected due to the shallow-water setting of the depositional environment, which is largely controlled by runoff and local tectonics besides the climatic perturbations.The available data show a marked negative Oxygen Isotopic Excursion (OIE) of ~1‰ shown by the δ 18 O bulk record across the interval from 112.5 m to 122 m that is followed by rapid recovery to pre-excursion values.Similar OIEs have been documented in deep-sea cores from the Indian, Atlantic, Pacific, and Southern Oceans, as well as from land-exposing marine successions in China, Turkey, Italy, United Kingdom, and North America (e.g., Bohaty & Zachos 2003;Jovane et al. 2007;Villa et al. 2008;Luciani et al. 2010;Spofforth et al. 2010;Dawber & Tripati 2011;Toffanin et al. 2011;Boscolo Galazzo et al. 2013;Methner et al. 2016;Giorgioni et al. 2019;Shi et al. 2019).The recorded OIE has been interpreted mainly as a global temperature signal suggesting a transient global warming of 4-6 °C in both surface and deep waters (Bohaty & Zachos 2003;Bohaty et al. 2009;Bijl et al. 2010).The δ 18 O record from Sealza shows more negative values with respect to those obtained from the oceanic sites (e.g., Bohaty et al. 2009) because the on-land successions may be affected by diagenetic or meteoric processes, which usually lower the δ 18 O values (e.g., Marshall 1992;Schrag et al. 1995;Frank et al. 1999).Additionally, the isotope shift largely, yet not perfectly, correlates with a sudden shift in lithology, which might indicate differential diagenetic pathways for the lithotypes (e.g., Giraldo Gómez et al. 2017), thus revealing different isotopic signal.The isotopic values obtained for the isolated taxa seem coincident with the bulk, thus they neither give additional information on specific temperature differences on water column versus seafloor settings, nor exclude the lithological variation as a bias for the isotopic signal.Available δ 13 C records show considerable geographic and bathymetric variability but are commonly characterized by rising rather than declining δ 13 C values during the initial gradual warming (Bohaty & Zachos 2003;Bohaty et al. 2009;Boscolo Galazzo et al. 2014;Borelli et al. 2014).These records imply that there is no evidence for a sudden release of isotopically light carbon during the MECO warming phase.A transient ~0.5‰ negative Carbon Isotope Excursion (CIE) (~50 kyr) during the warming peak occurs at some but not all sites (Bohaty & Zachos 2003;Edgar et al. 2010;Boscolo Galazzo et al. 2014;Giorgioni et al. 2019).At the Sealza section, the δ 13 C curve does not show the same trend recorded previously, except for a negative CIE corresponding to the OIE interval displayed by δ 13 C bulk , δ 13 C Subbotina and δ 13 C Cibicidoides .We cannot establish whether these shifts correspond to those previously recorded at this stage.It is known that in shallow-water environments, changes in δ 13 C values may be controlled by the organic matter content (Oehlert & Swart 2014).
The MECO interval in our section could also be supported by the occurrence of Zone E12, marker O. beckmanni, which largely corresponds to the MECO interval (Sexton et al. 2006;Edgar et al. 2007;Bohaty et al. 2009), even though Edgar et al. (2010) demonstrated an earlier appearance (~500 kyr) in the equatorial Atlantic than in the subtropics (40.5 versus 41.0 Ma).This latitudinal diachroneity is attributed to the poleward expansion of warm surface-waters during the onset of the MECO, which created favorable conditions at higher latitudes for this potentially thermophilic species.At the Sealza section, the correct identification of O. beckmanni base and top is hampered by the scarce abundance and poor preservation of planktic foraminiferal assemblages.The species O. beckmanni first occurs just above the onset of the OIE at 113.5 m.
Biostratigraphic data from calcareous nannofossils at the Sealza section are in good agreement with the previous literature.Generally, the onset of the MECO is associated with the extinction of Sphenolithus furcatolithoides (Boscolo Galazzo et al. 2013;Toffanin et al. 2013;Messaoud et al. 2021), which in this section has been observed only just above the base of the OIE.At Sealza, the base of O. beckmanni seems to occur just below the top of S. furcatolithoides, while in other sections (e.g., Alano and Varignano, NE Italy), it is significantly above (Luciani et al. 2020).However, it must be remarked that at ODP Site 1051 (Blake Nose, Western North Atlantic), the base of O. beckmanni virtually coincides with the top of S. furcatolithoides (Edgard et al. 2010;Agnini et al. 2014;Luciani et al. 2020).Thus, at Sealza, the base of O. beckmanni appears to be comparable with the data from the western North Atlantic.

Paleo-water depth reconstruction
A first approximation to estimate the paleodepth of the Sealza section derives from the P/B ratio data and the regression proposed by Van der Zwaan et al. (1990) for living foraminifera.The low values of the P/B ratio recorded, especially in the lower part of the section, are associated with paleodepths <50 m, while in the upper part (MECO interval), the increase of the P/B ratio values indicates paleodepths up to ~100 m (Fig. 4).The species Uvigerina eocaena and U. cf.peregrina are recorded in lower neritic environments (Lamb & Miller 1984;Barbin & Keller-Grünig 1991).The species Cibicidoides lobatulus thrives in shallow waters, closely linked to near-bottom currents useful for its nutritional needs (Barbin & Keller-Grünig 1991;Altenbach et al. 1999;Schweizer et al. 2009).Other species, such as Heterolepa dutemplei and Melonis affinis, inhabited in a wide bathymetric range but are specially recorded in shallow waters through the stratigraphic time here investigated (Barbin & Keller-Grünig, 1991;Szarek 2001).
According to Briguglio et al. (2023), the LBF recorded in Sealza are well in agreement with the bathymetric range here proposed.Most of the samples have a rich LBF fauna, pointing to shallowwater settings within the photic zone, as LBF lived in symbiosis with photosynthetic microalgae.Toward the uppermost part of the sections, the LBF assemblages indicate a deeper setting characterized by depths that get closer to the oligophotic range, thus around 50 to 80 meters water depth.

Bottom waters reconstruction based on benthic foraminiferal paleoecology
Benthic foraminifera represent one of the most widely applied groups in paleoceanographic and paleoenvironmental reconstructions in the geological record, as their assemblages can provide useful information on several environmental parameters at the sea floor, such as organic matter, oxygen availability, depth, salinity, and temperatu-re (e.g., Van der Zwaan et al. 1999;Gooday 2003;Murray 2006;Jorissen et al. 2007).Previous investigations were done on paleoenvironmental reconstruction and biostratigraphy across the MECO from Tethyan successions but in deep-water settings (e.g., Boscolo Galazzo et al. 2013;Boscolo Galazzo et al. 2016;D'Onofrio et al. 2021).The investigation of the benthic foraminiferal assemblages at Sealza section, coupled with isotopic records and total organic carbon (TOC), provides, for the first time, evidence of the paleoenvironmental changes across the MECO in a shallow-water setting, and provides new insights on the resilience of both planktic and benthic communities in neritic environments.
The abundance of Cibicidoides in the pre-ME-CO interval (lower part of the measured section; Fig. 3) suggests a well-oxygenated seafloor because this genus is recognized as an epifaunal benthic foraminifera living in well-oxygenated environments (e.g., Speijer 1994;Speijer & Schmitz 1998;Giraldo-Gómez et al. 2018 a,b ).These conditions are also supported by the rich LBF fauna (e.g., Briguglio et al. 2023).
The marked decrease of Cibicidoides across the MECO interval suggests oxygenation reduction and a rise in the organic flux, as confirmed by abundant infaunal taxa (Uvigerina and Bolivina), which suggest an enhanced nutrient supply (Fig. 3).Similar findings were reported for the initial MECO phase at the Alano section and interpreted as a shift from more oligotrophic to fully mesotrophic conditions (Boscolo Galazzo et al. 2013).In marginal continental basins, eutrophication and deposition of organic-rich sediments are commonly associated with the MECO warming and directly afterward (e.g., Spofforth et al. 2010).The decrease in oxygen content during the MECO caused an increase in organic matter preservation at the seafloor, providing environmental conditions favorable for Uvigerina, which is known as an infaunal genus, usually living in fine-grained sediments (Van der Zwaan et al. 1986) and considered an index of carbon-rich and oxygen-poor conditions (e.g., Sen Gupta & Machain-Castillo 1993;Kaiho 1994;Thomas & Gooday 1996;Schmied et al. 1997;Van der Zwaan et al. 1999;De Rijk et al. 2000;Hess & Kuhnt 2005;Kawagata et al. 2006).The slight increase in TOC content within the MECO (Fig. 5) supports this hypothesis.
The occurrence of the infaunal Bolivina exclusively across the MECO interval (Fig. 3), also supports an increase of organic matter at the seafloor together with lower oxygen concentrations (Kaiho 1994;Thomas 2007).The abundance of the epifaunal Heterolepa follows the same trend as that of Cibicidoides within the MECO, although it is rare in the pre-MECO, possibly due to a competition with Cibicidoides.Heterolepa is considered to inhabit oxygenated waters as well (e.g., Van der Zwaan 1982;Rögl & Spezzaferri 2002;Russo et al. 2022), and its decrease during the MECO supports weaker oxygenation at the seafloor.
The shift to meso-eutrophic conditions at the MECO is possibly related to an enhanced hydrological cycle and supply of terrigenous flux, which is plausible in the context of the active tectonic regime coupled by climatic perturbation, similar to what has been documented by geochemical proxies in different Tethyan sections (e.g., Luciani et al. 2010;Rego et al. 2018;Peris-Cabré et al. 2023).Our data from Rock-Eval analysis confirm the terrestrial origin and increase of organic matter deposited in the basin during the MECO (Fig. 5).It is reasonable to suppose that fluvial dynamics played an important role in the erosional and depositional regimes of the river mouths linked to the precipitation increment.Nonetheless, the presence of marly limestone suggests that the coastline was distally located, as indicated by the fine material deposited.
Our results indicate that the benthic foraminiferal communities from the shallow-water setting of Sealza were sensitive to the MECO perturbances that induced low oxygen conditions and high nutrient input, similarly to the communities of the deep-water settings of marginal continental basins.

Surface water conditions based on planktic foraminiferal paleoecology
Planktic foraminifera are highly sensitive to several environmental changes, such as variation in water column stratification, density, chemistry, temperature, and nutrient supply (e.g., Corfield & Norris 1998;Quillévéré & Norris 2003;Wade & Bown 2006;Luciani et al. 2007).Nonetheless, the impact of the MECO on planktic foraminiferal assemblages is still poorly investigated, although the climatic perturbations of the lower Paleogene have been extensively documented by geochemical, sedimentological, and isotopic data.As expected, the Sealza shallow-water succession was dominated by LBF, and planktic foraminifera are not abundant, as confirmed by the PFN and the P/B ratio (Fig. 4).
The increase in the number of genera of planktic foraminifera, total abundance, and higher values of the P/B ratio detected across the MECO interval (Fig. 4) are related to the deepening of the environmental setting.Within this interval, the significant increment in abundance of Acarinina and the occurrence of Morozovelloides and O. beckmanni likely suggest an increase in surface water temperature.These taxa are indeed recognized as mixedlayer, oligotrophic warm-water indicators that dominated the tropical and subtropical assemblages in the late Paleocene to middle Eocene (with Morozovella preceding Morozovelloides) (e.g., Boersma et al. 1987;Pearson et al. 1993Pearson et al. , 2001;;Edgar et al. 2010).The absence of the genus Globigerinatheka, which is a common component of middle Eocene assemblages, appears unusual, especially considering that several species of this genus should co-occur with O. beckmanni which is considered the end member of the Globigerinatheka curry-euganea lineage (Proto Decima & Bolli 1970).The genus Globigerinatheka is regarded as a symbiont-bearing mixed layer dweller as recording depleted δ 18 O and enriched δ 13 C values (Boersma et al. 1987;Pearson et al. 1993;Premoli Silva & Jankins 1993;Pearson et al. 2001;Wade & Kroon 2002;Wade 2004), thus possibly comparable, for its ecological requirements and significance, with Acarinina.However, stable isotope values in Globigerinatheka commonly indicate calcification in deeper waters (Wade 2004;Premoli Silva et al. 2006;Sexton et al. 2006;Burgess et al. 2008;Boscolo Galazzo et al. 2014), suggesting that the genus had late-stage calcification, with CaCO 3 crusts forming deeper in the water column, possibly close to the thermocline (Pearson et al. 2001;Wade & Kroon 2002;Premoli Silva et al. 2006); that seems also supported by boron isotope data (Pearson & Palmer 1999).Unfortunately, stable isotope paleobiology is yet unknown for the species G. curry, G. euganea, and O. beckmanni.To explain the occurrence of O. beckmanni, though with very rare specimens, and the absence of Globigerinatheka, we hypothesize a differential response of gametogenesis with respect to the shallow-water character of the depositional environment here observed: Orbulinoides beckmanni probably completed its life cycle in shallower water, whereas Globigerinatheka might have preferred a deeper setting for gamete release.Test recrystallization and poor preservation of O. beckmanni in our samples prevent targeted stable isotope analysis to confirm this hypothesis.Nonetheless, this first record of planktic foraminifera in a shallow-water environment highlights such ecological requirements and calls for a better knowledge of O. beckmanni paleoecology.Across the MECO interval, the parallel decrease of the cold-water index genus Subbotina (e.g., Sexton et al. 2006) across the MECO interval reinforces the evidence of upper water column warming.The relative high abundance of Subbotina at the basal MECO despite the warming may be related to the meso-eutrophic conditions and elevated nutrient supply at the thermocline (D'Onofrio et al. 2021).This could also explain the absence, also in this portion of the section, of the oligotrophic Globigerinatheka and Acarinina.The dominance of Subbotina at the top of the section in the sample above the covered interval may be interpreted as a return to cooler conditions.
The decline in abundance of Acarinina in the upper MECO interval could have different explanations.Edgar et al. (2013) document that the symbiotic relationship of Acarinina was temporarily reduced for ~100 k.y.during the peak of the MECO at Sites 748 and 1051 (Southern Ocean and midlatitude North Atlantic, respectively).The authors relate this bleaching episode as linked to the ecological stress induced by increased sea-surface temperatures.We cannot establish whether a bleaching episode caused the decrease in abundance of Acarinina at the Sealza section.However, a permanent middle Eocene decline in the abundance of acarininids, specifically of the species larger than 150 µm, well before their evolutionary extinction level occurred at the Bartonian-Priabonian boundary (Agnini et al. 2021) in the Tethyan realm (Luciani et al. 2010(Luciani et al. , 2020;;D'Onofrio et al. 2021).Although only a single sample of the studied section is attributable to the post-MECO and cannot be considered conclusive to reconstruct foraminiferal resilience during the aftermath of the event, the very low abundance of Acarinina (Fig. 4) suggests that, even at the Sealza setting, this genus did not recover.The causes of this persisting decline in abundance are not yet established.However, a link to the climatic change and paleoceanographic variations that occurred during the MECO, or just following it, seems implied.Further analyses are needed to establish whether the permanent decrease in abundance of Acarinina is globally recorded.
The changes in planktic foraminiferal abundance recorded in the shallow-water setting of Sealza prove to be significantly different with respect to those analyzed from deeper environments.In the Tethyan Alano (northeastern Italy) and Baskil middle-bathyal sections (eastern Turkey), deposited in marginal continental basins (Luciani et al. 2010;D'Onofrio et al. 2021), planktic foraminiferal assemblages record a marked decline of acarininids and Morozovelloides within the MECO, coupled with an increase in subbotinids.Changes in salinity and a marked increase in eutrophic conditions related to freshwater input due to an enhanced hydrological cycle have been invoked as having impacted the habitat of the specialized, oligotrophic mixed-layer dweller acarininids.On the contrary, the eutrophic genus Subbotina may have benefitted from high food delivery to the thermocline, and this could explain the apparent inconsistency of the increase in this cold-water index genus at the Tethyan sections.In the Sealza assemblages, highly opportunistic and eutrophic taxa such as Chiloguembelina, Jenkinsina, and Pseudoglobigerinella bolivariana, do not occur, ever they were recorded in high abundance across the MECO interval from the Alano section (Luciani et al. 2010;D'Onofrio et al. 2021).This evidence, together with the increase of warm, specialized index taxa, could attest to a minor degree of eutrophication in the Sealza setting.This scenario is also supported by the occurrence at Sealza by the LBF communities, which should have been heavily impacted by excess eutrophication.

Ecological resilience of the Sealza shallow-water communities across the MECO
Our results on benthic and planktic foraminifera highlight that the assemblages moved to a different state during the MECO perturbation.As detailed above, oxygenation at the sea floor probably decreased, as substantiated by a drop in abundance of Cibicidoides and Anomalinoides, paralleled by an increase in taxa tolerating deoxygenated conditions and high nutrient input, such as Uvigerina and Bolivina and by the increase in TOC values.However, in the upper part of the MECO interval, the benthic community recovered the genera abundance of the pre-MECO, except for Anomalinoides.In the upper water column, planktic foraminifera show marked changes in the assemblages, mainly related to the intense MECO warming.The decline of large acarininids, which started in the upper MECO interval, is also recorded above the MECO, even though we can provide just one sample from this interval.Remarkably, this permanent decline appears to coincide with the Tethyan deep-water record (Luciani et al. 2010;D'Onofrio et al. 2020).Considering that the benthic and planktic communities moved to a new state, they proved to be not resilient to the MECO perturbance.Interestingly, the LBF also show changes in assemblages across the MECO interval, moving from rich LBF communities to assemblages dominated by small nummulitids and operculinids (Briguglio et al. 2023).It is difficult to unravel the significance of these changes in LBF because they could be related to the deepening of the depositional setting (Briguglio et al. 2017;Eder et al. 2016Eder et al. , 2017Eder et al. , 2018) ) rather than to the MECO temperature and nutrient increase (Seddighi et al. 2015;Hohenegger et al. 2019, Roslim et al. 2019,).
In conclusion, as recorded from the deep-water settings, the MECO event permanently impacted the benthic and planktic communities, although did not induce extinctions.The tipping point to undergo a permanent new regime shift was reached, thus preventing the recovery after the disturbance, wherein the communities proved to be highly unstable and finally not resilient.

suMMary and conclusIons
The identification of the MECO and its impact on both benthic and planktic faunas is still not fully understood, and most of the records and interpretations rely on deep-water successions.This study offers the remarkable opportunity to analyze the MECO from the shallow-water setting of the Sealza section (NW Italy), to evaluate the resilience of the studied groups to this environmental perturbance, and to compare shallow-water assemblages with those from deep-water settings.
We summarize below our main results: The MECO interval was targeted by using the calcareous plankton biostratigraphy and the oxygen stable isotope signal that records a marked negative shift.The lithological expression of the MECO at the studied section is represented by variations from silty marls embedded within biocalcarenite to bio-calcirudite beds rich in LBF, corals, and mollusks.The recorded lithological modifications evidence that a somewhat drastic change in the neritic inputs happened.However, correlating this evidence with the MECO event is not immediate, as the regional tectonic regime during the middle Bartonian was the main trigger of the ramp drowning, and therefore this must have played a role in the lithological variations.Nevertheless, the MECO negative shift in the isotopic signal synchronous with the lithological variations allows us to attribute mainly the silty marls deposition to enhanced hydrological cycles.This was enhanced by the temperature increase, that provided more carrying capacity of the riverine system to shed more material onto the offshore.
Our record on benthic and planktic foraminifera highlights that the communities moved to an unstable state during the MECO perturbation, as recorded de-oxygenation at the seafloor and an increase in the planktic foraminiferal warm-water index genus Acarinina.This genus significantly declined in abundance in the upper part of the MECO interval and did not recover in the post-MECO, as recorded in the deep-water Tethyan record.To explain the absence of Globigerinatheka in our record, we hypothesize a differential response of gametogenesis with respect to the shallow-water character of the depositional environment observed at Sealza.Orbulinoides beckmanni probably completed its life cycle in shallower water, whereas Globigerinatheka might have preferred a deeper setting for gamete release.
With the exception of the acarininid decline, changes in planktic foraminiferal assemblages are proved to be markedly different from those recorded from the bathyal Tethyan Alano succession, which documents a decrease in warm-water index taxa and a pronounced shift to eutrophic conditions, as proved by the marked increase of opportunistic and highly eutrophic taxa, that was not observed at Sealza.This dissimilarity suggests that the nutrient input and consequent eutrophication due to the enhanced hydrological cycle were less pronounced at the Sealza setting, as also substantiated by the occurrence of the oligotrophic LBF.These highlights suggest that the riverine system was rather far from the Sealza depositional setting.
The MECO perturbance at Sealza permanently impacted the benthic and planktic communities that exceeded the tipping point to move to a new regime, thus proving to be not only resilient but also not recording extinctions.In turn, the LBF demonstrate to be more resilient, as their changes in assemblages appear to be more related to the deepening of the depositional setting than to the MECO perturbation.
This study proved that only a multidisciplinary approach and a reliable stratigraphic constraint can tackle complex environmental interpretations during climatic crises, which is especially important when the focus is on the resilience of marine taxa.Our scenario on the impact on foraminiferal assemblages of the MECO warming event may offer a significant deep-time perspective to the IPCC or other policy-relevant organizations such as the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services (IPBES) that still insufficiently consider the paleontological contribution to past biological changes (e.g., Kiessling et al. 2023).

Fig. 2 -
Fig. 2 -Carbon (δ 13 C) and oxygen (δ 18 O) isotope data from bulk organic matter (red curve: five-point movement average) plotted against the stratigraphic log of the Sealza section and biozones.Large benthic foraminiferal zones are from Serra-Kiel et al. (1998); planktic foraminiferal zones follow the zonation scheme by Wade et al. (2011); calcareous nannofossils zonation are according to Fornaciari et al. (2010) and Agnini et al. (2014).The yellow-shaded band outlines the Middle Eocene Climatic Optimum (MECO) interval.Blue curves represent carbon and oxygen stable isotopes measured on the benthic foraminifera genus Cibicidoides, while green curves represent stable isotopes measured on the planktic foraminifera genus Subbotina.

Fig. 3 -
Fig.3-Abundance of benthic foraminiferal genera (>63 μm fraction) before, across, and after the MECO interval (yellow-shaded band) of the Sealza section plotted against lithology, oxygen isotope curve (δ 18 O) and benthic foraminiferal assemblage's parameters (H' and D').BFN: benthic foraminiferal number (specimen/1g sediment); H': Shannon index; D: dominance.Others include the minor components of the assemblage.The green line at 205 m represents the abundance of genera in sample SE-33.

Fig. 5 -
Fig.5 -A) Total organic carbon (TOC), hydrogen, and oxygen indices (HI, OI) across the MECO interval (yellow-shaded band), plotted against lithology, biozones and oxygen isotope curve.Samples with TOC values lower than 0.2 wt% have been excluded for the interpretation of the HI and OI indices of this study; see dashed lines green and blue of the HI and OI indices, respectively; B) the scatter plot of hydrogen and oxygen indices (HI/OI) indicates organic matter origins.Type I indicates a strong marine component, while type III indicates a strong terrestrial component.Samples with TOC values higher than 0.2 wt% (rectangles in paleo-violet red) have been used for the interpretation of the HI and OI indices of this study.