Understanding the spatio-temporal variability, the magnitude, and the different expressions of Quaternary orbital and millennial-scale paleoclimatic changes across regions is a frontier challenge of modern paleoclimatology (e.g. EPICA community members, 2006). Addressing this issue requires the acquisition of regionally representative high-resolution and well-dated records of climatic variability.

In this framework, the Fucino paleolake in central Italy (Fig. 1a) should add a nodal point in the western, currently vacant area of a network of long terrestrial Mediterranean records (including e.g. the Dead Sea (Neugebauer et al., 2014), Lake Van (Litt and Anselmetti, 2014), Lake Ohrid (Wagner et al., 2009), and the Tenaghi Philippon (Pross et al., 2015)). Indeed, among the central Italy intermountain tectonic depressions, it is probably one of the oldest and the only one that hosts a continuous, lacustrine succession since the late Plio–Quaternary (Fig. 1b). The site is also ideal because of its proximity to Quaternary peri-Tyrrhenian volcanic centres (Fig. 2) that on occasion deposited tephras in the basin that serve today as important chronological marker beds (e.g. Giaccio et al., 2012, 2015; Regattieri et al., 2015). This is a crucial requirement for comparing intra- and inter-regional paleoclimatic records, based on different dating methods, for evaluating the temporal relationship between them and with respect to the main climatic forcing (e.g. orbital).

With this purpose, an 82 m long core was recovered from the central eastern Fucino Basin (F1–3 in Fig. 1) in June 2015. The major goals of this drilling were to (i) collect a continuous record back to MIS (marine isotope stage) 6, (ii) assess the quality of the lacustrine sedimentary succession, (ii) explore the sensitivity of different paleoclimatic proxies, and (iii) document the presence of widespread tephras useful for regional to extra-regional correlation and their suitability for 40Ar/39Ar dating.

The Fucino Basin (Fig. 1) is one of the largest intermountain depressions formed during the Plio–Quaternary extensional phase along the Apennine Chain (e.g. D'Agostino et al., 2001) within the earlier fold-and-thrust-belt system (e.g. Patacca and Scandone, 2007). Its opening was mainly driven by a major 110–130∘ N trending fault system (Galadini and Galli, 2000; Fig. 1a), which led to the formation of a typical half graben with up to ∼ 900 m of thick Quaternary deposits in the hanging wall of the master faults (Cavinato et al., 2002; Fig. 1b). Prior to the drainage, undertaken first by Romans in 1st century AD and then at the end of the 19th century (e.g. Galadini and Galli, 2001), the Fucino Basin hosted the largest lake (∼ 150 km2, 20 m maximum water depth) of peninsular Italy.

In the past decades, several cores were drilled in the Fucino Basin for scientific and geotechnical purposes (Fig. 1). A preliminary palynological study of the 200 m long GeoLazio core (GL in Fig. 1) led to the conclusion that the upper 65 m of the succession would span the last 130 ka (Follieri et al., 1986). However, the dating of a thick tephra layer at ∼ 101 m depth yielded an age of 540 ± 9 ka (Follieri et al., 1991), hardly compatible with this chronological framework. Later, Narcisi (1994) analysed two foiditic tephra occurring in the GeoLazio and Telespazio cores (TS in Fig. 1), at ∼ 12–14 m depth, which, on the basis of their peculiar glass major element composition, were attributed to the eruptive units Albano 5–7 (Giaccio et al., 2007), from a cluster of Colli Albani eruptions dated between 40 and 36 ka (Giaccio et al., 2009). Finally, the tephrostratigraphic study of a 30 m deep borehole (SP in Fig. 1a) drilled close to the GeoLazio site, and in a comparable range of the isochrones of the Quaternary infill (Fig. 1a), allowed for the recognition of some relevant tephra markers spanning between ca. 18 and ca. 160 ka (Authors' unpublished data; Fig. 3). All of these chronological constraints lead us to a reinterpretation of the chronology of the GeoLazio pollen record, spanning the MIS 1–9 interval, suggesting that the Fucino succession extends continuously at least back to ∼ 540 ka (Fig. 3).

Both cores are composed of grey lacustrine calcareous marl, with a variable proportion of clay. A preliminary 82 m long composite succession is based on visual correlation after core opening and using the line scan images from the Itrax core scanner (Cox Analytical, Sweden) and the Corewall software package (Corelyzer 2.0.1). Core sections that looked disturbed from the coring process were not included in the core composite. Although, theoretically, the drilling depths should have avoided gaps between the individual runs in each hole, overlapping sequences in the lower part of the F3 hole indicated gaps of around 20 cm between the individual runs of the F1 hole due to core expansion after recovery. We therefore assumed 20 cm gaps between the F1 sequences, where there were no overlapping F3 sequences; i.e. for the depths > ∼ 55 m. Based on the core composite stratigraphy, Ca, K, Ti, Rb, and Sr data from XRF scanning were compiled to produce an overview profile.

At least 16 centimetric to decimetric thick and relatively coarse-grained tephra layers were identified along the composite core, several of which are also clearly marked by prominent spikes in K, Rb, and Sr (Fig. 4). A provisional correlation of some of the tephras has been tentatively established based on their stratigraphic order and lithological features (e.g. overall tephra colour, thickness, grain size, assortment of the mineral and lithic components, and colour and shape of the juvenile clasts) compared to the local tephra successions from the Fucino core SP and the chronologically recalibrated GeoLazio core (Fig. 3) and cores from the surrounding basins of Campo Felice (Giraudi and Giaccio, 2015) and Sulmona (Giaccio et al., 2012; Regattieri et al., 2015) (Figs. 3, 4). Although the lack of robust geochemical data renders it premature to propose an age model based on tephrochronology alone, general tephrostratigraphic consistency allows us to ascribe the investigated successions to the MIS 6–1 period. The resulting sedimentation rate would range between ∼ 0.3 and 0.6 mm a-1, with a mean of ∼ 0.45 mm a-1, which is substantially higher than that of the GeoLazio/SP succession (∼ 0.2 mm a-1; Fig. 3).

Geochemical and biogeochemical analyses from core catcher material show significant variability, with major trends consistent between the different proxies and with XRF data for Ca and Ti (Fig. 5). There is a strong (r=0.95) negative correlation between Ca and Ti. Due to this strong anticorrelation and to the fact that Ca and TIC depth series show nearly identical patterns of variation, we assume that both proxies are mainly related by authigenic (i.e. biomediated) precipitation of calcite. The amount of calcite in the sediments depends on lake primary productivity, temperature, the input of Ca from the catchment (e.g. Gierlowsky-Kordesh, 2010; Vogel et al., 2010), and the preservation of calcite as a function of organic matter (OM) decomposition and bacterial CO2 release in the bottom waters (e.g. Müller et al., 2006). Also, TOC is mainly a function of changes in organic production in the lake and of changes in catchment vegetation (Leng et al., 2013), and, along with TIC, it tends to be higher during warmer and wetter periods (i.e. interglacials/interstadials). During these periods, calcite (TIC × 8.33) and organic matter (TOC × ∼ 2.5) derived from TIC and TOC may form up to > 90% of the total sediments (Fig. 5). This suggests the negligible contribution of clastic, minerogenic input from catchment erosion and is consistent with the low Ti concentration (e.g. Vogel et al., 2010). δ18OTIC of most Mediterranean lakes is instead considered a proxy for the balance of precipitation vs. lake evaporation (e.g. Roberts et al., 2008), with higher or lower values related to decreasing or increasing moisture conditions, respectively.

The TOC / TN ratio is often related to the source of the OM and generally reflects the proportion of aquatic (macrophytes and phytoplankton) vs. terrestrial plants, with higher values indicating prevailing allochthon (terrestrial) input for the OM (Meyers and Ishiwatari, 1995), i.e. enhanced development of vegetation in the lake catchment. However, some sections in the core indicate TOC / TN ratios as low as ∼ 2 (Fig. 5). As this is too low for natural substances, selective decomposition of OM could have taken place and may have affected the TOC / TN ratio. In addition, very low amounts of TOC or TN, which are close to the detection limits, may have biased the TOC / TN ratio. TS is related to variations in OM production and in redox conditions within the lake. In our record, the high positive correlation between TS and TOC (r=0.70) and their similar patterns suggest that higher S values are related to the development of anoxic conditions at the bottom of the lake due to increasing productivity and/or lake stratification.

Based on their paleoclimatic/paleoenvironmental significance, the consistent variations in all of the presented proxies (Fig. 5) perfectly match the timescale proposed by the preliminary tephrochronological framework (Fig. 4). Indeed, lower precipitation (higher δ18OTIC), reduced lake productivity (lower TIC, TOC, and TS), contraction of terrestrial vegetation (lower TOC / TN ratio), and increase in catchment erosion (higher Ti) correspond to the colder glacial conditions of the MIS 6 and MIS 4–2 (Fig. 5). During the glacial part of the MIS 5 and the interglacial MIS 5.5 and MIS 1, warmer and wetter climate conditions, promoting primary productivity and vegetation development, are instead apparent. Interestingly, although the temporal resolution of the proxy obtained from the core catcher (δ18O, TIC, TOC, TOC / CN, and TS) is very low (∼ 2 ka, based on the preliminary estimation of the sedimentation rate of ca. 0.45 mm a-1) the interstadial/stadial pattern during the MIS 5 is also recorded by concurrent variations in all the investigated proxies (Fig. 5). This notion is corroborated also by the similarities of the Fucino record with the high-resolution oxygen isotope time series from the Sulmona Basin (Regattieri et al., 2015, Fig. 4). Indeed, although the two series are correlated only by means of the preliminary tephrostratigraphic analyses (Fig. 4), across the four presumably common tephras, the F1–3 δ18OTIC profile seems to show the same pronounced millennial-scale fluctuations documented in the Sulmona record (Fig. 4).

The first multiproxy analyses performed on the newly acquired 82 m long core from Fucino lake sediments, along with the reinterpretation of the previous data, suggest that i.

the entire lacustrine succession in the Fucino Basin extends continuously at least up to ∼ 540 ka and, by extrapolating the sedimentation rate (0.4–0.5  mm a-1) to the maximum depth of ca. 900 m, possibly 2 Ma back ;

In order to explore the full potential of the Fucino succession, further investigations, including detailed tephra analyses, 14C and 40Ar/39Ar geochronology, and high-resolution biogeochemical, stable isotope, paleontological, and paleomagnetic analyses are in progress. Nevertheless, the present preliminary results make the Fucino Basin a good candidate for a deep-drilling project, which could provide one of the longest continuous records for studying, at both orbital and millennial scale, the Quaternary paleoclimatic history of the central Mediterranean area, possibly 2 Ma back.