A high-resolution record of landscape changes and land use over the last 5000 years in western Calabria (S. Eufemia Gulf,southern Tyrrhenian Sea,Italy)

Abstract

Pollen analysis of a marine core collected in the Gulf of S. Eufemia (Tyrrhenian Calabria, Italy) allowed reconstructing the regional changes in vegetation and land use over the last 5000 years. Pollen diagram zonation through Constrained Cluster Analysis highlighted three compositional zones whose boundaries mark the major changes that affected the vegetation structure. A dense forest cover with a few signs of human activities characterized the wide pollen source area from 5055 to 2700 BP (Zone 1). In this period, the Pre-Protohistoric communities were mainly concentrated on the Tropea Promontory where they had a significant local impact. Minor forest rarefactions at 5000–4800 BP, ca. 4400–4000 BP, ca. 3450–3150 BP were correlated to phases of climatic shifts toward aridity. From 2700 to 2000 BP (Zone 2a), a longer and more incisive period of forest decline was connected to a time of aridity that favored the intense activities of an increasing anthropogenic pressure. Indeed, important urban centers developed in the area during the Greek and Roman colonization. Diffuse deforestation and cultivation occurred from 790 BP (Zone 3), enhancing soil erosion and fluvial discharge as testified by the sudden increase in sedimentation rates. This disruption of the slope morphodynamics was connected to the collapse of territorial management following the end of the Western Roman Empire. Compositional Data Analysis, applied to a simplified pollen dataset, highlighted both a negative correlation between Abies and Fagus and a close similarity between the AP/NAP curve and the Axis 1 scores of the Relative Variation Biplot.

Keywords

anthropogenic indicators compositional data analysis deforestation marine core pollen analysis

Introduction

Human-environment interactions around the Mediterranean basin represent a crucial topic widely debated by the scientific community (e.g. Bevan et al., 2019; Roberts et al., 2019; Walsh et al., 2019 and references therein) due to the many environmental/climatic and cultural changes that have affected the central Mediterranean, and particularly the Italian Peninsula, during the Middle-Late Holocene (e.g. Bevan et al., 2019; Giraudi et al., 2011; Mackay et al., 2005). This is a period characterized by mild climate conditions interrupted by brief cool/dry episodes called RCC (rapid climate change) by Mayewski et al. (2004). Among these, the most widely recognized are the 8.2 ka BP (e.g. Rasmussen et al., 2007) and the 4.2 ka BP events (e.g. Railsback et al., 2018) that have been used to mark the boundaries among the Holocene stages (Walker et al., 2018). Despite their wide dissemination and recognition, they proved to be not always synchronous across time and space (e.g. Bini et al., 2019; Di Rita and Magri, 2019; Mayewski et al., 2004; Zanchetta et al., 2016). More recently, pollen data from the Tyrrhenian Sea have indicated that the Holocene has been affected by a millennial-scale climate variability, with a cyclicity of approximately 1860 years (Di Rita et al., 2018).

All these climate variations, together with the progressive increase in human presence since 5000 years BP, have caused major changes in vegetation cover and this topic represents another hotly debated issue (e.g. Di Rita et al., 2018; Ruiz and Sanz-Sánchez, 2020; Stoddart et al., 2019; Woodbridge et al., 2018 and references therein), especially with regard to the intensity and the extension of vegetation changes in the Mediterranean area. For example, timber has always been the economic basis of past populations but the real value of this exploitation is not yet clarified (Harris, 2013b). Certainly, according to Mercuri et al. (2019), the low/high impact of land uses depends on the scale (space) and duration (time) of the human presence/action in a given territory. The human impact may have different effects on the environment, not always detectable over a large area, far from settlements or productive areas.

In this regard, palynology is widely recognized to be one of the most appropriate method for reconstructing past landscapes, assessing climate and anthropic pressure and related vegetation changes (e.g. Mercuri et al., 2010; Woodbridge et al., 2018). To achieve these aims, palynology has to be supported by other methods, closely related to pollen analysis, such as a careful definition of the anthropogenic indicators (Behre, 1981; Brun, 2011 and references therein) and a quantitative analysis of microcharcoals that can help to understand the regime of fires in the past. In addition, the archeological knowledge of the area under consideration is a key aid to understanding vegetation changes. Indeed, many studies in the Mediterranean (e.g. Roberts et al., 2019) and particularly in the central and southern Tyrrhenian areas (e.g. Di Lorenzo et al., 2021a, 2021b; Mensing et al., 2015; Mercuri, 2014; Russo Ermolli et al., 2018; Sadori et al., 2004; Stoddart et al., 2019) have used a multidisciplinary approach in which archeology and palynology support each other in the interpretation of societal and/or climatic dynamics.

Spatial resolution of pollen data and the reconstruction of past environments are dependent on the pollen source area (basin catchment) and the basin size (Jacobson and Bradshaw, 1981), which are strictly correlated. If the sedimentary basin is very large (which means also a wide catchment), such as a marine basin or a big lake, pollen derived from regional sources (several 100 m and more) will dominate the spectra. Conversely, pollen from local and extra-local sources will be under-represented. If the basin (and its catchment) is small, only pollen derived from local (within 20 m) and extra-local (between 20 and several 100 m) sources will enter the basin and local vegetation will be over-represented in the spectra. Therefore, integrating regional pollen data with local pollen data is extremely important in order to assess the complexity of vegetational changes and their possible causes. Sometimes, in local records coming from densely populated territories, it is much more common to recognize the signs of anthropogenic impact (e.g. anthropogenic indicators and microcharcoals), while the regional records allow us to identify more accurately wider environmental/climatic changes. But, the interpretation of pollen data is not always so simplistic because it must be remembered that a changing climate may lead both to directly changing vegetation and land cover composition (e.g. Di Rita et al., 2018; Jalut et al., 2009; Mercuri et al., 2019; Roberts et al., 2019) and to changing societal strategies (e.g. Cremaschi et al., 2016; Gogou et al., 2016; Izdebski et al., 2016; Mazzini et al., 2016; Roberts et al., 2019). In this regard, different societies can respond in different ways to the same specific climate change (Weiberg et al., 2016) or also the same society seems resilient to climate change at one point in time but not at another (Sadori et al., 2016).

Following the above considerations, the main aims of this paper are (1) to provide a regional record of vegetation changes over the last ca. 5000 years; (2) to compare regional and local data from the same catchment area; (3) to clarify the reasons (natural or anthropic) of vegetation changes.

To achieve these goals we have performed a high-resolution pollen analysis of a marine core taken in the Gulf of S. Eufemia (Tyrrhenian Calabria, Italy) whose wide catchment area has been inhabited in continuity since Prehistory by numerous populations in almost all geographical contexts. The obtained pollen results have been compared with the available local data from the same catchment (Di Lorenzo et al., 2021a; Russo Ermolli et al., 2018).

Geological and geomorphological setting

The S. Eufemia Gulf and the Lamezia Plain represent the offshore and onshore western edge of the Catanzaro Trough (E-S Isthmus in Figure 1), a Neogene-Quaternary depression developed in the central Calabrian Arc (Brutto et al., 2014). This depression is filled by a thick (up to 2000 m) Plio-Quaternary sequence of marine to continental deposits, overlying strongly deformed Miocene units (Longhitano et al., 2014; Monaco et al., 1996; Tortorici et al., 1995; Westaway, 1993).

Figure 1.

Geological sketch of the S. Eufemia Gulf area with position of sites cited in the text. Sites located outside this area are reported in the frame at the top of figure, where Calabria is indicated in dark gray.

Starting from the Late Miocene-Pliocene, the area has been dissected by highly dipping NE-SW and N-S oriented normal faults that caused the uplift of the Calabrian block. This uplift, coupled with the sea level changes, caused the formation of marine terraces at different elevation all along the coast and in the inner sector of the S. Eufemia Plain (Filocamo et al., 2009; Loreto et al., 2013; Tortorici et al., 2003). In agreement with Jacques et al. (2001) and Presti et al. (2013), these fault systems has also played a relevant role in the recent seismotectonic processes controlling the geodynamic of the central Calabrian Arc, representing the source of the main destructive earthquakes occurred in the area. Consequently, the continental shelf and upper slope of western Calabria are largely affected by gravitative processes (slides, turbidity currents, etc.), intervening both within canyons and in non-channelized sectors (Gamberi and Marani, 2006).

The Lamezia Plain is drained by two main rivers flowing into the gulf, the Amato and Angitola Rivers (Figure 1), showing a braided pattern in their widespread distal part. The piedmont zone of the plain is characterized by different generations of entrenched alluvial fans, Late Pleistocene-Holocene in age (Ruello et al., 2017; Russo Ermolli et al., 2018). The present coastal plain is the result of alluvial sedimentation and of typical beach and back-ridge sediments and landforms.

Archeological setting

Around the Gulf of S. Eufemia, there are traces of intense agricultural exploitation of the piedmont areas of the Lamezia Plain from the Neolithic (Russo Ermolli et al., 2018) and of the Poro Plateau (Figure 1) during the Eneolithic (Di Lorenzo et al., 2021a and references therein). This area continued to be extensively used for agricultural and grazing purposes in the Bronze Age, when the Punta di Zambrone coastal village (Recent Bronze Age) became an important center for the trade exchanges with the Aegean communities (Jung and Pacciarelli, 2021). First with the emergence of Iron Age proto-urban villages (Matarese, 2017; Pacciarelli, 2001; Pacciarelli et al., 2017) and then with the Greek colonization (colonies of Hipponion − seventh cent. BC, Medma − sixth cent. BC, Terina − sixth-fifth cent. BC), the exploitation of the fertile promontory intensified (Russo Ermolli et al., 2018 and references therein).

The gradual romanization of the area (third-second cent. BC) led to a new agrarian landscape organization with the emergence of varies colonies (Valentia, Tempsa, Nocera Terinese), the construction of important roads (via Annia Popilia) and a progressive increase of large villas (first-second cent. AD) both in internal and coastal position (Sangineto, 1994, 2013). During the Roman Period, intensive commercial activities between the Ionian and Tyrrhenian coasts were facilitated by the Isthmus of S. Eufemia-Squillace (E-S Isthmus in Figure 1), the shortest path and lowest isthmus (260 m a.s.l) of Calabria (Givigliano, 1978). A century after the end of the Western Roman Empire, in the sixth cent. AD, the ruralization of the territory increased even if the papal influence certainly guaranteed the privileged development of Tropea and Nicotera that became important port centers progressively eclipsing the ancient Roman harbor of Valentia (Sogliani, 2012).

A large exploitation of the landscape is also attested during the Lombard dominion (sixth-seventh cent. AD), when timber from the Calabria mountains was used by the Pope for the Patrimonium Sancti Petri (Gregorio Magno – Registrum Epistularum, AD 599). In the Byzantine period, rural villages, developed around fortified sites, are attested in the areas immediately north and south of the Poro Plateau (Di Muro, 2011). Moreover, this territory assumed military and defensive characteristics, which increased with the advent of the Normans in the second half of the 11th cent. AD.

With the spread of Greek orthodoxy, important monastic centers developed and it was around the monastery named after S. Eufemia that the inhabited pseudonym arose, as it happened in many other places in Calabria. During the Swabian dominion of Federico II (13th cent. AD), the medieval settlement of Vibo Valentia (attested from the 10th cent. AD) increased in importance with the built of an imposing castle. The city, now called Monteleone, will continue to grow even in the Angevin age (Sogliani, 2012).

Vegetation and climate setting

A Mediterranean climate characterizes the Calabrian region, with different microclimates based on altitude and proximity to the coast. Specifically, the catchment of the S. Eufemia Gulf includes different altitudinal belts typified by different climate and vegetation. The higher altitudinal belts, reached on the Sila Massif, have a cold-temperate climate type with average annual temperatures between 7°C and 12°C and average annual rainfall between 1000 and 1100 mm and 1400 and 1700 mm (https://www.parcosila.it). In the lower altitudinal belts, such as the Lamezia Plain, the climate is Mediterranean with hot summers, an average annual temperature of 16.3°C and an average annual rainfall of 1046 mm (https://it.climate-data.org/europa/italia/calabria-451/).

In terms of vegetation cover, turkey oak and beech forests, that prefer deep, clayey, and acidic soils, cover the montane belts of the Sila and Coastal Range where mowing and grazing grasslands are also present. At higher elevations, beech forests often host a relict population of fir (Abies alba Mill.; Bernardo and Gangale, 2017). The most characteristic tree species of the Sila Massif is Pinus nigra J.F. Arnold, of which the oldest individuals form the so-called “Giants of Sila” reserve (Bernardo and Gangale, 2017).

The coast is dominated by pine and oak forests with cultivated and grazing lands. Typical species include: Olea oleaster Hoffmanns & Link, Pistacia lentiscus L., Pistacia terebinthus L., Laurus nobilis L., Arbutus unedo L., Myrtus communis L., and Briza maxima L. (Klee et al., 2021). After the reclamation from the swamps carried out in the first half of the last century (http://www.area.cs.cnr.it/imseb/malaria/bonifica/), the plain has been used for intensive cultivation with fruit and olive trees (Caridi et al., 1996).

Material

This research is focused on the analysis of part of a sedimentary core (C4 in Figure 1), collected in the outer continental shelf of the S. Eufemia Gulf (Lat 38°46.609′N; Long 16°10.007′E), at 82 m water depth. This core, that recovered a stratigraphic succession of 4.18 m beneath the seafloor (bsf), was studied by Cosentino et al. (2017) through a multidisciplinary approach including benthic foraminifera, sedimentology, geochemistry, tephrostratigraphy, and radiocarbon analyses. The latter indicated that the age of the cored succession spans from 11.1 ka to the Present. An abrupt faunal turnover occurring at 296 cm depth was considered coincident with an erosional surface that marks a stratigraphic gap of about 3000 years, between ca. 8000 and 5000 BP. A similar gap was also documented in other Mediterranean areas (i.e. the Gulf of Taranto, Pepe et al., 2014) and was explained by a major change in the deep-water circulation pattern along the eastern Tyrrhenian margin during the Holocene climatic optimum. This might have caused, in turn, non-deposition and/or erosion at the seafloor, due to the activity of strong bottom currents, under relatively low sedimentation rates (Cosentino et al., 2017).

Sedimentological analysis of C4 core, performed by Cosentino et al. (2017), allowed distinguishing three main facies associations (A, B, C in Figure 2). Facies A (from bottom of core to 260 cm bsf) consists of poorly sorted sandy silt with bioclasts and lithoclasts. Two erosional surfaces are present at 400 and 296 cm depth. At 281 and 275 cm, two thin beds of reworked volcaniclasts also occur. Facies B (from 260 to 200 cm bsf) is made of relatively homogeneous, bioturbated, very poorly sorted clayey sandy silt. At 226 cm bsf, a few centimeters thick level of reworked white pumices was correlated with the well-known Pompeii eruption of AD 79. Facies C (from 200 cm bsf to the top) consists of grayish poorly sorted mud (clayey silt-clayey sandy silt), with interbedded lenses rich in bioclasts, lithoclasts and volcaniclasts. The sandy fraction is generally fine to very fine-grained.

Figure 2.

Core C4: sedimentology and dating (modified after Cosentino et al., 2017).

For the purposes of the present study, pollen analysis was limited to the core interval above the erosional surface at 296 cm depth, which represents a continuous sedimentary record covering the last ca. 5000 years.

Methods

Radiocarbon dating and age model

The chronological framework of C4 core, already established by Cosentino et al. (2017) on the basis of six radiocarbon analyses, two tephra layers, and two biostratigraphical markers, was enriched with two further ¹⁴C analyses performed in the upper interval of the core, object of the present study. In particular, Nucula remains were collected at 200–202 and 252–254 cm depth and pretreated at iCONa (Isotopic mass spectrometry laboratory) of the Department of Environmental, Biological and Pharmaceutical Sciences and Technologies of University of Campania. Therefore, radiocarbon dating was performed by Accelerator Mass Spectrometry using the dedicated beam line of the HVEE 3 MV Tandem accelerator installed at the INFN-LABEC laboratory in Florence (Chiari et al., 2021). All the radiocarbon ages (both old and new) of the considered core interval (296–0 cm depth) were calibrated using the marine 20.14C curve (Heaton et al., 2020) with the program CALIB REV8.2 (Stuiver and Reimer, 1993). An age-model based on linear interpolation between median probabilities of dated levels was constructed with MATLAB2019.

Pollen analysis

Pollen analysis was undertaken on 75 silty samples collected from the C4 core interval 296–0 cm depth. About 5–6 g of sediment were treated with chemical (HCl 20%, HF 40%, hot HCl 10%) and physical (10 μm ⩽ sieving ⩽ 200 μm, ZnCl₂ floating or LST Heavy Liquid at density 1.8) procedures in order to concentrate pollen grains in the residue. One Lycopodium tablet was added to each sample in order to calculate pollen and microcharcoal concentration. Determinations and counts were carried out under a light microscope at 400×, 500×, and 1000× magnification, with the support of pollen atlases (Beug, 2015; Reille, 1992, 1995) and reference pollen material (DiSTAR of University Federico II di Napoli; ISEM of Université de Montpellier). Due to the sieving adopted during sediment treatment, only microcharcoal fragments ranging between 11 and 200 μm are found in pollen slides. Microcharcoals consist in black particles with sharp edges, completely opaque and angular (Conedera et al., 2009; Tinner and Hu, 2003). In the present study, following Sadori et al. (2004), we counted microcharcoal particles ⩾20 µm, considering the smaller ones as representing the background signal of regional fires. Due to the general low pollen concentration of marine sediments, a counting of 150 grains/sample (at least) was considered valuable when a diversity of 20 taxa was achieved.

A detailed pollen diagram was computed with percent variation of all taxa plotted against depth. Spores, indeterminate grains, dinoflagellates, and microcharcoals were excluded from the pollen sum for the calculation of AP and NAP taxa percentages. Marine cores are often rich in Pinus pollen due to its morphology that makes it easily transportable by wind over long distances; for this reason, this taxon is sometimes excluded from the AP sum so as not to mask possible changes in vegetation. Contrary to this commonly used procedure, Pinus was not excluded from our AP sum because its quantity, not particularly abundant, proved to have a negligible influence on the rate of other APs.

In addition, a synthetic pollen diagram was computed through TILIAGraph (Grimm, 2004) with taxa group percentages (see Table 1) plotted against age. The used chronological model (calibrated age BP and BC/AD) was based on linear interpolation between dated levels.

Table 1.

Composition of taxa groups as plotted in the synthetic pollen diagram.

Taxa/Groups	Group composition	Ecological meaning
Montane forest	Abies	cool-humid
Montane forest	Fagus	cool-humid
Pinus	Pinus spp.	none
Deciduous forest	Quercus, Carpinus, Ostrya, Tilia, Acer, Ulmus, Fraxinus ornus, Hedera, Platanus	temperate
Wet Woodland	Alnus, Corylus, Populus, Fraxinus, Salix	humid soil/waterlogged
Marsh/Water plants	Cyperaceae, Typha, Lythrum	humid soil/waterlogged
Mediterranean forest	Quercus ilex, Forsytia, Phillyrea, Pistacia, Cistus, Ziziphus, Coriaria mirtifolia, Ericaceae, Tamarix, Ficus	warm-dry
Arboreal Pollen (AP)	All tree and shrub taxa	forest cover rate
Non Arboreal Pollen	All herb taxa	open landscape – dry
Tree crops	Olea	crops
Tree crops	Juglans, Vitis, Castanea
Anthropogenic herbs	Cereals
	Rumex, Plantago, Urtica pilulifera, Urticaceae, Lamiaceae, Apiaceae, Brassicaceae, Mercurialis, Artemisia	ruderal
	Cichorieae, Asteroideae, Ranunculaceae, Fabaceae	grazing
Coprophilous fungi	Various spp.	grazing
Monolete spores	Various spp.	fluvial discharge / erosion
Trilete spores	Various spp.	fire / erosion
Microcharcoals	⩾20 µm	fire / erosion

Compositional data analysis

Compositional data analysis (CoDA) methods adopted in this work, and carried out with MATLAB2019, include Constrained Cluster Analysis (CCA) (Grimm, 1987) based on Aitchison distance performed to obtain compositional zones (Di Donato et al., 2009), Relative Variation Biplot (RVB), which are based on singular values decomposition of log centered data (Aitchison and Greenacre, 2002) and Geometric Mean Bar Plot (GMBP) (Martín-Fernández et al., 2015). Short remarks on these techniques can be found in Di Donato et al. (2022a). We considered both covariance and form RVB, which provide a better representation on variables and objects, respectively.

Coherently with the CoDA balances approach (Egozcue and Pawlowsky-Glahn, 2005), we considered an ilr variable contrasting arboreal and non-arboreal pollens, expressed as ${ilr}_{AP / NAP} = \sqrt{0.5} \times \log (A P / N A P)$ as a counterpart of the classical AP/TOT. An ilr variable was also constructed to describe the logratio variation of Abies and Fagus, expressed as: ${ilr}_{Ab / Fa} = \sqrt{0.5} \times \log (A b i e s / F a g u s)$ . Following Daunis-i-Estadella et al. (2011), these additional variables were added ex-post in the RVB. Finally, we performed an R mode clustering of the parts of the pollen assemblages, which was based on Ward’s method applied to the covariance CoDa biplot loadings (Di Donato et al., 2022b). It is worth note that the covariance biplot loadings provide least squares approximation of the variation matrix, which for a D parts compositional data set is a square matrix expressed as: T_D,D with $t_{i j} = var (\log \frac{x_{i}}{x_{j}})$ (Aitchison, 1986). The Ward algorithm applied on these loadings minimizes at each step the subcomposition variabilities, coherently with the definition of variation matrix given above.

The original dataset includes several rare taxa and a quite large amount of zero values. By converse, CoDA methods require log ratio transformations that can be applied only to strictly positive data. Consequently, the CoDA was performed on 16 parts representing the most abundant/significant taxa or group of taxa which were obtained by means of amalgamation (Aitchison, 1986; Greenacre, 2020) of rarer taxa. Afterward, the zero values still present in the dataset were replaced following Martín-Fernández et al. (2003). Amalgamation was also employed to construct the ${ilr}_{AP / NAP}$ variable described above.

Results

Radiocarbon dating and age model

Radiocarbon analysis indicates that the age of the core interval 296–0 cm bsf, studied through pollen analysis, starts from about 5055 cal yr BP up until Present. The radiometric ages obtained in this portion of the core are summarized in Table 2. The age/depth model indicates that the apparent sedimentation rates show a sudden increase from c. 200 cm (Figure 3). Two further chronological constraints, based on biostratigraphical markers (Cosentino et al., 2017), fall in proximity of the segment that connects the last dated level with the Present, enhancing the validity of the reconstructed age model.

Table 2.

Results of radiocarbon dating. Calibration data set: marine 20.14C (Heaton et al., 2020).

Depth (cm)	Lab code	Material	Radiocarbon age yr BP	cal yr BP	cal yr BP	Median probability cal yr BP	Reference
Depth (cm)	Lab code	Material	Radiocarbon age yr BP	1 sigma	2 sigma	Median probability cal yr BP	Reference
160–162	DSH5760	Cerastoderma sp.	994 ± 23	386–508	306–541	442	Cosentino et al. (2017)
200–202	Fi4336	Nucula sp.	1430 ± 15	751–882	686–927	813	This paper
226–228	DSH5757	Nucula sp.	2194 ± 32	1540–1686	1455–1765	1612	Cosentino et al. (2017)
252–254	Fi4337	Nucula sp.	2750 ± 55	2174–2379	2085–2504	2293	This paper
288–290	DSH7031	Nucula sp.	4777 ± 27	4774–4937	4674–5022	4843	Cosentino et al. (2017)

Figure 3.

Age-depth model for C4 core. Red asterisks indicate the dated levels and their age (median probability) in cal ka BP. Dark blue numbers indicate sedimentation rates (cm/a) between dated levels. At the right side is indicated the position of two biostratigraphical markers that define the Little Ice Age interval (after Cosentino et al., 2017).

The pollen diagram

The analyzed samples were quite rich in pollen grains apart from the most recent levels that resulted very poor or barren. Pollen sums (AP + NAP) range from 150 to 351 grains; 72 taxa were identified; concentration values range from 12,000 to 7000 grains/g of sediment. Three main compositional pollen zones were identified in the detailed diagram on the basis of CCA (Figure 4a and b). The two uppermost samples collected at 6 and 64 cm were excluded from the zonation and from the detailed diagram yet were included in the synthetic diagram in the discussion section:

Figure 4.

(a) Detailed pollen diagram with Arboreal Pollen and AP/NAP percentages plotted against depth (cm). CCA (Constrained cluster Analysis) defines three Zones (1, 2, and 3); Zone 2 is further subdivided in two subzones (2a and 2b). Calibrated ¹⁴C ages (in bold) are expressed as median probability in cal yr BP. The investigated portion of the log is displayed on the left. (b) Detailed pollen diagram with Non Arboreal Pollen and NPP percentages plotted against depth (cm). Dashed lines indicate the separation among the Zones defined through CCA (see Figure 4a). Calibrated ¹⁴C ages (in bold) are expressed as median probability in cal yr BP. The investigated portion of the log is displayed on the left.

Zone 1 (292–260 cm)

This zone is represented by 17 samples. In these layers, the AP are always over 70% indicating a rather forested environment mainly dominated by deciduous Quercus, Quercus ilex, Abies, and Pinus. In particular, Quercus ilex presents values around 15% all along the zone a part from a lower amount (6%) at 282–284 cm and higher values (20%) at 270 cm. The other Mediterranean taxa show a very low (1–2%) and sporadic presence all over the zone. Only the Ericaceae family is well represented, even though the percentages never exceed 5%, with an increase in correspondence of the Q. ilex decrease at 282–284 cm. With regard to the deciduous forest, Quercus is the most representative taxon that seems to possess stable percentages all over the zone (20–22%), even though various rapid decreases are evidenced at 288, 284, and 270 cm. From this last level, a slight decrease of deciduous Quercus is evident, which continues into Zone 2. All other deciduous trees, including the wet woodland elements, present low values all along the zone, a part from Carpinus and Ulmus that are more continuously represented. Pinus percentages are very high all over the zone even though they show a slight decrease at 272–270 cm. The mountain trees are well represented by Abies that shows high values (up to 20%) all along the zone, even if an incisive decrease is recorded around 270 cm. Fagus is scarce but shows a slight progressive increase from 274 cm. Concerning the herbs, the Cichorieae dominate the assemblages with a value of c. 10%, excluding the very low amount recorded at 272 cm. This subfamily is followed by the Asteroideae and Poaceae. To underline the presence of cereals at 280 and 260 cm. All other herbs are poorly represented over the zone. Marsh taxa are mainly represented by Cyperaceae, of which only a few are present throughout the entire zone. Monolete and trilete spores are abundant over the entire zone, with fluctuating values, whereas the dinoflagellates are more abundant (up to 30%) in the lower part of the zone and decrease down to values below 10% in the upper part. Coprophilous fungal spores usually occur with almost stable values (5–7%), even if they slightly increase in the upper part of the zone. High percentages of microcharcoals are present in almost all the zone, but show three negative peaks at 286, 282, and 270 cm. It is interesting to note that the first and second microcharcoal decrease take place in correspondence with the increase of deciduous Quercus, while the third decrease occurs in correspondence with an increase in Q. ilex.

Zone 2 (258–200 cm)

This zone is subdivided, at 241 cm, in two subzones 2a and 2b, represented by 9 and 21 samples, respectively. In subzone 2a, AP values oscillate around 70%; from 240 cm (2b) they stabilize at 80%, while in the upper part of the subzone (from 202 cm) a rapid decrease is recorded, reaching values of c. 58% in the uppermost sample. Quercus ilex shows a slight decrease in subzone 2a (250–246 cm), then an increase in 2b, with percentages reaching peaks (c. 20%) at 236 and 216 cm. Afterward it records a constant decrease toward the end of the subzone. In correspondence with the Q. ilex decrease, an increase in the Ericaceae family occurs, with some Pistacia. To underline a stable presence (with low percentages) of Olea between 244 and 230 cm. Other Mediterranean elements show very low values (less than 1%). The same trend of the Q. ilex curve seems evident also for deciduous Quercus, though it shows two peaks at 230 (45%) and 224 cm (c. 38%). Carpinus remains at constant low values throughout the zone, whereas Ostrya is continuously present in the bottom part of the zone up until 224 cm, then it almost disappears. Other deciduous and wet woodland tree taxa show low percentages and are not representative of this assemblage. Pinus percentages are stable but reduced with respect to Zone 1, while Abies shows stable values (10%) and achieves a peak at 220 cm depth. Fagus continues to develop in almost all the zone, though its decrease is evident from 206 cm. Concerning the herbs, the opposite trend of Cichorieae and Asteroideae curves with respect to the AP percentages is very significant. Other herb values do not seem to change with respect to Zone 1, even if cereals (always in low percentages) are almost constantly present all over the zone. Marsh assemblage is dominated by fluctuating values of Cyperaceae, higher than in the previous zone. Monolete spores show stable values all along the zone, while trilete spores register a progressive increase, reaching c. 42% in the upper part of the zone. Coprophilous fungal spores are stable up until 230 cm, then they show a rapid increase followed by a progressive decrease interrupted by two positive peaks at 220 and 202–200 cm. Percentages of dinoflagellates oscillate around a mean of 10%, while microcharcoal values are rather high, with some negative peaks, and show a general increase with respect to Zone 1.

Zone 3 (198–122 cm)

This zone is represented by 24 samples. AP percentages show a progressive decrease reaching the lowest value of the entire diagram at 130 cm (40%). Q. ilex continues its progressive and definitive decrease up to nearly disappear from 160 cm, while Olea confirms its discontinuous presence all along the zone. The other Mediterranean taxa seem to progressively appear, such as Cistus, which is very poorly represented in the other zones. Ericaceae do not seem to change their percentage with respect to Zone 2 even if they nearly disappear in the upper samples. Deciduous Quercus continues its decrease, but shows a peak at 160 cm depth. Other deciduous taxa are very poorly represented, such as also all wet woodland elements. Pinus maintains constant values, while Abies almost disappears in this zone followed by Fagus. On the contrary, the most representative herbs (Poaceae, Cichorieae, and Asteroideae) increase progressively. Cereals are almost always present but with low percentages. With regard to marsh taxa, only the Cyperaceae family is representative of this assemblage with an increase at 130 cm. Trilete spores and coprophilous fungal spores seem to increase in this zone while monolete spores maintain the same value as in Zone 2. Concerning microcharcoal, in the lower part of the zone they are usually abundant (c. 85%), while from 140 cm they reach lower percentages.

Compositional data analysis

The compositional changes recorded across the succession are synthetized by the RVB and the GMBP (Figures 5 and 6). The variance accounted by the first 2 axes in the RVB is around 50% of the total. This implies a certain degree of approximation in the relationships evidenced by the analysis. In the RVB are also shown the two ilr coordinates related to the Abies/Fagus and the AP/NAP variability. As regards the “classical” AP/TOT and the ilr_AP/NAP, their interpretation can be expected to be generally not too different. However, their relationship is obviously not linear, since the ilr_AP/NAP is basically a logit function, and their nonlinear relationship is particularly evident for high and low AP values. The ilr_AP/NAP is more coherent with the CoDA approach applied in the data analysis and as a real variable obtained from an original closed dataset, may be more suitable for, as an example, regression applications. A more in-depth discussion of this topic is beyond the scope of this article; anyway, the 70% value in the AP/TOT, a threshold allowing open from forested landscapes to be discriminated, corresponds in the ilr_AP/NAP to a 0.6 value.

Figure 5.

(a) Covariance relative variation biplot. A confidence ellipse is drawn for each group of samples defined by the CCA. The ilr_AP/NAP and ilr_Ab/Fa are added as supplementary variables. Note that the latter is almost parallel to the link between Abies and Fagus, also represented in red, which also represents the Abies/Fagus logratio variability. HER: Herbs; TRI: Trilete spores; MED: Mediterranean taxa; JCV: Juglans, Castanea, Vitis; WAT: Water plants; MON: Monolete spores; WW: Wet Woodland taxa; DEC: Deciduous forest taxa; (b) Axis 1 and 2 factor scores along the pollen intervals are compared with ilr_AP/NAP and ilr_Ab/Fa, respectively.

Figure 6.

(a) Constrained Cluster Analysis, y scale in cm; (b) R-mode cluster analysis; both clusters are divided into two subclusters (1a–1b and 2a–2b), Within cluster 1, Abies is a single element subcluster; (c) Geometric mean bar plot, showing the distribution of the part within the Compositional Zones defined by the CCA. The position along the y axis of the different parts is related to their overall abundance expressed in terms of geometric mean (as an example, Pinus is characterized by high mean abundance and low variability; Fagus is characterized by lower average abundance and higher variability across the intervals).

In the covariance biplot (Figure 5a) the Axis 1, which accounts for 37% of total variability, contrasts woody taxa, located at its positive side, with herbs, Ericaceae and Olea located at the negative side. In can be noted that the ilr_AP/NAP vector is oriented toward the positive side of Axis 1 with a very low angle (Figure 5a). Overall, this indicates that the main source of variance in the dataset can be related to climate and in particular to humidity rate that characterize the difference between trees (high humidity) and herbs/Mediterranean plants (low humidity). Accordingly, factor scores for Axis 1 in the form biplot, which are correlated with ilr_AP/NAP with r = 0.81, can also be interpreted as a humidity/aridity index. In fact, the two curves, side by side in the graph in Figure 5b, show a very similar trend and highlight two maxima at around 3.7 ka and between 2.0 and 1.0 ka BP with a decreasing phase in between, and a marked decrease after 1.0 ka BP.

It can be also noted that woody taxa such as Abies, DEC, Quercus ilex, and Fagus, which in the R-mode cluster analysis are grouped in the cluster 1 (Figure 6b), also spread along Axis 2, in relationship with a secondary source of variance (about 13% of total). The highest logratio variability, which strongly influence the Axis 2 factor scores, involves Abies and Fagus and seems largely independent from ilr_AP/NAP (the correlation coefficient between ilr_AP/NAP and is ilr_Ab/Fa r = 0.21). The low abundance values recorded in general for Fagus, which in the interval 3 also correspond to low Abies abundances (Figure 6c), suggest some caution in the interpretation of the ilr_Ab/Fa, as small variations in the counts can result into quite large variations in their ratio. However, woody taxa show a slightly distinct distribution. As shown by the GMBP (Figure 6c), Abies has well above-average abundance in Zone 1, average abundance in Zone 2 and below-average abundance in Zone 3. In fact, within the cluster 1, this taxon joins the group at high distance, as a single element subcluster. Quercus ilex and DEC have above-average abundance in Zones 1 and 2 and below-average abundance in Zones 3, while Fagus has slightly below-average abundance in Zones 1 and 3 and above-average abundance in Zone 2 (Figure 6c). Axis 2 factor scores and ilr_Ab/Fa values are strictly correlated (with r = 0.86) and their curves, side-by-side in Figure 5b, show higher values between 4.4 and 3.8 ka BP. The lowest values are recorded from 2.3 to 1.8 ka BP. The closeness of Herbs, Ericaceae and Olea variable points on the negative side of Axis 1, indicates a reduced variability of their mutual logratios. These taxa, which in the R-mode cluster analysis are grouped in the cluster 2a (Figure 6b), have above-average abundance in Zone 3 (Figure 6c). Quercus, Pinus and WetWood, included in the cluster 2b, are located closer to the center of the biplot, showing a lower variability across the Zones.

Discussion

In order to better visualize the vegetational changes along the analyzed record, the discussion will be based on the synthetic diagram of Figure 7, in which the taxa proposed in Table 1 are plotted against the reconstructed age and associated archeological periods. Any age model involves the possibility of chronological imperfections and thus the identification of a certain event should necessarily be regarded as falling “around” the reconstructed age.

Figure 7.

Synthetic pollen diagram with taxa or groups of taxa (see Table 1) plotted against age. Gray strips correspond to phases of AP decrease. The ilr_ap/NAP curve is side by side to the AP/NAP curve: the value 0.6 corresponds to 70% of AP. Also the ilr_ab/Fa curve is included in the frame in order to show its fluctuations with respect to the AP/NAP oscillations. At the right side, ages expressed in yr cal BC/AD are flanked by the division into archeological periods; E: Eneolithic; B: Bronze Age; I: Iron Age; G: Greek Period; R: Roman Period; M: Medieval Period; Md: Modern Age.

The classical AP/NAP curve is flanked by the ilr_ap/NAP curve and by the ilr_ab/Fa curve for the purpose of showing their trend with respect to the AP/NAP curve.

From 5055 to 2700 cal yr BP (Zone 1)

From the base of the diagram (ca. 5055 BP) up to ca. 2700 BP, the high AP percentages fluctuating around 80% indicate that a mostly closed landscape characterized the pollen source area, likely represented by the wide catchment basins of the Amato and Angitola Rivers and of the other minor watercourses flowing into the S. Eufemia Gulf (Figure 1). Within this interval, distinct above-average-abundance of the tree taxa that make up cluster 1 are recorded (Figure 6c), together with below-average-abundance of Mediterranean and herb taxa included in cluster 2a. This period of dense forest development, associated with mild and wet winters in northern Europe (Olsen et al., 2012) and in the southern Mediterranean (Magny et al., 2013), is marked by some minor lowering of AP percentages. A first slight decrease in AP is centered at around 4900 BP (Figure 7) and is marked by a negative peak of Abies followed by a decrease of the oak forest. The concomitant slight increase in Mediterranean plants and herbs suggests that this forest decline could be due to a local phase of increased aridity probably corresponding to the one recorded around the same time at Lake Trifoglietti (Joannin et al., 2012), located somewhat further north in the Coastal Range (Figure 1). The local character of this phase, which probably affected mainly mountainous areas, is suggested by its absence in the Gulf of Gaeta (Figure 1) where this period shows a dense forest development associated with a phase of NAO positive values (Di Rita et al., 2018). The anthropogenic activity during the Eneolithic period (5600–4100 BP) was mainly concentrated on the Poro Plateau (Figure 1) where deforestation, pasturing and cereal crops were clearly highlighted by local pollen data (L1 core in Figure 8; Di Lorenzo et al., 2021a). As common in marine sediments, the occurrence of anthropogenic indicators is very scarce in the C4 core where just some coprophilous fungal spores coupled with herb increase could be considered evidence of inland grazing activities (Cugny et al., 2010 and references therein). The certain absence of fir on the Poro Plateau (maximum altitude 600 m) further supports the climatic character of this phase: people living there could not have had an impact on the fir forests likely located on the Sila and Serre Mountains, as today (Pignatti, 1982), and on the Coastal Range, as shown by the Trifoglietti pollen data (Joannin et al., 2012).

Figure 8.

Comparison of marine (C4) and continental (SL1, L1) data from the S. Eufemia Gulf area. For core location, see Figure 1.

A second slight decline of the AP curves is evident between ca. 4400 and 4000 BP, and is marked by a negative peak of Abies at ca. 4100 within a general stability of the other forest types. The increase in herbs is coupled with the first recovery of a few cereal grains and a stable presence of coprophilous fungal spores, both indicative of anthropogenic activities. Also during this period, the intensive agricultural and grazing practices were concentrated on the Poro Plateau where the final natural filling of the Lacco pond (L1 in Figures 1 and 8) was accelerated by the climatic interference of the 4.2 event (Di Lorenzo et al., 2021a). This widely recognized climatic shift is not very clearly recorded in the C4 core as in other pollen records from the Mediterranean area, located between 39° and 43° of latitude (Di Rita and Magri, 2019). At Lake Trifoglietti, a dry episode was particularly accentuated from 4000 to 3600 cal BP (Joannin et al., 2012) while in the Gulf of Gaeta, the 4.2 event is visible and associated to a slight negative North Atlantic Oscillation (Di Rita et al., 2018).

Thereafter, a general revival of the forest cover occurs, preceding a new phase of major Abies reduction at around 3300 BP (3150–3450 BP; Figure 7). This phase, as the 1 at 4900 BP, can hardly be attributed to the impact of Bronze Age communities that were essentially distributed on the Tropea Promontory. A further climate shift toward aridity is recorded around 3500–3800 BP at Lake Trifoglietti where the Abies decline is coupled with the increase of Fagus. This negative correlation between Fagus and Abies is evident all along the C4 record, as indicated by the ilr_Ab/Fa curve (Figure 7), and can be explained with the different response of these taxa to water stress (Rita, 2014). In the Gulf of Salerno, a progressive deciduous forest decline starts from this moment (Di Donato et al., 2008; Russo Ermolli and di Pasquale, 2002), and a drastic Abies decline, already started at ca. 5000 BP, occurs.

This trend toward aridity during the second half of the Holocene, which greatly disadvantaged the fir tree, was also highlighted by pollen-inferred precipitation patterns in the south-central Mediterranean showing dryer conditions from 6500 to 4500 cal yr BP, and a period post 4000 cal yr BP with dry conditions (Peyron et al., 2013, 2017). The same results emerge from a study by Magny et al. (2013) that recognizes arid conditions during the Late Holocene in Mediterranean contexts located South of 40°N.

From 2700 to 790 cal yr BP (Zone 2)

Starting from ca. 2700 BP and until about 2000 BP (Zone 2a), AP fluctuates around 70% indicating a more open landscape due to a general decline of all forest types coupled with an increase in herbs and pines. This forest decline is probably the response to the general aridification of the south-central Mediterranean in the second half of the Holocene, whose effects were already visible in the preceding period. It is indeed within this time interval that Mayewsky’s RCC between 3500 and 2500 ka (Mayewski et al., 2004) falls.

The clear climatic character of this forest decline is also attested in the Gulf of Gaeta, where a marked AP reduction, following the Bond event 2 (Bond et al., 1997), is coupled with a negative NAO oscillation (Di Rita et al., 2018), in the Gulf of Salerno, where the lowering in deciduous trees is coupled with the increase in Mediterranean taxa (Russo Ermolli and di Pasquale, 2002) and in south-eastern Sicily where this phase is correlated to the so-called “2.8 event” (Michelangeli et al., 2022).

An additional response to this phase of increased aridity in the S. Eufemia area can also be seen in the rise of beech, beginning to replace fir in the montane forest association. The ilr_ab/Fa curve shows a negative trend that starts at 2650 and culminates in a negative peak at c. 1900 BP (Figure 7).

The concomitant rise in coprophilous fungal spores and cereals suggest assuming a permanent exploitation of the land between the end of the Iron Age and the beginning of the Roman period that was probably favored by the climatic/environmental opening of the forest cover. Starting from c. 2500 BP, local reduction of wet environments, attested in the Lamezia coastal plain (Russo Ermolli et al., 2018), caused the decrease of the wet woodland and the resulting increased availability of arable land (SL1 in Figure 8).

Rising anthropogenic pressure is clearly evidenced by the archeological record that shows the development of proto-urban centers during the Iron Age, the emergence of important towns during the Greek period and the new agrarian landscape organization during the romanization of the area (third-second cent. BC).

Between 1980 and 790 BP (Zone 2b), which correspond to the Roman and Early Medieval periods, the forest cover shows a phase of stability around 80–90% probably due to a climatic improvement that especially favored the deciduous forest development. In fact, this interval is characterized by an above-average-abundance of woody taxa belonging to cluster 1a (Figure 6c). Increased forest cover is also attested in both the Salerno and Gaeta Gulf records during the same time interval. Tree ring–based reconstructions of central European summer precipitation and temperature over the past 2500 years indicate the occurrence of wet and warm summers during periods of Roman and Medieval prosperity (Büntgen et al., 2011). Along the same period, Abies shows a new and definitive decline after its last peak at 1400 BP. A similar decreasing trend of Abies is recorded at the Trifoglietti site (Joannin et al., 2012) and in the Salerno Gulf (Russo Ermolli and di Pasquale, 2002).

Such a forest cover stability in central and southern Italy, during a period of high human pressure and intensive exploitation, demonstrates that the Romans managed their woods very cautiously. Uncontrolled deforestation started after the decline of the Western Roman Empire, in agreement with Harris’ remarks on the history and economy of woods and timber in the ancient Mediterranean (Harris, 2011, 2013a, 2013b). In contrast to this general trend, the decrease in AP percentages recorded at Lamezia from 2200 BP (SL1 core in Figure 8; Russo Ermolli et al., 2018) could indicate, in addition to natural limitations, either the exploitation of local resources or more extensive timber production at a regional scale. In particular, the greater availability of firs on the Calabria reliefs, with respect to other regions of southern Italy, was probably the reason for this local over-exploitation. It is rather well known that fir was a favored timber for shipbuilding (Neapolis shipwrecks; Allevato et al., 2010) and furniture (Herculaneum; Moser et al., 2018).

From 790 cal yr BP (Zone 3)

Starting from 790 BP, a general drastic decrease of forests starts and continues rapidly, reaching values around 40-50% (open landscape, sensu Favre et al., 2008). This compositional change is clearly evident in both the RVB, in which samples belonging to Zone 3 are located on the negative side of Axis 1 (Figure 5a), as well as in the GMBP (Figure 6c), where above-average-abundance is recorded for taxa belonging to cluster 2a, and below-average-abundance for taxa belonging to cluster 1. At the same time, apparent sedimentation rates in the C4 core start to increase (see Figure 3), as also evidenced by the common presence of sand lenses (Figure 2). The progressive rise in spores and microcharcoals (Figure 7) suggests the beginning of a phase of abundant fluvial intake and flood events. Thus, deforestation induced enhanced soil erosion and detrital arrivals to the sea through fluvial discharges.

The increase in sediment flux rates has been a common feature throughout the Mediterranean area when a substantial and sustained increase in rural population occurred and woodland were converted to farmland and pasture (Walsh et al., 2019). Despite a higher climatic variability characterized the period from the third to the seventh century AD (Büntgen et al., 2011), it has been clearly demonstrated that the presence/absence of man played a crucial role in influencing the slope dynamics (e.g. Amato et al., 2021; Walsh et al., 2019).

Despite the scarce archeological evidence for this time interval in the S. Eufemia area, several historical sources testify to the intensive exploitation of woods from Calabria (Russo Ermolli et al., 2018; Sogliani, 2012). This evidence of increased human pressure is preceded by intensive morphodynamic changes in the Lamezia Plain. Here, a further progradation of the coastline (around 300 m) and the increase in fluvial discharge is testified by the development of superimposed alluvial fans, which caused the burial of the town of Terina and of the Acconia Roman villa during the Late Antique-Early Medieval period (Russo Ermolli et al., 2018). The same drastic AP decline is visible at Salerno and Gaeta at ca. 1400 and 1200 BP, respectively. These chronological differences in the beginning of intensive deforestation could be linked to the different demographic density and consequent diachronic impact on the landscape of the different regions.

Geoarchaeological studies in southern Italy confirm a period of dramatic instability during the Late Antiquity, that triggered the floods of Velia (Amato et al., 2021; Russo Ermolli et al., 2013) and the significant coastline progradation of Neapolis (Russo Ermolli et al., 2013, 2014). Erosion increase induced by human activities has been proved in many areas worldwide (Montgomery, 2007 and reference therein), but the relative impact of humans and climate on erosion is a still complex and debated question (Walsh et al., 2019 and reference therein).

Starting from the same moment, all tree crops and cereals show an increase in the C4 record indicating their intensive cultivation. The beginning of Olea cultivation falls in a period of general decrease of Q. ilex, testifying to its anthropogenic meaning and chronologically corresponds to that featured in the core of Lamezia (Russo Ermolli et al., 2018). Other data coming from the same region (Tropea, Laos, and Canolo Nuovo) witness the occurrence of olive crops in the Middle Age (Amato et al., 2012; Caramiello and Zeme, 1994; Schneider, 1985). The presence of Juglans is recorded since the Roman Imperial period, testifying to a probable first form of cultivation that seems to emerge more timidly in the Lamezia sequence (Russo Ermolli et al., 2018).

Conclusions

Pollen data obtained from the C4 marine core help to clarify the causes of vegetation changes over the last ca. 5000 years in the S. Eufemia Gulf area. Comparison with continental core data from the Lamezia Plain and the Tropea Promontory highlighted the importance of having different resolutions of data from the same catchment area. Indeed, the vegetation changes, and in particular the AP decrease phases recorded in the S. Eufemia marine core, can mostly be interpreted as climatic drying phases (ca. 5000–4800 BP, ca. 4400–4000 BP, ca. 3450–3150 BP, ca. 2700–2000 BP) up to at least ca. 790 BP.

In fact, although human presence has been attested in the catchment area since the Neolithic period, it seems that the exploitation of woodland affected very restricted areas (ex: exploitation of the Tropea Promontory mainly in the Eneolithic and Bronze Age), so that anthropogenic indicators and local deforestation appear very diluted in the regional picture. The hypothesis of climatically-related dry phases relies on the fact that the species that is most affected by reduction is Abies alba, a tree that grows at high altitude (above 1000 m) and that was certainly absent from the Poro Plateau (ca. 500 m a.s.l.) where the prehistoric and protohistoric settlements were concentrated. On the other hand, fir was probably present, as today, on the Sila Mountains, which do not seem to have been frequented by man during these periods. The idea that pre-protohistoric communities, who had woods of the Tropea Promontory at their disposal, went to the Sila to recover fir timber is therefore unconvincing.

However, it must be emphasized that even if the main causes of the forest cover lowering are due to climatic reasons, the presence of anthropogenic indicators suggests a large-scale land use from 3000 BP (Iron Age) onward. The increase in population and the consequent increasing land use was also favored by aridification that had made various portions of the land, previously dominated by forests, more accessible for farming and grazing.

In addition to the brief dry phases recorded in the pollen record, a general trend of aridification is quite evident and is marked by the progressive decrease in Abies percentages until its almost complete disappearance in modern times. Also CoDA analysis has shown a negative correlation between Fagus and Abies all along the C4 record, suggesting a different response of these taxa to water stress. It is nevertheless important to emphasize that, starting from the Roman Age, the gradual decline of the fir tree is also related to the anthropic over-exploitation.

We therefore emphasize the importance of accurate archeological analysis that allows the precise location of settlements and provides insights into the economy and technology of the various communities over time in order to understand the nature the AP reduction phases.

Otherwise, from 790 BP onward, the drastic and sustained decrease in all forest types can be with certainty associated with anthropogenic deforestation, which is quite clearly seen in other marine records of the southern Tyrrhenian slope (e.g. Gaeta, Salerno), albeit with some chronological differences. Even in the Lamezia Plain record it seems that deforestation begins earlier (ca. 2200 BP), but the pollen data from this record appear to be strongly influenced by environmental changes in the coastal plain. In fact, the dominant wet woodland of the Lamezia core is much reduced in the S. Eufemia record.

Intensive deforestation, started at 790 BP, induced enhanced soil erosion and increased detrital arrivals through fluvial discharge. These phenomena could be linked to the synergic action of a worsening climate and the collapse of territorial management following the end of the Western Roman Empire. It is thus rather probable that the effects of intensive land use were enhanced by extreme climatic events creating a “window of opportunity” (Bintliff, 2002; Casana, 2008) in which episodes of severe erosion occurred.

Footnotes

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Halinka Di Lorenzo

Elda Russo Ermolli

References

Aitchison

(1986) The Statistical Analysis of Compositional Data. Monographs on Statistics and Applied Probability. London: Chapman & Hall Ltd (Reprinted in 2003 with additional material by The Blackburn Press).

Aitchison

Greenacre

(2002) Biplots of compositional data. Applied Statistics 51(4): 375–392.

Allevato

Russo Ermolli

Boetto

, et al. (2010) Pollen-wood analysis at the Neapolis harbour site (1st–3rd century AD, southern Italy) and its archaeobotanical implications. Journal of Archaeological Science 37: 2365–2375.

Amato

Cicala

Valente

, et al. (2021) Anthropogenic amplification of geomorphic processes along the Mediterranean coasts: A case-study from the Graeco-Roman town of Elea-Velia (Campania, Italy). Geomorphology 383(2021): 107694.

Amato

De Vita

Filocamo

, et al. (2012) Geomorphological and pedostratigraphical approach as a tool for understanding the archaeological landscapes and environments: the case study of the ancient Laos territory (Calabria, southern Italy). In: Geomorphic Processes and Geoarchaeology: from Landscape Archaeology to Archaeotourism, International Conference held in Moscow-Smolensk, Russia, 20–24 August 2012. Extended abstracts. «Universum», pp.20–24.

Behre

(1981) The interpretation of anthropogenic indicators in pollen diagrams. Pollen et Spores 13: 225–245.

Bernardo

Gangale

(2017) La Sila. In: Blasi

Biondi

(eds) La Flora in Italia. Roma: Ministero dell'Ambiente e della Tutela del Territorio e del Mare, Sapienza Università Editrice, pp.370–373.

Beug

(2015) Leitfaden der Pollenbestimmung Für Mitteleuropa und Angrenzende Gebiete. Munchen: Verlag Dr. Friedrich Pfeil.

Bevan

Palmisano

Woodbridge

, et al. (2019) The changing face of the Mediterranean – Land cover, demography and environmental change: Introduction and overview. The Holocene 29(5): 703–707.

10.

Bini

Zanchetta

Perşoiu

, et al. (2019) The 4.2 ka BP event in the Mediterranean region: An overview. Climate of the Past 15: 555–577.

11.

Bintliff

(2002) Time, process and catastrophism in the study of Mediterranean alluvial history: A review. World Archaeology 33(3): 417–435.

12.

Bond

Showers

Cheseby

, et al. (1997) A pervasive millennial-scale cycle in North Atlantic holocene and glacial climates. Science 278: 1257–1266.

13.

Brun

(2011) Anthropogenic indicators in pollen diagrams in eastern France: A critical review. Vegetation History and Archaeobotany 20: 135–142.

14.

Brutto

Muto

Loreto

, et al. (2014) The tectonic evolution of the western Catanzaro Trough (Calabria, South Italy), preliminary results. In: GNGTS conference 2022, pp.43–47.

15.

Büntgen

Tegel

Nicolussi

, et al. (2011) 2500 years of European climate variability and human susceptibility. Science 331: 578–582.

16.

Caramiello

Zeme

(1994) Analisi palinologiche in terreni dello scavo del Palazzo Vescovile a Tropea . In: Di Giangi

Lebole

Sabbione

(eds) Scavi Medievali in Calabria: Tropea 1, Rapporto Preliminare, vol. XXI. Firenze: Archeologia Medievale, pp.368–370.

17.

Caridi

Aramini

, et al. (1996) Carta dei Suoli della Piana di S. Eufemia Lamezia. ARSSA - Programma Operativo. Reg. CEE 2052/88Ob.

18.

Casana

(2008) Mediterranean valleys revisited: Linking soil erosion, land use and climate variability in the northern levant. Geomorphology 101: 429–442.

19.

Chiari

Barone

Bombini

, et al. (2021) LABEC, the INFN ion beam laboratory of nuclear techniques for environment and cultural heritage. The European Physical Journal Plus 136: 1–28.

20.

Conedera

Tinner

Neff

, et al. (2009) Reconstructing past fire regimes: Methods, applications, and relevance to fire management and conservation. Quaternary Science Reviews 28: 555–576.

21.

Cosentino

Molisso

Scopelliti

, et al. (2017) Benthic foraminifera as indicators of relative sea-level fluctuations: Paleoenvironmental and paleoclimatic reconstruction of a holocene marine succession (Calabria, south-eastern Tyrrhenian Sea). Quaternary International 439: 79–101.

22.

Cremaschi

Mercuri

Torri

, et al. (2016) Climate change versus land management in the Po Plain (Northern Italy) during the Bronze Age: New insights from the VP/VG sequence of the Terramara Santa Rosa di Poviglio. Quaternary Science Reviews 136: 153–172.

23.

Cugny

Mazier

Galop

(2010) Modern and fossil non-pollen palynomorphs from the Basque mountains (western Pyrenees, France): The use of coprophilous fungi to reconstruct pastoral activity. Vegetation History and Archaeobotany 19: 391–408.

24.

Daunis-i-Estadella

Thió-Henestrosa

Mateu-Figueras

(2011) Including supplementary elements in a compositional biplot. Computational Geosciences 37(5): 696–701.

25.

Di Donato

Esposito

Garilli

, et al. (2009) Surface–bottom relationships in the Gulf of Salerno (Tyrrhenian Sea) over the last 34 kyr: Compositional data analysis of palaeontological proxies and geochemical evidence. Geobios 42: 561–579.

26.

Di Donato

Esposito

Russo-Ermolli

, et al. (2008) Coupled atmospheric and marine palaeoclimatic reconstruction for the last 35ka in the sele Plain–Gulf of Salerno area (southern Italy). Quaternary International 190: 146–157.

27.

Di Donato

Pawlowsky-Glahn

Egozcue

, et al. (2022b) Preliminary findings in Ward R-mode clustering method for Compositional Data. In: Proceedings of the CoDaWork2022 (eds Thomas-Agnan C and Pawlowsky-Glahn V), Toulouse, June 27–July 1, 2022, Asociaciòn para Datos Composicionales, Girona, Spain, pp.32–38.

28.

Di Donato

Sgarrella

Sprovieri

, et al. (2022a) High-frequency modification of the central Mediterranean seafloor environment over the last 74 ka. Palaeogeography Palaeoclimatology Palaeoecology 593: 110924.

29.

Di Lorenzo

Jung

Pacciarelli

, et al. (2021a) Human impact and landscape changes between 3000 and 1000 BC on the Tropea Promontory (Calabria, Italy). The Holocene 31(6): 926–942.

30.

Di Lorenzo

Aucelli

Corrado

, et al. (2021b) Environmental evolution and anthropogenic forcing in the garigliano coastal plain (Italy) during the holocene. The Holocene 31(7): 1089–1099.

31.

Di Muro

(2011) Il popolamento rupestre in Calabria. In: Menestò

(ed.) Atti Del IV Convegno Internazionale Sulla Civiltà Rupestre, Le Aree Rupestri dell’Italia Centro-Meridionale Nell’ambito Delle Civiltà Italiche: Conoscenza, Salvaguardia, Tutela. Savelletri Di Fasano (BR), 2009. Fondazione CISAM, Spoleto, pp.181–198.

32.

Di Rita

Lirer

Bonomo

, et al. (2018) Late Holocene forest dynamics in the Gulf of Gaeta (central Mediterranean) in relation to NAO variability and human impact. Quaternary Science Reviews 179: 137–152.

33.

Di Rita

Magri

(2019) The 4.2 ka event in the vegetation record of the central Mediterranean. Climate of the Past 15: 237–251.

34.

Egozcue

Pawlowsky-Glahn

(2005) Groups of parts and their balances in compositional data analysis. Mathematical Geology 37(7): 795–828.

35.

Favre

Escarguel

Suc

, et al. (2008) A contribution to deciphering the meaning of AP/NAP with respect to vegetation cover. Review of Palaeobotany and Palynology 148: 13–35.

36.

Filocamo

Romano

Di Donato

, et al. (2009) Geomorphology and tectonics of uplifted coasts: New chronostratigraphical constraints for the Quaternary evolution of Tyrrhenian North Calabria (southern Italy). Geomorphology 105: 334–354.

37.

Gamberi

Marani

(2006) Hinterland geology and continental margin growth: The case of the Gioia Basin (southeastern Tyrrhenian Sea). Geological Society Special Publication 262: 349–363.

38.

Giraudi

Magny

Zanchetta

, et al. (2011) The holocene climatic evolution of Mediterranean Italy: A review of the continental geological data. The Holocene 21: 105–115.

39.

Givigliano

(1978) Sistemi di comunicazione e topografia degli insediamenti di età greca nella Calabria Brettia. Edizioni “Il Gruppo”, Cosenza.

40.

Gogou

Triantaphyllou

Xoplaki

, et al. (2016) Climate variability and socioenvironmental changes in the northern Aegean (NE Mediterranean) during the last 1500 years. Quaternary Science Reviews 136: 209–228.

41.

Greenacre

(2020) Amalgamations are valid in compositional data analysis, can be used in agglomerative clustering, and their logratios have an inverse transformation. Applied Computing and Geosciences 5: 100017.

42.

Grimm

(1987) CONISS: A FORTRAN 77 program for stratigraphically constrained cluster analysis by the method of incremental sum of squares. Computational Geosciences 13: 13–35.

43.

Grimm

(2004) Tilia graph v. 2.0. 2. Illinois state museum, research and collections center.

44.

Harris

(2011) Bois et déboisement dans la Méditerranée antique. Annales HSS, January-March 2011, no. 1, pp. 105–140.

45.

Harris

(Ed) (2013a) The Ancient Mediterranean Environment Between Science and History. Columbia Studies in the Classical Tradition. Leiden-Boston: Brill.

46.

Harris

(2013b) Defining and detecting Mediterranean deforestation, 800 BCE to 700 CE. In: Harris

(ed.) The Ancient Mediterranean Environment Between Science and History, Columbia Studies in the Classical Tradition. Leiden-Boston: Brill, p.39.

47.

Heaton

Köhler

Butzin

, et al. (2020) Marine20—The marine radiocarbon age calibration curve (0–55,000 cal BP). Radiocarbon 62: 779–820.

48.

Izdebski

Pickett

Roberts

, et al. (2016) The environmental, archaeological and historical evidence for regional climatic changes and their societal impacts in the eastern mediterranean in late Antiquity. Quaternary Science Reviews 136: 189–208.

49.

Jacobson

Bradshaw

RHW

(1981) The selection of sites for paleovegetational studies. Quaternary Research 16: 80–96.

50.

Jacques

Monaco

Tapponnier

, et al. (2001) Faulting and earthquake triggering during the 1783 Calabria seismic sequence. Geophysical Journal International 147: 499–516.

51.

Jalut

Dedoubat

Fontugne

, et al. (2009) Holocene circum-Mediterranean vegetation changes: Climate forcing and human impact. Quaternary International 200: 4–18.

52.

Joannin

Brugiapaglia

de Beaulieu

, et al. (2012) Pollenbased reconstruction of holocene vegetation and climate in southern Italy: The case of Lago Trifoglietti. Climate of the Past 8: 1973–1996.

53.

Jung

Pacciarelli

(2021) The settlement of Punta di Zambrone in its local and Mediterranean context during the Recent Bronze Age. In: Jung

(ed.) Punta Di Zambrone I. 1200 B.C.E. A Time of Breakdown – A Time of Progress in Southern Italy and Greece. Austrian Academy of Science Press, Wien, OREA, pp.67–112.

54.

Klee

Zach

Thanheiser

(2021) Archaeobotanical investigations at Punta di zambrone (Calabria, Italy). In: Jung

(ed.) 1200 B.C.E. A Time of Breakdown – A Time of Progress in Southern Italy and Greece, vol.17. Austrian Academy of Science Press, Wien, OREA, pp.157–214.

55.

Longhitano

Chiarella

Muto

(2014) Three-dimensional to two-dimensional cross-strata transition in the lower pleistocene Catanzaro tidal strait transgressive succession (southern Italy). Sedimentology 61(7): 2136–2171.

56.

Loreto

Fracassi

Franzo

, et al. (2013) Approaching the seismogenic source of the Calabria 8 September 1905 earthquake: New geophysical, geological and biochemical data from the S. Eufemia Gulf (S Italy). Marine Geology 343: 62–75.

57.

Mackay

Battarbee

Birks

, et al. (2005) Global Change in the Holocene. London: Hodder Arnold.

58.

Magny

Combourieu-Nebout

de Beaulieu

, et al. (2013) North–south palaeohydrological contrasts in the central Mediterranean during the holocene: tentative synthesis and working hypotheses. Climate of the Past 9: 2043–2071.

59.

Martín-Fernández

Barceló-Vidal

Pawlowsky-Glahn

(2003) Dealing with zeros and missing values in compositional datasets using nonparametric imputation. Mathematical Geology 35(3): 253–278.

60.

Martín-Fernández

Daunis-i-Estadella

Mateu-Figueras

(2015) On the interpretation of differences between groups for compositional data. SORT-Statistics and Operations Research Transactions 39(2): 231–252.

61.

Matarese

(2017) Il centro fortificato indigeno di Torre Galli (Drapia, VV): Una proposta di inquadramento preliminare dei reperti ceramici dell’abitato. In: Pacciarelli

Cicala

(eds) Atti Del Convegno Internazionale, Centri Fortificati Indigeni Della Calabria Dalla Preistoria All’età Ellenistica. Naus editoria, Napoli, pp.485–487.

62.

Mayewski

Rohling

Curt Stager

, et al. (2004) Holocene climate variability. Quaternary Research 62: 243–255.

63.

Mazzini

Gliozzi

Galaty

, et al. (2016) Holocene evolution of Lake Shkodra: Multidisciplinary evidence for diachronic landscape change in northern Albania. Quaternary Science Reviews 136: 85–95.

64.

Mensing

Tunno

Sagnotti

, et al. (2015) 2700 years of Mediterranean environmental change in central Italy: A synthesis of sedimentary and cultural records to interpret past impacts of climate on society. Quaternary Science Reviews 116: 72–94.

65.

Mercuri

(2014) Genesis and evolution of the cultural landscape in central Mediterranean: The ‘where, when and how’ through the palynological approach. Landscape Ecology 29: 1799–1810.

66.

Mercuri

Florenzano

Burjachs

, et al. (2019) From influence to impact: The multifunctional land use in Mediterranean prehistory emerging from palynology of archaeological sites (8.0-2.8 ka BP). The Holocene 29(5): 830–846.

67.

Mercuri

Florenzano

Massamba N'siala

, et al. (2010) Pollen from archaeological layers and cultural landscape reconstruction: Case studies from the Bradano valley (Basilicata, southern Italy). Plant Biosystems 144(4): 888–901.

68.

Michelangeli

Di Rita

Celant

, et al. (2022) Three millennia of vegetation, Land-Use, and climate change in SE Sicily. Forests 13: 102.

69.

Monaco

Tortorici

Nicolich

, et al. (1996) From collisional to rifted basins: An example from the southern Calabrian arc (Italy). Tectonophysics 266: 233–249.

70.

Montgomery

(2007) Soil erosion and agricultural sustainability. Proceedings of the National Academy of Sciences of the United States of America 104(33): 13268–13272.

71.

Moser

Nelle

Di Pasquale

(2018) Timber economy in the Roman age: charcoal data from the key site of Herculaneum (Naples, Italy). Archaeological and Anthropological Sciences 10: 905–921.

72.

Olsen

Anderson

Knudsen

(2012) Variability of the North Atlantic oscillation over the past 5,200 years. Nature Geoscience 5: 808–812.

73.

Pacciarelli

(2001) Dal Villaggio Alla Città. La Svolta Protourbana Del 1000 a.C. nell’Italia Tirrenica. Firenze: All’Insegna del Giglio.

74.

Pacciarelli

Iaia

Matarese

(2017) Torre Galli: Le campagne di scavo 2012-2013. In: Pacciarelli

Cicala

(eds) Atti Del Convegno Internazionale, Centri Fortificati Indigeni Della Calabria Dalla Preistoria All’età Ellenistica. Naus editoria, Napoli, pp.337–350.

75.

Pepe

Bertotti

Ferranti

, et al. (2014) Pattern and rate of post-20 ka vertical tectonic motion around the Capo Vaticano Promontory (W Calabria, Italy) based on offshore geomorphological indicators. Quaternary International 332: 85–98.

76.

Peyron

Combourieu-Nebout

Brayshaw

, et al. (2017) Precipitation changes in the Mediterranean basin during the holocene from terrestrial and marine pollen records: A model–data comparison. Climate of the Past 13: 249–265.

77.

Peyron

Magny

Goring

, et al. (2013) Contrasting patterns of climatic changes during the holocene across the Italian peninsula reconstructed from pollen data. Climate of the Past 9: 1233–1252.

78.

Pignatti

(1982) Flora d’italia. Bologna: Edagricole.

79.

Presti

Billi

Orecchio

, et al. (2013) Earthquake focal mechanisms, seismogenic stress, and seismotectonics of the Calabrian arc, Italy. Tectonophysics 602: 153–175.

80.

Railsback

Liang

Brook

, et al. (2018) The timing, two-pulsed nature, and variable climatic expression of the 4.2 ka event: A review and new high-resolution stalagmite data from Namibia. Quaternary Science Reviews 186: 78–90.

81.

Rasmussen

Vinther

Clausen

, et al. (2007) Early holocene climate oscillations recorded in three Greenland ice cores. Quaternary Science Reviews 26: 1907–1914.

82.

Reille

(1992) Pollen et spores d’Europe et d’Afrique du nord. Laboratoire de botanique historique et palynologie URA CNRS 1152, Marseille.

83.

Reille

(1995) Pollen et spores d’Europe et d’Afrique du nord. Supplément 1. Laboratoire de botanique historique et palynologie URA CNRS 1152, Marseille.

84.

Rita

(2014) Sensibilità al clima dell’abete bianco e del faggio in boschi misti dell’Appennino Meridionale. Italian Journal of Forest and Mountain Environments 69(2): 95–101.

85.

Roberts

Woodbridge

Palmisano

, et al. (2019) Mediterranean landscape change during the holocene: Synthesis, comparison and regional trends in population, land cover and climate. The Holocene 29(5): 923–937.

86.

Ruello

Cinque

Di Donato

, et al. (2017) Interplay between sea level rise and tectonics in the Holocene evolution of the St. Eufemia Plain (Calabria, Italy). Journal of Coastal Conservation 21: 903–915.

87.

Ruiz

Sanz-Sánchez

(2020) Effects of historical land-use change in the Mediterranean environment. The Science of the Total Environment 732: 139315.

88.

Russo

Ermolli E

di Pasquale

(2002) Vegetation dynamics of southwestern Italy in the last 28 kyr inferred from pollen analysis of a Tyrrhenian Sea core. Vegetation History and Archaeobotany 11: 211–220.

89.

Russo

Ermolli E

Romano

Ruello

(2013) Human-environment interactions in the southern Tyrrhenian coastal area: Hypothesis from Neapolis and Elea-Velia. In: Harris

(ed.) The Ancient Mediterranean Environment Between Science and History, Columbia Studies in the Classical Tradition, vol.39. Leiden-Boston: Brill, pp.213–232.

90.

Russo

Ermolli E

Romano

Ruello

, et al. (2014) The natural and cultural landscape of Naples (southern Italy) during the Graeco-Roman and Late Antique periods. Journal of Archaeological Science 42: 399–411.

91.

Russo Ermolli

Ruello

Cicala

, et al. (2018) An 8300-yr record of environmental and cultural changes in the Sant’Eufemia Plain (Calabria, Italy). Quaternary International 483: 39–56.

92.

Sadori

Giraudi

Masi

, et al. (2016) Climate, environment and society in Southern Italy during the last 2000 years. A review of the environmental, historical and archaeological evidence. Quaternary Science Reviews 136: 173–188.

93.

Sadori

Giraudi

Petitti

, et al. (2004) Human impact at Lago di Mezzano (central Italy) during the Bronze Age: A multidisciplinary approach. Quaternary International 113: 5–17.

94.

Sangineto

(1994) Per la ricostruzione del paesaggio agrario delle Calabrie romane. In: Ettis

(ed.) Storia Della Calabria Antica, vol. 2. Rome: Età italica e romana, pp.557–593.

95.

Sangineto

(2013) Roma Nei Bruttii. Città e Campagne Nelle Calabrie Romane. Rossano: Ferrari.

96.

Schneider

(1985) Analyse palynologique dans l’Aspromonte en calabrie (Italie méridionale). Cahiers Ligures Préhistoire et Protohistoire 2: 279–288.

97.

Sogliani

(2012) Repertorio delle fonti scritte per la ricostruzione della vicenda insediativa di Vibo Valentia tra Tarda Antichità e Medioevo. In: D’Andrea

(ed.) Vincenzo Nusdeo. Sulle Tracce Della Storia. Studi in Onore Di Vincenzo Nusdeo Nel Decennale Della Sua Scomparsa. ADHOC Edizioni s.n.c., Vibo Valentia, pp.271–335.

98.

Stoddart

Woodbridge

Palmisano

, et al. (2019) Tyrrhenian central Italy: holocene population and landscape ecology. The Holocene 29(5): 761–775.

99.

Stuiver

Reimer

(1993) Extended ¹⁴C data base and revised CALIB 3.0 ¹⁴C Age Calibration Program. Radiocarbon 35: 215–230.

100.

Tinner

(2003) Size parameters, size-class distribution and area-number relationship of microscopic charcoal: Relevance for fire reconstruction. The Holocene 13: 499–505.

101.

Tortorici

Bianca

de Guidi

, et al. (2003) Fault activity and marine terracing in the Capo Vaticano area (southern Calabria) during the middle-Late Quaternary. Quaternary International 101-102: 269–278.

102.

Tortorici

Monaco

Tansi

, et al. (1995) Recent and active tectonics in the Calabrian arc (Southern Italy). Tectonophisics 243: 37–55.

103.

Walker

Head

Berkelhammer

, et al. (2018) Formal ratification of the subdivision of the holocene Series/Epoch (Quaternary System/period): Two new global boundary stratotype sections and points (GSSPs) and three new stages/subseries. Episodes 41(4): 213–223.

104.

Walsh

Berger

Roberts

, et al. (2019) Holocene demographic fluctuations, climate and erosion in the Mediterranean: A meta data-analysis. The Holocene 29(5): 864–885.

105.

Weiberg

Unkel

Kouli

, et al. (2016) The socio-environmental history of the Peloponnese during the holocene: Towards an integrated understanding of the past. Quaternary Science Reviews 136: 40–65.

106.

Westaway

(1993) Quaternary uplift of southern Italy. Journal of Geophysical Research 98: 21741–21772.

107.

Woodbridge

Roberts

Fyfe

(2018) Pan-mediterranean Holocene vegetation and land-cover dynamics from synthesized pollen data. Journal of Biogeography 45: 2159–2174.

108.

Zanchetta

Regattieri

Isola

, et al. (2016) The so-called “4.2 event” in the central Mediterranean and its climatic teleconnections. Alpine and Mediterranean Quaternary 29: 5–17.