Geoarchaeology and castlescapes: heritage management case studies in Spain and the eastern Baltic

This paper promotes the application of geoarchaeology—joint studies using historical, archaeological and heritage approaches— to the conservation and management practice of castles and landscapes in the context of ‘castlescapes’. Using case studies from recent research on medieval castles in frontier regions of the eastern Baltic and Spain, it demonstrates how geoarchaeology can create synergies between on-site and off-site environments and between cultural and natural heritage and draw valuable information from soils and sediments about the changing form and function of spaces within castles, and about the links between these spaces and activities in their hinterlands. Geoarchaeological approaches can also illuminate the diachronic biographies that hide from visitors in the buried archaeology of castles, which to most visitors would be blank cavasses, but which can provide snap-shots of castle life in the context of a wider landscape. Castles are commonly publicly recognised as being important historical monuments, but from a heritage perspective they are often presented in isolation from their associated historical territories, and often (especially in frontier regions) appropriated within modern politics, which has influenced both heritage management decisions and research frameworks.


Introduction
This article describes a geoarchaeological approach to the study and presentation of castles and their landscapes in two frontier regions (Figure 1): the eastern Baltic, in the context of the Northern Crusades which began in earnest in the thirteenth century, and Iberia, in the context of the 'Reconquista', a term describing the incremental Christian conquest over many centuries of al-Andalus, the Islamic polity created in the eighth century. The term 'Reconquista' was popularised in late nineteenth century historiography to promote a nationalist agenda of unification (Ríos Saloma 2013), but since the 1980s many Spanish scholars have rejected the term in favour of thinking of a series of regional 'conquests' and 'colonisations'; it retains some international currency, however, and remains in use within conservative sectors of Spanish society (Torró 2000;García-Sanjuán 2016). Both the eastern Baltic and Iberia regions were shaped by conquests in the formative medieval period, supported by colonisation and economic intensification, mirrored across the frontiers of Latin Europe (Bartlett 1993). Following the conquests, castles were built or rebuilt to manage the annexed territories, and today, they are amongst the most important and striking historical monuments in these regions, including the largest castle in Europe, in what was medieval Prussia, formerly known as Marienburg, today Malbork in Poland.
As a concept, the term 'landscape' can define a geographical area or tract of land, an assemblage of ecosystems in mutual interaction, and in cultural terms as a person's perception of their physical world (Hoffman 2014). The ESF/COST Science Policy Briefing Landscape in a Changing World and the European Landscapes Convention share the view of landscape being 'not simply the environment, but the world "as being perceived be people", or the environment's human element' (Council of Europe 2000; ESF/COST 2010). If the concept of landscape is considered as a major force for social and environmental change or continuity, as the ESF briefing suggests, then European castle landscapes embody these forces. Landscape in a Changing World also sets the agenda for future integrated landscape research.
This paper seeks to reconnect castles with their landscapes through the application of geoarchaeology, which applies earth science techniques, such as soil micromorphology and physio-chemical analyses, to archaeological research questions. Geoarchaeological techniques can be applied to create synergies between the buried archaeology of castles with sediments from their historical territories or hinterlands and between the cultural and natural heritage. They allow the preservation conditions of buried archaeological deposits to be assessed to inform heritage management decisions (Canti and Huisman 2015), but also provide important taphonomic information, when used with other sources of environmental evidence, such as faunal remains, plant macro-remains, and palynological studies (Banerjea et al. 2020a), to reconstruct the history of land use within a castle's hinterland Stivrins et al. 2015;. Geoarchaeology can also aid the interpretation and presentation of castle landscapes within their formal management structures and inform future research frameworks and practice.
The castles discussed here, as elsewhere in Europe, have been the focus of many surveys and excavations, and have long-represented important historical narratives. From a heritage perspective, however, they have largely been disconnected from their associated historical territories. We argue that the links between on-site activities and those in the hinterland that are revealed by geoarchaeological research are important in defining the castle landscape itself from both a scholarly and a heritage perspective, and for presenting the past to the public in ways that are relevant for the future identity of these iconic monuments. Our research underlines the importance of incorporating the hinterland into management decisions for the standing remains, while situating these decisions within the political context of heritage legislation, particularly exploring their role in the creation of heritage narratives and identity politics, such as ethnic nationalism. This is visible elsewhere in Europe, such as in post-Soviet Croatia where castles form part of a national identity emphasising resistance to outsiders (Harrison 2013), but it is particularly emphatic in the case of historical frontier regions in which castles are iconic monuments of ideologically driven conflicts, power relations and processes of colonisation during a formative period in European history (Bartlett 1993). These conflicts have been appropriated within modern politics and steer heritage management decisions and research frameworks.

Political considerations in policy development
In Poland and Spain, medieval castles are embedded within complex geopolitical contexts and nationalist agendas, and, as such, have been viewed as symbols of authoritarian and/ or colonising regimes. In the eastern Baltic, the crusades waged against native pagan societies for almost a hundred years resulted in their disenfranchisement and the reorganisation of the conquered territories into the two Christian polities of Livonia and Prussia, both dominated by the Teutonic Order (Pluskowski et al. 2018;Pluskowski 2019a). In former East Prussia, the Teutonic Order's castles were one of the key elements in building national and state identity, and during and after World War II, the Teutonic Order became synonymous with German militarism (Pluskowski 2012). Once Prussia had been partitioned between Poland, Lithuania and the Soviet Union, these castles became unwanted, foreign heritage. This was also the case in Latvia during both its interwar period of independence and the following five decades of Soviet occupation.
In Estonia, however, research in the folklore archives of the Estonian Literary Museum in Tartu by one of the present authors (Valk) suggests that despite a history of German and Soviet occupations, perceptions of these monuments differ. Castles are not connected to medieval German militarism but are simply ruins from the past, a part of the local cultural landscape. Memories of castles are quite vague: castles are only briefly noted as destroyed in wars, connected with 'Swedish time' (which ended between 1704/1710 in Estonia, the Swedes being the last political power to possess castles, or being associated with romantic legends about knights.
In Poland, the perceptions of the Teutonic Order's striking red brick castles-and the cultural heritage of former East Prussia more widely-was the result of Poland's 'Third Partition' of 1795-1797, the aftermath of World War I and, above all, the trauma of World War II. For this reason, it has been problematic in Poland to consider castles regarded as 'German' as National Heritage. Even so, Teutonic Order castles were protected under Polish law as monuments as early as the late 1940s, initially until 1962 by pre-war Polish law that protected state-owned monuments, and thereafter by laws of 1962 and 2003. In 1989, however, the ownership of castles was transferred from the national state to local governments and private individuals. The heavy financial obligation rested primarily on local governments and private owners, and led many municipalities to sell castles to private owners, usually investors, and typically at a very low price. The law requires castle owners to finance archaeological excavations in the case of planned ground disturbance, which led to the excavation of castles by small private archaeological companies selected on the criterion of cost, rather than any research design.
There are some similarities in the management of castles in former Livonia (present day Estonia and Latvia) and Prussia (present day North-Eastern Poland and the Kaliningrad oblast of Russia), but also some differences. Some castles and castle sites in Latvia and Estonia are privately owned, but they are most often owned by the state or local municipalities, and castles and ruins preserved above ground are protected by the law. The importance of archaeology in the research of castles in Latvia and Estonia cannot be overestimated not only because most of the castles are ruins and some have no above ground remains ( Figure 2), but also because only a very limited amount of documentary records has survived; Prussian castles are associated with a relatively larger quantity of medieval written sources. Most castles are in ruinous states or have no above ground remains and across the eastern Baltic, some castles or their ruins were damaged or destroyed as a result of conflicts arising from German and Soviet occupations during the first and second World Wars. In Latvia and Poland the perception of castles as unwanted heritage, arising from these conflicts, has negatively impacted the extent to which they are the subject of research, except for large-scale rescue excavations in Latvia after 1960.
It is a similar situation in Spain, where the Spanish Civil War impacted on perceptions and the preservation of castles as well as the extent of research, although again since 1960 there have been some larger excavations. A Spanish association for protecting medieval fortifications has a list of more than 10,000 fortifications across Spain, of which at least 3000 must be considered medieval castles (Schnell 2009). The 'National Plan of Defensive Architecture' drafted by the Spanish Cultural Heritage Institute, forms part of the national policy for Spanish historical heritage as far as information, preservation and restoration are concerned (Rodríguez 2014). The fragmented territorial model of the Spanish State results in the overlapping of different rules, legislation and permissions that affect research inside castles, despite efforts at cataloguing and protection from the central government (Carrión 2015). The complicated, heterogeneous and fragmented situation is the legacy of the Franco regime and its law of 1949 which placed all castles under legal protection as national Spanish heritage, with an 'architect-curator' ' in charge of 'conservation' (López-Muñiz 2017). All fortifications were immediately designated in this way, but in practice only the castle's architecture was protected by heritage management officials and regional legislators, and the provision of specific and integrated research agendas and management strategies became difficult.
Researchers have traditionally focused on the fortified architecture of the late Middle Ages due both to the preference for rammed-earth constructions during the Islamic period, which causes difficulties with conservation and identification, and to the place of the idea of the Christian Reconquista in the founding discourse of the Spanish nation and its heritage (Moreno 2016). With very few exceptions, most regional studies concern monumental and defensive architecture and their integration into analyses of the settlement pattern (e.g. Palacios 2019, 442-444). This perception of castles as an expression of the successful political and cultural imposition of Spanish culture onto the previous Islamic polity has affected the type of archaeological research carried out on these monuments (Palacios 2019), causing a focus that is still largely (despite recent development in methods, concepts and themes) on defensive architecture rather than on questions relating to daily life, procurement and provisioning of raw materials, or how their construction and modifications impacted on the surrounding landscapes (Valor and Guiérrez 2014, 148-175). There is a similar architectural focus to castle research in Poland, initiated by Kajzer (1996, 139-148) and elsewhere in Europe. In the United Kingdom, despite the upsurge in research on the social and landscape contexts of castles, English Heritage's (now Historic England's) Regional Research Frameworks have still tended to subsume castles under the heading 'Defence and Warfare' (Creighton 2009).
In both the eastern Baltic and Spain there is no specific legislation for managing or protecting a castle's landscape. In Estonia and Latvia protected monuments are surrounded by 'protection zones' where development and activities are regulated, and these can include the protection of views and landscapes. Recent laws-Estonia's Heritage Protection Act 2019 and Poland's 2003 Act 'on the protection and care of monuments'-include some general principles for protecting cultural landscapes but essentially remain guidelines and generic ideals, a striving towards a 'new thinking' that still awaits translation into practice. In Poland, geoscientific approaches have focused primarily on examining the geomorphological conditions for a castle's location in the landscape. In Spain, castles are studied from the perspective of architecture and traditional approaches to the study of material culture (ceramics, metal artefacts and in a few cases faunal or plant remains), but the heritage law of the Franco regime did not apply to the buried archaeology.

Life-histories of castles and the value of buried deposits
The 'life-history' or biography of a building or monument refers to the dynamic chronologies of its establishment, development, architectural modification, abandonment and post-abandonment, as well as its diachronic use. It is most effectively examined using a geoarchaeological approach to interpret formation processes at high resolution (e.g. Banerjea et al. 2015Banerjea et al. , 2017Banerjea et al. , 2020bBanerjea et al. , 2020cBorderie et al. 2020), but until very recently such research has been rare. Soils within profiles or unexcavated areas in and around monuments have been investigated only infrequently, and without a specific geoarchaeological research focus, except to assess the geological and geomorphological aspects for military strategies in castle location (Mariani et al. 2019). Geoarchaeology has been applied in only three published studies, of Juslibol (Zaragoza), Alfama (Teruel) and Castillejo de los Guájares (Granada), but in all three cases, as in Poland, their focus is more on geomorphology than buried archaeology (Burillo et al. 1981;Peña-Monné et al. 2014;Forlín and Gerrard 2017). Recent micromorphological research in Molina de Aragón, Guadalajara (Banerjea et al. 2020b) and within the 'Landscapes of (Re)Conquest' research project (García-Contreras  has however pioneered the application of geoarchaeological approaches, particularly soil micromorphology, to castle landscapes in Spain. The potential of geoarchaeological analysis at castle sites is thus beginning to be realised (Banerjea 2019; Banerjea and Badura 2019; Banerjea et al. 2020a,b;Borderie et al. 2020;Brandolini et al. 2018) and in what follows, its benefits and value is demonstrated through stages in the life-histories of four castles in the Baltic and Iberia, at Elbląg, Karksi and Cēsis in the Baltic and Molina de Aragón in Spain. Geoarchaeological approaches to classify construction materials according to composition and production techniques, and to identify source materials with a view to providing chronological phases for the standing remains, are applied to understand the development of monuments and the acquisition of materials (Hayward 2019), or for conservation purposes (Gutierrez-Carrillo et al. 2019).
Excavated sequences at Elbląg, Poland, and Karksi, Estonia contained well-preserved, well-stratified and semi-waterlogged deposits with exceptional preservation of organic remains (for full micromorphology results see Banerjea 2019; Banerjea and Badura 2019; Banerjea et al. 2020a). The sediments at the base of the profile at Karksi were deposited rapidly between A. D. 1266 and 1290 (Brown 2019a). Timbers at the base of the deposit sequence from Elbląg produced a dendrochronological date of A. D. 1245 (Fonferek 2019) and is one the earliest dated Teutonic Order timber castles in Poland. The sequence not only represents the very earliest occupation and use of the outer ward, but also provides a unique insight into castle function and use in the early years of Teutonic Order's rule. The microstratigraphic investigations, augmented by integrated macro-and micro-botanical proxies, showed local clearance activities (Karksi), in situ evidence for animal husbandry and livestock alimentation (both sites), refuse remains from middening activities (Karksi) and the destruction of the timber castle (Elbląg).
Excavations at Elbląg in 2012 and 2013 not only confirmed the militaristic character of the site, but also contributed valuable information on the provisioning and storage systems of the early castle (Fonferek 2019). Geoarchaeological analysis of the profile from the 2013 excavation (Banerjea et al. 2020a) revealed that the outer ward of the early castle was first used to keep small herbivores, probably sheep or goats, then to stable larger herbivores, which, on the basis of the horse furniture recovered during excavation and documentary evidence relating to the use of the outer bailey of Teutonic Order castles, were probably war horses. The sequence showed regular flooding events, attested by depositions of alluvial sediment, including before the destruction of the timber stronghold prior to rebuilding in stone. The Old Town of Elbląg, is located on the River Elbląg. In the present day, the city is a port on the river, which flows into the freshwater Vistula Lagoon, from where the Strait of Baltiysk gives passage to Baltic. In 2012, geological and palaeoecological investigations of sediments which underlie the courtyard of the Museum of Archaeology and History in Elbląg documented the evolution of this area from a relatively deep lake, Lake Drużno, through a shallowing water body, to a waterside zone, where the River Elbląg flowed into the Vistula Lagoon, by the time the timber castle was erected (Nitychoruk et al. 2016). The soil profile obtained during excavation in 2013 (Banerjea et al. 2020a) show that the outer bailey of the timber castle flooded on several occasions. There appears to be evidence for a 'brush wood' platform being laid following one of these flooding events, prior to the use of the area to stable horses. It is plausible that the repeated flooding of this area by the River Elbląg was a reason why the area, which does not seem to have been incorporated into the new brick castle, was abandoned.
Geoarchaeological and palaeoecological evidence from Karksi, Estonia, revealed how the earliest inhabitants at the site modified the local environment and how areas of the site would have appeared during the construction of the castle (Banerjea et al. 2020a;Banerjea and Badura 2019). This new phase reflected the central role of castles in reorganising conquered tribal landscapes . The sediments at the base of one profile were overlaid by deposits that probably dated to the late fourteenth or fifteenth century, which formed below the first preserved pavement of the High Castle in which artefacts dating to the fifteenth century were recovered (Valk et al. 2012). Soil micromorphology defined the formation processes of the sequence during the thirteenth and fourteenth centuries. Analysis of the thirteenth century deposits show that livestock were stabled in the area that was to become the High Castle, prior to and during the removal of vegetation, probably for its construction, as evidenced by a reduction in tree pollen in the pollen record (Banerjea et al. 2020a). The clearance of vegetation provided leaf fodder for the animals. Microstratigraphic analysis (see Banerjea et al. 2020a) shows that a spring was dammed using redeposited clay to create a small pond as a watering hole for livestock, which also contained fish at one point, possibly used as a food source by occupiers during the construction of the castle. Midden material infilled the pond when it fell out of use. This was rich in artefacts and refuse, but only very few pottery sherds (Valk et al. 2012). It also contained plant macro-remains (Banerjea et al. 2020a), including traces of cereals (rye, barley, oat, wheat), strawberry seeds and hazelnut shells, charred wood, egg shell, fish bone, burnt and unburnt bone; fragments of coprolites belonging to small herbivores such as sheep or goat were also trampled into it. Microstratigraphic analysis of other deposits dating to the fourteenth century (see Banerjea and Badura 2019) showed that they formed in an external area and comprise poorly formed soils, trampled or frost-affected layers, and a levelling surface, which is overlaid by occupation deposits containing anthropogenic refuse and which could indicate a point when there is an intensification of activity from the newly built castle.
At Cēsis, our understanding of the development of the castle landscape has been increased by geoarchaeological analysis at three locations: the stronghold on Riekstu Hill, the outer bailey of the stone castle, and the eastern range (Figure 3). Micromorphological methods were applied to examine potential destruction deposits associated with the earthwork and timber fortress of Riekstu Hill (see Banerjea and Badura 2019) to understand its composition, whether it consisted of multiple structures that were destroyed, or whether it represents multiple rebuilding and destruction events or a single event. The analysis examined the validity of the suggestion of religious assimilation and peaceful coexistence between the Teutonic order of the Brothers of the Sword and the Vends, who are documented as voluntarily converting to Christianity in 1206 (Banerjea and Badura 2019;Pluskowski et al. 2019b) and supports the hypothesis that Cēsis was a place of coexistence between the Vends and the Sword Brothers. Micromorphology shows that the timber fortification of the hillfort was only destroyed once, and that subsequent thin horizons containing charred material were formed because of erosion and trampling processes, rather than due to several destruction or burning episodes. The erosion and trampling events were interspersed with periods of surface stabilisation where pedogenesis began, which indicates periodicity in their formation. This shows that the timber structures were not repeatedly destroyed and rebuilt prior to the construction of the masonry fortifications on Riekstu Hill. The Order's stone castle at Cēsis (Wenden) was begun (by the Brothers of the Sword) in 1218 and overlapped with the occupation of the adjacent the Vends' stronghold on Riekstu Hill .
The occupation of a castle site can be dynamic. At Molina de Aragón, soil micromorphological data from profiles within the citadel and the church of Santa Catalina, or Cristo de las Murallas, dating from the twelfth century ( Figure 4) show alternating periods of habitation occupation, abandonment and post-abandonment occupation (Banerjea et al. 2020b;García-Contreras et al. 2016). These microstratigraphic data show the changing nature of the occupation within this area of the citadel, from Islamic to Christian Micromorphological thin-sections from samples from the citadel, collected from strata between a destruction deposit overlying the earlier Islamic floor and occupation, and post-medieval features such as a nineteenth century drain (Arenas 2008, 26 and 50) show a series of plaster floor surfaces with associated occupation residues containing herbivore dung; livestock were present in this area and the dung was also used as fuel. Units of discarded material and trampled floor surfaces (Banerjea et al. 2020b) represent the (Re)Conquest phase, c. 1154 (Arenas 2008), and are overlaid with a series of mortar and gypsum plaster floors (Banerjea et al. 2020b), probably related to modifications inside the courtyard during the thirteenth and fourteenth centuries (Arenas 2008). The other inclusions within these occupation residues represent domestic activities. This contrasts with the deposit below the destruction horizon, overlying the earlier probably eleventh century Islamic plaster floor (Arenas 2008), which contained very low abundances of occupation residues associated with domestic activities and suggests that the plaster floor below was cleaned regularly. The deposit overlying the floor plaster shows features of soil forming processes which indicate that this area was abandoned and left without a roof for some time prior to being covered with the thick deposit of destruction material. Experimental archaeological research has shown that in Northern Europe occupation deposits can turn to soil within 2.5 years after being left without a roof , but no such experiment has been conducted in a Mediterranean context.

Agriculture and land-use
Buried archaeological deposits within castles can be related to the contemporaneous transformation of the associated landscape. Palynology has been crucial in reconstructing the vegetation and land-use changes associated with the impact of the Baltic Crusades, particularly at the regional scale, with geoarchaeology, arguably, playing a supporting, but important role in detecting more localised, immediate changes such as soil erosion and specific events relating to vegetation clearance and animal husbandry. Reliable chronologies are a fundamental part of palaeoecological reconstruction from lake and peat bogs . High resolution dating is particularly necessary in the medieval period; long pollen sequences covering several thousand years often lack the sample and chronological resolution required to resolve rapid or short-term changes in vegetation occurring at the sub-centennial to decadal scale (Brown and Pluskowski 2014). Although rotational farming was already present in north-western Estonia before the crusades (Pluskowski 2019b), in Livonia and parts of eastern Prussia, pre-Crusade agricultural practices survived for centuries and trade was not as well developed as elsewhere . Colonisation in the eastern Baltic (Livonia) was limited largely to the major urban centres such as Riga, where their establishment involved new concentration of people and animals (Banerjea et al. 2017;Brown et al. 2017), or key rural centres such as the Teutonic Order castle at Trikata, located on one of the main north-south roads through Livonia that probably held a key provisioning role (Stivrins et al. 2015;. There was limited colonisation of the rural hinterlands, and native populations continued to follow indigenous patterns of land-use (Stivrins et al. 2015;Brown and Pluskowski 2020). In comparison with geochemical enrichments and macrobotanical evidence from middening activities in Riga during the thirteenth and fourteenth centuries, the data from sediments within the moat (mid-fourteenth century) and first bailey of the stone castle at Cēsis indicate quite a low intensity of occupation, without any large build-up of effluent and refuse as seen at Riga (Banerjea and Badura 2019).
Geochemistry and palynology can produce complementary datasets for reconstructing land-use around castles. Palaeoenvironmental data from Lake Araiši, eight km south-west of Cēsis, revealed that the most significant changes in vegetation and environment occurred between A. D. 780-1050 with the establishment and occupation of a Late Iron Age lake-dwelling, but that after the crusades the form and type of agriculture continued as it previously had. There is little evidence for increased land-use associated with the founding of the castle during the first half of the fourteenth century or with the subsequent post-medieval manor. Similarly, geochemistry shows an increase in Titanium (a reliable indicator of soil erosion and anthropogenic trace elements, Boyle 2001) which suggests Late Iron Age soil erosion and contamination from activities associated with a nearby castle of the Cēsis commandery (Stivrins et al. 2015).
At Radzyń Chełmiński, Poland, the close association between the lake, the stronghold occupied from the eleventh to the early thirteenth century, and the later fourteenth century castle and medieval town presents the ideal opportunity to reconstruct past vegetation and land-use dynamics within successive phases of archaeologically and historically documented phases of colonisation and land-use . Geochemical analysis of cores from the Castle Lake identified changes in the landscape associated with the earliest phase of land-use in the seventh/eighth century associated with Slavic colonisation. Geochemical analysis detected subtle land-use changes before they were observed in the standard pollen abundance record, and can reflect activity at a very local level in comparison with the palynological data , which is harder to identify due to the bias in tree pollen and its wider dispersal; a sharp peak in Titanium enrichment starting c. A. D. 750 represents soil erosion and a decrease in tree pollen from tree removal is not observed until A. D. 800-900. The modelled pollen data correct for this bias and show an increase in meadow land and cereals from A. D. 700 to 800 , which is consistent with the geochemical evidence for soil erosion.
In Spain, differences in climate (affecting pollen preservation) and agricultural practices give geoarchaeology a different role to play than in the eastern Baltic in understanding land use and landscape management such as terracing associated with castles in this region. Palynological studies in Spain have to date been conducted primarily on peat bogs and lakes located in mountainous landscapes at mid-to high-altitude, with the aim of investigating long-term Holocene vegetation history, climate and human activity (e.g. Although there are a significant number of palaeoenvironmental studies from Spain, with a notably concentration in the Spanish Central System, the chronological resolution is usually insufficient to provide a nuanced interpretation of medieval land-use and vegetation change at the sub-centennial scale appropriate for comparison with the historical and archaeological record, as undertaken in the Baltic. For example, Castro et al. (2020), discussing the Late Iron Age period and medieval climatic optimum, report 22 radiocarbon dates, but only five for the last 2000 years. Many sequences have one or two radiocarbon dates corresponding to the medieval period, or none at all. This is not intended as a criticism of existing palynological research, but rather an observation that a focus on the medieval period requires a higher chronological resolution than is available in many cases.
The data from these mid-and high-altitude pollen studies suggest a varied picture of vegetation change and human activity occurring in the context of important climate and cultural changes (e.g. In al-Andalus, they have been studied as refuges for peasants, fortified granaries or administrative centres; during their Christian phases, their political and military roles are studied. In neither case are castles or their inhabitants explicitly attributed a proactive agency in shaping and transforming the surrounding landscape. This is one of the aims of the new 'Landscape of (Re)Conquest' research project (García-Contreras  where the geoarchaeological sampling strategy at Molina de Aragón encompasses areas within the castle, adjacent terraces and irrigation system (Figure 4(B)). Palynological studies will be conducted as part of this project, but the lakes and wetlands from which cores will be taken, are situated outside of the valley of Molina (Figure 4(A)).

Presenting castle landscapes
In the castle landscape, monuments can be related to historically defined territories, and particularly in the eastern Baltic, with sacred landscapes (Pluskowski et al. 2018(Pluskowski et al. , 2019c. In Iberia, the castle landscape is synonymous with a shift in the overall pattern of rural settlement (Pluskowski et al. 2011). In the eastern Baltic, many castles only survive as ruins or have no above ground remains, which poses the question as to what extent ruins should be considered 'a castle'? Thus, the protection of castle landscapes is regulated by different general principles depending on type and level of survival.
The castle landscape can be articulated as a 'heritagescape' (Garden 2006). The term offers a common language to examine heritage sites, and a coherent methodology to analyse them, both as individual places and in terms of their relationship to other heritage sites. While it is an established ontology, 'heritagescape' continues to be revisited as a springboard for articulating new ways of preserving, protecting (Lopes et al. 2019) and promoting (Penrose 2020; Burlingame 2019) heritage, and as a basis for developing analytical, interpretive and operational models in tourism (Fagence 2014). A 'heritagescape', like any other landscape, is the set of individual components that, together, create a particular place. Whilst heritagescape incorporates some of the basic constructs of landscape theory, it is a particular idea that relates wholly to heritage sites. The framework of administrative protection such as that of a UNESCO Geopark provides a wider context in the ethnographies of particular heritage and/or tourism sites, or for particular countries' or regions' tourism development endeavours linking both natural and cultural values (Di Giovine 2009). A Geopark has a formal structure that can create and maintain interdisciplinary understandings of the landscape.
In that sense, by using geoarchaeological and other palaeoenvironmental proxies, links between the heritage site (i.e. the standing remains of the castle) and other related sites and features in the landscape (such as, at Molina de Aragón, the activities within the castle and those on associated terraces and irrigation systems) can be more effectively demonstrated, linking landscape research with policy, practice and knowledge exchange (Figure 4). In the research framework of a 'castlescape', these should be investigated and presented together in future signage, rather than separately, with interpretations beyond a focus on the geomorphological context of archaeological sites ( Figure 5). For a site such as Molina de Aragón, the presentation of the development of the castle and the history of its territory can be synergised with the geoheritage of the Molina and Alto Tajo UNESCO Global Geopark (Saiz et al. 2015) thus linking geological heritage and all other aspects of the area's natural and cultural heritage and enabling the creation of geo-educational itineraries of the landscape , fundamental features of a Geopark.
For historical landscapes, collaborations between physical sciences, humanities and social sciences can be more effective in connecting landscape research with policy and practice. Geo-educational itineraries for the 'castlescape' within a Geopark, which is arguably an ideal setting, can provide models that can borrowed by castle museums and visitor centres outside of a Geopark. Such itineraries can be integrated into heritage trails and cultural routes that include castles such as the Hansa network (a Council of Europe 'Cultural Route') that includes Cēsis Castle. Geoarchaeological and palaeoenvironmental research enables stories within the soils to be told: stories about contested pasts, some of which still provoke an emotional response, and stories about other aspects of cultural identity such as food and wine tourism (Metro-Roland and Soica 2019), where soils play an important role in past, present and future production.
The county of Molina, a long-established historical territory in central Spain (Figure 1), provides an exceptional opportunity for investigating the heritage of the dynamics of frontiers, colonisation and social reorganisation during the Middle Ages. Over several centuries, political control of the territory fluctuated: first part of an Islamic polity (A. D. 711-1128) and subsequently of various Christian authorities (first a semi-independent lordship linked to the Castilian crown under the rule of the Lara Family, a short period within the Kingdom of Aragon, A. D. 1366 -1375, and thereafter a part of the Crown of Castile (Estepa 2006, García-Contreras et al. 2016. The centre of power in the region consisted of a fortified hilltop complex in one of the principal mountain ranges in Iberia. The different phases of evolution of the main fortification are directly related to landscape changes identified across its territory, and to the spectrum of cultivated fields, mines, quarries or forests that were exploited. All this is evident from analysis of both the raw materials used in the standing architecture and the buried archaeology, and from spatial analysis using GIS with contributions from geoarchaeology, from soil micromorphology to palynology. On the other hand, the institutions in charge of managing both natural and cultural heritage are based in the territory's own capital.
As discussed above, geoarchaeological analysis in castles has successfully demonstrated how inhabitants of castle sites began to modify their local environment and adapt and manage it. At Karksi, Estonia, for example geoarchaeological evidence shows clearance of birch vegetation prior to the construction of the High Castle and the damming of a spring to create a small pond-cum-watering hole for livestock; at Elbląg, Poland it has been shown how inhabitants responded to repeated flooding of the outer bailey. Such analyses also provide snap-shots of how different areas of the castle landscape looked at specific points in time by revealing change through time, activities within different spaces, areas of abandonment and reuse and interactions with wider agricultural practices by understanding what livestock were kept in the castle, where and how they were fed. A modern visitor to these castles might well see these spaces as blank areas, yet geoarchaeology provides information to colour these areas and enhance their visibility within and a wider castle landscape and heritagescape.
'Heritage' itself has very little to do with the past, and its practice should fundamentally be concerned with assembling and designing the future though dialogue; thus cultural heritage conservation could become part of a broader environmental agenda (Harrison 2015). The scientific importance of palaeosols (French 2003) is not novel in archaeology. However, their protection as cultural and geoheritage (Costantini and L'Abate 2009) and their potential for informing dialogues in the interpretation of cultural landscapes (Conway 2010) is a concept which is starting to gain some momentum (Messaeroli et al 2020; Garrido Machado Botelho 2020) For example, in the UK, sediment profiles can be highlighted as a 'Site of Special Scientific Interest' (SSSI) and a Geopark provides an obvious administrative framework in which to highlight soil and sediment profiles to be protected as palaeoenvironmental archives of cultural scientific potential, arguably an 'in situ' archive . With the recognition that 'change and transformation has been reframed as an integral element of heritage ' (DeSilvey et al. 2020, 360), it is not argued here that castle landscapes (or any cultural landscapes) are frozen in time, but where under threat, the potential of soils and sediments is explored and they are geoarchaeologically investigated. Soils, sediments and stratigraphy are the backbone of the archaeological record, yet geoarchaeological techniques to investigate them still remain underused resources (Goldberg 2008;Goldberg and Aldeias 2018).
Geoarchaeological data from a castle landscape can illustrate historic processes such as anthropogenic pollution and localised soil erosion (e.g. Brown et al. 2015), and better understand the formation, use and erosion of abandoned terraces and thereby inform appropriate conservation (Bevan and Conolly 2011; Stanchi et al. 2012). When castles and landscapes are adapted or repurposed, modern owners or users would benefit greatly from the type of information that can be gained from prior geoarchaeological and palaeoenvironmental examination of the onsite soils, sediments and features (such as agricultural terraces) that lie beyond the monument but are still part of the castlescape. A detrimental, but fortunately isolated, example of the lack of such prior research is the Teutonic Order's castle in Giżycko (formerly Lötzen), where the transformation of the castle into a hotel in 2009-2011 without any archaeological investigations, led to the original castle buildings and their surroundings being covered by a new structure stylised as Gothic architecture, and arguably to the devastation of a historical castlescape. Today, the new buildings are often considered by tourists as a 'stylishly restored' castle (Pluskowski 2012, 355, 364). Garden (2006) argues that a 'heritagescape' should have clearly defined boundaries and we argue that for a castle's 'heritagescape' those boundaries should encompass its hinterland, or its landscape, which in the case of the Teutonic Order's castles in the eastern Baltic, for example, is the extent of their districts, and for a castle such as Molina de Aragón is its dynamic territory defined by multiple shifting frontiers. Geoarchaeology creates synergies between on-site and off-site environments and between cultural and natural heritage, and thus enables proper understanding of such large and extensive complexes. Soils and sediments are the backbone of the archaeological record and contain valuable information about the changing form and function of castles and associated land-use. These narratives ought to play a pivotal role in the creation of joint studies using historical, archaeological and heritage approaches for the conservation and management practice of castles and landscapes. The 'castlescape' is a particular type of 'heritagescape', often with many buildings, such as towers, associated with the castle, with a range of spaces within the castle each with their own complex life-histories, and with extensive territories beyond the castle, such as agricultural systems. The boundaries of hinterlands, and frontiers around the castle, can be fluid and so the presentation and management of the castle landscape should accommodate this.

Conclusion
The abstract concept of 'castlescape' is not a physical dimension that could be measured in basic cartesian terms, but rather represents an effort to situate fortified sites within the context of their broader landscapes and environments, not only in terms of visual space or their economic catchment area but also in relation to their role in the reconfiguration of territories, land-use and settlement patterns both in the past and present. This requires the integration of both on-site and off-site perspectives when considering an individual site's heritage. In summary, the castlescape is the cultural landscape associated with the biography of the castle. For this reason, there is no single model which can serve as a template for every castle; instead, it is important to begin with the specifics of each site. This highlights the value of local, biographical studies alongside broader, regional and inter-regional approaches.
In this regard, the papers in the special issue on 'Forgotten Castle Landscapes' of which the present paper is part, present many case studies for linking castle landscapes with heritage management. They range from embedded, active museum engagement with a range of research strands (Runge 2021), to castles at risk from climate change (Kerr 2021), to the challenges of presenting castle landscapes to the public (Bizri et al. 2021), and to the challenges of conducting landscape research on urban fortifications ). These cases show that palaeoenvironmental investigations should of course be embedded within research project designs for castlescapes at the planning stage, rather than as an afterthought; such integration of palaeoenvironmental data with survey and museum engagement will not only enrich our understanding of castle landscapes, but can bring them to life for a broader audience, and inform management decisions regarding what is heritage at risk and how to investigate it.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Notes on Contributors
Rowena Y. Banerjea is a Researcher at the University of Reading (UK) in geoarchaeology, formation processes of the archaeological record, and frontier landscapes in medieval Europe. Rowena worked as Forum Coordinator to establish the National Heritage Science Forum (2011)(2012)(2013) under the auspices of the Centre for Sustainable Heritage, University College London (UK).
Guillermo García-Contreras Ruiz is Senior Lecturer at the Universidad de Granada (Spain) where he teaches Medieval and Postmedieval Archaeology. He is specialist in Islamic Archaeology, Landscape Archaeology and settlement patterns in relation with natural resources exploitation. Currently he is Co-PI of the international project 'Landscapes of (Re)conquest'.
Gundars Kalniņš studied history at the University of Latvia and graduated from the University of Leicester (UK) where he acquired a MA degree in Archaeology and Heritage. Since 2005, he has worked at Cēsis Museum where he is responsible for the conservation, research and interpretation of the Cēsis Castle.
Maciej Karczewski is Associate Professor at the University of Białystock (Poland). His research interests include settlement and environmental archaeology of the southeast Baltic in the 1st-15th centuries from a multidisciplinary perspective. His research also covers historical archaeology, especially the archaeology and heritage of WWI, and the history of archaeology and archaeological museology in the areas of Poland, Lithuania, Belarus, and former East Prussia.
Aleks Pluskowski is Associate Professor of Medieval Archaeology at the Department of Archaeology at the University of Reading, UK. His research focuses on the relationship between nature and culture in medieval Europe, particularly within frontier societies associated with crusading, colonisation, cultural encounters and religious transformations. He is currently the PI of the 'Landscapes of (Re)conquest' project.
Heiki Valk is a lead researcher at the University of Tartu, Estonia. His research concerns the final stage of the Iron Age, the time of the Baltic crusades and the later medieval period in the Eastern Baltic region. The main research interests concern hillforts, as well as social and cultural aspects of the archaeological record.
Alexander Brown is Principal Geoarchaeologist at Wessex Archaeology (UK) and a visiting academic at the University of Reading (UK). His research uses palaeoecology to investigate vegetation and land-use change, particularly associated with the conquest and colonisation of European medieval frontier landscapes. His research also includes woodland history and wetland and intertidal geoarchaeology.