**1. Introduction**

In long-term anthropized landscapes, the biodiversity has been preserved over time thanks to agricultural, pastoral and silvicultural practices that we would define today as sustainable. The high biodiversity, maintained during the long interaction between natural processes and human history, is a peculiarity of certain types of landscapes and agrarian systems that, for this reason, are recognized as High Nature Value (HNV) farmlands [1]. The strong connection between biodiversity conservation and the maintenance of traditional agricultural activities has been the concept behind the development of the definition of HNV farmlands in Europe [2]. In particular, seminatural pastures and woods are essential elements of traditional and historical agricultural landscapes for the maintenance of the high degree of naturalness [3]. High presence of seminatural areas, extensive mosaic landscapes and areas hosting species of conservation concern are elements for the definition of an HNV farmland [4]. In most cases, the HNV farmlands are "historical landscapes" and are intended to be landscapes that are long-standing in a certain territory without or with gradual changes [5,6]. However, what is it meant by "long-standing" in a given area? The understanding of a historical landscape is based on the analysis of natural factors and historical dynamics that have determined its characteristics and on the interpretation of human–environment interactions. In fact, the landscape as a biological and cultural system is the result of processes of changes that have determined a stratified and dynamic pattern over time, notably also in the Mediterranean [7,8].

Vegetation studies are useful to describe landscape patterns and transformations. The structure and floristic compositions of plant communities are strictly connected to the environmental set-up of the ecosystems. For this reason, the pool of community species has been recognized as an indicator of environmental conditions and the characterization of the ecosystems [9–11]. Phytocoenoses are not only the result of natural processes but also derive from the anthropic action that changes their characteristics. Secondary formations such as shrubs, pasture meadows, ruderal and weed communities are examples of vegetation derived from sylvo-pastoral and agricultural activities [12] and in many cases are of considerable phytogeographic interest [13–15]. The structure, diversity and floristic composition of these communities have been influenced by the current and historical use of the territory. In particular, fire, deforestation and grazing are the anthropic activities that have influenced the pattern of natural plant communities present in rural landscapes [16–18]. In summary, the characteristics of a plant association—such as the result of interactions between natural and human factors—are fundamental keys to understanding the landscape. However, this is not su fficient; to gain a wider view of the landscape, it is necessary to interpret the mosaic of associations generated by the anthropic disturbance in terms of series of vegetation [12,19–21]. The study of the landscape from the perspective of vegetation series allows us to interpret the territorial mosaic both in terms of spatial patterns (di fferent associations that share the same ecology) and in the temporal dynamics (the phytocoenosis that occurs over time in relation to anthropic activities). In fact, the spatial distribution of a vegetation series is influenced by environmental factors (e.g., climate, lithology, landforms), which change over the long term, often measured with the geological time scale. The anthropic factors do not change the entire spatial distribution of a series but modify, in a shorter time scale, the plant associations (the di fferent stages) that compose it. However, the diachronic analysis of the landscape in terms of vegetation series—as for all ecosystem ecology analyses [22]—is limited only to a time-range of just over a century [12].

The improvement of the temporal depth in the analysis of landscape dynamics is possible through a Historical Ecology approach that integrates historical and archaeological data with ecological information [23] and specifically—in this "vegetation series perspective"—with the dynamic–catenal phytosociology (see Rivas-Martínez [20]). This multidisciplinary perspective makes it possible to study the interaction between human societies and biophysical environments while also considering long-term dynamics [24,25].

The reconstruction of past landscape dynamics and the impact of agricultural, pastoral and forestry practices on vegetation during the long term can be significantly improved by archaeobotanical data [26]. The impact of anthropic activity on vegetation over the centuries can be assessed using pollen [27,28] but also archaeobotanical macroremains [29,30]. In particular, the study of the charcoals found in archaeological excavations provides a fundamental tool for the reconstruction of past landscapes and for interpreting strategies for the exploitation of forest resources by human communities [31–33].

Archaeobotanical studies and analyses of wood charcoals for the reconstruction of paleoenvironments are a practice in most current archaeological projects in the Mediterranean area [34–36], although in Sicily the picture is still rather limited [37–39]. The presence of wood remains, often charred, in archaeological contexts is filtered by the process of human selection—for different purposes—of the plant resources available in the surrounding environment [40–42].

The Sicani Mountains, in central-western Sicily, have several historical landscapes with a high degree of biodiversity and a diversified territorial mosaic rich of peculiar phytocoenosis [43–45] that have been recognized by the European Environment Agency as HNV farmland (Figure 1) [46].

**Figure 1.** (**A**) Case study area and distribution of High Nature Value (HNV) farmland in Sicily (Source: European Environment Agency [46]). (**B**) Location of archaeological site (Contrada Castro) on topographic base map by the Italian Geographic Military Institute (aut. n. 4848 27/07/1998). The U.T.M. grid, in purple, has an interval of 1000 metres. (**C**) Topographical map of the location of the archaeological excavation in Contrada Castro (Redrawn from: Regional Geographical Information System of Sicily—[47]).

Recent research in the Castro-Giardinello area (Corleone, Palermo) [48,49], the study area of this paper, has shown that a small area of about 400 hectares preserves a biodiversity of about 502 taxa of vascular flora that have also been an important resource in traditional food habits [50]. As a result of the dynamics of the renaturalization which has taken place in the last 60 years, this sector of the Sicani Mountains (Figure 2) retains a high degree of naturalness and a low anthropic impact [10]. This interesting biodiversity is linked to a territory with long-term human occupation as evidenced by a recent archaeological survey [48].

The aim of this paper is to detect the relationship between the geobotanical characteristics of the landscape (phytosociological analysis of current vegetation dynamics) and the effective historical use of wood resources (anthropological analysis on charcoal finds from the Medieval phases of Contrada Castro site) in order to understand the long-term suitability and sustainability of this historical landscape and HNV farmland in Sicani Mountains district (central-western Sicily).

(**a**)

**Figure 2.** Sicani Mountains landscape: Farmland in Contrada Giardinello (**a**) and olive groves (**b**) near Pizzo Castro (Corleone). The area depicted in these photos is free of anthropic elements (buildings, roads, electric lines, etc.). Photos by Pasquale Marino.

### **2. Archaeological Background of the Site**

Since 2015, the Castro-Giardinello area located in the Sicani Mountains district, in the southern part of the territory of Corleone (central-western Sicily), has been investigated by archaeological survey and excavation (directed by the Soprintendenza BB.CC.AA. of Palermo) within the project "Harvesting Memories" focused on the investigation of long-term landscape trajectories and human dynamics [49]. In particular, the project area is located in the Bona Furtuna LLC estate (a 100% organic farm) between the valley of the Giardinello creek and the south-western side of Barra ù mount (maximum height 1475 m a.s.l.) (Figures 1b and 2). The long-term anthropic occupation of this area has been detected by a field survey that identified various sites dating back to the Bronze Age, classical, medieval and postmedieval periods [48].

For its topographical position on a hill-top plateau (715–713 m a.s.l.) and its remarkable presence of pottery, the site of Contrada Castro has been chosen for archaeological excavations from 2017, and these are still ongoing (Figure 3). The site extends over a flat, raised, east-west plateau (0.54 ha) on brown marls substrate. To the north, it is adjacent to a sinkhole that separates it from the steep slopes of Pizzo Castro, and to the south, there is an almost vertical slope towards the valley of the Giardinello river (Figure 1b).

**Figure 3.** The hill-top plateau of the archaeological site of Contrada Castro (Corleone, Palermo, Sicily). Drone photo by Filippo Pisciotta.

In this area, the surveys have highlighted the presence of dispersed ceramic sherds which indicated a potential occupation during the Early Middle Ages (Byzantine and Islamic periods, 7th–11th centuries AD), while rare fragments of black-gloss pottery also indicated an earlier occupation during an archaic/classical period [48]. The initial hypothesis was therefore related to the possible presence of a site that was intensely populated during the Early Middle Ages and subsequently abandoned and never again permanently resettled. The first archaeological seasons have confirmed this hypothesis [48]. The earlier period of the site is documented by some wall remains associated with ceramic finds of the late 6th–5th centuries BC. On the collapses of these structures, some evidence of a reoccupation of the site are related to two perinatal burials dated by radiocarbon analysis between the 7th and the mid-8th centuries AD (sigma 1 65%: AD 662–AD 778). Later, during the late 8th–9th centuries AD, this area was further exploited for the installation of a building dedicated in the first phase to the production of ceramics and tiles, as evidenced by the discovery of two circular kilns. The chronology of these

structure is also based on radiocarbon analysis of *Pistacia terebinthus* wood charcoals (sigma 1 65%: AD 774–AD 878) from a burnt layer related to the productive activity of one of the two kilns. This structure was reused after a short time, changing its function because one of the pottery kilns was converted into a food oven (possibly for bread). The collapse of this building occurred during the 9th century AD. On the ruins of this building, which is no longer visible, another building was built with two phases, which can be placed chronologically between the first half of the 10th and the 11th century AD. For this period, the chronology was confirmed by pottery chronotypology and by radiocarbon analysis of an *Equus asinus* bone (sigma 1 65%: AD 965–AD 1042) [48]. The largest number of archaeobotanical findings comes from the medieval layers of the site and can be subdivided into two periods: Period 1, late 8th–9th centuries AD; Period 2, 10th–11th centuries AD.

### **3. Materials and Methods**

### *3.1. Vegetation Data*

The syn-phytosociological study of the vegetation allowed us to define a forest landscape model (current and potential) of the Sicani Mountains, which was necessary for the interpretation of anthracological data from the Contrada Castro sequence. The analysis of landscape patterns in terms of vegetation series followed the approach outlined in detail by Bazan et al. [21].

The vegetation surveys were conducted based on the classical phytosociological approach [51] including recent methodological advances [19]. A total of 24 relevés, in 100 m<sup>2</sup> plots, were created during the spring 2019 (Table S1). The survey was carried out with a Stratified Random Sampling by dividing the study area into different physiognomical homogeneous contexts, based on the stages of the vegetation series defined by Bazan et al. [12] and randomly sampling within each of these smaller areas [52].

In order to ge<sup>t</sup> a more complete picture of the floristic composition and structure of the forest and preforest associations, the relevés of Gianguzzi et al. [53] were also considered. The complete data set consisted of 106 relevés which were summarized in a synoptic table (Table 1).

The mean coverage of each woody species (phanerophytes and nanophanerophytes) within each association was calculated by transforming the Braun-Blanquet cover-abundance scale into a quantitative form (0–100%) as proposed by McNellie et al. [54]. A mean coverage curve was constructed for the associations of the vegetation series describing our forest landscape model.

The syntaxonomical nomenclature followed the "International Code of Phytosociological Nomenclature" [55]. Species identification was done using the identification key of Pignatti et al. [56] and for the nomenclature, the online database "The Plant List" [57], was used. The complete names of specific and subspecific taxa of this paper, including the authors' citations of plant names, are given in Table 1.

The bioclimatic data followed the framework developed for Sicily by Bazan et al. [58]; other ecological information was obtained from Gianguzzi et al. [59].

**Table 1.** Synoptic table of the associations of vegetation series. Column numbers: 1 = *Ampelodesmo mauritanici-Quercetum ilicis*; 2 = *Sorbo torminalis-Quercetum ilicis*; 3 = *Oleo oleaster-Quercetum virgilianae*; 4 = *Sorbo torminalis-Quercetum virgilianae*; 5 = *Euphorbio characiae-Prunetum spinosae*; 6 = *Roso siculae-Prunetum spinosae*; 7 = *Roso corymbiferae-Rubetum ulmifolii*; 8 = *Crataegetum laciniatae*; 9 = *Ulmo-Salicetum pedicellatae*. Roman numbers indicate the presence classes of species in the phytosociological relevés of Table S1 as follows: I = 0–20%, II = 20–40%, III = 40–60%, IV = 60–80%, V = 80–100%. The values of diagnostic species of association and alliance are in the grey boxes.




*Euphorbia ceratocarpa* Ten.

*Fedia graciliflora* Fisch. & C.A. Mey.

*Ficaria verna* Huds.

*Ostrya carpinifolia* Scop.

*Rumex thyrsoides* Desf.

*Salvia argentea* L.

*Scandix* 

*Festuca arundinacea* Schreb.

*Senecio squalidus* L. subsp.

Arcang.

*Hyoseris radiata* L.

*Piptatherum*

*Sanguisorba minor* Scop.

*Pteridium aquilinum* (L.) Kuhn subsp. *aquilinum*

 *miliaceum* (L.) Coss.

*pecten-veneris* L.

*microglossus* (Guss.)


 . . . . III . I . II

 .

 .

 . . . . II . . . .

 . . . II . . . . .

 . . . . I . I . .

. . . . . . I.

 . . . . I . . . .

 . . . . . I . . .

 . . . . . I ..

 . . . . . . . I .

 . . . . I . . . .

. . . . . I ..

 II .

 II .

......

......

> .

> .

> .

**Table 1.** *Cont*.


