Domestic and Productive Earthen Architecture Conserved In Situ in Archaeological Sites of the Iberian Peninsula

Abstract

For past societies on the Iberian Peninsula, one of the most prolific architectures was earthen construction, with a wealth of typologies and solutions derived from the legacy of local construction and materials. However, its study within the field of archaeology has been limited. The challenges posed by conservation, archaeological identification and social recognition have traditionally limited the dissemination of this type of architecture. Its low profile is perceived as fragile once the original protections collapse. The ethnological information preserved and linked to past ways of life and societies is crucial to the interpretation of the cultural development handed down over generations by different communities and now transmitted to the general public through archaeological sites. This research aims to provide an overview of the main earthen vestiges from domestic, productive and funerary architecture dating from the prehistoric, protohistoric, Roman and medieval periods. For this, a bibliographical review and data collection through fieldwork were conducted for numerous case studies. These included the classification of materials, compositions and construction techniques, creating an integrated comprehensive database with information on geography, measurements and the general state of conservation.

1. Introduction

In their transition to sedentary lifestyles, past societies relied most heavily on the use of earth as a construction material. This material, which was easy to source and handle, could be found in abundance in any type of habitat. This inspired a series of processes of observation, experimentation and improvement of permanent and durable constructions, knowledge of which was handed down over generations. These constructions were produced by the people for the people [1], without architects [2], prioritize economy, the efficient use of resources, functionality and resistance to aesthetics [3].

Overall, the climate and geology of the Iberian Peninsula made it an ideal location for earthen architecture to thrive over almost the whole territory, as can be deduced from evidence from different periods: prehistory [4,5]; protohistory, also boosted by orientalizing influences [6]; the Roman period [7,8]; and the Middle Ages [9]. Since the late 19th century, the search for ethnological and historical information that it can provide and its interpretation as a non-renewable document for sustainability [10] have all promoted growing interest in the processes of documenting and interpreting the technique [11]. It has also prompted the production of different studies and general classifications of techniques, excavations and interpretations since the end of the 20th century [12,13]. These have been developed in keeping with international research [14,15,16], although the establishment of up-to-date databases has been infrequent, even from the public administration.

Archaeology has made it possible to access the different physical remains conserved and to identify the architecture and construction techniques used. There are four major primary groups—mixed structures (Figure 1a), cob (Figure 1b), adobe (Figure 1c) and rammed earth) (Figure 1d)—which also prompt reflection on form and function. However, the highly variable state of conservation of this type of heritage makes this classification a complex task. This is due to both its nature and the wide range of subvariants of construction processes and traditions rooted in different territories or, hypothetically, in specific societies and cultures. Although monumental and defensive structures are more widely studied and recognized by society, this construction style is not limited to the building envelope, and it also covers aspects such as fixed furniture, production systems and burial rituals in the domestic space. While shown to be highly variable compared to other types of architecture, relatively few cases have been identified to date.

In the case of envelopes, although until now roofs have only appeared in a state of collapse, they have been classified based on the diameters of plant imprints burnt into fragments following fire damage. In addition, stone plinths have usually been conserved in walls and, in some of the best-case scenarios, so have one or two floors of earthen elevation. Such is the case, for example, with Cancho Roano, Contrebia Belaisca, Bílbilis and Casas del Turuñuelo. Most frequently, these are less than one metre from the springing (plinth or ground level) [17].

Numerous examples of domestic elements have been classified, with self-construction and adaptation to the space resulting in unique characteristics and layouts. These can be classified into main groups, depending on use, and include benches (Coll del Moro (Gandesa) [18] and Castellet de Banyoles (Tivissa) [19]; shelves for vessels (Cabezo Redondo (Villena) [20]; hearths (Tossal de Manises (Alicante) [21]; platforms (Cerro de San Vicente (Salamanca) [22]; water deposits (Plaza de Moros (Villatobas) [23]; ponds and kilns (Lépida Celsa (Velilla de Ebro) [24]; and workshops and paving.

Many funerary element typologies have also been documented. These include simple graves (Cerro Santuario (Baza)) [25]; central pillars (Tútugi (Galera)) [26]; perimeter walls (Doña Blanca, (El Puerto de Santa María)); outer rings (El Castillo, (Castejón)) [27]; and structures such as those with corbel vaults (Castejón de Arguedas (Arguedas)) [28], with roofs usually found in a state of full collapse.

In addition to kilns for domestic use, the economic activities of the different societies—such as pottery or ironwork—have left behind a specific earthen construction heritage sharing similar problems. While frequent in different societies and historical periods, one particularly notable example is that of the technologies introduced in the Late Iron Age by the inhabitants of the Eastern Mediterranean. These survived until Roman times and beyond, gradually transitioning to a form of hierarchized production that was in contrast with the domestic production of the Bronze Age [29].

In response to the need for high temperatures to transform materials, specific pieces and elements were developed for production kilns. One of the most widely studied innovations is that of kilns with vertical flues, which usually include a combustion chamber below, horizontal compartmentation or a grate and a firing chamber above these. Numerous variants can be found within this typology: with a system supporting a grate, with a central pillar (with axial prolongation or freestanding [30], seen in kilns 2 and 1 in S3-Camposoto (San Fernando), respectively); with perpendicular side walls or rows of walls (La Jericó (Herrera del Pisuerga) [31]) (Figure 2d); or with low radial walls. Differently shaped systems have also been documented: U systems with a compartmented combustion chamber (El Olmo and Mas del Moreno) [32] (Figure 2a); omega systems; circular or oval (in the Roman villa El Ruedo (Almedinilla)) [33]; and quadrangular or rectangular (El Monastil (Elda)) [34]. The construction process often took advantage of the natural earth stratum when selecting the location for the combustion chamber and used a wide variety of pieces. This is the case of the corbel vaults built with plano-convex adobe bricks in kiln 1 in Torrealta (San Fernando) [29]; the steps at the base of the praefornium in La Milagrosa (San Fernando) [35]; and grates with radial adobe bars in Pajar del Artillo (Santiponce) [36].

However, not all this variety is suitably valorized or physically represented with visible structures. Given the vital importance of physical documents as tools for transmission to the wider public, this lack of visibility negatively affects the dissemination and social recognition of this type of architecture, compared with that of the use of carved stone masonry in varying degrees. By ensuring maximum representation of examples of structures conserved in situ in different states of preservation or replacement, their importance and increased value can be highlighted within these enclaves, where they can be maintained and transmitted as part of a cycle of cultural use.

This broad approach to these techniques, employing general concepts whilst steering clear of the detailed and independent study of specific territorial cases, limits a general overview to the search for chronological parallels. This could lead to a more comprehensive joint examination highlighting a richer selection of characteristics. However, as far back as the 1990s, some publications already favoured the presentation of collections of samples, including metrics [37,38], where scope and antiquity make it possible to update a selection of sites still visitable in their original contexts. This collective discussion of techniques, tools and measurements is still relevant in the academic field of archaeology and architecture, lacking consensus on hypothetical connections with specific societies, in stark contrast with the consequent spontaneous need for timely construction solutions.

The main objective of this study Is a clear, painstaking examination of the state of the art and different aspects of earthen structures on the Iberian Peninsula. For this, a database is created following the bibliographical review and data collection in fieldwork for the case studies conserved in situ in archaeological settings. This should boost the analysis and cross-referencing of geographical information systems to detect the patterns and density of case studies while serving as the basis for subsequent vulnerability and risk studies. This general overview has led to debates on a series of statistical observations and reflections on the materials and techniques employed, classified by typology or historical period, while also revealing the main issues or case studies to be addressed during the cultural transfer.

2. Methodology

The methodology selected consists of two main phases: the search and selection of the case studies and the statistical management of data.

2.1. Case Studies

The starting point is the identification of case study locations based on various indirect sources, including national archives (Archive of the Spanish Cultural Heritage Institute (IPCE) of the Ministry of Culture); national archaeological collections (Excavaciones Arqueológicas de España, EAE; Noticiario Arqueológico Hispánico, NAH); regional collections (Serie de Trabajos Varios del Museo de Prehistoria de Valencia, MUPREVA; Anuario Arqueológico de Andalucía, AAA); and other sources, such as repositories (Calaix-GENCAT; the SOS-Tierra research project; publications of the Archaeological Museum of Alicante, MARQ); and dossiers, reports or articles on the territory. The dispersion of the information throughout the Iberian Peninsula has resulted in a more extensive compilation of publications and journal articles, accounting for 46% of the total. Remaining sources are found in lower percentages: 17% came from the IPCE, 13% from national collections, 12% from regional repositories, 10% from regional series of studies and the remaining 5% from research projects and museums (

Table 1. Dispersion of sources consulted and case studies of interest identified.

Source	n	%	Source	n	%
MUPREVA	9	5%	Publications	79	46%
MARQ	2	1%	SOSTierra Research Project	7	4%
IPCE-Map library	10	6%	Calaix	20	12%
IPCE-Projects	17	10%	AAA-Anuario Arqueológico de Andalucía	8	5%
IPCE-Photo library	2	1%	EAE-Excavaciones Arqueológicas de España	9	5%
IPCE-Inventory	0	0%	NAH-Noticiario Arqueológico Hispánico	14	8%

).

As the research requires examination of remains conserved in situ, geographical limitations were a conditioning factor in regions such as Galicia, Principado de Asturias and Cantabria. Earthen constructions are extensively documented in these areas [39,40], but the predominance of half-timbered techniques greatly limits preservation. Given the organic nature of timber, which did not survive in situ, no cases of half-timber were found, which in turn limited verification tasks to the elements observed in calcined and collapsed fragments. No reliable conclusions have been reached on the absence of other construction techniques in these spaces, possibly due to material and climate issues (abundance of forest masses and high average rainfall) [41]. The social issues (scope of societies) offer a comparison of Celtiberian and Iberian expansion, the latter widely found in the Spanish Levant, with a well-documented culture of adobe and cob [13].

It Is Impossible to establish clear limits for the case study timeframe given the limited number of domestic cases and the inclusion of techniques, such as rammed earth, which appear to be more recent [42]. However, the greater popularity of cases of Iberian origin and the difficulties in characterizing techniques, such as cob compared to adobe, suggest increased representation of the protohistoric period.

Efforts have been made, when examining typologies, to prioritize structures that are vernacular, domestic, productive and funerary and are at greater risk of loss or invisibilization. The high quality of construction of earthen monumental and defensive architecture, inherently larger and thicker, enjoys a privileged position in terms of the social recognition, impact and priority of intervention, which facilitates its survival.

Care has been taken to minimize errors in classifying terminology due to the misuse of synonyms [43], and, whenever possible, a macrovisual review has been provided in cases of ambiguous historical descriptions. The most common confusions involve misuse of the terms adobe, tapia and tapial (adobe, rammed earth and formwork, respectively) generally used to describe amasado (cob).

Based on the above, 170 case studies in different states of conservation and context were preselected (Figure 3) (Appendix A), aiming to represent all the provinces within the territory and recognizing the limitations mentioned above. Field visits were made to 121 of these based on their individual characteristics, collecting data to update information, rapidly changing with this type of heritage at times of unfinished musealization.

2.2. Characteristics and Statistical Management of Information

Information was collected and organized through fiches with a hierarchized sequence to allow the visual interpretation of general and specific characteristics [44]. This article focuses particularly on the urbanistic and territorial situation, featuring typology, use, ownership, urban location, historical period of the structures and geographical location. Given that these conditioning factors can affect the survival of these structures in varying degrees, they were taken into account in the data collection process.

In addition, an urban location can affect the conservation and valorization process, showing the contrast between musealized sites in urban settings and abandoned sites in isolated and non-urbanized natural spaces. Natural events, such as flooding, tend to be occasional and irregular, with the potential to cause different effects on flat surfaces or in settings with evacuation systems vulnerable to overflow caused by torrential rain. This also affects society, as access to human and economic resources reflected in maintenance and cultural and tourist operations can all vary. Observing the locations where these archaeological sites are usually found and their symbiosis with their surroundings is vital to understanding interventions.

Furthermore, assigning typologies can showcase the volume of disseminated finds, which originally served a specific function. Used as the basis for future risk and conservation studies, it can also aid reflection on natural problems, for example, in settings where seismic movements can be minimized by structural geometry and bracing. On social issues, they either help or hinder the general public’s access to information about a broader range of constructions incorporating these materials and techniques.

Ownership may also be a factor of interest when cross-referencing the state of conservation of different case studies. The budget guaranteed by the public administration can facilitate access to resources for the conservation and enhancement of the enclaves, although this continues to be a latent problem in the sector. Given the lack of additional support measures, sites such as Illa de Banyets (El Camepello, Alicante), acquired at a later stage, have lost a high volume of earthen construction since being uncovered.

In addition, the original function of these archaeological remains has given way to a new typology, used for documentation and research (aimed at the specialist public) or for cultural or exhibition purposes (aimed at the general public). During the intermediate phases of excavation, these two uses are often compromised, despite the increased number of initiatives offering guided tours after the different annual excavation campaigns. Different forms of damage can also be observed, depending on their specific predominance.

Ultimately, any examination of the use of the different earthen construction techniques in past societies also requires the approximate dating of any finds. Current knowledge and tools can be applied to those with earlier dating, which are susceptible to review. In this way, a general overview of the situation on the Iberian Peninsula can be deduced from the compilation of periods identified by a series of professionals in charge of the different enclaves.

The text also examines the architectural situation, featuring the identification of the material by element, construction technique, metrics, stabilizing agents, complementary use with other systems and specific features of the structure, when these values are known.

In this regard, depending on the structures conserved, earth can be found in different spots within the same site. Whereas, in monumental and defensive construction, the demand for greater resistance or wealth of construction of many societies resulted in a predominant use of carved stone, in the domestic and production spheres, earth was a key element. Stone masonry was usually limited to use for plinths or additions to load-bearing walls in order to ensure minimum damage to sensitive areas, thus saving on materials that were more expensive to extract and handle.

As regards construction techniques, the classification divides these into four main groups: half-timber techniques, cob, adobe and rammed earth. According to extensive research from the first international classifications [45], a large number of subvariants with these well-established techniques can currently be found on the Iberian Peninsula. However, the territorial misuse of synonyms in the 20th century [43] has resulted in issues that have hindered classification tasks and are still ongoing.

Therefore, some interesting decisions have recently been made to homogenize the terminology [46,47], establishing separate uses for the words adobe and tapial (formwork) to refer to amasado (cob). This also applies to the use of components and terms with negative connotations, such as arena (sand), arcilla (clay) or barro (mud) when referring to earth. Even so, the difficulties of preservation and display in situ can lead to the underrepresentation of this wide range of techniques, meaning that only a small fraction of these can be observed.

Each of these techniques is analyzed independently based on characteristics such as the thickness of construction elements or bricks, bonding or execution systems.

Nevertheless, earthen constructions display several weak points, particularly in the base and upper section, as the material is frequently combined with other types of structures, and only the lower section is conserved following the collapse of the roof. Thanks to the wide range of solutions employed, these elements help improve the quality of the construction and the response to rising damp and major flooding episodes. Identifying the frequency of recurrence of these may be relevant to the classification of these systems.

A general overview, resulting from the use of GIS and statistical management software and filtered by theme of interest, reveals the recurring presence of different cases, through heatmaps, analysis and visual examination. Thus, case density is obtained from the heatmaps for construction techniques using QGIS software and a 70km radius. The results are complemented by interpretations obtained from Real Statistics, software developed by C. Zaiontz, carrying out a Kernel density estimation (KDE) of random variables with a bandwidth of 0.8, overcoming the limitations of histograms for aspects, such as recorded metrics, to test the probability of their appearance in specific ranges or periods. This can be relevant when proposing a response to the question of metrics and their implications, as different experts still consider this variable to Be of interest when finally establishing intersocietal correlations.

3. Results

Case studies have been identified from a series of perspectives, offering a context, a general and detailed framework of the enclave and construction techniques. The geographical distribution of the territory studied has been analyzed, observing dispersion, whilst analyzing the historic, urbanistic, typological, use, architectural and construction characteristics according to the bibliography consulted and a macrovisual review. Microscopic studies have also been considered in the course of these processes.

3.1. Historical, Urbanistic, Typological and Use Characteristics

3.1.1. Geographical Distribution

Geographical distribution was observed from the total number of case studies, including the 170 preselected ones, and considering all communities except Galicia, Principado de Asturias and Cantabria, mentioned above. While this dispersion appears to be logical, with a tendency to show higher case study percentages in certain communities, some territories display a characteristic accumulation of documented remains. Although these include the Ebro valley and the south of Alicante, this may not necessarily be due to strategic issues of proximity to the sea, commercial ports or major rivers.

Case studies worth noting include the vast territory of Andalucía, with 19% of cases; the Levant, with 19% in Catalonia, 18% in Comunidad Valenciana, 11% in Aragón and 5% in Región de Murcia; and a lower representation inland, with 10% in Castilla y León, 6% in Castilla-La Mancha, 4% in Comunidad de Madrid, 3% in Comunidad Foral de Navarra, 2% in Extremadura, 1% in the Basque Country and 1% of cases in the La Rioja. Cases are also found in the south, east and north of Portugal as well as in 3 different districts (Bragança, Evora and Faro), making up 2% of the total.

3.1.2. Urban Location

According to the sample, 69% of case studies are in isolated locations, while 21% are in built-up and unbuilt areas and have an infrastructure for water evacuation, as in the cases of the Roman villa El Ruedo (Almedinilla) (Figure 4b), Tossal de Manises (Alicante) and Casa del Mitreo (Mérida). 10% of cases are in built-up plots, occupying the ground or basement floors of the construction (Domus Avinyó (Barcelona)) (Figure 4d). They are also found in historic town centres or below pedestrian level, as in El Molinete (Cartagena) (Figure 4c). Although archaeological finds are often uncovered during the construction of infrastructures, this architectural typology has mainly been conserved in natural surroundings far from urban nuclei.

3.1.3. Typology

Non-defensive earthen architecture in archaeological sites shows viewers a variety of constructions dedicated to domestic life (Figure 5a) (Figure 5b), pottery production (Figure 5c) and death rituals (Figure 5d). In the sample obtained, 41% of the structures, such as walls or fragments of roof, are linked to the dwelling envelope. Furthermore, 25% are linked to production activity, including kilns; 33% are associated with domestic elements, as in the case of fixed furniture; 5% are linked to funeral architecture, found in different types of burial sites; and 3% correspond to a completely religious typology. Finally, despite not being a subject of study, 2% of cases also incorporate at least one structure above ground, which could be linked to defensive typology. It should be noted that earthen residential architecture and earthen productive architecture are both present in 4% of sites, while in 21% of sites, productive architecture can be found in isolation. As with the rest of the factors, these figures are very broad-ranging in relation to the current condition of these enclaves, and the discovery of new sectors could alter the number of case studies.

3.1.4. Ownership

In contrast, attempts have been made to identify general ownership of the different case studies. In this respect, the results obtained show that 64% of cases are publicly owned, like Contrebia Belaisca (Botorrita, Zaragoza) (Figure 6a), compared to 21%, which are privately owned, like the site of Tos Pelat (Moncada) (Figure 6b). In 16% of cases, this information could not be confirmed.

3.1.5. Use

For most of the cases of the sample valorized, except those where the remains found have been either deliberately or accidentally destroyed [48], research and exhibition purposes should be assessed. This ensures occasional scientific use while protecting cultural transmission of the knowledge they hold. However, not all archaeological sites offer optimum conditions for conservation or valorization, and, for a number of reasons, there are high levels of re-burial, which lessen loss, abandonment and even destruction.

Thus, 51% of cases are visible for cultural and exhibition use (Figure 7a); 24% are not visible and are buried; 4% are not visible as the excavation is underway (Figure 7b); 12% have been temporarily abandoned (Figure 7c); and 10% have been destroyed (Figure 7d) or dismantled due to geographical location or the lack of measures for supporting the earthen structures.

3.1.6. Historical Period

Based on the selected sample, 4% of structures can be dated to prehistory (Figure 8a), 53% to protohistory (Figure 8b), 29% to Roman times (Figure 8c) and 6% to the Middle Ages (Figure 8d). More precise dating processes show that 6% of cases can be dated to the Bronze Age, 7% to the Late Bronze Age, and 18% to Iron Age I and II. Although the medieval era has been selected to include rammed earth case studies, most structures can be dated to protohistory, followed by the Roman period, which largely continued these construction traditions. Therefore, the lower representation of this medieval period does not necessarily equate to a lower number of case studies, and it is merely the result of bibliography and typology. However, correlations can be established between the techniques identified and their frequency. The lack of prehistoric cases conserved could be due to their greater antiquity and, thus, the greater probabilities of collapse. A more widespread use of timber and earth for load-bearing structures also makes situ conservation practically impossible.

These percentages, referring to all case studies, have been compiled and observed for the different construction techniques. It should be noted that the profuse adobe structures in isolated locations can be outnumbered by cob and rammed earth, although, proportionally, these are in the minority in built-up settings.

Notable inequalities can also be found in the remaining factors, such as the notable presence of cob in residential structures, as it is more likely to be found in domestic furniture than in the walls themselves. Furthermore, an abundant use of adobe has also been documented in most productive elements, although it is also found in the natural terrain excavated. However, rammed earth has almost been relegated to residential and structural use. This is only to be expected, as it is harder to execute, less adaptable when used on joints and smaller construction elements as well as elements thought to have been added at later stages.

As regards ownership, while all systems display a balanced public–private ratio, according to the research, ownership remains unknown in higher percentages in the cob case studies documented. In terms of use, a similar pattern can be observed, with 50% of the buried case study typologies compared to approximately 50% of the musealized ones. Nevertheless, a significantly higher number of cases of cob construction is believed to have been lost In relation to the overall figures for the technique, while most examples of rammed earth have survived (

Table 2. Proportions of historical, urbanistic, typological and use characteristics in case studies.

Characteristics	Earthen Construction Technique
Urban Location	All Techniques		Cob		Adobe		Rammed Earth
Isolated	117	68.8%	21	12.4%	101	59.4%	17	10.0%
Urbanized plot	36	21.2%	8	4.7%	33	19.4%	4	2.3%
Built plot	17	10.0%	3	1.7%	15	8.8%	3	1.8%
Typology	All Techniques		Cob		Adobe		Rammed Earth
Residential	69	40.6%	12	7.0%	53	31.2%	14	8.2%
Domestic	56	32.9%	14	8.2%	36	21.2%	7	4.1%
Productive	42	24.7%	6	3.5%	38	22.4%	0	0.0%
Funerary	8	4.7%	0	0.0%	8	4.7%	0	0.0%
Religious	5	2.9%	0	0.0%	2	1.2%	3	1.8%
Defensive	4	2.4%	0	0.0%	4	2.4%	0	0.0%
Ownership	All Techniques		Cob		Adobe		Rammed Earth
Public	109	64.1%	17	10.0%	95	55.9%	18	10.5%
Private	35	20.6%	5	2.9%	33	19.4%	4	2.4%
Unknown	28	16.5%	10	5.9%	23	13.5%	2	1.2%
Use	All Techniques		Cob		Adobe		Rammed Earth
Exhibition/cultural	86	50.6%	13	7.6%	76	44.7%	14	8.2%
Closed (under excavation)	6	3.5%	0	0.0%	6	3.5%	0	0.0%
Closed (buried)	40	23.5%	8	4.7%	37	21.8%	10	5.8%
Abandoned	21	12.4%	6	3.5%	17	10.0%	0	0.0%
Destroyed	17	10.0%	5	2.9%	13	7.6%	0	0.0%
Historical Period	All Techniques		Cob		Adobe		Rammed Earth
Prehistoric	28	16.5%	13	7.6%	17	10.0%	1	0.6%
Protohistoric	97	57.0%	15	8.8%	92	54.1%	11	6.4%
Roman	47	27.6%	4	2.4%	45	26.4%	8	4.7%
Middle Ages	9	5.3%	0	0.0%	7	4.1%	4	2.3%

Sites can display different techniques in a single location (elements ≥ sites). Percentages are shown for the total number of sites (170).

).

3.2. Architectural and Construction Characteristics

The general and detailed architectural classification of the structures found is key to identifying the most frequent systems and subvariants. In addition, the location of the earthen material, the monolithic or built construction system, its measurements, the presence or absence of stabilizing agents and the combinations with other materials and techniques can potentially determine different responses. Thus, a more precise representation of the different behaviours, degradations and intervention methodologies is obtained. In turn, ensuring maximum precision in sample classification can help form a solid database for further research.

3.2.1. Presence of Earth

The sample studied shows a distribution of 67% of cases using earth-on-wall elevations raised on plinths in different materials and 12% at the base of the wall, due to the absence of a plinth, which is more common in internal compartmentations. Additionally, 49% correspond to domestic elements (Figure 9a), usually at ground level and with little elevation; 29% to production structures, especially pottery kilns (Figure 9b); 7% to funerary elements; 59% to paving (Figure 9c); and 36% to protective renderings of the load-bearing structure (Figure 9d). These are usually irregular in appearance as this is the first layer to be sacrificed.

3.2.2. Construction Techniques

The different construction techniques identified in the case studies belong to the four major families identified in recent years, both within and outside the peninsula.

The structures resulting from the mix of water and earth covering the timber substructure, also used as internal supporting reinforcement, are considered examples of mixed techniques (torchis or técnicas mixtas). This technique was widely observed on the Iberian Peninsula from as early as the Neolithic [49], and was also found in Roman times [50]. It survives to this day thanks to half-timber, although the only archaeological remains found conserved in situ are collapsed fragments. Therefore, all the cases observed are interpretative reconstructions used to disseminate and identify some of the usual subvariants. This is the case with woven reed (Figure 10a) and braided reed (Figure 10b), found in 4% of the case studies, with thicknesses of around 27 cm, as seen in Castellón Alto (Galera, Granada) and Numancia (Garray, Soria).

Furthermore, the construction technique of cob (bauge or amasado/pared de mano) used plastic applications of simple earth to form monolithic walls, which, despite often appearing as piles of units, were individual homogeneous masses [51]. This technique was predominant in the 3rd and 2nd millennia, along with mixed techniques [52]. In this case, layers of earth, naturally stabilized with plant fibres or anthropically with lime, were applied in heaps, usually in easy-to-manipulate spherical shapes. Walls could be made either with or without formwork, being later smoothed off using forks, planes and other tools. In this classification, this group includes elements linked to domestic fixed furniture, such as hearths, kilns, shelves for vessels and benches all built in the same way.

The identification of cob in archaeological sites has been widely debated in the field, as it has been traditionally considered adobe or rammed earth. Thus, this has contributed to both the protohistoric standardization of adobe and its underrepresentation in the sample. Nineteen percent of all the sites feature some form of cob structure, including La Olmeda (Pedrosa de la Vega, Palencia) (Figure 11a) and Caramoro I (Elche, Alicante) (Figure 11b). Walls around 37–45 cm thick are found, as well as ones that are 25–30 cm thick for interior compartmentations and 60–80 cm thick in the case of domestic structures. Moreover, the paving category was filtered, and 48% of cases were identified as compacted earth and 16% as clay.

Adobe (brique crue or adobe) is the most widely documented variant in the group of earthen walls. It is normally made with raw earth units mixed with straw and coarse sand to improve resistance and durability. It is serially manufactured with moulds and pieces dried in the open air, suggesting a possible quick execution process, although moulds were not necessarily involved in many of the variants documented. Although terminological issues have obscured discussion of the origins of the term, there is conclusive evidence of its existence in the Late Bronze Age and Early Iron Age [6], when it was linked to orientalizing influences in the south and southeast of the Iberian Peninsula. Subsequently, from the second half of the 6th century BC, it spread across the peninsula and has been documented in some settlements predating the Phoenicians [13]. Among the different case studies selected and reliably dated to these periods, only Cerro de la Mota (Medina del Campo, Valladolid) has conserved and musealized remains in situ.

Given the extremely variable placement of units, in keeping with the high number of variants, solutions can be achieved that can satisfy construction requirements in terms of resistance and thickness. Load-bearing walls, for example, tend to use stretcher and header bonds for good adhesion and resistance, as seen in El Turuñuelo (Guareña, Badajoz) [53]. Header configurations are found in walls of one foot in Puntal dels Llops (Olocau, Valencia) [54]; double stretcher configurations are found in Cancho Roano (Zalamea de la Serena, Badajoz) [55]; and to a lesser extent, square ones, with a better seismic response, are found in Mas Castellar (Pontós, Girona) [56]. When thinner constructions are sought, stretcher configurations can be seen, as in the case of the upper floors or the partition walls of Tossal de Sant Miquel (Lliria, Valencia) [57]; in the funerary sphere, in El Castillo (Castejón, Navarra) [58]; or in shiner bond, as in La Celadilla (Ademuz, Valencia). These thicknesses can also be achieved with stretcher configurations in smaller modules (15–20 cm wide), as in Puntal dels Llops (Olocau, Valencia) [59].

The system is also found in other contexts, including paving, in Castellet de Banyoles (Tivissa, Tarragona) [60] or smaller domestic elements, such as benches with stretcher configurations in Coll del Moro (Gandesa, Tarragona) [18], kilns in La Jericó (Herra del Pisuerga, Palencia) [31] or in Mas del Moreno (Foz-Calanda, Teruel) [32], platforms like those in Cerro de San Vicente (Salamanca) [22] or hearths or water deposits. In some enclaves, although information is not conclusive, research suggests the possible construction of corbel domes in the dwellings in El Turuñuelo (Guareña, Badajoz), using lightweight modules that were easier to manipulate [61]. It was also used for coating in Alto de la Cruz (Cortes, Navarra) [62] and was more common in productive architecture, as seen in kiln 3 in Torrealta, kiln 4 in Camposoto (San Fernando, Cádiz) [29] and in La Cabrera (Torredonjimeno, Jaén) [63].

Finally, these pieces have successfully adapted special geometries for concrete elements, usually structural ones in productive settings. A notable example is that of large plano-convex elements documented as a system to support combustion chambers in Pajar del Artillo (Santiponce, Sevilla) [36]. Other pseudotriangular structures are found, such as kiln 4 in Torrealta (San Fernando, Cádiz), as well as tegulae inserts, as in the production sector in El Ruedo (Almedinilla, Córdoba) [33].

Adobe has historically been the most widely recognized construction system, with at least one structure in 88% of the case studies. Moreover, the measurements collected provide an overview of the different construction scenarios, using probability to identify those most frequently repeated in specific periods. This is particularly complex in the case of adobe, given its variability (Figure 12) and numerous conditioning factors, including mould size, type of execution, construction culture, spontaneity and simple needs to cope with construction issues. In order to combine the metrics documented by the different specialists (Appendix B), the probable measurement ranges for adobe bricks put forward by the KDE have been observed: prehistoric (44/52 × 26/30 × 7/10 cm), protohistoric (36/43 × 19/25 × 10/12 cm) and Roman (40/51 × 29/40 × 10/11 cm) (Figure 13).

Rammed earth (pisé or tapia), the result of using a rammer to compact layers of earthen elements within a wooden formwork (tapial), was conclusively identified on the Iberian Peninsula in the Roman domus in Ampurias [42]. Although its origin is still subject to debate, on an international scale, it can be dated back to the Punic period in Carthage [64]. While the range of subvariants currently known for earthen construction with rammed earth is very high [9], the variability found in the archaeological case studies selected is significantly lower. This includes rammed earth walls in Rábita Califal (Guardamar del Segura, Alicante) (Figure 14a), as well as walls in masonry and gypsum or with gypsum reinforcement, such as those found in the Mosque of Cortijo del Centeno (Lorca, Murcia) and Medina Siyasa (Cieza, Murcia) [65] (Figure 14b).

According to the sample, 15% of case studies feature at least one historic structure in rammed earth, dated in 9% of cases to the Roman period and in 6% to the Islamic Middle Ages. While the range of metrics is much wider, given the smaller database sample, these structures can be found in a range of 30 to 80 cm in Roman times, peaking between 30 and 50 cm. This metric increased slightly to 100 cm in the Middle Ages, although 50 cm remains be the most common measurement. As these ranges only varied relatively recently, measurements of between 40 and 90 cm can be found in rammed earth constructions in Comunidad Valenciana [66].

3.2.3. Stabilizing Agents

In 86% of case studies, it was not possible to identify stabilizing agents, as information on their composition is limited by the general lack of microscopic studies on the fragments found. This lack of information can mostly be attributed to much of the prospecting carried out predating the existence of solid professional knowledge of earthen construction. This means that the assessment is currently being called into question, although the widespread use of plant elements in their composition is assumed.

Cases with the confirmed use of stabilizing agents are distributed as follows: plant elements were used in 12% of cases; lime additions in 2%; stones, such as aggregate or other smaller stones, were used in 2% due to the limitations of the moulds or joints. Carbons, ceramic and gypsum are each found in 1% of case studies. It should be noted that according to these studies, no stabilizing agents are found in 1% of the sample, which are composed solely of raw earth.

3.2.4. Complementary Techniques and Materials

According to the case studies, the most common complementary solution was the use of stone masonry. It is found in 86% of the total samples, acting as a plinth or stem wall in 53% of cases, as seen in Casa de los Grifos (Alcalá de Henares, Madrid). However, as documented in 17% of cases, fired brick was also frequently used. Stone masonry was generally found together with other materials from the start of the 1st century BC [13], as confirmed in 1% of plinths from Roman times or later, in El Molinete (Cartagena, Murcia) and Contrebia Belaisca (Botorrita, Zaragoza). This setting is displayed with the same percentages in the opus caementicium, visible in sites such as Els Munts (Altafulla, Tarragona) [67]. Finally, stone slabs are documented in 7% of the sample, specially positioned for thresholds and other systems; while ashlar and timber are each found in 5% of samples, although the latter is not conserved in situ, as mentioned above. Despite mainly being combined with other techniques in plinths, stone slabs are not found in the remaining 46% of cases. This is the case particularly in most domestic furniture and in thinner interior partitions, such as those in la Rábita Califal (Guardamar del Segura, Alicante) [68], where certain types of wall sprang from the ground itself. This absence can be attributed to a number of general factors: the smaller size of elements, material availability, savings or quality of construction. These factors tend to be noticed mostly in the supporting system of the building. For the most part, these plinths are under 100 cm high, with numerous examples of around 50 and 23 cm. Their thickness tends to be in line with the documented elevations, and most of the sample displays thicknesses of between 30 and 50 cm, with peaks at 40 and 50 cm (Figure 15).

These architectural statistics can also be observed by establishing a direct relation between the construction techniques and their degree of exposure, location, stabilization and complementation. In keeping with the descriptions above, some additional results to be extracted include the higher degree of exposure of rammed earth compared to the total case studies or the notable burial and destruction of cob.

Furthermore, the greater presence of cob in lower sections of walls is in direct correlation to its greater presence in domestic elements, which tend to be low and have no plinth. When examining similar executions in rendering and continuous paving, it can be observed in 92 and 95% of case studies, respectively. In contrast, limited use is made of adobe with no plinth, which appears complemented in up to 89% of the case studies complemented. Its lower presence in rendering and paving of less than 15% should also be noted. Rammed earth is only detected in the execution of walls.

Finally, the proportionally high combination of rammed earth and fired brick in the sample should also be noted, as should the exclusive use of adobe with ashlar and its predominance with flat stone. Although, statistically, the higher number of cases of this technique is also widely complemented with masonry, especially in relation to cob, andthe use of rammed earth is remarkably only 20% lower in the use of these plinths compared to the total case studies selected for the individual techniques (

Table 3. List of architectural and construction characteristics in case studies.

Construction Technique	Sites		Visible		Buried		Collapsed
Mixed techniques	6	3.5%	0	0.0%	0	0.0%	6	100.0%
Cob	32	18.8%	12	37.5%	11	34.3%	10	31.2%
Adobe	149	87.6%	68	45.6%	55	36.9%	41	27.5%
Rammed earth	24	14.1%	12	50.0%	10	41.6%	2	8.3%
Compacted earth pavements	82	48.2%	41	50.0%	31	37.8%	22	26.8%
Clay pavements	27	15.8%	11	40.7%	9	33.3%	9	33.3%
Unidentified	1	0.6%	1	100.0%	0	0.0%	0	0.0%
Presence of Earth	Sites		Cob		Adobe		Rammed Earth
Wall elevations	114	67.0%	9	7.8%	102	89.4%	23	20.1%
Wall basing	21	12.3%	4	19.0%	14	66.6%	3	14.2%
Domestic elements	84	49.4%	18	21.4%	71	84.5%	1	1.2%
Production structures	50	29.4%	4	8.0%	46	92.0%	0	0.0%
Funerary structures	12	7.0%	2	16.6%	10	83.3%	0	0.0%
Pavements	100	58.8%	92	92.0%	13	13.0%	0	0.0%
Renderings	61	35.8%	58	95.1%	3	4.9%	0	0.0%
Stabilizing Agents	Sites		Cob		Adobe		Rammed Earth
Plant	20	11.7%	5	25.0%	16	80.0%	4	20.0%
Lime	4	2.3%	1	25.0%	3	75.0%	1	25.0%
Ceramics	3	1.7%	0	0.0%	2	100.0%	1	33,3%
Gypsum	1	0.6%	0	0.0%	1	100.0%	0	0.0%
Small stones	4	2.3%	1	33.3%	2	66.6%	1	33.3%
Carbons	2	1.1%	1	50.0%	0	0.0%	1	50.0%
No stabilizing agents	1	0.6%	0	0.0%	1	100.0%	0	0.0%
Unknown	141	82.9%	23	16.3%	127	90.0%	21	14.8%
Complementary Techniques	Sites		Cob		Adobe		Rammed Earth
Masonry	147	86.4%	6	4.1%	128	87.0%	16	10.8%
Fired brick	15	8.8%	1	6.6%	10	66.6%	4	26.6%
Flat stone or stone slabs	11	6.4%	0	0.0%	10	90.9%	1	9.0%
Ashlar	8	4.7%	0	0.0%	8	100%	0	0.0%

Sites can display different techniques or structures in a single location (elements ≥ sites). The percentages for the columns Visible, Buried, Collapsed, Cob, Adobe and Rammed earth are calculated in relation to the subtotal of case studies (Sites column) for each of the characteristics listed.

).

4. Discussion

The overview of earthen architecture constructions in archaeological sites on the Iberian Peninsula provides a broad variety of case studies and construction techniques. This makes it possible to identify some of the most frequent characteristics, documented separately for their independent and combined examination.

Based on the sample observed, the most common urbanistic profile is that of an isolated context that is neither urbanized nor built up. These are usually found in small mounds in strategic locations and are the result of the material collapse of the habitats and the passing of time. These spaces account for almost a third of the samples, are mostly publicly owned, and are dedicated to exhibition or cultural use. This solution is the main strategy supporting the conservation and operation of these heritage enclaves, which are not suited for residential or administrative use. Another strategy is the re-burial of the remains or granting professionals and academics priority access over the general public, postponing valorization until a later stage. It should be noted that the state of abandonment or destruction is seen in a fifth of the sample, where preservation is at great risk and the sites may be irrecoverable.

Furthermore, the most common visitable typological profiles on the Iberian Peninsula are those with residential functions, mostly structural (41%) or with fixed furniture (33%). Of these, only 4% are located in the same area as earthen production structures, as in the case of Ampurias (La Escala, Girona). Less than 1% are found together with funerary structures. This expands the joint vision of the general use in different sectors, characteristic of these techniques. It also enables the valorization of finds with mostly independent typologies but with wide distribution across the whole peninsula.

In the specific analysis of structures, in the case studies, the observed earthen techniques are identified in wall elevations, paving and domestic elements. The most prevalent combination is that of uncarved stone plinths over fired brick or ashlar. In addition, there are only a few cases of walls built with earth from the base, something found more frequently in fixed furniture, where it usually springs from the flooring inside the rooms.

The recurring construction techniques have provided information on the different forms of conservation represented. While mixed structure systems are completely lost in situ, reconstructions can be found in smaller geographical settings, clustered in the provinces of Granada (Andalucía) and Soria (Castilla y León) (Figure 16).

In contrast, adobe is widely identified throughout the Iberian Peninsula (Figure 17). In this case, the sample displays a higher density of exposed and visible cases in the Levant, with a greater concentration in Catalonia and Alicante, as well as in other parts of the provinces of Zaragoza, Valencia, Badajoz and Jaén. However, the locations with the highest number of buried structures analyzed in the sample are the provinces of Zaragoza, already mentioned, and Murcia, Madrid, Valladolid and Seville. These are thus concentrated in the central Iberian Peninsula, most notably the valley of the river Ebro or the area surrounding the estuary of the river Segura. Although these elements can remain unaltered thanks to the earthen protection of natural substrate, they entail the introduction of new social risk factors, potentially muting and hindering the dissemination of this knowledge.

Cob (Figure 18) and rammed earth (Figure 19) are the second and third most observed categories for monolithic walls, while the first includes indoor fixed furniture featuring no brick constructions. In spite of this, fewer cases of cob walls have been identified, and rammed earth walls are recorded in domestic buildings from Roman times. Although both can provide solutions with a wealth of varied construction techniques, their presence is quite limited, and some variations, such as ball cob or gypsum-reinforced rammed earth, are especially scarce in domestic constructions.

The limitation of the sample and the variability of the fixed furniture also constrain conclusive results in terms of metrics and dimensions, as the main purpose of these was to provide precise answers to individual construction needs. However, peaks have been recorded for 50-cm-thick rammed earth walls in Roman times, while cob walls between 35 and 45 cm thick have also been found.

Adobe is the most frequently found construction technique for the periods and cases consulted. It is overwhelmingly found in the sample, possibly because of standardization in the Iron Age and its practicality and ease of execution. This may also be due to overuse of the term and more straightforward identification of the pieces. Found in all types of bond, adobe is also the only technique used to build kilns combined with the removal of earth from the terrain. This earth is then used to render and construct low walls, pillars or grates. Adobe is far more frequently represented than cob or rammed earth monolithic systems.

Although probability studies seemingly reflect a representative attribution of metrics and dimensions for the different historical periods, no clear distinction has been identified that will allow the use of these characteristics to study the scope and construction traits of particular societies. In general, pieces that are over 30 cm wide and 10 cm thick have been documented from the Roman period, although length barely differs. Protohistoric pieces are considerably longer and broader in relation to the other periods, at times reaching lengths of 40 cm and widths of 20 cm. Case study samples from the Bronze Age show that these pieces, which are between 37 and 50 cm long, are also thinner (around 7 cm). Although width is highly variable, the most common range tends to be 30–35 cm. Nevertheless, given the general high variations in metrics, it is advisable to err on the side of caution. Considering that these chronological variations may not be conclusive but rather just another factor based on construction needs as dictated by individual situations, buildings or pieces of furniture to be executed, this approach becomes significant.

It should be stressed that there is work still to be done in these classifications, conducting microscopic studies, such as micromorphology [69], petrography [70] or XRD [71]. Those are needed to analyze the composition of the different subvariants, including the presence of lime, gypsum or carbon in the original mixes.

In up to four-fifths of the sample, this information is yet to be completed but should enrich the original assumed use of plants. This can often remain unnoticed in macrovisual surveys as the decomposition of fibres caused by water runoff on surfaces during degradation processes leaves spaces that are then filled in.

5. Conclusions

The overview provided in this analysis is a documentary and statistical basis for the current conservation situation of earthen architecture, capable of combining the results and finds in current society. The keen interest in representing this current state of the art in the field of study is motivated by the dispersion of the data, which tend to be managed and presented on a regional or individual basis. After collating different forms of data, information can be handled in GIS, enabling search and filtering tasks to be carried out in relation to context and architectural characteristics.

Although this could be considered a limited sample for the purposes of obtaining correlations in an extremely complex context, a degree of convergence can be observed between case studies. This establishes a solid basis for developing future case studies for risk prevention and minimization. Thus, this type of architecture is characterized as generally isolated, public and mostly involving domestic and productive typologies that can be visible or buried as well as widely developed in protohistory. The construction richness observed sheds light in the different uses and degree of conservation of techniques, such as cob, adobe or rammed earth on the Iberian Peninsula, depending on architectural or technological needs. It also highlights the compromised exposure of certain systems, such as mixed techniques.

Furthermore, a partial response, limited to the selected sample, is provided for the metrics of this type of heritage. Kernel density estimations provide a visual synthesis of thickness probabilities according to the database and also provide brick lengths and widths, identifying major groups by period. Although there is notable interest in the field of archaeology in linking metrics to periods, societies and locations, a single potentially high and variable metric should be borne in mind, as it only answers to the adaptation of constructions according to the needs dictated by the circumstances.

The difficulties of classification and the legacy of the lack of professional knowledge of these types of techniques when studying the past societies of the Iberian Peninsula compromise the accuracy of its scope. The undocumented loss due to the absence of relief measures calls for a growing multidisciplinary involvement in identifying these systems and joining forces to offer new reflections, backed by geographical scope and recurring cases. Likewise, monitoring and reviewing information on systems from different periods has helped update their in-situ implementation and dissemination. Their lack of protection is directly associated with the existence of rapid and intense changes in the state of preservation over short periods of time. This in turn entails the loss or temporary or long-term burial of physical documents in situ.

Reflection on case studies and pressing needs currently presents the main challenge as the tasks of prevention, planning and intervention [72]. The general identification of issues, complemented by the painstaking study of individual case studies, constitute fundamental pillars to be developed in parallel to the suitable characterization, both present and future, of finds based on the new knowledge generated. The consideration of sustainability requirements [73] for this type of heritage, as well as the selection of conservation methods that address the contradiction that stems from retaining authenticity without maintenance, are necessary to its survival. By paying as much attention as possible to minimum interventions, the use of materials analogous to the original ones or recurring surface consolidation can help consolidate and enhance future studies transmitted to coming generations.