Central Mediterranean tephrochronology for the time interval 250–315 ka derived from the Fucino sediment succession

In the lacustrine succession F4‐F5 of the Fucino Basin, central Italy, 20 visible tephra layers were identified in the time interval 250–315 ka (Marine Isotope Stages 8–9). Fifteen of them contained suitable material to explore their volcanic sources. Among these tephra some well‐known eruptions and eruptive sequences of the Roman and Roccamonfina volcanoes were identified, such as the Tufo Giallo di Sacrofano and the Lower White Trachytic Tuff, respectively. Furthermore, the sediment succession documents a more complex eruptive history of the Sabatini, Vulsini, Colli Albani and Roccamonfina volcanic complexes during the investigated period, as inferred from previously undescribed tephra deposits. Single‐crystal‐fusion 40Ar/39Ar dating of two of the inspected tephra layers combined with two already published tephra ages provided the basis for a Bayesian age‐depth model. The modelled tephra ages allow chronological constraining of so‐far undefined eruptions of the Sabatini (272.5±4.7, 281.8±4.7, 308.5±2.8, 312.8±2.1 ka), the Vulsini (311.7±2.3, 311.9±2.3 ka) and the Colli Albani (301.0±3.6 ka) volcanic districts. Two tephra layers of an undefined volcanic source from the Roman volcanoes have modelled ages of 309.5±2.7 and 310.5±2.6 ka. The new 40Ar/39Ar and modelled ages were further used for a reassessment of the timing of already known and dated eruptive units, such as the Tufo Giallo di Sacrofano (40Ar/39Ar: 289.3±4.8 ka). Tephra tentatively correlated with the Valle Santa Maria, Case Pisello and the White Trachytic Tuff Unit E3 or Unit F offer modelled ages for these eruptions of 296.6±3.9, 301.8±3.5 and 303.6±3.4 ka, respectively. The results complete the tephrostratigraphical investigations of the c. 425 ka old F4‐F5 record, extend the Mediterranean tephrostratigraphical framework and provide a significant contribution for improving knowledge on Italian volcanic explosive activity.

Tephrostratigraphy and tephrochronology are powerful tools, which find a broad range of application in archaeology, palaeoclimatology and geosciences (e.g.Lowe 2011).Due to the synchronous deposition of volcanic ash (tephra) in different sedimentary environments, tephra layers are used to set sedimentological records in a stratigraphical context and to investigate their inter-site relationships.Tephra layers also enable dating of sedimentary successions, either by radiometric dating of suitable minerals of a tephra layer or by transferring the ages of equivalent tephra layers between different sites.An outstanding advantage of tephrochronology is that chronologies developed by this method are independent of the interpretation of proxy data utilized for tuning approaches, so that, e.g.circular reasoning can be avoided and spatial and temporal evolution and relationships of different environmental systems can be assessed (Zanchetta et al. 2016).This is of fundamental importance, as a precise and accurate temporal framework is crucial for assessing palaeoclimatic questions and identifying short-and long-term palaeoclimatic changes and their drivers.Combining the reconstructed volcanic activity with information on palaeoenvironmental changes can also provide insights into the interactions of volcanic activity and palaeoclimate (Lavigne et al. 2013;Kutterolf et al. 2019).Moreover, tephrostratigraphy can provide valuable information for assessing volcanic hazards in terms of the frequency and magnitudes of past eruptions (Sulpizio et al. 2014;Albert et al. 2019).
However, for the successful application of tephrostratigraphy and tephrochronology allowing the transfer of reliable ages between sites, the construction of a robust tephrostratigraphical framework with extended knowledge of the explosive volcanic history of a region and a comprehensive volcanic glass geochemical data set of the different eruptions are both needed.The Mediterranean region offers suitable conditions for tephrostratigraphy and tephrochronology, which has led to the development of a comprehensive and continuously expanding tephrostratigraphic framework (Keller et al. 1978;Paterne et al. 2008;Zanchetta et al. 2011;Wulf et al. 2012;Insinga et al. 2014;Petrosino et al. 2015;Giaccio et al. 2015aGiaccio et al. , 2017b;;Vakhrameeva et al. 2018Vakhrameeva et al. , 2021;;Leicher et al. 2019Leicher et al. , 2021;;McGuire et al. 2022;Monaco et al. 2022b).However, the tephrostratigraphic framework >100 ka is still fragmentary.
The lacustrine succession of the Fucino Basin, central Italy (Fig. 1), has been proven to be one of the richest Mediterranean Pleistocene tephra records and greatly contributed to a more complex tephrostratigraphical framework (Giaccio et al. 2017b(Giaccio et al. , 2019;;Del Carlo et al. 2020;Monaco et al. 2021Monaco et al. , 2022a;;Leicher et al. 2022).Its geographical position in good range downwind of most volcanic systems of central Italy and its continuous sedimentary succession make it a corner stone to develop a reference section for volcanic activity, such as successfully shown for the last 190 ka, and provided also detailed palaeoenvironmental information (Giaccio et al. 2017b;Di Roberto et al. 2018;Mannella et al. 2019).The F4-F5 sediment succession obtained in 2017 from Fucino Basin extends this work for the last c.425 ka (cf.Fig. 1; Giaccio et al. 2019) and has already greatly improved the tephrostratigraphy for the intervals 0-253 ka (Giaccio et al. 2019;Monaco et al. 2022a) and 313-425 ka (Monaco et al. 2021;Leicher et al. 2022).These results contributed towards a better knowledge of the precise timing and stratigraphical succession of Italian major volcanic eruptions, but also for preceding or following associated minor eruptive events within the Middle Pleistocene.Based on the obtained tephrochronological information, the Fucino sediment succession can be precisely dated, which is an essential prerequisite for investigating its regional palaeoenvironmental and palaeoclimatic history and allows aligning it with other Mediterranean palaeoenvironmental archives.
Here, we complement the tephrostratigraphical and tephrochronological investigations of the F4-F5 record for the pending interval 250-315 ka (marine isotope stages (MIS) 8-9), presenting new glass geochemical and chronological data of macroscopic visible tephra layers and their integration within the regional tephrostratigraphical framework.

Fucino Basin
The Fucino Basin is located ~650 m a.s.l. in the Abruzzo region in central Italy and is the largest extensional tectonic basin of the central-southern Apennine chain (Fig. 1A, B).The basin opened along the E-W, NE-SW and NW-SE oriented Fucino Fault System starting from the Late Pliocene to Lower Pleistocene (e.g.Galadini & Galli 2000;D'Agostino et al. 2001;Giaccio et al. 2012;Amato et al. 2014).Seismic investigations of the Fucino Basin revealed a semi-graben architecture with an increasing sediment infill of up to ~900 m thickness from the west towards the depocentre in the east (Patacca et al. 2008).The Plio-Quaternary deposits unconformably overlay both Late Messinian terrigenous deposits and the Meso-Cenozoic carbonate basement (Cavinato et al. 2002;Giaccio et al. 2019;Mondati et al. 2021).The Plio-Pleistocene sedimentation is believed to have started before 2.0 Ma (Giaccio et al. 2015b) and 3.0 Ma (Mondati et al. 2021) and appears to be continuous, at least in the central part of the basin (Giaccio et al. 2017b(Giaccio et al. , 2019;;Mannella et al. 2019).The basin hosted Lake Fucino (Lacus Fucinus) until the lake was artificially drained, starting from the 1st-2nd century CE and completed by the end of the 19th century for agricultural reasons.

Italian volcanic activity of the Middle Pleistocene
The Pleistocene-Holocene Italian volcanism includes the volcanic activity of the Roman Province (Vulsini, Vico, Sabatini and Colli Albani volcanic districts), the Volsci Volcanic field and the Roccamonfina volcano (former Ernici-Roccamonfina Volcanic Province), the Campanian Province (Neapolitan volcanoes of Somma-Vesuvius, Campi Flegrei, Ischia and Procida and the buried Campanian Plain volcanism), Mount Vulture, the Aeolian Arc, the Sicily Province (after Peccerillo 2005;Peccerillo & Frezzotti 2015;Cardello et al. 2020; see also Fig. 1C).For the period investigated here (250-315 ka), explosive volcanic activity is mainly known from the Roman Province and Roccamonfina volcano and only subordinately from the Campanian Province and the Volsci Volcanic field (Fig. 1C).
In the investigated time period, the Vico volcano activity was characterized by effusive activity of the Lago di Vico lava Formation, which led to the construction of a stratovolcano (258-305 ka; Perini et al. 2004), before an
In the Colli Albani Volcanic district (CAVD), the Faete Phase (De Rita et al. 1988), during which the Tuscolano-Artemisio peri-caldera ring fracture system and the intra-caldera Faete stratovolcano formed (Giordano et al. 2006(Giordano et al. , 2010)), occurred at c. 250-308 ka (Marra et al. 2003) and was characterized mainly by scoria cones and lava flows, but also included the Castiglione maar succession with an estimated age of c. 285 ka (Marra et al. 2003).
With respect to Campanian volcanoes, there is still a lack of knowledge of volcanic products older than the Campanian Ignimbrite eruption (c.40 ka; Giaccio et al. 2017a), as those have been buried by Late Pleistocene-Holocene volcanic activity.The only accessible deposits for the investigated time interval discovered so far are the Seiano ignimbrite deposits south of the Campanian Plain, the ages of which range from c. 250 to 290 ka (Rolandi et al. 2003).

F4-F5 sediment record
In 2017, at the Fucino F4-F5 drill site close to the centre of the basin (Fig. 1B, latitude 42°00 0 06.22 00 N, longitude 13°32 0 17.79 00 E) two parallel core successions were recovered reaching a field depth of 87.00 and 87.75 m, respectively.Seismic information (Cavinato et al. 2002) and sedimentation rate estimations based on previous drilling campaigns at adjacent sites (GeoLazio and SP cores, ~20 cm ka À1 ; Giaccio et al. 2015bGiaccio et al. , 2017b) ) suggested the F4-F5 drill site as ideal for recovering older (>190 ka) sediments at a relatively shallow depth as compared to the F1-F3 (0.45 mm a À1 ) in the main depocentre of the lake (Giaccio et al. 2017b;Mannella et al. 2019).Individual (1.5-m-long) core segments from both boreholes were spliced to a 98.11-m-long composite profile (F4-F5 record) based on core images, XRF data and palaeomagnetic information (Giaccio et al. 2019).Sediments are dominated by grey-whitish lacustrine calcareous marls alternating with darkish lacustrine clays and frequently intercalated, macroscopically visible tephra layers (n = 140).The variability of the calcareous content in the F4-F5 sediments is mainly related to variations in precipitation of endogenic calcite, which depends on the lake's primary productivity (Mannella et al. 2019).The productivity in turn is mainly controlled by temperature and hydrology, which are related to glacial-interglacial and sub-orbital climatic variability (Mannella et al. 2019).
Here, the remaining interval 250-315 ka is investigated.

X-ray fluorescence (XRF) core scanning
Individual core halves of the F4-F5 record were scanned using an Itrax XRF scanner (Cox Analytical Systems, Sweden) at the Institute of Geology and Mineralogy of the University of Cologne, Germany.XRF scans were made as described in Giaccio et al. (2019) using a Chromium tube set at 55 kV and 30 mA with a dwell time of 10 s and a step-size of 2.5 mm.

Tephra identification and sample preparation
Tephra layers in the F4-F5 Fucino record were identified during visual core description and subsequent inspection of high-resolution line-scan images.Tephra layers were sampled over their entire thickness as a bulk sample and, if different subunits were identified, a subsample was taken from each subunit (Table 1).If tephra layers of thickness <0.5 cm were mixed with substantial amounts of lake sediments, samples were sieved and treated with HCl (10%, room temperature <12 h) to remove carbonates and enrich the glass fraction.

Sampling of the proximal Tufo Giallo di Sacrofano (TGdS) deposits
The TGdS was sampled at two sites at the northern rim of the Sacrofano caldera of the Sabatini Volcanic District.At site A (42°8 0 34.00 00 N, 12°23 0 45.00 00 E), TGdS deposits are separated by a palaeosol from underlying local scoria cone deposits and sample TGDS-1 was taken from the basal part of the TGdS deposits represented by ~1-2 m thick blackish, well-sorted, cm-sized scoria of the initial phase of the eruption.TGDS-2 is a sample from the overlying basal part of the main eruptive phase represented by yellowish, partly zeolitized ash deposits of several-m-thickness.At site B (42°8 0 5.00 00 N, 12°23 0 45.00 00 E), only the TGdS products of the main phase are exposed on top of Monte Ficoreto scoria cone deposits.Samples TGDS 3-4 were sampled from the middle and the bottom, respectively, of this >10-m-thick succession.

EPMA-WDS
Major and minor elements of individual glass fragments were analysed by electron microprobe wavelength dispersive spectroscopy (EPMA-WDS) to determine the geochemical fingerprint of the respective tephra layers of the Fucio succession and proximal TGdS samples.All samples were measured at the University of Cologne (Germany), whereas one sample (TF-47) was also measured at the Istituto di Geologia Ambientale e Geoingegneria of the Italian National Research Council (IGAG-CNR, Rome, Italy).A JEOL JXA-8900RL electron microprobe equipped with five wavelength dispersive spectrometers was used for analysis at the University of Cologne.The operation conditions were set to 12-kV accelerating voltage, 6-nA beam current and 5-10 lm beam diameter.Full details of calibration and measuring conditions are given in Leicher (2021).Analysis at IGAG-CNRwere performed with a Cameca SX50 electron microprobe equipped with five wavelength dispersive spectrometers and operated with an accelerating voltage of 15 kV, a beam current of 15 nA, a beam diameter of 10 lm and a counting time of 20 s per element (full details of calibration and measuring conditions are given in Leicher & Giaccio (2021)).Analytical differences caused by the different EPMA-WDS settings of the IGAG-CNR and University of Table 1.Tephra layers identified in the time interval 250-315 ka of the F4-F5 succession and information about their position within the record, appearance, petrological composition according to the TAS-diagram (Le Bas et al. 1986)  During microprobe analysis at the University of Cologne, MPI-DING glasses (ATHO-G; StHs6/80-G) were used as secondary standards to evaluate the accuracy and precision of measurement sessions based on preferredvalues from Jochum (2006).Meanvalues for precision (relative standard deviation %) and accuracy (bias of mean to preferred reference value in %) of analyses are respectively up to 1.4 and 1.1% for elemental concentrations >60 wt%, up to 3.2 and 2.6% for 25-5 wt %, up to 8.2 and 9.2% for 5-1 wt%, up to 11.3 and 10.0% for 1-0.2 wt% and 61 and >18% for 0.2-0.1 wt%.
The EPMA laboratory at IGAG-CNR used the Kakanui augite USNM 122142 (Jarosewich 2002) and rhyolite RLS132 glass from the United States Geological Survey (Huebner & Woodruff 1985) as secondary reference material prior to sample measurements to evaluate data quality.The mean analytical precision and accuracy are respectively up to 1.8 and 0.2% for element abundances >50 wt%, up to 2.7 and 3.8% for element abundances between 20-10 wt%, up to 5.8 and 9.5% for element abundances between 9-4 wt%, up to 3.0 and 5.3% for element abundances 2-1 wt%, up to 40.8 and 18.1% for element abundances 1.0-0.1 wt% and up to 23.3 and 54.4% for element abundances <0.1 wt%.
Only EPMA-WDS geochemical analyses of glass fragments with analytical totals >90 wt% were considered and normalized to 100% on a water-free basis, excluding volatiles (Cl, SO 3 and F).The tephra layers were classified according to their geochemical glass composition using the total alkali vs. silica (TAS) classification system (Le Bas et al. 1986).
40 Ar/ 39 Ar dating Two tephra layers (TF-48/-61, see Table 1 for depths) were selected for dating and subsequently sieved using mesh sizes of 500, 350 and 250 lm.For tephra TF-48, the 500-350 and 350-250 lm fractions were further separated using a FRANTZ magnetic separator at the University of Cologne.All aliquots were washed with distilled water before transparent, inclusion-free sanidine and/or leucite crystals were handpicked under a binocular and sent to the Laboratoire des Sciences du Climat et de l'Environnement (CEA, CNRS UMR 8212, Gif-sur-Yvette, France).At CEA, crystal aliquots underwent an additional purity screening and were selected based on available crystal sizes for irradiation and dating (500-350 lm fraction for TF-61 and 350-250 lm fraction for TF-48).Between 20 and 30 crystals for each tephra layer were irradiated at the CLICIT facility of the Oregon State University TRIGA reactor before individual minerals were measured by means of single crystal 40 Ar/ 39 Ar fusion at the CEA.Neutron fluence J factor was calculated using co-irradiated Alder Creek sanidine standard ACs-2 associated with an age of 1.1891 Ma (Niespolo et al. 2017) according to the K total decay constant of Renne et al. (2011) (k e.c.= (0.5757AE0.016)910À10 a À1 and k b À = (4.9548AE0.013)910À10 a À1 ).J-values are the following: TF-48 (J = 0.00055581AE0.00000139;TF-61 (J = 0.0005606AE0.00000045).A detailed description of analytical settings can be found in Leicher et al. (2022) and full analytical data for each sample can be found in Table S1.

Age-depth modelling
Age-depth modelling for the studied interval of the F4-F5 succession was performed using the software package Bacon v. 2.5.8 (Blaauw & Christen 2011) within the open-source statistical environment R (R Core Team 2022) and used four available 40 Ar/ 39 Ar ages of tephra layers identified within the sediment succession.The entire succession was divided into 5-cm vertical sections to model individual accumulation rates at a 99% confidence interval, which provide the basis for the agedepth model.Major sedimentological changes identified during core description and within the XRF-downcore data were considered via the 'boundary' function within the model to account for different sedimentation rates during glacial and interglacial conditions.This major lithological change (carbonate and clastic content) was identified at about 53.1 m correlated depth (m c.d.) and marks the change from interglacial (MIS 9) to glacial (MIS 8) conditions (Giaccio et al. 2019;Mannella et al. 2019).Based on the age-depth model, an individual age with a 2r uncertainty was calculated for each tephra layer identified.

Results
40 Ar/ 39 Ar dating results and age-depth modelling The results of 40 Ar/ 39 Ar dating of individual tephra layers are presented as probability diagrams (Fig. 2).Weighted mean age uncertainties are reported at 2r, including J uncertainty, and were calculated using Isoplot 4.1 (Ludwig 2009).Inverse isochrones of individual samples, although imprecise for tephra TF-48, are characterized within uncertainties by an atmospheric 40 Ar/ 36 Ar initial intercept, suggesting that dated crystals are without detectable excess argon.The full inverse isochrones data set can be found in Table S1.A total of 15 individual sanidine and leucite crystals of TF-48 were dated.The main population, constituted by 13 crystals (Fig. 2), was used to calculate a weighted mean age of 289.3AE4.8ka (MSWD = 0.1, p = 1.0).Two older crystals were also found (356.0AE16.4,454.0AE23.4ka).The age of tephra layer TF-61 was determined based on 15 leucite single crystals, which share within uncertainties the same age (Fig. 2).This allowed the calculation of a robust and precise weighted mean age of 313.3AE1.0ka (MSWD = 0.3, p = 1.0).
The two new 40 Ar/ 39 Ar ages and the previously obtained ages for TF-62 (313.5AE1.4ka; Leicher et al. 2022) and TF-43 (253.4AE0.8 ka;Monaco et al. 2022a) provided the basis for age-depth modelling of the studied interval (Fig. 2).Overall, the model suggests continuous sedimentation with a mean accumulation rate of 67.6 a cm À1 (14.8 cm ka À1 ) during glacials and 41.1 a cm À1 (24.3 cm ka À1 ) during interglacial conditions.The modelled ages of the tephra layers that have not been directly dated are provided in Table 1.All 40 Ar/ 39 Ar ages of tephra layers discussed from literature were recalculated according to ACs at 1.1891 Ma (Niespolo et al. 2017) and the total K decay constant of Renne et al. (2011) and are summarized in Table 2.

Tephra identification, morphological description and compositional features
Within the interval 49.00-60.61m c.d. of the F4-F5 sediment succession 20 layers containing pyroclastic material were identified by visual inspection.The position of these tephra horizons and their characteristic lithological features, i.e. thickness, colour, morphological appearance, are given in Table 1.Often, tephra occurs as massive discrete layers or single or arrays of lenses intercalated in fine grained lake sediments and vary in thickness between 0.4 and 10.5 cm.Bioturbation, sediment load, or drilling disturbance structures are present, but had a minor effect on the preservation of most of the tephra layers with respect to determining the position of the isochrone or internal structures.Tephra layers TF-43 and TF-62, framing the here investigated interval, were previously analysed and correlated with the Canino eruption from the VVD (Monaco et al. 2022a) and the Magliano Romano Plinian Fall (MRPF) eruption of the SVD (Leicher et al. 2022), respectively.Of the remaining 18 tephra layers, 13 contained sufficient amounts of fresh glass for detailed geochemical fingerprinting.The other five tephra layers, in which glass was absent or affected by devitrification processes, are found in the interval 49.02-56.37m c.d.This partially extends the interval from ~46-49 m c.d., in which methane-sediment interactions have caused alteration of volcanic glass as described by Monaco et al. (2022a).However, similar to the interval described before, not all tephra layers were affected by alteration.The full data set of EPMA-WDS analyses is available at the EarthChem repository (Leicher et al. 2023).The individual petrological classification of tephra layers according to the total alkali vs. silica (TAS) diagram (Le Bas et al. 1986) is summarized in Table 1 and visualized in Fig. 3.

General compositional features of investigated tephra layers
Tephra layers show the entire suite of rocks known from peri-Tyrrhenian volcanoes, but are dominated by tephritic to phonolitic compositions (n = 17).Only one tephra had a phonolitic-trachytic composition.Based on their compositional differences and positions within the CaO/ FeO vs. Cl diagram, the tephra layers are discussed according to their proposed volcanic source.The majority of compositions suggest an origin from the Roman volcanoes (Sabatini Volcanic District (SVD), Vulsini Volcanic District (VVD), Colli Albani Volcanic District (CAVD)), but there is also evidence of activity from the Roccamonfina volcano (RMF) (Fig. 3).
From the CAVD the crater forming event causing the Castiglione maar succession has an estimated age of c. 285 ka (Marra et al. 2003), thus being much younger.Therefore, TF-53 is most likely related to the generic explosive activity of the Faete phase leaving a precise correlation to a specific eruption pending.

Roman tephra with mixed SVD/VVD compositions
Tephra layers TF-54, TF-57 and TF-58 have a compositional spectrum in the TAS diagram ranging from tephrites to phonolites.This heterogeneous composition is also expressed in their overlapping position within the CaO/FeO vs. Cl diagram, occupying the fields of the Roman volcanic districts Sabatini, Vulsini and Vico.In addition to uncertainties concerning the precision of the data themselves, the CaO/FeO ratio itself is less distinctive for this compositional range, as also the data defining the composition fields within the diagram suggest a natural overlap in the composition of the three Roman volcanoes.
For the Vico volcano only effusive products for the time interval 305-258 ka are known (Perini et al. 2004), and compositions related to older and younger explosive events are of more evolved character following a phonolitic-trachytic-rhyolitic evolution (Pereira et al. 2020;Monaco et al. 2021Monaco et al. , 2022a)).The F4-F5 tephra layers form two age groups 302 ka (TF-54) and 310-312 ka (TF-57/TF-58), for which potential eruptive equivalents of the SVD and VVD are discussed below.
TF-54 (modelled age 301.8AE3.5 ka).-TF-54 has a tephriphonolitic-phonotephritic composition with some minor phonolitic components, and a mean alkali ratio of 1.4.The position within the CaO/FeO vs. Cl diagram plots both in the SVD and VVD fields, with a tendency towards lower, VVD-like, CaO/FeO ratios.For TF-54 a potential eruptive equivalent is the Case Pisello eruption of the VVD, dated at 293.8AE4.0ka (Brocchini et al. 2000).
TF-57/-58 (modelled ages 309.5AE2.7,310.5AE2.6 ka).-Tephra TF-57 has a tephritic-phonotephritictephriphonolitic composition, whereas the composition of TF-58 also includes phonolitic components with mean alkali ratios of 1.6 and 1.9, respectively.Based on their close position to the underlying MRPF deposits, these tephra layers were tentatively related to the post volcanic activity within the SVD, as no dated explosive activity has been reported for the VVD so far (Marra et al. 2020b).TF-58 is the thicker (7 cm) of the two tephra layers and likely represents a major eruptive event.Similar to the other SVD-like tephra layers (TF-56, -61) the specific origin from within the SVD cannot fully be resolved at present.Besides the geochemical and chronological similarities to the MRPF, the ages of the Fucino tephra layers overlap with the youngest age of the Tufo die Bracciano eruption (310.0AE5.0ka; Sottili et al. 2010).Indications of contemporaneous volcanic activity are also found in palaeo-surfaces from the Tyrrhenian coastal area, where an unknown volcanic event is dated at 310.2AE4.3ka (KW-1; Marra et al. 2023).

Roccamonfina tephra
TF-55 (modelled age 303.6AE3.4ka).-TF-55 has a phonolitic-trachytic composition with a low alkali ratio (mean = 0.8) and low CaO values (mean = 1.0 wt%).Its position within the CaO/FeO vs. Cl diagram plots is in the field of Ischia; however, Ischia's known volcanic activity only dates back to 150 ka (Sbrana et al. 2018).Glass compositions similar to TF-55 have been observed for tephra layers TF-64 and TF-64a (c.316 ka), which are correlated with the volcanic products of the White Trachytic Tuff (WTT) Unit D of the Roccamonfina volcano (312.9AE4.0ka; Giannetti & De Casa 2000;Leicher et al. 2022).The younger WTT Unit E has besides its basal unit (UE) three distinct subunits (E1-3), of which E1 was imprecisely dated in proximal deposits between 301 and 310 ka (Giannetti & De Casa 2000) and correlated with an ash layer of the Bojano basin (312.1AE5.0ka; Amato et al. 2014;Leicher et al. 2022).However, the composition of E1 together with that of the undated WTT units UE and E2 only partially overlap with TF-55 (Fig. 6), as they have a generally more evolved trachytic character.For the younger Unit E3 (301.5AE8.0ka) and Unit F (292.5AE14.0ka) of the WTT no glass geochemical data are available, but these are the best potential equivalents from a temporal perspective.However, Ballini et al. (1991) described less evolved compositions for the younger WTT deposits.This is also observed for tephra layers found in the Lake Ohrid succession (OH-DP-0997/-1055: c. 230-241 ka; Leicher et al. 2019), which are related to the younger WTT activity and are supposed to have similar compositions to TF-55.Therefore, TF-55 most likely represents an equivalent of the WTT series, whose exact correlation with one of the subunits still needs to be verified.
The 250-315 ka explosive history of Italian volcanism inferred from the Fucino record The investigations of the volcanic sources of the tephra layers identified in the F4-F5 site confirm the general pattern of volcanic activity observed in proximal settings   with a dominating influence of the Roman volcanoes (Fig. 7).The majority of tephra layers (n = 10) of the F4-F5 succession likely originates from the SVD.Two of those can be confidently correlated with known major explosive events (Sottili et al. 2010(Sottili et al. , 2019) ) of the Magliano Romano Plinian Fall (TF-62; Leicher et al. 2022) and the Tufo Giallo di Sacrofano eruption (TF-48).Directly above the tephra equivalent of the MRPF eruption (TF-62), a cluster of four tephra layers (TF-56/-57/-58/-61) suggest by the narrow time interval of their deposition between 308.5AE2.8 and 313.3AE1.0ka and the similarities in their geochemical signature an alignment with the post-MRPF volcanic activity.However, also a correlation of TF-58 with the Tufo di Bracciano (TdB) is possible from a chronological perspective, as the youngest age obtained for the TdB is 310.0AE5.0ka (Sottili et al. 2010).The TdB is a major eruptive event with a widespread dispersal of a thick pyroclastic succession (Sottili et al. 2010), for which two clusters of ages exist (Sottili et al. 2010;Pereira et al. 2017).The younger age cluster at 310.0AE5.0ka (Sottili et al. 2010) is potentially represented by TF-58, a 7-cm-thick tephra representing a larger eruptive event, whereas the older cluster at 319.1AE6.0 to 323.9AE2.0ka   Sottili et al. 2010).Following the TF-48/ TGdS eruption two tephra layers (TF-47: 281.8AE4.7 ka and TF-46: 272.5AE4.9ka) of the SVD are identified, which suggest a longer paucity and lower frequency of post-TGdS activity compared to the post-MRPF activity.The composition of TF-47 is quite similar to that of TF-48, which is also different compared to the post-MRPF eruptions with low evolved compositions.
Another major eruptive event of the SVD, the Pizzo Prato eruption (251.4AE16.0ka; Sottili et al. 2010), is not found within the Fucino sediment succession.It was identified in the Lake Ohrid succession (OH-DP-1175), with an older, but more precise age of 270.6AE4.9ka (Leicher et al. 2021) according to the age model.As Lake Ohrid is located >700 km apart from the SVD, the lack of evidence in the F4-F5 record is surprising, especially because of the presence of the numerous minor SVD eruptions identified.Unfavourable preservation conditions in the Fucino succession are unlikely, as TF-46 is of similar age and related to the SVD.This may suggest a more southern and curved dispersal axis for the main Pizzo Prato phase.The different composition of TF-46 compared with the composition of the main Pizzo Prato event suggests it to represent either a different phase of the eruption or a pre/post eruptive event.
Besides the previously identified Canino eruption (TF-43: 253.4AE0.8ka; Monaco et al. 2022a), three other tephra layers of the F4-F5 record are related to the activity of the VVD.TF-59/-60 are undefined eruptive events with an age of 311.7-311.9AE2.2ka, which are potentially also recorded in the youngest age cluster of CB4/5 in proximal deposits (Marra et al. 2020a).Tephra TF-54 (301.8AE3.5 ka) is tentatively associated with the Case Pisello eruption, which was dated at 293.8AE4.0ka (Brocchini et al. 2000).The absence of further VVD-like tephra layers between c. 253 and 302 ka in the F4-F5 record confirms the lack of major explosive events from the VVD during this period observed in its direct vicinity (Brocchini et al. 2000;Palladino et al. 2010;Marra et al. 2020b).
Tephra layer TF-53 (301.0AE3.6 ka) was associated with the explosive volcanic activity of the Delle Faete phase (c.250-308 ka; Marra et al. 2003) of the CAVD; however, a correlation with a specific eruption is pending.
Within the studied succession, only one tephra indicates an origin from the Roccamonfina volcano.Tephra TF-55 (303.6AE3.4ka) is associated with the White Trachytic Tuff, but a robust correlation with one of the subunits of WTT E or F is pending at present.However, also for the older WTT equivalents, identified in the section below (Leicher et al. 2022), a straightforward correlation with one of the proximal subunits was challenging, indicating the need for further detailed geochemical and chronostratigraphical investigations of the proximal deposits.
The 250-315 ka Fucino record in the tephrostratigraphical framework of the central Mediterranean So far, the central Mediterranean tephrostratigraphical framework for the interval 250-315 ka was based on fragmentary information from the Italian Peninsula.Information derived from continuous stratigraphical succession covering the entire interval was only provided by distal archives, which may record only the most widespread eruptions.At present, the Fucino F4-F5 record from central Italy provides the most detailed and well-dated succession providing 20 volcanic events for this interval (Fig. 7) and can act as a reference section to improve correlations between different archives.
Mediterranean sedimentary archives for which tephra layers within the investigated time frame have been reported are the Italian San Gregorio Magno (Munno & Fig. 8. Synthesis of proximal and distal information of the volcanic history of the Italian volcanoes and the resulting tephrostratigraphical framework of the central Mediterranean region.All 40 Ar/ 39 Ar ages from literature were recalculated according to ACs at 1.1891 Ma (Niespolo et al. 2017) and the total K decay constant of Renne et al. (2011).See also Table 2 for references and ages.The ages of tephra layers from the distal records are based on archive specific age-depth models (Lourens 2004;Piva et al. 2008;Vakhrameeva et al. 2019;Leicher et al. 2021).Data for the LR04 are based on Lisiecki & Raymo (2005) with the MIS substages of Railsback et al. (2015).XRF data of the Lake Ohrid DEEP site from Wagner (2019).The column 'Synthesis of Italian volcanic activity' combines all information.If robust correlations of equivalent eruptive events could be established and several ages were available, only the most reliable age is shown as discussed in the respective subsection.Petrosino 2007) and Bojano basins (Amato et al. 2014), outcrops along the Tiber river and its tributaries (Marra et al. 2016(Marra et al. , 2018(Marra et al. , 2023)), the marine records of KC01B/ ODP964A+B (Lourens 2004) and PRAD1-2 (Piva et al. 2008), and the terrestrial records from Lake Ohrid (Leicher et al. 2021) and Tenaghi Philippon (Vakhrameeva et al. 2019) (cf.Fig. 1A).A tephrostratigraphical overview combining distal archives and near-vent volcanic records is provided in Fig. 8.The oldest tephra layers (S1-S4, Campanian or Roccamonfina origin) found in the San Gregorio Magno basin (Munno & Petrosino 2007) have an estimated age between c. 241-263 ka (Ascione et al. 2013), but no equivalent tephra could be recognized in the Fucino record.Furthermore, the integration of 10 volcanic events dated in the different outcrops of the Tiber valley and its tributaries (Marra et al. 2016(Marra et al. , 2018(Marra et al. , 2023) ) can only be conducted from a chronological perspective, since geochemical data are not available.However, this would be of particular interest as these deposits were used to reconstruct sea-level changes (Marra et al. 2016(Marra et al. , 2023) ) and thus would provide tie points for the alignment with other palaeoenvironmental records.Some of these tephra ages overlap with the (undefined) eruptions identifiedwithin the F4-F5 record, such as the volcanoclastic deposits PdG-13 (247.3AE3.9ka; Marra et al. 2016) and R93-15H2 (252.2AE8.0ka; Marra et al. 2016) with the Canino equivalent .Besides the identification of the Pizzo Prato eruption in the Lake Ohrid succession (OH-DP-1175; Leicher et al. 2019), it also might be represented by PdG-S4+R95-04B (268.3AE3.8ka; Marra et al. 2016).AV-AF (277.0AE1.6 ka; Marra et al. 2016) has a similar age to that of the altered tephra layers TF-46 and TF-47.The Tufo Giallo di Sacrofano eruption (TF-48; 289.3AE4.8ka) was also identified within the Roman area (R93-28: 289.7AE2.0ka; Karner & Renne 1998;Marra et al. 2016).Tephra layers TF-51/-55 (297-302 ka) arise as potential equivalents of PdG-S19 (297.3AE2.2ka; Marra et al. 2016) and CdP (302.5AE9.0ka; Marra et al. 2018).
In marine records, tephra layers I-19 (c.270 ka, potential Pizzo Prato) and I-20 (c.313 ka, potential MRPF) were described in cores from the Ionian Sea (KC01B/ODP964A+B; Lourens 2004), but investigations for obtaining their geochemical composition were not successful (Vakhrameeva et al. 2021).Further, two tephra layers (PRAD5054: c. 290 ka and PRAD5233: c. 309 ka) identified in the PRAD1-2 record from the Adriatic Sea (Piva et al. 2008) represent potential equivalents from a chronological perspective of the TGdS (TF-48) and the post-MRPF activity identified in the Fucino succession (TF-56-TF-61).From the Tenaghi Philippon peat record in Greece, at least five tephra layers of Italian origin were identified between 289-315 ka, but their compositions suggest an origin from either the Neapolitan or Aeolian Arc volcanoes, thus differing from the tephra layers of the Fucino record.

Conclusions
Lithological, geochemical and chronological characteristics of 13 tephra layers identified in the MIS 8/9 interval of the Fucino F4-F5 sediment succession were investigated aiming at assessing their volcanic sources.Single crystal 40 Ar/ 39 Ar ages of four tephra layers set the basis for a Bayesian age-depth model for the interval 250-315 ka, which enabled indirect dating of the other tephra layers within the succession through interpolation.The investigation of this interval completes the tephrostratigraphical and tephrochronological investigations of macroscopic tephra layers of the F4-F5 sediment succession, which represents the first continuous and most tephra rich sediment succession of the last 425 ka within the central Mediterranean region.Identifying the volcanic sources and, in some cases, the specific equivalent eruptions or eruptive phases provides new insights into the eruptive history of Italian volcanism.This applies in particular to the major known eruptions of the Sabatini and Vulsini volcanic districts of the Roman Province and the Roccamonfina volcano.
The established tephrostratigraphy of the stratigraphically ordered Fucino succession represents a reference data set for reassessing the order and chronology of (minor) eruptive events not yet conclusively identified in proximal settings.It is also a key section for the subsequent extension of the central Mediterranean tephrostratigraphical framework for this period, which previously lacked a continuous sequence in close proximity to the volcanic sources.Similar to the tephrostratigraphical observations made for the F4-F5 record before, the amount of so far unidentified eruptive events suggests the need for further detailed investigations of near-vent successions.This would help to improve the geochemical and geochronological data basis especially for eruptions of lower magnitude, but highlights also the need for cryptotephra analysis of distal records, to link the available archives.10.26022/IEDA/112832.The XRF data that support the findings of this study are available from the corresponding author upon reasonable request. 40Ar/ 39 Ar original data are available in Table S1.

Fig. 2 .
Fig. 2. A, B. Relative probability diagrams of single crystal fusion 40 Ar/ 39 Ar ages of tephra layers TF-48 and TF-61.C-F.Age-depth model (C) based on available 40 Ar/ 39 Ar ages of tephra layers of the studied interval, modelled using Bacon v. 2.5.8 (Blaauw & Christen 2011).The red line indicates the position of the boundary function, which was set within the model to account for differences in the accumulation rate during glacial and interglacial conditions.D. Graph showing the fit of all Markov Chain Monte Carlo iterations of the model run.E, F. Prior (green) and posterior (grey) distributions of accumulation rate (E) and memory (F) for the modelled succession of F4-F5 along with the chosen input parameter.

Fig. 3 .
Fig. 3. Geochemical classification and overview of F4-F5 tephra layers investigated.A, C, E: TAS classification according to (Le Bas et al., 1986).B, D, F: CaO/FeO vs. Cl diagrams show the different volcanic origins based on Giaccio et al. (2017b).G, H, I: SiO 2 vs. K 2 O/Na 2 O diagrams of the respective tephra layers.Error bars represent the twofold standard deviation of replicate analyses of ATHO-G.

Fig. 4 .
Fig. 4. TAS diagram and bi-variate plots of tephra layers associated with the SVD.A, H, I.The full compositional spectrum of tephra layers TF-46/-47-/48 is shown together with potential equivalents from the SVD (new data of this study TGDS1-4 and TGdS data (from TGDS2-4) based on Giaccio et al. (2019) and Pizzo Prato data from Monaco et al. (2022a)) and the Lake Ohrid record (data from Leicher et al. (2021) and Wagner et al. (2023)).B, C, D, E, F. Bi-variate plots of compared tephra layers with SiO 2 composition >56 wt%.G, J. Bi-variate plots of compared tephra layers with SiO 2 composition <52 wt%.TAS classification according to Le Bas et al. (1986) and CaO/FeO vs. Cl diagram based on Giaccio et al. (2017b).Error bars represent the twofold standard deviation of replicate analyses of ATHO-G.

Fig. 6 .
Fig. 6.TAS diagram and bi-variate plots of tephra layers associated with the WTT of the Roccamonfina volcano and younger Roccamonfina products from the Lake Ohrid succession.A-H.Comparison of TF-55 with compositions of WTT Units D, E Base UE, E1-2 and E2 from Leicher et al. (2022) and tephra layers OH-DP-1053/-1055 from the Lake Ohrid succession (Leicheret al. 2021), which represent younger products (c.230-245 ka) of the Roccamonfina WTT stage.Error bars represent the twofold standard deviation of replicate analyses of ATHO-G.

Canino Latera Vulsini fields
Leicher et al. (2021)d 40 Ar/ 39 Ar ages.†= 40 ages of this study; ‡ = 40 Ar/ 39 Ar age fromMonaco et al. (2022a); ‡ ‡ = 40 Ar/ 39 Ar age fromLeicher et al. (2022).Tephra layers written in italics represent altered material.Cologne laboratories are discussed in detail inLeicher et al. (2021).This previous comparison revealed higher results for SiO 2 and MgO and lower for FeO, Mn, Na 2 O of the IGAG-CNR analyses relative to those obtained at the University of Cologne, whereas concentrations of TiO 2 , Al 2 O 3 , CaO and Cl do not show systematic deviations.With respect to sample TF-47, results of both laboratories fully overlap within the compositional spectrum of the sample, and so no correction was applied. Howevr, as observed within the previous comparison, IGAG-CNR results trend towards the overall upper range of SiO 2 and lower range of Na 2 O concentrations measured for TF-47.

Table 2 .
Renne et al. (2011)1740 Ar/ 39 Ar ages discussed in the text and in Fig.8.For a homogenized comparison, all ages were recalculated to ACs at 1.1891 Ma(Niespolo et al. 2017) and the decay constant ofRenne et al. (2011).Uncertainties are given at a 2r level including analytical and Jvalue uncertainties. N/A =no information about decay constant and mineral flux standards in original reference available for recalculation.