INTRODUCTION
The Central Cryptoporticus of the Palatine Hill represents a structure of extraordinary historical and architectural significance, yet its long-term stability is influenced by complex interactions between geological, environmental, and anthropogenic factors (De Angelis D’Ossat, 1956; Cecchi, 2011; Mancini et al., 2018). Given the presence of underground anthropogenic cavities and built structures, there are potential risks to the overlying archeological site, as well as to visitors accessing the site. These risks derive from structural instability caused by underlying cavities and deterioration processes linked to unfavorable geo-environmental conditions, such as rainwater infiltration or groundwater interception, and even seismic vulnerability (Moscatelli et al. 2011, 2014; Calabresi et al., 2013; Di Luzio et al., 2013). The progressive deterioration of foundation materials and structural elements highlights the need for a comprehensive investigation into the geological and geochemical characteristics of the site, forming the basis for conservation strategies aimed at mitigating structural instability (Gaudiosi et al. 2024).
The stratigraphy of the Palatine Hill, as described by Mancini et al. (2014, 2018), records repeated cycles of valley incision and infill from the Middle Pleistocene to the Holocene, resulting in a complex sequence of volcanic, alluvial, and anthropogenic deposits. According to Servizio Geologico d’Italia (2008) and Funiciello & Giordano (2008), these stratigraphic units include the tuff deposits of the Villa Senni formation and the muddy to sandy, volcaniclastic-rich alluvial sediments of the Aurelia and Fosso del Torrino formations, which together form the foundation of many ancient structures in the area. Given their role in supporting archaeological remains, the mineralogical composition, mechanical stability, and post-depositional transformations of these deposits are critical factors in evaluating the long-term preservation and structural integrity of the site (Schmertmann, 1991; Cardarelli et al., 2016; El Howayek et al., 2017).
Several studies have examined the weathering and diagenetic transformations of the volcanic tuffs used in Roman construction, particularly those of the Villa Senni pyroclastic formation including the Tufo Lionato member (cfr. “Tufo Lionato litoide” Auctt. - Jackson et al., 2005; Jackson & Marra, 2006; Jackson et al., 2014; Cardarelli et al., 2016; Brocato et al., 2019; Diffendale et al., 2019). Water infiltration, whether from direct precipitation, capillary rise, or groundwater interaction, plays a central role in their degradation. Specifically, hydration of zeolitised material, the alteration of glass fragments into clay minerals, and associated volumetric changes contribute to the progressive weakening of these foundation materials, making them highly susceptible to long-term deterioration (Jackson & Marra, 2006; Jackson et al., 2014, 2018; Cardarelli et al., 2016). Additionally, biological activity from plants and animals (i.e. root growth, burrowing, boring) can also contribute to the alteration of these deposits. These processes significantly affect the physical and mechanical properties of the building materials.
In alluvial fine-grained deposits, geochemical and hydrological factors influence their stability. Reactions such as carbonate precipitation and redoximorphic transformations can significantly modify the geochemistry and microtexture of these deposits, influencing moisture dynamics and, consequently, the long-term preservation of foundation soils and construction materials (Macphail & Goldberg, 2018; Artioli et al., 2020; Chen et al., 2022a, b). These sediments are periodically exposed to groundwater saturation or draining in combination with different water table depths, which diminishes their load-bearing capacity and contributes to structural weakening (Stephens et al., 1984; Macphail & Goldberg, 2018). Densification (caused by drying, consolidation, and compaction) and biological oxidation of these soils may cause subsidence (Stephens et al, 1984). These transformations are not limited to recent environmental changes but originate from the very earliest depositional stages. From initial sedimentation, these materials may undergone a complex history of pedogenesis, often involving multiple soil-forming cycles that progressively altered their microstructure, mineralogy, and geochemical signatures. This long-term evolution has shaped their current physico-chemical properties, conditioning their mechanical behavior and vulnerability to environmental stresses. Despite this, field studies tend to rely on macroscopic descriptors while overlooking the mechanical implications of such cumulative pedogenic processes, including overconsolidation and microstructural evolution (Robertson, 2016). This gap limits the full understanding of material behavior under stress and environmental change. The interaction between these inherited characteristics, ongoing post-depositional processes, and environmental stressors underscores the necessity to combine multiscale characterisation and continuous monitoring.
In this context, we investigate how the mineralogical and microstructural evolution of the Palatine Hill foundation units governs present-day mechanical heterogeneity and instability beneath the Central Cryptoporticus. Our specific objectives are to: (i) constrain the stratigraphic architecture and lateral variability of the main foundation units (anthropogenic cover, Aurelia formation floodplain deposits (AEL), Villa Senni formation volcaniclastic deposits (VSN), and Fosso del Torrino fluvial deposits (FTR); (ii) document the principal pedogenic and diagenetic transformations (carbonate redistribution, redox cycling, clay translocation, and glass alteration) that modify fabric and bonding; and (iii) integrate these observations into a conceptual degradation model linking water routing and moisture fluctuations to progressive weakening, differential settlement and cavity-roof vulnerability.
Methodologically, we combine lithofacies logging from three boreholes and one trench with bulk geochemistry (pXRF), bulk mineralogy (XRD; semi-quantitative phase proportions), petrography, and soil micromorphology. This multi-proxy approach provides a workflow for translating microscale alteration pathways into macroscale engineering behaviour and for identifying conservation priorities focused on hydrological control and targeted mitigation in the most compressible and voided foundation domains.
GEOLOGICAL AND STRATIGRAPHIC SETTING
The Palatine Hill lies in Rome’s historical centre (Fig. 1). The geological setting of the city (Fig. 2) record the interplay of several geological and climatic phenomena during the last three million years, such as: the post-orogenic extensional tectonics of the Latium margin postdating the Apennines compressional tectonics, the volcanic activity of the Colli Albani and Monti Sabatini volcanic districts, climate variations, and sea-level fluctuations (Cavinato et al., 1992; De Rita et al., 1993, 2002; Milli, 1997; Giordano et al., 2003; Mancini & Cavinato, 2005; Mancini et al., 2007; Milli et al., 2008; Marra & Florindo, 2014; Tentori et al., 2025).
- Aerial view of the archaeological area in the historic centre of Rome (Google Earth™), highlighting the Palatine Hill and the Central Cryptoporticus (outlined in red). The Colosseum, Via dei Fori Imperiali, and the Tiber River are labeled for spatial context. The Central Cryptoporticus, located within the imperial palace complex, is a key architectural feature for understanding the interplay between urban settling and stratigraphic development of the Palatine sector.
A 3D aerial view of Rome highlighting Palatine Hill, Central Cryptoporticus, Colosseum, Via dei Fori Imperiali, and Tiber River, with a scale bar.
- Geological map of the study area showing the location of the Palatine Hill (dotted black square) and the Central Cryptoporticus (red dot), and trace of the a-a’ cross section (modified from Di Salvo et al., 2020). The geological units represent formal lithostratigraphic and synthemic classifications from the Geological Map of Italy, sheet 374 Rome, 1:50,000 (Servizio Geologico d’Italia, 2008; Funiciello & Giordano, 2008).
A geological map of Rome highlighting major hills and the Tiber River, accompanied by a detailed color-coded stratigraphic legend.
During the Plio-Quaternary, the study area underwent a progressive accumulation of sedimentary units (Milli, 1997; Bordoni & Valensise, 1998; Giordano et al., 2003; Mancini et al., 2007; Tentori et al., 2022), recording a transition from fully marine to coastal-transitional and finally continental depositional environments. These deposits reflect two major phases of infill and landscape evolution: an initial syn-rift phase spanning from the Early Pliocene to the Early Pleistocene, and a subsequent syn-uplift phase that began in the latest Early Pleistocene and continues into the Holocene. The continental succession is closely tied to the fluvial dynamics of the Tiber River system and includes a variety of fluvial channel-belt and floodplain deposits. These are interlayered with pyroclastic products associated with Middle Pleistocene volcanic activity from nearby districts, particularly the Colli Albani and Monti Sabatini volcanic systems (Fig. 3) (De Rita et al., 2002; Mancini & Cavinato, 2005; Funiciello & Giordano, 2008, 2010; Marra et al., 2009, 2014).
- Simplified schematic representation of the stratigraphic and facies architecture of the Palatine Hill (after Mancini et al. 2014) and the adjacent Labicano and Velabro Valleys with pictures (a,b,c) of key outcrops (refer to legend in Fig. 2 for lithostratigraphic and facies codes). In both panels, stratigraphic surfaces I to V represent the unconformities defining the synthems of Funiciello & Giordano (2008).
A stratigraphic cross-section of Palatine Hill and the Colosseum, with three field photos labeled a to c showing corresponding geologic layers.
The stratigraphic framework of the Palatine hill’s subsurface consists, from top to bottom, of anthropogenic deposits, alternating fluvial sediments and pyroclastic units, overlying a buried geological substrate composed of marine clays (Mancini et al., 2018). The uppermost unit consists of heterogeneous anthropogenic deposits accumulated over the last 3,000 years, comprising construction materials such as masonry, bricks, ceramics, foundations, and substructures, locally reaching up to 20 m in thickness. Below the anthropogenic deposits lies a succession of alluvial and volcanic deposits. The alluvial fill, dating from the Middle Pleistocene to Holocene, occupies deeply incised palaeovalleys carved by fluvial erosion, primarily associated with the Tiber and Palaeo-Tiber river systems. These nested palaeovalleys, reaching depths of up to 50 metres, were progressively infilled with fluvial sediments and volcanic materials sourced from the nearby Colli Albani and Monti Sabatini volcanic districts. Overlying and adjacent to these deposits, Middle Pleistocene fluvio-volcanic sequences (650–300 ka) form extensive interfluve areas dominated by pyroclastic units (Mancini et al., 2018). At the base of the succession lies the geological bedrock, composed of a thick series of Pliocene marine clayey-sandy sediments known as the Monte Vaticano formation (Servizio Geologico d’Italia, 2008; Funiciello & Giordano, 2008). Although not exposed at the surface, this unit underlies the continental deposits and forms the structural foundation of the Palatine Hill.
FOUNDATIONS SOIL AND INSTABILITY PHENOMENA
The foundation soils of the Central Cryptoporticus, are situated at the transition between the Middle Pleistocene interfluve zones and palaeovalleys, creating a geologically complex and heterogeneous subsurface (see geological map in Figure 2 and cross section of Figure 4). Specifically, the foundation structures rest upon three distinct lithostratigraphic units: the floodplain muds of the Aurelia formation (AEL), the tuffs and pyroclastic rocks of the Villa Senni formation, “Tufo Lionato” member (VSN1), and the fluvial silty and sandy deposits of the Fosso del Torrino (FTR) palaeovalley fill (Figs. 3 and 4).
- Geological cross-section of the Palatine Hill (a) (modified from Mancini et al., 2018) and Internal (b,c) and external (d) views of the Central Cryptoporticus. The section trace is shown in Fig. 2.
A geological cross-section of the Palatine Hill with color-coded strata layers and three archaeological site photos.
The complex stratigraphic and geomorphic conditions at the site contribute to recurring instability phenomena that compromise the structural integrity of the Central Cryptoporticus. Previous studies have documented structural failures attributed to foundation instability (Calabresi & Scarpelli, 1998; Croci & Biritognolo, 1998), emphasizing the site’s sensitivity to subsurface processes. The instability phenomena affecting this archaeological site are mainly due to lateral contact between the rigid lithoid tuffs and the ductile, inconsistent and highly compressible clay-rich deposits infilling the palaeovalleys. The lithological contrast, in fact, generates differential settlement and progressive subsidence, with consequent localised deformations, rotational movements and overturnings within the Central Cryptoporticus (Moscatelli et al., 2011).
Beyond lithological transitions, hydrogeological factors play a key role in undermining foundation stability. The high porosity of pyroclastic rocks facilitates water infiltration, and fluctuations in the groundwater table cause volume changes in fine-grained sediments. These include swelling and shrinkage of clay-rich layers, resulting in differential settlements. Simultaneously, the capillary rise and associated moisture fluctuations accelerate chemical weathering in volcanic deposits, gradually reducing their mechanical strength and load-bearing capacity.
An often-overlooked contributor to instability is the influence of pedogenic processes on soil mechanical behavior. Physical and chemical alterations of soil microstructure, such as carbonate cementation and aging, significantly affect strength and stiffness but are rarely accounted for in standard geotechnical models (Demars, 1982; Hauser et al., 2025). Soils with a developed microstructure, known as structured soils, are typically stronger and more resistant to deformation, even under small pressures compared to younger, unstructured soils, even when subjected to the same stress conditions and void ratios (Leroueil & Hight, 2003; Cotecchia & Chandler, 1997). However, when these microstructural bonds are disrupted, soils become “destructured,” resulting in pronounced declines in mechanical performance (Amorosi & Rampello, 1998). Aging processes, which encompasses both time-dependent processes (pure aging) and environmental modifications such as wetting-drying cycles, freeze-thaw activity, and groundwater fluctuations (Schmertmann, 1991) produce stratigraphic horizons with highly variable mechanical responses even within apparently homogeneous deposits. This is particularly relevant at archaeological sites like the Palatine Hill, where burial conditions, time-dependent diagenesis, and near-surface environmental exposure all contribute to microstructural evolution.
Compounding these natural challenges is the extensive network of artificial cavities develops beneath the Palatine Hill, further compromising the stability of the monuments (Mocchegiani Carpano, 1985). These cavities are distributed throughout both the geological bedrock and anthropogenic layers, with their aging and progressive deterioration increasing their susceptibility to collapse and subsidence. The presence of voids within the subsurface amplifies the effects of differential settlement and structural weakening, making them a key factor in the ongoing instability affecting the site.
Although the influence of geological, hydrogeological, and anthropogenic factors on structural stability is recognised, it remains insufficiently documented. A comprehensive mineralogical and geochemical characterisation of the foundation soils therefore represents an essential step for understanding the mechanisms governing mechanical stability, weathering susceptibility, and chemical alteration. By integrating these analyses, this research seeks to clarify the foundation conditions beneath the Central Cryptoporticus, generating critical data to guide conservation efforts and ensure the long-term preservation of the Palatine Hill’s archaeological heritage.
METHODS
This study is based on the facies and stratigraphic analysis of the Palatine Hill subsoil, where the Cryptoporticus archaeological site is located. Lithological and textural data were gathered from three coring boreholes (1MS, S1, and S2) and one stratigraphic trench (Figs. 5 and 6). These boreholes provided essential information for constructing schematic correlation panels, which illustrate the lateral and vertical stratigraphic relationships between the different lithological units (Fig. 7). The stratigraphy of the Palatine Hill, as defined in detail by Mancini et al. (2018), was used as a framework for this study, as it is considered a key site for Middle Pleistocene-Holocene continental deposits in the Rome basin. Additionally, to better characterize the foundation materials, 25 samples were collected from the boreholes and the stratigraphic trench (Fig. 5). These samples were investigated using a combination of portable X-ray fluorescence (pXRF), X-ray diffraction (XRD), petrographic thin-section analysis, and micromorphological characterisation. The pXRF and XRD analyses were performed on 16 samples from the S1 and S2 boreholes and from the trench, whereas the 1MS core was used exclusively for petrographic thin-section observations in order to document primary textures of the VSN deposits. Together, these methods allowed to constrain the mineralogical composition, textural characteristics, and pedogenetic features of the foundation materials. A complete sampling inventory, including depth, stratigraphic attribution, macroscopic and lithological description, and the analytical methods applied to each sample, is provided in Table 1.
- Stratigraphies intercepted by the S1, S2, and 1MS boreholes and within the trench site used for petrographic, micromorphological, X-ray diffraction (XRD), and X-ray fluorescence (XRF) analyses of foundation soil.
Stratigraphic log for 1MS, S1, and S2 boreholes and a trench site, displaying subsurface sediments with a detailed rock legend.
- Representative core photographs from 1MS, S1, S2, and the stratigraphic trench, illustrating the main lithofacies associations of the studied foundation units (h, AEL, VSN, and FTR). The AEL deposits show clear pedogenic horizonation, with CaCO3-depleted horizons (Ab) in the upper part of the unit and scattered carbonate nodules in the lower, carbonate-enriched horizon (Bk). An enlarged view of the VSN deposits at ~8 m depth highlights sparitic calcite veins occurring within the VSN unit. The location of core 1MS is shown in Fig. 4, while cores S1, S2, and the trench were collected from the interior pavement level of the Cryptoporticus.
A collage of core samples and trench site photos illustrating lithofacies associations of foundation units h, AEL, VSN, and FTR.
- Stratigraphic scheme in the Central Crytpoporticus area compared with stratigraphic logs of the S1, S2 and 1MS boreholes.
A cross-section showing subsurface strata around a cryptoporticus, comparing a normal layer succession against a cavity collapse model.
| Sample_ID | Site | Depth_m_bgs | Stratigraphic_unit | Horizon/facies | Lithological_class | Grain-size/fabric | Macroscopic_description | pXRF | XRD | Thin_section |
|---|---|---|---|---|---|---|---|---|---|---|
| CC1 | Trench | 0.3 | h (Anthropogenic cover/backfill) | Backfill (sand–pebble; mixed masonry) | Anthropogenic backfill / masonry | Sand–pebble; locally cobble-sized masonry clasts; silty–clayey matrix | Mixed backfill: sand–pebble volcaniclastic material with masonry fragments; variable carbonate content | x | x | x |
| CC2 | Trench | 0.55 | VSN (Villa Senni Fm.) | VSN? altered/reworked & pedogenised volcaniclastic | Altered/reworked and pedogenised volcaniclastic deposit | Silty clay to sandy silt; reworked volcaniclastic matrix; locally friable | Brown to yellowish-brown volcaniclastic sediment with variable clay enrichment and secondary carbonates; pedogenised/reworked | x | x | x |
| CC3 | Trench | 0.65 | VSN (Villa Senni Fm.) | VSN? altered/reworked & pedogenised volcaniclastic | Altered/reworked and pedogenised volcaniclastic deposit | Silty clay to sandy silt; reworked volcaniclastic matrix; locally friable | Brown to yellowish-brown volcaniclastic sediment with variable clay enrichment and secondary carbonates; pedogenised/reworked | x | x | x |
| CC4 | Trench | 0.75 | VSN (Villa Senni Fm.) | VSN? altered/reworked & pedogenised volcaniclastic | Altered/reworked and pedogenised volcaniclastic deposit | Silty clay to sandy silt; reworked volcaniclastic matrix; locally friable | Brown to yellowish-brown volcaniclastic sediment with variable clay enrichment and secondary carbonates; pedogenised/reworked | x | x | x |
| CC5 | Trench | 0.93 | VSN (Villa Senni Fm.) | VSN altered/pedogenised volcaniclastic | Altered/reworked and pedogenised volcaniclastic deposit | Silty clay to sandy silt; reworked volcaniclastic matrix; locally friable | Poorly lithified, friable, ash-rich pozzolanic material (VSN-derived); locally altered | x | x | x |
| CC6 | S2 borehole | 5,8 | AEL (Aurelia Fm.) | AEL upper horizon (oxide-rich/Bw?) | Pedogenised floodplain mud (soil/paleosol) | Clayey silt (locally silty clay); massive to weakly structured | Pale yellow to brown clayey silt with pedogenic features (mottling/redoximorphic nodules; carbonate accumulation in calcic horizons) | x | x | x |
| CC7 | S2 borehole | 6,5 | AEL (Aurelia Fm.) | AEL lower calcic horizon (Bk) | Pedogenised floodplain mud (soil/paleosol) | Clayey silt (locally silty clay); massive to weakly structured | Pale yellow to brown clayey silt with pedogenic features (mottling/redoximorphic nodules; carbonate accumulation in calcic horizons) | x | x | x |
| CC8 | S2 borehole | 6,8 | AEL (Aurelia fm.) | AEL lower calcic horizon (Bk) | Pedogenised floodplain mud (soil/paleosol) | Clayey silt (locally silty clay); massive to weakly structured | Pale yellow to brown clayey silt with pedogenic features (mottling/redoximorphic nodules; carbonate accumulation in calcic horizons) | x | x | x |
| CC9 | S2 borehole | 12,1 | VSN (Villa Senni fm.) | VSN pedogenised/altered volcaniclastic interval | Altered/reworked and pedogenised volcaniclastic deposit | Silty clay to sandy silt; reworked volcaniclastic matrix; locally friable | Brown to yellowish-brown volcaniclastic sediment with variable clay enrichment and secondary carbonates; pedogenised/reworked | x | x | x |
| CC10 | S1 borehole | 3,5 | h (Anthropogenic cover/backfill) | Backfill (sand–pebble; mixed masonry) | Anthropogenic backfill / masonry | Sand–pebble; locally cobble-sized masonry clasts; silty–clayey matrix | Mixed backfill: sand–pebble volcaniclastic material with masonry fragments; variable carbonate content | x | x | x |
| CC11 | S1 borehole | 6,1 | h (Anthropogenic cover/backfill) | Masonry/travertine-rich level | Anthropogenic backfill / masonry | Sand–pebble; locally cobble-sized masonry clasts; silty–clayey matrix | Carbonate masonry (travertine/limestone) fragment; very carbonate-rich | x | x | x |
| CC12 | S1 borehole | 8,5 | VSN (Villa Senni fm.) | VSN? altered/reworked & pedogenised volcaniclastic | Altered/reworked and pedogenised volcaniclastic deposit | Silty clay to sandy silt; reworked volcaniclastic matrix; locally friable | Brown to yellowish-brown volcaniclastic sediment with variable clay enrichment and secondary carbonates; pedogenised/reworked | x | x | x |
| CC13 | S2 borehole | 4,5 | h (Anthropogenic cover/backfill) | Backfill (sand–pebble; mixed masonry) | Anthropogenic backfill / masonry | Sand–pebble; locally cobble-sized masonry clasts; silty–clayey matrix | Mixed backfill: sand–pebble volcaniclastic material with masonry fragments; variable carbonate content | x | x | x |
| CC14 | S2 borehole | 16,5 | FTR (Fosso del Torrino fm.) | Alluvial sand–silt alternations (crevasse/overbank) | Alluvial sand–silt deposits (floodplain/crevasse-splay) | Interbedded laminated sand and silt; locally silty sand | Yellow to light brown laminated sands interbedded with silt; quartz–feldspar-rich with minor volcanic heavy minerals | x | x | x |
| CC15 | S2 borehole | 18 | FTR (Fosso del Torrino fm.) | Alluvial sand–silt alternations (crevasse/overbank) | Alluvial sand–silt deposits (floodplain/crevasse-splay) | Interbedded laminated sand and silt; locally silty sand | Yellow to light brown laminated sands interbedded with silt; quartz–feldspar-rich with minor volcanic heavy minerals | x | x | x |
| CC16 | S2 borehole | 15,1 | FTR (Fosso del Torrino fm.) | Alluvial sand–silt alternations (crevasse/overbank) | Alluvial sand–silt deposits (floodplain/crevasse-splay) | Interbedded laminated sand and silt; locally silty sand | Yellow to light brown laminated sands interbedded with silt; quartz–feldspar-rich with minor volcanic heavy minerals | x | x | x |
| 1MS0 | 1MS borehole | 1,3 | AEL (Aurelia fm.) | Pedogenised floodplain mud (AEL) | Pedogenised floodplain mud (soil/paleosol) | Clayey silt (locally silty clay); massive to weakly structured | Pale yellow to brown clayey silt with pedogenic features (mottling/redoximorphic nodules; carbonate accumulation in calcic horizons) | x | ||
| 1MS1 | 1MS borehole | 1,9 | AEL (Aurelia fm.) | Pedogenised floodplain mud (AEL) | Pedogenised floodplain mud (soil/paleosol) | Clayey silt (locally silty clay); massive to weakly structured | Pale yellow to brown clayey silt with pedogenic features (mottling/redoximorphic nodules; carbonate accumulation in calcic horizons) | x | ||
| 1MS2 | 1MS borehole | 2,5 | VSN (Villa Senni fm.) | VSN altered/pedogenised volcaniclastic | Altered/Pedogenised volcaniclastic deposit | Silty ash matrix with fine lapilli/lithic fragments (variable) | Weathered volcaniclastic material; primary pyroclastic textures partly obscured by pedogenesis/alteration | x | ||
| 1MS3 | 1MS borehole | 3,3 | VSN (Villa Senni fm.) | VSN altered/pedogenised volcaniclastic | Altered/pedogenised volcaniclastic deposit | Silty ash matrix with fine lapilli/lithic fragments (variable) | Weathered volcaniclastic material; primary pyroclastic textures partly obscured by pedogenesis/alteration | x | ||
| 1MS4 | 1MS borehole | 4,3 | VSN (Villa Senni fm.) | VSN Pozzolanelle facies (pozzolanic tuff) | Pozzolanic tuff (poorly lithified pyroclastic deposit) | Ash-rich tuff: fine ash matrix with lapilli-sized juvenile/lithic clasts | Friable, ash-rich pozzolanic tuff (Pozzolanelle); locally pedogenised/altered; porous matrix-supported fabric | x | ||
| 1MS5 | 1MS borehole | 5,1 | VSN (Villa Senni fm.) | VSN Pozzolanelle facies (pozzolanic tuff) | Pozzolanic tuff (poorly lithified pyroclastic deposit) | Ash-rich tuff: fine ash matrix with lapilli-sized juvenile/lithic clasts | Friable, ash-rich pozzolanic tuff (Pozzolanelle); locally pedogenised/altered; porous matrix-supported fabric | x | ||
| 1MS6 | 1MS borehole | 6,4 | VSN (Villa Senni fm.) | VSN Pozzolanelle facies (pozzolanic tuff) | Pozzolanic tuff (poorly lithified pyroclastic deposit) | Ash-rich tuff: fine ash matrix with lapilli-sized juvenile/lithic clasts | Friable, ash-rich pozzolanic tuff (Pozzolanelle); locally pedogenised/altered; porous matrix-supported fabric | x | ||
| 1MS7 | 1MS borehole | 10,2 | VSN (Villa Senni fm.) | VSN Tufo Lionato facies (lithoid tuff) | Lithoid tuff (welded/lithified pyroclastic rock) | Lithified tuff; juvenile/lithic clasts in cemented ash matrix | Massive, coherent lithoid tuff (Tufo Lionato); welded/cemented; fractures/veins locally carbonate/zeolite-cemented | x | ||
| 1MS8 | 1MS borehole | 11,5 | VSN (Villa Senni fm.) | VSN Tufo Lionato facies (lithoid tuff) | Lithoid tuff (welded/lithified pyroclastic rock) | Lithified tuff; juvenile/lithic clasts in cemented ash matrix | Massive, coherent lithoid tuff (Tufo Lionato); welded/cemented; fractures/veins locally carbonate/zeolite-cemented | x |
Micromorphological and Petrographic Analysis
In addition to visual examination of sediment texture and composition, petrographic and micromorphological characterisation was carried out to investigate lithofacies and pedogenetic features (Figs. 8 and 9). Micromorphological analyses followed the guidelines of Stoops (2021), which facilitated the detailed description of features within the soil. These observations provided insights into post-depositional and polyphase pedogenic processes affecting the samples.
Eight thin-section photomicrographs, A to H, showing soil microstructures like Fe-nodules, carbonate lithics, and coatings with scale bars.
- Representative microphotographs in plain-light and cross polar view of soil samples from the AEL and pedogenised VSN units. Sample labels include CC-numbered samples analyzed by XRF and XRD, and additional thin sections from the 1MS core for petrographic documentation only. A-B) Orthic Fe-oxyhydroxide nodules with sharp boundaries embedded in a brown micromass composed of micrite, clays, and fine quartz; silt-sized quartz and feldspar grains occur both inside and outside the nodules, indicating in situ formation. Silt and sand-sized grains are monomineralic grains of quartz, feldspar and pyroxene; C) Orthic Fe-nodule with concentric layers and sharp boundaries and micritic limestone fragments in crystallitic micritic b-fabric. The outer layers of the iron-nodule diffuse into the groundmass and is indicative of an in situ formation process; D) Concentric impregnative orthic aggregate nodule precipitated around a pyroxene grain. The presence of an intact weatherable mineral such as pyroxene suggest that the nodules protected the mineral from further weathering. E) Aggregate nodule formed by an aggregation of small nodules in crystallitic b-fabric; F) Root-related hypocoating developed in a sericitic–calcitic crystallitic b-fabric. G) Polygenetic multilayered Fe-bearing nodule associated with clay coatings and clay domains. The internal part of the multi-layered nodule is composed of an iron oxyhydroxides nucleus impregnating the groundmass, clays and silt-sized quartz grains. This nucleus is detached from a cortex formed by multiple thin layers of clay and hydroxides of the same mineralogical composition; H) Micritic limestone fragment in a micritic-clayey micromass with scattered silicate grains and local oxide impregnations; I) Thick, laminated to crescent-shaped illuvial clay coatings and clay infillings lining voids and grain margins. Note the etched pyroxene grain; J) Laminated to crescent-shaped clay coatings lining voids, indicating clay illuviation; K) Dark clay-rich laminae, clay infilling, and sparitic carbonate infilling; L) Carbonate-rich microfabric with sparitic calcite infillings and spiritic to micritic coatings around pores and clasts, locally associated with clay infillings; M) Carbonate coatings and infillings around rounded pores, consistent with calcified root-related voids and secondary carbonate precipitation; N) Altered volcaniclastic fragment (leucititic lava) and orthic micritic carbonate nodule trapping mineral grains (including pyroxene), within a micrite and clay-rich micromass; O) Fragmented clay infilling and relict clay coatings within the groundmass, indicating illuviation and post-depositional disturbance; P) Pedoturbation features related to clay swelling and shrinking, associated with domains of oriented clay.
Eight thin-section photomicrographs I to P, showing microscopic soil features like clay coatings, carbonate infillings, and scale bars.
Eight thin-section photomicrographs A to H showing altered volcanic grains, etched pyroxene, and zeolite cements with scale bars.
- Representative thin sections photomicrographs in plain-light and cross polar view of pyroclastic rock samples from the VSN unit (samples 1MS 4,5,6,7, and 8, and CC5, CC9, and CC10, CC12, which were also analyzed by XRF and XRD). A) Crystal-rich tuff with abundant phenocryst fragments (clinopyroxene and feldspar) dispersed in a fine ash matrix with localised alteration of the groundmass ; B) Altered leucitite lava, crystals of pyroxene, sanidine, and biotite embedded in an ash-rich matrix locally cemented by zeolite; C) Highly weathered pyroxene and scoriaceous juvenile fragment with rounded and elongated vescicles. Glass fragments are mainly yellow sideromelane (mafic glass) and orange palagonite (the alteration product of sideromelane); D) Pervasive zeolite cementing altered pyroxene crystals in a clayey matrix; E) Enlarged view of leucitite rock fragment consisting essentially of leucite phenocrysts in a fine-grained groundmass composed mainly of glass, feldspar and pyroxenes; F) Microlite-rich volcanic rock fragment consisting of small crystals of leucite, clinopyroxene, and k-feldspar (mainly sanidine). G) Preferentially oriented and elongate scoria clasts defining a weak eutaxitic to flow-aligned fabric, consistent with compaction and incipient welding; H) Vesicular juvenile clasts with elongated vescicles filled by zeolite cement; I) and J) Vescicular juvenile lapilli and vitric ash matrix with abundant vesicles and pores (black, epoxy filled) and locally palagonitised or clay altered shard margins. Rock fragments include mainly limestone and leucititic lavas; and crystal fragments, mainly leucite, clinopyroxene, sanidine, biotite, cemented by calcite. Secondary calcite is precipitated by meteoric water in voids and veins, and replaces silicates through a solution–precipitation process; K) Pervasive zeolite and calcite cement; L) Zeolite and calcite cement filling equant and elongate vescicles within juvenile clasts; M) Rounded micritic limestone xenolith with alteration rim in a clay-rich groundmass; N) Micritic limestone xenolith and secondary sparitic carbonate cement; O) and P) Sparite cement filling veins and fractures.
Eight thin-section photomicrographs, I to P, showing carbonate lithic clasts, secondary calcite, and sparite-filled fractures with scale bars.
X-ray Fluorescence (XRF) Analysis
Bulk chemical analysis was performed on untreated samples using a Bruker Tracer 5i handheld XRF spectrometer, utilizing two integrated calibration settings: GeoMining and GeoExploration (Table SM1). The GeoMining calibration is optimised for high accuracy in major and minor elements, while GeoExploration is designed for trace elements. The instrument operates in three phases (30 kV, 50 kV, 15 kV) to cover a broad range of detectable elements. The calibration settings were chosen according to the sample composition, allowing for comprehensive chemical analysis. Portable XRF provides semi-quantitative concentrations for selected elements; volatile components (e.g., H2O, CO2) and Loss on Ignition (LOI) cannot be determined and are therefore not included. Accordingly, the XRF dataset is used to evaluate relative geochemical trends among stratigraphic units rather than absolute bulk-rock compositions. Elemental concentrations were converted to major-oxide wt% and renormalised to 100% on an anhydrous basis (LOI excluded) prior to covariation plots and multivariate analysis (Table SM1 and Figs. 10 and 11).
- Binary covariation diagrams showing major-element relationships for the analyzed samples after recalculation to an anhydrous basis (LOI excluded) and normalisation to 100 wt%. Plots illustrate CaO–SiO2, Al2O3–SiO2, MgO–SiO2, and total Fe2O3–SiO2 relationships. The diagrams highlight compositional differences among stratigraphic units and emphasize alteration trends within VSN-derived materials, including variable Ca enrichment linked to carbonate redistribution and relative enrichment in Al- and Fe-bearing phases associated with pedogenesis.
Four geochemical scatter plots showing CaO, Al2O3, MgO, and Fe2O3 versus SiO2 for color-coded geologic formations.
- Principal component analysis (PCA) biplot of anhydrous, 100 wt% normalised major-oxide compositions derived from portable XRF data (Table 1). PCA was performed on selected major oxides after standardisation and is used here as an exploratory tool to summarize covariation patterns among bulk chemical components. The first principal component (PC1) primarily reflects the contrast between carbonate-dominated compositions (CaO) and silicate-rich components (SiO2 and Al2O3), whereas the second principal component (PC2) captures variability associated with mafic and accessory major oxides (Fe2O3, TiO2, MnO, and MgO).
A PCA biplot of anhydrous-normalized major oxides showing PC1 versus PC2, with samples grouped by color-coded geological units.
Principal Component Analysis (PCA) was applied to the anhydrous, 100 wt% normalised major-oxide dataset to summarize dominant compositional gradients and covariation among samples. PCA was performed on the correlation matrix (i.e., using z-score standardised variables). Sample scores and variable loadings are reported as a biplot (Fig. 11). To support interpretation of the oxide-based PCA, we additionally calculated a set of compositional indices (e.g., SiO2/Al2O3, SiO2/Fe2O3, Si/Ses, Si/R, and base/resistant and base/acid ratios; definitions in Table SM2) and performed a companion PCA on the standardised index matrix (Fig. SM1).
X-ray Diffraction (XRD) Analysis
To assess the bulk mineralogy of the samples, XRD analysis was conducted on sample, micronised under 70 microns in size by vibrating rotary cup mill, using a Bruker D2 Phaser desktop diffractometer, equipped with LYNXEYE XE-T detector and operating in a Bragg-Brentano θ/2θ geometry. The acquisition conditions were set at 30 kV and 10 mA, and scans were performed from 0° to 80° 2Θ with a count time of 2 seconds per step. The XRD data were processed using DIFFRAC.EVA software, combined with the Crystallography Open Database (COD) to identify the mineral phases. Mineral proportions (Fig. SM2) are semi-quantitative estimates derived from Bruker DIFFRAC.EVA based on relative XRD peak intensities of the identified phases (COD database). Values are normalised to 100% of the identified crystalline phases and are intended to illustrate relative compositional contrasts rather than absolute modal abundances. The identifiers of the COD can be freely obtained (http://www.crvstallographv.net/cod/) since COD is an open-access database. Because XRD was performed on bulk powdered samples, identification of clay minerals is qualitative and does not resolve mixed-layer phases or illite crystallinity; detailed clay mineralogy (e.g., <2 μm oriented mounts with glycolation/heating) is beyond the scope of this study.
Through these combined techniques, this study provides a detailed understanding of the stratigraphy, mineralogy, and pedogenetic processes affecting the geological deposits in the study area, offering insights into the historical development of the site’s foundation materials. The information obtained from the analytical techniques was used to investigate the intrinsic characteristics of the deposits and the processes that influenced their susceptibility to weathering.
RESULTS AND DISCUSSION
Macroscopic description of sampled stratigraphic intervals
Anthropogenic deposits (h)
Anthropogenic deposits from S1, S2 and 1MS boreholes and the stratigraphic trench consist of backfill material and remnants of Roman-age structures (masonry), reaching a maximum thickness of 8.50 m in the studied boreholes (Fig. 5 and 6). The backfill comprises sand- to pebble-sized clasts within a volcaniclastic silty-clayey matrix, while the masonry primarily consists of large rock fragments and masonry material, including travertine, volcanic tuff, and lava.
Aurelia formation (AEL)
This unit comprises ~1–4 m of pale-yellow clayey silt that displays well-developed pedogenic overprint. In the upper part, horizons are locally CaCO3-depleted (Ab) and show mottling, root traces, and orange–red Fe3+-oxide cutans. Down-profile, the unit grades into a carbonate-enriched horizon (Bk) characterised by scattered to locally more abundant subspheroidal CaCO3 nodules, together with persistent mottling and redox-related features (Fig. 6). These sediments, exposed on the Palatine Hilltop (Fig. 3), represent fine-grained floodplain deposits that have undergone pedogenesis under dominantly oxidizing conditions with episodic saturation and redox fluctuations.
Villa Senni formation (VSN)
The Villa Senni formation (VSN) in the study area comprises a volcaniclastic succession that includes both primary pyroclastic deposits and variably altered to pedogenised derivatives of the same parent material. Where primary textures are best preserved (notably in the 1MS core), the unit can be subdivided into an upper Pozzolanelle facies and a lower Tufo Lionato facies. The Pozzolanelle facies consists of poorly lithified, friable, ash-rich pozzolanic material with a porous, matrix-supported fabric, whereas the Tufo Lionato facies is represented by a massive, coherent lithoid tuff characterised by a higher degree of lithification and cementation.
In the 1MS borehole, the VSN is well developed and reaches a total thickness of approximately 10 m, of which the Tufo Lionato accounts for ~5 m (Figs. 5 and 6). Here, the Tufo Lionato is a dark red to orange-brown lithoid tuff with an ash-rich matrix containing sparse phenocrysts (e.g., leucite, biotite and clinopyroxene), rare limestone lithics, and grey to black scoriaceous juvenile clasts. Induration is heterogeneous, ranging from strongly cemented intervals, locally associated with zeolitisation and calcite veining, to more weakly consolidated, scoriaceous zones that grade upward into pozzolanic material (Figs. 5 and 6). These features are consistent with deposition from dense pyroclastic flows and subsequent early diagenetic cementation.
In contrast, the VSN encountered in the S1 and S2 boreholes and at the trench site is predominantly represented by altered/reworked and pedogenised volcaniclastic material derived from the VSN parent deposits (Fig. 5). Primary lithoid and pozzolanic textures are commonly obscured, and the deposits are characterised by friable matrices, pervasive discoloration, carbonate concretions and coatings, Fe-Mn oxide nodules/hypocoatings, mottling, and locally developed pedogenic horizons. Macroscopically, these intervals appear as brown to yellowish-brown volcaniclastic sediments with variable fine-matrix enrichment and carbonate accumulation, reflecting pedogenic overprint and carbonate redistribution under shallow subsurface conditions.
At the trench site, the VSN shows a vertical differentiation between a lower volcaniclastic interval and an overlying horizon of reworked/pedogenised volcaniclastic material, separated by diffuse boundaries rather than sharp lithostratigraphic contacts (Fig. 5). The upper pedogenised interval is marked by increased carbonate accumulations, finer textures, and secondary structures consistent with in situ soil development and repeated moisture fluctuations. Similar alteration features are observed in S1 and S2, where VSN-derived horizons occur beneath anthropogenic deposits and above older sedimentary units.
Overall, the combined macroscopic evidence indicates that the Villa Senni formation includes a primary lithoid and pozzolanic Tufo Lionato facies preserved in the 1MS borehole, whereas much of the VSN exposed in shallow subsurface contexts and at the trench site corresponds to pedogenised and reworked volcaniclastic material derived from the same volcanic parent deposits. This distinction is critical for interpreting subsequent mineralogical and geochemical data, particularly where altered horizons may differ significantly from primary pyroclastic lithotypes.
Fosso del Torrino formation (FTR)
The FTR consists of yellow to light brown laminated sands composed of quartz, feldspar, muscovite, and volcanic ferromagnesian minerals such as augite, biotite, and magnetite. These sands are interbedded with decimetre-thick silt layers (Figs. 5 and 6). The observed alternation of sand- and silt-dominated intervals, together with pervasive lamination, is consistent with deposition from low-energy overbank processes punctuated by episodic higher-energy flooding events, typical of distal floodplain and crevasse-splay environments. This interpretation is in agreement with the regional stratigraphic framework proposed by Mancini et al. (2018), who interpreted the FTR in the Cryptoporticus area as predominantly out-of-channel and crevasse-splay deposits later modified by pedogenesis.
Petrographic and micromorphological observations
Petrographic and micromorphological observations were carried out on a subset of samples that were also analyzed by XRF and XRD, as well as on additional thin sections collected from the 1MS core to document vertical textural variability (Table 1, and Figs. 8 and 9). Samples examined in thin section and directly comparable with the geochemical and mineralogical datasets include CC2, CC3, CC4, CC5, CC6, CC7, CC8, CC9, CC10, CC12, and CC13. Additional thin sections from the 1MS core at depths of 1.30, 1.90, 2.50, 3.30, 4.30, 5.10, 6.40, 10.20, and 11.50 m were analyzed exclusively for petrographic characterisation and are used to support textural and micromorphological interpretations.
Soil Micromorphology of AEL and VSN pedogenised deposits
The petrographic and micromorphological analysis of the Aurelia formation sedimentary deposits and volcaniclastic pedogenised intervals of the VSN Formation (samples CC2, CC3, CC4, CC5, CC6, CC7, CC8, CC9, CC10, CC12, and CC13), reveals a complex assemblage of coarse mineral constituents, micromass components, and various pedofeatures indicative of significant post-depositional processes (Fig. 8).
Coarse Mineral Constituents
The coarse fraction of the samples is dominated by detrital monomineralic grains, primarily quartz, feldspar, pyroxenes, and mica (Fig. 8A,B,C,D,N). These grains are predominantly silt- to sand-sized and together with volcanic rock fragments and siliciclastic–carbonate lithic fragments, point to a mixed sedimentary and volcanic provenance. Rare gastropod shell fragments suggest biogenic influences during sedimentation. The coarse mineral components often show weathering features, including intense etching and alteration, particularly in volcanic-derived materials (Fig. 8I), pointing to early post-depositional pedogenic transformations.
Micromass
The micromass is mainly composed of clay minerals and micritic carbonate (Fig. 8E,F,I,N). The undifferentiated b-fabric is observed as a uniform and fine-grained structure (Fig. 8P), suggesting little to no directional alignment in the clay particles. In contrast, the striated b-fabric exhibits subtle layering or linear structures, indicating some level of mineral alignment or compaction. This could reflect the processes of settling or pedogenetic alteration over time.
The sericitic and calcitic crystallitic b-fabric (Fig. 8C,E,F) shows evidence of fine crystal formation, with sericite (a fine-grained mica) and calcite crystals developing within the micromass. These crystals typically form during pedogenesis and indicate secondary mineralisation linked to groundwater fluctuations. Granostriated b-fabric represents a granular structure with fine, striated layers that may have resulted from physical processes like compression or shearing. The porostriated b-fabric (Fig. 8O) reveals a porous structure with striated alignments, which may indicate void spaces created during pedogenic processes, perhaps due to organic activity or evaporation.
Common pedofeatures
Redoximorphic Features
Iron and manganese hydroxide nodules are commonly observed within the groundmass (Fig. 8C,D,E), displaying varying morphologies such as concentric, layered, and dendritic forms. These nodules are indicative of redoximorphic mobilisation and reprecipitation associated with wet-dry and oxidation-reduction cycles. The different shapes reflect varying stages of mineral accumulation, with dendritic forms indicating rapid precipitation, while concentric and layered forms suggest slower, more stable conditions for the deposition of these minerals. Fe-Mn oxide nodules reflect the weathering of iron and manganese minerals in the soil, contributing to the formation of these oxide-rich features within the sediment matrix. Fe- and Mn-hydroxide coatings, hypocoatings, and quasicoatings are found on mineral grains and other components (Fig. 8P), forming thin layers on mineral surfaces. Hypocoatings represent coatings that form beneath the surface of the grains, indicating a later stage of deposition, while quasicoatings are partial and irregular, reflecting incomplete precipitation.
Calcium Carbonate Features
Calcite nodules are observed in both orthic and anorthic forms, with concentric and layered structures. These nodules are a result of the precipitation of calcium carbonate, likely from groundwater or biological activity. Similar calcite coatings (Fig. 8M), hypocoatings, and quasicoatings are observed on mineral grains, indicating that calcite has precipitated over time. Calcite depletion hypocoatings represent areas where calcite has been selectively leached or dissolved, leaving behind voids or weaker layers within the sediment, further suggesting the influence of weathering processes. Infillings of sparitic calcite are also present (Fig. 8M), filling voids and fractures within the sediment. These infillings represent calcite deposition from groundwater or percolating fluids by meteoric water in voids and veins through a solution–precipitation process.
The presence of tubiform pores (Fig. 8M) suggests calcification processes, which may have been influenced by biological activity such as root systems or microbial precipitation.
Clay coatings/infillings
Clay infillings and clay coatings are common, with clay particles filling void spaces or forming thin layers on mineral grains. The presence of illuvial clay (Fig. 8I,J) suggests that fine-grained clay particles have been translocated from upper horizons and accumulated in lower soil layers, a process known as illuviation.
Pedofeatures in volcanic-rich sediments and pedogenised tuff
Volcanic components and their alteration products in soil horizons indicate the influence of pedogenetic processes on volcanic parent material (e.g., volcanic tuff), whether in situ or reworked. During the initial stages of weathering, especially of the volcanic glass, extensive accumulations of non-crystalline or poorly crystalline Si-Al components, such as allophane and imogolite, are produced.
Petrographic description of VSN pyroclastic deposits
Texture
The Villa Senni formation samples examined in thin section (1MS core samples 4,5,6,7, and 8; Fig. 9) consist of poorly sorted pyroclastic-flow deposits characterised by a wide grain-size ranging from fine ash (< 1/16 mm) to lapilli-sized fragments (2-64 mm). The coarse fraction (approximately 10–20 vol%) is represented by clasts ranging from ~500 μm to >2 mm and includes juvenile vescicular clasts (scoria), volcanic lithic clasts, and accidental carbonate lithics (micritic limestone) (Fig. 9A,B,C,D,M). Juvenile scoriaceous fragments display rounded to elongate vesicles, locally deformed, consistent with rapid degassing during magma fragmentation and emplacement within a pyroclastic flow (Fig. 9G,H).
Ash matrix
The deposits are supported by an ash matrix consisiting of finegrained interstitial fraction (fine ash, crystal fragments, and altered vitric material) (Fig. 9A, B). This ash matrix is extensively altered and locally replaced by secondary zeolites, clay minerals, and calcite cement (Fig. 9D,I,J,K,L,N). The ash matrix represents the most reactive component of the deposit and is the principal locus of post-depositional mineralogical transformation.
Pyroclastic components
The primary components of the VSN pyroclastic deposits include juvenile clasts, consisting of scoriaceous clasts and glass shards (Fig. 9A, B, G, H), crystal fragments, dominated by leucite, clinopyroxene, sanidine, and subordinate biotite (Fig. 9A–D), lithic clasts, subdivided into volcanic lithics (lava fragments) and accidental lithics, mainly micritic limestone fragments derived from the sedimentary substrate (Fig. 9M, N, J). The presence of carbonate lithics indicates incorporation of substrate material during magma ascent or pyroclastic-flow propagation.
Diagenetic cementation
Lithification of the VSN pyroclastic deposits is primarily controlled by post-depositional zeolitisation of the the vitric ash matrix (Fig. 9D,K), driven by the circulation of meteoric and/or groundwater shortly after emplacement. Zeolitisation involves the replacement of volcanic glass by zeolite minerals under under low-temperature hydrous conditions, progressively increasing framework cohesion and mechanical strength.
Calcite cement locally fills pores, fractures, and vesicles and replaces silicate phases through solution-precipitation processes (Fig. 9I–L, N). Calcite cementation commonly preserves primary pyroclastic textures and occurs as void-filling sparite and fracture cement.
Post-depositional alteration
Juvenile clasts, volcanic lithics, and carbonate lithics commonly show evidence of alteration. In the scoria clasts, volcanic glass is variably transformed into palagonite and allophane, forming alteromorphs that preserve the original clast morphology (Fig. 9C,G,H,I,L). Yellow sideromelane represents relatively unaltered glass, whereas orange palagonite reflects hydration, oxidation, and Fe enrichment during alteration. These processes are enhanced by interaction with meteoric water and soil-derived solutions.
Another observed mechanism contributing to post-depositional weakening is the hydration-induced swelling of zeolite and clay minerals derived from glass alteration (Fig. 9A,B,D,M). Expansion of these hydrous phases generates microfractures and disrupts grain contacts, leading to reduced cohesion and shear strength at both the microscale and engineering scale.
XRF Chemical Analysis
Portable XRF measurements were acquired on 16 samples from S1, S2, and the trench site to evaluate bulk compositional contrasts among anthropogenic cover, sedimentary units, and VSN-derived volcaniclastic materials. Major-oxide compositions were used semi-quantitatively to compare stratigraphic units and evaluate relative carbonate enrichment versus silicate framework contributions (Figures 10 and 11).
Summary statistics and major-oxide covariation
In the anhydrous, 100 wt% normalised oxide dataset (Table SM1), SiO2 ranges from 0.91 to 72.02 wt% (mean 46.01 wt%), Al2O3 from 1.10 to 22.29 wt% (mean 10.17 wt%), CaO from 1.58 to 92.02 wt% (mean 24.60 wt%), and total Fe as Fe2O3 from 0.13 to 24.52 wt% (mean 9.97 wt%). MgO varies from 0.00 to 7.99 wt% (mean 3.12 wt%), broadly tracking the relative contribution of mafic minerals versus carbonate-rich components.
The binary covariation diagrams (Fig. 10) highlight first-order mixing between carbonate-rich and silicate-rich endmembers. CaO decreases systematically with increasing SiO2, consistent with variable carbonate dilution/addition (primary lithic carbonate and/or secondary carbonate cement) superimposed on volcaniclastic and siliciclastic frameworks. VSN-derived materials span the widest compositional range, indicating heterogeneous degrees of carbonate enrichment and matrix alteration, whereas FTR samples cluster toward higher SiO2 and moderate Al2O3-Fe2O3, consistent with quartz-rich compositions.
The PCA biplot (Fig. 11) provides an exploratory synthesis of these relationships. PC1 primarily expresses the carbonate versus silicate contrast (CaO opposed to SiO2-Al2O3), whereas PC2 reflects variability in mafic-associated oxides (Fe2O3-TiO2-MnO-MgO). The first two components explain 51.1% (PC1) and 19.4% (PC2) of the total variance (70.6% cumulative), and eigenvalues indicate that the dominant gradients are captured by a limited number of components (Table SM1). Together, Figs. 10 and 11 support the interpretation that the principal geochemical variability beneath the Cryptoporticus is driven by carbonate redistribution/cementation and by variable preservation and alteration of the silicate framework, particularly within VSN-derived horizons. Consistent with these trends, compositional indices (Table SM2) show systematic variation among units: silica ratios (e.g., SiO2/Al2O3, Si/R) increase toward FTR samples, whereas VSN-derived horizons span the widest range, reflecting variable preservation/alteration of the silicate framework and carbonate enrichment. A companion PCA performed on the standardised index dataset (Fig. SM2) reproduces the same firstorder geochemical gradients, supporting the robustness of the carbonate–silicate contrast inferred from the covariation plots.
XRD analysis for the studied stratigraphic intervals
This section presents the XRD analysis results for bulk powdered samples from the trench and the S2 and S1 boreholes (Figure SM2). Mineral proportions represent semi-quantitative estimates of the identified crystalline phases and are normalised to 100%; they are used to emphasize down-profile trends in carbonate accumulation, zeolitisation, and alteration of volcaniclastic components. Because analyses were performed on bulk powders, clay minerals are treated here at a reconnaissance level and the data do not resolve mixed-layer phases (e.g., I/S), hydroxy-interlayered vermiculite, or clay crystallinity indices; therefore, the absence of these phases in the bulk diffractograms should not be interpreted as evidence that they are not present.
Trench site
The trench profile records a mineralogical progression from the anthropogenic cover to volcaniclastic substrate units. The anthropogenic sample (CC1) is dominated by quartz and feldspar, with subordinate calcite and mica, consistent with mixed backfill material. The underlying interval sampled at CC2-CC4 is interpreted as altered/reworked and pedogenised volcaniclastic material: it retains a quartz-feldspar framework but shows variable enrichment in carbonates and a mica/clay minerals signal, consistent with pedogenic overprint and carbonate redistribution; in this shallow setting, carbonate phases may reflect a combination of lithic carbonate inputs and secondary cementation, rather than primary volcanic mineralogy. The basal trench sample (CC5) differs from the overlying assemblages by a stronger juvenile/volcanic signature, expressed by clinopyroxene (diopside) together with feldspar and quartz, consistent with a volcaniclastic parent material in which primary silicates are largely preserved while alteration is preferentially focused on the vitric fraction. The absence of a strongly zeolite-rich assemblage at this level supports interpretation of CC5 as VSN-derived volcaniclastic material (locally altered) rather than a fully developed, strongly zeolitised lithoid tuff facies.
S2 core
In S2, the uppermost sample (CC13) shows a mixed assemblage dominated by quartz and feldspar, with substantial calcite and mica/clay minerals, consistent with reworked anthropogenic/backfill material. Samples from the Aurelia formation interval (CC6-CC8) are characterised by high carbonate contents (calcite) associated with feldspar and mica, and locally minor kaolinite, consistent with pedogenised fine-grained floodplain deposits containing carbonate accumulation. Kaolinite is detected locally in bulk XRD, but detailed interpretation of clay transformation pathways (e.g., mica to vermiculite or mixed-layer I/S) cannot be evaluated from bulk-powder XRD alone.
The VSN-derived interval (CC9) is distinguished by the appearance of analcime together with feldspar, quartz, and calcite, indicating low-temperature alteration of the volcaniclastic/vitric fraction while preserving primary silicate phases. Deeper samples assigned to the FTR interval (CC14-CC16) are dominated by quartz and calcite with persistent feldspar and minor diopside, consistent with alluvial sand-silt deposits containing both carbonate components and a subordinate volcanic heavy-mineral contribution.
S1 core
The S1 core shows strong mineralogical heterogeneity tied to anthropogenic mixing and alteration of volcaniclastic material. Sample CC11 is composed almost entirely of calcite, consistent with a carbonate-rich masonry/travertine component within the anthropogenic cover. In contrast, CC10 contains abundant zeolites (phillipsite and chabazite-K) with feldspar, minor clinopyroxene (diopside), and subordinate calcite, consistent with reworked volcaniclastic material in which the vitric component has been altered to zeolites; subordinate calcite may reflect secondary cementation and/or incorporation of carbonate debris within the anthropogenic sequence. The deeper S1 sample (CC12) contains abundant feldspar and quartz with associated calcite and dolomite and mica, supporting attribution to a pedogenised/altered volcaniclastic horizon derived from VSN parent material, where secondary carbonates and clay-bearing phases coexist with preserved primary silicates.
ALTERATION AND DEGRADATION PROCESSES BENEATH THE CRYPTOPORTICUS
This section integrates the stratigraphic observations (Figs. 6 and 7), micromorphology and petrography (Figs. 10 and 11), and the geochemical and mineralogical results (Figs. 8, 9 and SM2) to derive a model for alteration and progressive mechanical degradation of the volcanic-sedimentary substrate beneath the Cryptoporticus. Figure 12 synthesizes the key pedofeatures and alteration pathways inferred from the combined datasets and is discussed below.
- Pedofeatures (left) versus degradation mechanisms (right) in the Cryptoporticus foundation. Principal diagnostic fabrics across the studied lithofacies include Fe/Mn–rimmed clay-iron coatings, illuvial clay cutans, calcite nodules and vertic slickensides in the AEL; clay cutans, calcite and Fe–Mn nodules in the weathered VSN; and calcite-filled fractures in the unweathered VSN. Dashed horizons mark transitions between units. The same stratigraphy is shown beneath the Cryptoporticus vault, now emphasising active pathways: meteoric infiltration and daily moisture fluctuations at the surface; oxidation, carbonate dissolution–precipitation and shrink–swell within the clay-rich AEL; water-table rise and fall that links surface weathering with deeper zones; zeolite hydration and softening in the tuff; and the potential collapse of an ancient quarry cavity, with secondary calcite sealing of cracks. Together, the panels illustrate how inherited pedogenic fabrics control present-day hydro-chemical processes and define the depth-dependent zones of mechanical weakness that threaten the monument.
A diagram comparing stratigraphic pedofeatures with structural degradation processes around an ancient cryptoporticus.
Reference framework and practical constraints
In the study area, the Villa Senni formation (VSN) includes both (i) relatively preserved primary pyroclastic deposits (lithoid Tufo Lionato facies in the 1MS core) and (ii) variably altered, reworked and pedogenised volcaniclastic material (trench, 1MS, S1, and S2 cores). The 1MS lithoid tuff provides the main textural characters for the primary volcanic substrate because it preserves juvenile clasts, crystal fragments and a vitric ash matrix, together with diagenetic zeolite and calcite cementation (Fig. 11). In contrast, the shallow subsurface beneath and around the Cryptoporticus is dominated by altered/pedogenised horizons where primary pyroclastic textures are partially obscured by clay-carbonate micromass development, redoximorphic features, and secondary carbonate redistribution (Fig. 10).
A direct whole-rock mineralogical-geochemical comparison between relatively preserved and strongly altered VSN deposits is limited by the absence of XRD and pXRF analyses for the 1MS core. Consequently, the 1MS thin sections are used as a reference for preserved primary pyroclastic textures and early diagenetic cementation. These observations are compared with petrographic and micromorphological characteristics from the trench, S1, and S2 cores, together with their corresponding bulk XRD and pXRF signatures. Because XRD was performed on bulk powders, clay minerals are treated here at a reconnaissance level (presence/relative enrichment), and mixed-layer relationships and clay crystallinity cannot be assessed without dedicated clay-fraction (e.g., <2 μm oriented) analyses.
Alteration indicators and style
Textural indicators
The 1MS lithoid facies retains a framework of juvenile vesicular pyroclasts and volcanic lithic fragment, together with abundant crystal and phenocryst fragments, supported by a vitric ash matrix (Fig. 9A–C). The dominant post-depositional imprint is lithification through secondary zeolite cementation and local calcite cement/veining that fills pores and fractures (Fig. 9D, K, O–P), whereas diagnostic pedogenic coatings and redoximorphic features are minor or absent. In contrast, the shallow VSN beneath the Cryptoporticus is characterised by pervasive micromass modification (clay-micrite mixtures), abundant clay coatings and infillings, carbonate nodules/coatings and void infillings, and Fe-Mn nodules/hypocoatings (Fig. 10). These features document open-system fluid-solid interaction and repeated moisture fluctuations, consistent with pedogenised volcaniclastic horizons and their associated secondary micromass development.
Mineralogical indicators
Bulk XRD results indicate enrichment of secondary phases within the shallow altered intervals, including widespread carbonate minerals and a detectable clay-mineral component, together with locally abundant zeolites/analcime and preserved volcanic phenocrysts (pyroxene and feldspars) in VSN-derived materials (Fig. SM2). We emphasize that, with bulk-powder XRD, the clay fraction cannot be resolved into specific clay species or mixed-layer assemblages; therefore, clay minerals are discussed here in terms of relative occurrence and enrichment rather than detailed clay-mineral systematics. Petrographic observations show that pyroxene phenocrysts are commonly preserved but locally exhibit surface etching, reflecting incipient interaction with percolating fluids. The coexistence of largely intact phenocrysts with zeolite/analcime and clay-bearing micromass is consistent with alteration focused primarily on the vitric fraction and ash matrix, rather than pervasive replacement of the primary silicate framework.
Geochemical indicators
Geochemical indicators. The portable XRF major-oxide dataset (Table SM1; Figs. 10 and 11) shows substantial variability in CaO, SiO2, and Al2O3 across the succession. High CaO values primarily track heterogeneous carbonate inputs and redistribution (anthropogenic masonry/backfill fragments, carbonate nodules, and secondary carbonate cement), whereas SiO2 and Al2O3 reflect the relative contribution of silicate-rich volcaniclastic/siliciclastic material versus fine, clay-bearing micromass. The broad compositional spread of VSN-derived samples in the PCA biplot (Fig. 11) is consistent with variable carbonate enrichment/dilution and matrix alteration acting on broadly similar volcaniclastic parent materials, rather than discrete, compositionally distinct sources. Given the semi-quantitative nature of pXRF data and the absence of LOI, these patterns are interpreted as relative compositional contrasts that corroborate the mineralogical and microstructural evidence for carbonate redistribution and ash-matrix alteration. This interpretation is further supported by compositional indices and a companion PCA on the standardised index dataset (Table SM2; Fig. SM1).
Taken together, the combined datasets show that the substrate beneath the Cryptoporticus is best described as a vertically and laterally heterogeneous assemblage. Pedogenic and diagenetic transformations (carbonate redistribution, redox cycling, finematerial translocation/illuviation, zeolitisation and subsequent hydration) overprint volcanic parent material to variable degrees and occur at different spatial scales (thin section to core scale). This heterogeneity is central to understanding local weakening and differential response beneath the monument. Although clay minerals are detected in bulk XRD, the specific clay assemblage is not resolved at the bulk scale; the micromorphological record therefore provides the principal basis for interpreting clay-related processes (coatings, infillings, vertic fabrics).
Process-based model for substrate degradation
Moisture cycling and redoximorphism
The presence of Fe-Mn nodules, impregnative features and Fe-hypocoatings/quasicoatings (Fig. 8C-E, P) indicates repeated changes in oxygenation state driven by fluctuating moisture conditions (Rabenhorst et al., 1998; Lindbo et al., 2010; Stoops et al., 2010). These redox cycles promote episodic mobilisation and reprecipitation of Fe and Mn and are a robust indicator of alternating wet and dry periods within the altered/pedogenised horizons. Even where redox features contribute only weakly to bulk chemistry, they remain mechanically relevant because they record repeated wetting-drying conditions that promote microfabric reorganisation and local weakening (Fig. 12).
Carbonate dissolution–precipitation and cement heterogeneity
Micritic nodules, coatings, laminar crusts and sparitic void and fracture infillings (Fig. 8K–M) document repeated dissolution and reprecipitation of CaCO3 along preferential pathways (pores, root channels, fractures) (Durand et al., 2010). XRD corroborates the widespread presence of carbonate phases. Carbonate precipitation can locally increase stiffness by cementing grain contacts and filling voids, but it also amplifies heterogeneity by concentrating cement along discrete pathways and leaving adjacent domains comparatively weak. This “patchy cementation” architecture is a key component of the degradation framework summarised in Fig. 12.
Clay translocation and shrink-swell deformation
AIlluvial coatings and fine-material infillings observed in thin section (Fig. 8I–J), together with porostriated/striated fabrics (Fig. 8O), record translocation of fine material and repeated shrink-swell cycles within clay-rich micromass. In the AEL and in pedogenised VSN horizons, these processes provide a direct pathway from hydrological forcing (seasonal moisture change, capillary rise, local ponding) to mechanical weakening via fissuring, loss of particle bonding and progressive disruption of the soil and volcaniclastic fabric (Fig. 12).
Zeolite alteration and hydration-related weakening in volcanic-rich horizons
In the VSN pyroclastic deposits, vitric ash matrices are variably altered to zeolites and clay-bearing secondary micromass (Fig. 9D, K). Because these phases are hydrous and moisture-sensitive, hydration-dehydration cycles can generate microfractures and disrupt grain contacts, translating moisture fluctuations into progressive fabric degradation. This mechanism is particularly relevant where zeolite-rich domains coincide with fracture/void networks and carbonate redistribution pathways, creating a coupled system where wetting events both hydrate the mineral framework and mobilize solutes (Fig. 12).
Integration and implications
Figure 12 emphasizes that degradation beneath the Cryptoporticus is not controlled by a single process. Instead, it reflects coupling between (i) water routing and moisture variability, (ii) open-system geochemical transformations (carbonate redistribution and redox cycling), and (iii) mineralogically controlled volume-change phenomena (shrink-swell clays and hydrous zeolites). This framework links microtextures and pedofeatures (Figs. 10 and 11) to the mineralogical and chemical datasets and provides a possible explanation for strong small-scale heterogeneity in substrate properties beneath the monument.
CONCLUSIONS
This study integrates stratigraphy, bulk geochemistry (pXRF), bulk mineralogy (XRD), petrography, and soil micromorphology to clarify how foundation materials beneath the Central Cryptoporticus evolve and why they display strong small-scale variability relevant to stability. Four principal conclusions emerge:
1) AEL behaves as a structured, polyphase palaeosol rather than a single floodplain mud. Illuvial clay coatings, vertic fabrics/slickensides, Fe-Mn redox features, depletion zones, and carbonate nodules generate a centimetre-scale mosaic of contrasting porosity, permeability, and stiffness. Because shrink-swell fissuring and redox/carbonate cycling are moisture-driven and can remain active under current hydrological conditions, AEL is highly sensitive to changes in water routing and saturation state.
2) VSN comprises two end-member petrofacies with contrasting durability. A weak, pedogenised volcaniclastic facies shows pervasive matrix modification (glass/clast alteration to clay-rich micromass), whereas a more lithified facies is cemented by zeolites and calcite. Zeolitisation is associated with increased lithification, but the same hydrous framework and associated clay formation can promote moisture-related softening and microfracturing under wetting, increasing the likelihood of local strength loss where saturation persists.
3) Hydrological coupling is the main “activator” of degradation across units. Meteoric infiltration, capillary rise, and vertical percolation connect AEL and VSN, simultaneously promoting shrink–swell deformation, zeolite/clay hydration, carbonate leaching and redistribution, and frictional weakening along discontinuities. Secondary CaCO3 precipitation can locally re-cement pores and microcracks, but this self-healing is spatially patchy and unlikely to offset continuing moisture-driven weakening.
4) Differential settlement and cavity-related instability are governed by sharp stiffness contrasts and moisture sensitivity. The highest collapse/subsidence potential occurs where hydric weakening coincides with thin tuff roofs or compressible pedogenised horizons and where anthropogenic cavities intersect the more vulnerable domains. Effective mitigation therefore hinges on (i) controlling surface infiltration and downward water translocation, (ii) targeted stabilisation across the most compressible/voided foundation sectors, and (iii) long-term hydro-geotechnical monitoring to capture seasonal to event-scale moisture changes.
More broadly, the study demonstrates that combining stratigraphy, geochemistry, and micromorphology provides a transferable framework for linking microscale alteration pathways to macroscale engineering behaviour at heritage sites founded on mixed volcanic–alluvial successions.
ELECTRONIC SUPPLEMENTARY MATERIAL
This article contains electronic supplementary material which is available at: https://doi.org/10.3301/IJG.2026.17.