Quaternary International xxx (xxxx) xxx Contents lists available at ScienceDirect Quaternary International journal homepage: www.elsevier.com/locate/quaint The mobility of shepherds in the Upper Pyrenees: A spatial analysis of pathways and site-location differences from medieval times to the 20th century David Garcia Casas a, *, Ermengol Gassiot Ballbè b a Institute of Heritage Sciences, Spanish National Research Council, Edificio Fontán, Bloque 4 Monte Gaiás, Santiago de Compostela, 15707, Spain b Prehistory Department, Autonomous University of Barcelona, Edifici B, Cerdanyola del Vallès, 08193, Spain A R T I C L E I N F O A B S T R A C T Keywords: This paper investigates ancient pastoral mobility in a high mountain area of the Pyrenees. Firstly, modern Pyrenees transhumant routes were analysed using GIS tools in order to understand possible determinant factors (such as Pastoralism terrain slope and altitude, water courses and possible nodes) in the layout of a route used for seasonal livestock Transhumant routes movements. The observations obtained were then used to model optimal paths which may have been used by Least-cost paths GIS ancient shepherds. Subsequently, an analysis was made of the spatial relationship between the simulated paths and the dispersion pattern of archaeological sites in two time periods: Late Antiquity-Medieval (3rd-14th cen­ turies AD) and Modern-Contemporary (18th-20th centuries AD). The results show significant differences in the spatial distribution of the sites throughout both periods in terms of accessibility and proximity to possible pathways. This variability provides information concerning historical changes in the social structure of pastoral alpine landscapes over long periods of time. 1. Introduction the social and economic relationships between regions at different levels of altitude.These seasonal movements have been the object of study by This paper focuses on ancient livestock routes across alpine and ethnographers, anthropologists, geographers, historians and archaeol­ subalpine zones and how these seasonal movements lent structure to ogists since the beginning of the 20th century. Transhumance has been human settlements in high mountain landscapes in the past. Seasonal regarded as a social phenomenon which has connected different regions pastoralism is a socioeconomic activity in which shepherds move flocks and societies since Late Prehistory. For instance, seasonal livestock between upland pastures in the summer and lowland pastures in the movements have been proposed as an explanatory factor of cultural winter. The objective of this movement is to make use of seasonal var­ diffusion during the Bronze Age in the Iberian Peninsula (Jimeno, 2001). iations in vegetation to feed flocks. These kinds of herder activities have Other studies have considered that this activity could provide an been of great economic and social importance in different rural regions explanation for the position of certain megalithic elements (Chapman, of Southern Europe. A great deal of variation can be found within this 1979; Galán and Ruiz-Galvéz, 2001). general strategy depending on the region. In the mid-20th century, this Furthermore, over the last twenty years, archaeological research in activity underwent a decline due to industrialisation, the withdrawal of the principal mountain chains of Europe has undergone significant primary activities and changes in the economy of rural zones. However, development, particularly in the Pyrenees (Gassiot Ballbè, 2016; Orengo despite this downturn, the activity still exists in certain areas, particu­ et al., 2014; Rendu et al., 2016) and the Alps (Della Casa, 2013; Walsh larly in Spain, France, Italy and Greece (Costello and Svensson, 2018). et al., 2014). Paleoenvironmental and archaeological research has Seasonal livestock movements have shaped the landscape of the rural shown that upland grazing areas in the Pyrenees have been exploited by areas in which transhumance took place. Place names, buildings such as pastoral groups since the fourth to third millenniums BC. (Galop, 2016; huts and enclosures, bridges and water sources are the materialisation of Gassiot Ballbè et al., 2014; Laborda et al., 2018). The spatial coincidence this activity. These movements have constituted a significant element of on a regional scale of prehistoric archaeological sites with modern the economy of the regions in question and have provided structure to transhumant settlements may suggest the dawn of a transhumant model * Corresponding author. E-mail address: David.garcia-casas@incipit.csic.es (D. Garcia Casas). https://doi.org/10.1016/j.quaint.2023.07.007 Received 10 January 2023; Received in revised form 26 June 2023; Accepted 26 July 2023 1040-6182/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Please cite this article as: David Garcia Casas, Ermengol Gassiot Ballbè, Quaternary International, https://doi.org/10.1016/j.quaint.2023.07.007 D. Garcia Casas and E. Gassiot Ballbè Quaternary International xxx (xxxx) xxx of pastoralism at the end of the Neolithic. However, other studies reject 2. Regional settings this timeless interpretation. While “transhumance” is a modern term used to describe modern pastoral practices (Rendu, 2006), certain The research takes two scales of analysis into consideration (Fig. 1). archaeological studies based on fieldwork data in different mountain The larger scale (Fig. 1b) is used to analyse the well-documented areas reject the idea of continuity from early herder occupations of transhumant pathways extending from the Ebro Basin to the upland upland zones at the end of Prehistory to modern transhumant pasto­ Pyrenean grasslands where the flocks graze in summertime. Fig. 1 b ralism. According to these studies, detailed examination of material shows the path selected for this study from among all the transhumant remains reveals discontinuities and a succession of different pastoral routes ethnographically recorded in the area (Fig. 2). The area shown in practices over this large timescale (Carrer and Angelucci, 2017; Gassiot the figure comprises the east of the province of Huesca in the region of Ballbè et al., 2020; Le Couedic et al., 2016). Aragon and the province of Lleida in Catalonia. The starting point has The aim of this paper is to investigate seasonal pastoralism move­ been chosen as it is the confluence of three minor routes. The arrival ments in the past in upland zones and to explore their possible spatial point is situated in the Val d’Aran, near the French Border. The large relationships with the location of archaeological sites. To achieve these area between these two points is characterised by a steep altitude objectives, our research focused, first of all, on a well-documented his­ gradient, passing from 200 to 300 m.a.s.l. in the Ebro basin to 3000 m.a. torical transhumant pathway from the lowlands of the north of Spain to s.l. in the Upper Pyrenees. In the southwestern part of the path, the Ebro the Central Pyrenees. In this step, the goal was to understand the ra­ basin and Cinca Valley are characterised by a Continental- tionality of the route in an attempt to discover why the transhumant Mediterranean climate with low rainfall (300–350 mm/yr), high inso­ route runs where it does. This provided the basis for the next step, which lation and aridity. On the other hand, the Axial Pyrenees are mainly consisted of modelling possible pastoral paths in the National Park of characterised by an Atlantic climate, with cold winters, mild summers Aigüestortes and Estany de Sant Maurici (PNAESM) via the use of digital and high rainfall throughout the year. Annual precipitation exceeds tools. This is an upland zone in which previous research carried out by 1000 mm/yr, while average temperatures range from − 4 ◦ C to 2 ◦ C in our group has recorded nearly 380 archaeological sites related to live­ winter and 13 ◦ C to 18 ◦ C during summer. stock activity over different periods of time (Garcia Casas, 2018; Gassiot The smaller and most important scale of analysis is the National Park Ballbè, 2016). The second part of this study is devoted to exploring the of Aigüestortes i Estany de Sant Maurici (PNAESM) (Fig. 1a), where the spatial differences in the location of archaeological sites from two archaeological sites selected for analysis are located. The PNAESM, different time periods: Late Antiquity and Medieval (3rd-14th centuries which coincides with the upper part of the transhumant route, covers an AD) and the Modern-Contemporary period (18th-20th centuries AD). It area of almost 40,000 ha in the Pyrenean area of the province of Lleida takes up a previous analysis that, in similar terms, established significant at the eastern extreme of the granite batholith of Maladeta. Some of the differences between sites associated with livestock practices from the peaks reach an altitude of 3000 m.a.s.l and the valley bottoms are, at Late Neolithic and Contemporary periods (Gassiot Ballbè et al., 2020). their lowest points, over 1500 m.a.s.l. The relief shows traces of glacial The results contribute towards improving knowledge of the historical action by the geological substratum and the alpine orogeny, with a dynamics and mutations of seasonal pastoralism and can provide an landscape dotted with small lakes, cirques and U-shaped valleys insight into the social organisation of high-mountain Pyrenean terri­ delimited by peaks that are generally quite steep and rocky. The tories in pre-modern times and their evolution towards the traditional orientation of the valleys reveals significant variations, especially in agrarian and transhumant landscapes of the 20th century. terms of humidity and precipitation, a fact which, along with their Fig. 1. a. (Left). The transhumant pathway analysed in a regional scale. Fig. 1 b (Right). Location of the study zone. The Aigüestortes and Estany de Sant Maurici National Park (PNAESM) and the archaeological sites analysed. 2 D. Garcia Casas and E. Gassiot Ballbè Quaternary International xxx (xxxx) xxx Fig. 2. Transhumant routes (cabaneres) at the beginning of XX century according the data collected by Ramon Violant i Simorra (R.V.S). Map adapted from data provided by Ignasi Ros. Some of toponyms are highlighted because they are mentioned again in the Fig. 8. 3 D. Garcia Casas and E. Gassiot Ballbè Quaternary International xxx (xxxx) xxx altitude, has a significant effect on ecosystems. very different to that of modern transhumance (Garcia Casas and Gassiot Ballbè, 2022). The other three sites are rock-shelters with enclosures in 3. Materials and methods the interior of which open area excavations or test pits have been carried out. The chronology of the occupations is inferred by AMS dating, with 3.1. The case study: archaeological data and previous research in the the exception of the three grouped settlements, in which the chronology PNAESM is inferred by morphological similarities in the architecture and number of structures (Garcia Casas, 2018). Table 1 shows the radiocarbon dates Since 2004, an archaeological research project has focused on the of archaeological sites analysed (Garcia Casas et al., 2022; Garcia Casas identification and mapping of archaeological remains in the PNAESM and Gassiot Ballbè, 2022; Gassiot Ballbè, 2016). If there are more than with extensive surface surveys employing a transect strategy to cover the one date in the same phase the table shows the maximum interval taking majority of the accessible terrain (around 90% of the whole PNAESM in account. All the radiocarbon samples are charcoal. The results have area). The fieldwork recorded 378 archaeological sites (Gassiot Ballbè been calibrated to 95% probability with the IntCal20 curve using the et al., 2016). Many of these are open-air structures built in dry-stone, OxCal program (Reimer et al., 2020). mainly huts and enclosures interpreted as folds. Specifically, sites with For the contemporary period, 17 archaeological sites were selected, one or more open-air structure comprise almost 65% of the total (Gassiot with radiocarbon dates being provided for only one case (SO-007A). The Ballbè, 2016). Furthermore, there are rock-shelters with traces of human others were selected due to their morphological similarities with occupation sometimes associated with open air structures and, to a transhumant settlements of the early 20th century as described by eth­ lesser degree, ring-stones which could be tumuli and deposits of arte­ nographers (Garcia Casas, 2013; Vilà i Vilà Valentí, 1950; Violant i facts. Most of the archaeological sites are located between 2000 and Simorra, 2001) or as a result of modern materials related with the 2400 m.a.s.l. In general terms, the spatial distribution of the sites is archaeological site being found on the surface. In this article, similar. Half of them are located less than 75 m away from a water “contemporary” refers to a chronology estimated between the 18th and source in grassland areas (Gassiot Ballbè et al., 2016). The combination 20th centuries (Gassiot Ballbè and Garcia Casas, 2014). of grass and water is a key factor in modern husbandry and it is probable The routes of the transhumant pathways were extracted from the that this was also the case in the past. The information on the surface official cartography of the autonomous government of Catalonia was complemented by that from the excavations, which consisted of (Departament d’Acció Climàtica, Alimentació i Agenda Rural, 2012). both small test pits and excavations of larger areas. Using radiocarbon This study pays particular attention to one such pathway (Fig. 1b). dates of the samples collected, a large sequence from the last 10,000 Cabanera is a Catalan word used to refer to a transhumant pathway and, years has been obtained. for reasons of simplification, in this paper the selected path shall be Late Antiquity and the Middle Ages are, together with the last two referred to as the cabanera. This pathway was chosen due to the fact that centuries, the best represented periods in terms of archaeology. For the it is the subject of a previous study (Nadal Subirà et al., 2010), the au­ first sample, archaeological sites were selected with a chronology thors of which followed a transhumant flock at different times between ranging from the 3rd to the 14th centuries AD. In this group, seven 2004 and 2007, recording the route taken with GPS in order to create a ensembles of between three and seven open-air structures were identi­ shapefile. This provides a more precise route than the regional cartog­ fied, with at least one of them a hut and two enclosures interpreted as raphy. The route appears in the ethnographical studies by Violant i folds. Another ten are large ensembles of more than eight structures. Simorra, written in the 1930s (2001: 77), and the geographer Vilà i Vilà These kinds of settlements seem to be places where different groups of Valentí (1950) (see Fig. 2). shepherds with several flocks would come together (Garcia Casas et al., In this paper, two terrain datasets with different resolutions are used. 2022). This demonstrates a social organisation of pastoralism which is The first is the European Digital Elevation Model provided by the Table 1 Archaeological sites from Late Antiquity and Medieval Age selected for the analysis. Where available the chronological interval is presented. For a better description of radiocarbon results see (Garcia Casas and Gassiot Ballbè, 2022; Garcia Casas et al., 2022) see SPM1 for more data about radiocarbon results. Other settlements have been dated for architectural typology as described in (Garcia Casas et al., 2022). Types: LS: Livestock settlement, GS: Grouped settlement. RO: Rockshelter associated to open-air structures. Site (ID) Name Type Archaeological intervention Dated by Interval Phase N. Dates AA-024A Gerber I GS Test pit Radiocarbon 415–571 cal AD 1 AA-026N Gerber II LS Test pit Radiocarbon 603–774 1 AA-043 Conjunt del Mig RO Test pit Radiocarbon 612–775 cal AD 1 ESP-015 Portarro RO Open-area excavation Radiocarbon 890–1020 cal AD 3–4 1 for each phase 709–952 cal AD ESP-016 Pletiu de la Coveta RO Test pit Radiocarbon 579–660 cal AD 1 ESP-024 Pletiu de Subenuix LS Test pit Radiocarbon 250–414 cal AD 1 NA-026 Pletiu deth Pòrt de Caldes II LS Test pit Radiocarbon 126–346 cal AD 1 NA-076 Poblat della Passada deth Nebot GS Surface record Typology – 1 NA-079 Conjunt del circ de Rius I GS Surface record Typology – NA-084 Despoblat del Tuc deth Lac Redon GS Test pit Radiocarbon 213–361 cal AD 1 NA-090 Despoblat del Lac Tort de dalt GS Test pit Radiocarbon 255–423 cal AD 1–2 1 for each phase 1301–1408 cal AD NA-091 Conjunt del Plan de Rius LS Test pit Radiocarbon 1025–1160 cal AD 1 SO-006 Conjunt de l’Estany Gran de Mainera GS Surface record Typology – VB-001A Palanco de Llacs GS Test pit Radiocarbon 1175–1273 cal AD 1 VB-014 Cova del Sardo RO Open- area excavation Radiocarbon 885–1032 cal AD 3 2 VB-017 sup Despoblat i pletiu de la Cova (Conjunt superior) GS Open-area excavation Radiocarbon 890–1020 cal AD 1–2 3 for 2 phases 990–1154 cal AD 1225–1290 cal AD VB-019 Despoblat de Casesnoves GS Test pit Radiocarbon 1215–1280 cal AD 1 VB-085 Conjunt de la Pleta d’Erdo GS Test pit Radiocarbon 545–642 cal AD 1 VB-088A Despoblat de Port de Rus GS Test pit Radiocarbon 1277–1392 cal AD 1 VB-098 Conjunt de les Cometes de Casesnoves IV LS Test pit Radiocarbon 436–637 cal AD 1 4 D. Garcia Casas and E. Gassiot Ballbè Quaternary International xxx (xxxx) xxx European Union’s Copernicus Land Monitoring Service (European et al., 2015). However, such documentary sources have limitations, Union, Copernicus Land Monitoring Service, 2022) with a spatial reso­ particularly in terms of chronology, as sources about transhumant lution of 25 m. The second is taken from the website of the Institut pathways were most commonly written between the 15th and 18th Cartogràfic I Geològic de Catalunya (Institut Cartogràfic i Geològic de centuries. For previous chronologies, mention of the routes of the Catalunya, 2022). The geospatial information regarding hydrology was pathways is rare, or completely non-existent before the year 1000. The extracted from the official cartography of the government of the assumption that these paths existed even before prehistoric times is Autonomous Community of Catalonia (Departament de Medi Ambient y problematic. Thus, historical documentation has chronological limita­ Agència Catalana de l’Aigua, 2015). All the GIS analysis was performed tions. Moreover, trends in mountainous zones tend to be misleading on in ArcGIS 10.5. and the statistical analysis in R package. maps, with only certain paths (presumably the most important ones) being represented. For instance, in our study area, only two transhumant 3.2. Methodology paths are documented crossing the area of the PNAESM in a north-south direction. However, if the distribution of settlements in the PNAESM This research has two main goals: to identify the paths used by (Fig. 1a) is examined, it can be observed that there was mobility through shepherds in the past and to examine the relationships between these many more valleys. paths and the archaeological sites. In order to achieve these objectives, Seasonal mobility in terms of pastoralism can vary according to the analysis was divided into two parts according to each goal. The first different types of practices. A great deal of research has been carried out concerns the identification of seasonal pastoralism mobility and, on seasonal livestock movements in transhumant and nomadic shep­ particularly, the identification of shepherds’ paths in mountain land­ herds (Chang and Tourtellotte, 1993; Cribb, 1991). To summarise scapes. The second part seeks to assess possible relationships between briefly, in the case of the former, the routes are longer and more these paths and pastoral sites and whether these change over time. established, with the route taken being more or less the same every year, Movement is one of the basic mechanisms used by humankind to only changing due to unexpected circumstances, and with only one or a explore, interact and manage the space around them (Ingold and Ver­ few specialised shepherds making the trip. In the latter case, the gunst, 2008). Paths are the materialisation of this mobility in the land­ movements have more variability within an established framework and scape, weaving together the disparate elements of daily life (Snead et al., large families move with the flocks. 2009:1). In addition to connecting people and places, paths also provide This paper explores other routes potentially used by flocks and structure to the landscape around them. Thus, the identification of paths shepherds to cross the PNAESM using the GIS-based modelling analysis is an important issue when seeking to understand the way in which a called Least Cost Pathways (LCP). The modelling of human movement given area is organised. The study of movements, trails and paths is a via GIS-based methods has been the subject of much research in the field growing field within landscape archaeology (White and Surface-Evans, of archaeology (e.g White and Surface Evans, 2012; Herzog, 2014; Polla 2012; Polla and Verhagen, 2014), both as a theoretical paradigm to and Verhagen, 2014; among others). Previous studies (Gietl et al., 2007) understand them as components influencing the formation of landscapes have shown that the resulting route can vary depending on the algo­ and for the development of new methodologies to enable the study of rithm and GIS software package used. Furthermore, algorithms some­ ancient movement, such as Remote Sensing and Geographical Infor­ times give priority to the shortest paths, even if they go through rough mation System-based modelling (Saintenoy et al., 2017). The applica­ zones in which human movement would be dangerous without climbing tion of these technologies to archaeological research has made a great equipment. contribution to the analysis of human mobility in the past. One of the The starting point of this paper is that there are certain universal topics addressed by the research has been the analysis of the routes of factors which affect human movement, and physical constraints that ancient paths, it can be mentioned (Gassiot Ballbè et al., 2020; Lanen slow down, hinder or even completely prevent circulation. Slope is one et al., 2015; Herzog, 2017; Parcero-Oubiña et al., 2019) as examples. of the main constraints modelled in GIS-based mobility analyses, but it is For our case of study, the transhumant pathways, there are usually not the only one (e.g., it is impossible to walk in bodies of water such as no stone causeway constructions which can be observed on pedestrian lakes or great rivers). However, the importance of physiological factors surveys or stratigraphically excavations as is the case of other ancient is different depending on the subjects who are moving (Fonte et al., routes studied archaeologically (eg. Roman Roads or Inca paths) (Fonte 2017) For instance, the gradient of slopes has a different effect on et al., 2017; Mendez-Quiros Aranda and Saintenoy, 2021). Therefore, commercial routes used by chariots than on the movement of shepherds. their traces in the landscape are more ephemeral and difficult to detect In addition, there are differences in social and cultural values which in archaeological surveys or excavations. Their markers are usually affect subjective behaviour and decision-making when choosing a given milestones, sources of water for livestock, territorial boundary markers route (Llobera, 2020). There are socio-political factors which cannot be or traces of rock modification to open up passable paths. An example of neglected. Land ownership and usufruct regimes affect accessibility to this is shown in Fig. 3 in which it is possible to appreciate work carried certain pieces of land and spaces. Likewise, the safety of people or goods out on the rocks in order to facilitate movement. However, such marks in movement is also a factor which can influence mobility. In both cases, are rare and are not sufficient to reconstruct the route of transhumant the archaeological determination of these factors is extremely complex pathways. The study of transhumant pathways is frequently approached and requires a great deal of information which is generally difficult to via the study of historical documentation (Fairen et al., 2006; Burri obtain and is practically absent in our study area. In future, these factors should be incorporated into more complex studies of the mobility of people and herds. To achieve this, more specific research programmes will be required (such as the models developed by M. Le Couédic in the Pyrenean Ossau valley) (Le Couedic, 2010). Bearing this in mind, and for the reasons already pointed out in Gassiot Ballbè et al. (2020), the present study begins by considering orography as the main factor in the determination of optimal paths. Subsequently, other spatial constraints are added, such as water courses and the possible existence of nodes in the route. A wide range of functions have been developed to mathematically Fig. 3. Traces of an ancient road in the PNAESM. (Right) carved cross in a express the influence of terrain gradient on human mobility in the past mountain pass between two rocks with traces to be arranged as a passageway (see Verhagen, 2018 or Herzog, 2014 for reviews about different (Left). Photo: Grup d’Arqueologia de l’Alta Muntanya. studies).Two of the main algorithms used in archaeological research 5 D. Garcia Casas and E. Gassiot Ballbè Quaternary International xxx (xxxx) xxx regarding human mobility in the past were tested: the Tobler Function addressed: the definition of the departure and arrival points of the paths. (Tobler, 1993; White, 2015) and the Llobera and Sluckin Function In this step, places with historical or archaeological evidence of popu­ (2007). In addition, another function was tested which has been used lation in Late Antiquity or the Middle Ages were chosen (Gassiot Ballbè with success in a previous study on Neolithic mobility in the PNAESM et al., 2020). Table 2 and Fig. 11 show the points selected to be con­ (Gassiot Ballbè et al., 2020). In this step, the anisotropic versions of nected as well as the type of evidence of human occupation and its these algorithms were used in order to make a comparison with a chronology. The majority of the points coincide with present-day vil­ returning pathway. Consequently, it was necessary to take into account lages and the evidence is based on documentation regarding the foun­ the positive and negative gradient slopes in both directions. dation of local churches or, in town in the Val d’Aran, the discovery of The first is algorithm is described in Sluckin and Llobera (2007), archaeological materials suggestive of permanent populations. As an exception, the location of Ares on the mountain pass of La Bonaigua is M(sw) = 2.635 + 17.73sw + 42.37 s2w - 21.43s3w + 14.93s4w worthy of mention. This was the main point of access to the Val d’Aran where S is the gradient terrain obtained using the equation s = Tan before the construction of the Vielha Tunnel in the 1950s. It can be noted ((“slope in degrees” * 3.1416)/180). Slope in degrees is obtained using that the chronological evidence of settlements outside the park does not the “Slope” command in Arcgis from a Digital Elevation Model. match exactly with the chronology of the archaeological sites. In other The outcome of this function is expressed in energy consumed by a words, some historical evidence of the foundation of churches is dated to walker travelling in hilly terrain. For the purposes of the present study, it the 9th and 11th centuries, clearly later than sites from Late Antiquity. is better to obtain the travel cost in time units. Thus, the results were However, it must be noted, on the one hand, as some historians (Adell converting to time by dividing the outcome by 12800. This can be and Riu Barrera, 1994) argue, that the foundation of medieval churches visualized as equivalent time intervals, based on the energy consumed at took place in villages with a long history (i.e., human settlement was a standard walking distance of 5 km/h on flat terrain. prior to the first documentary evidence). On the other hand, there are The second function tested is the Tobler’s algorithm (Tobler, 1993; locations in valleys which connect with other valleys where there is White, 2015): evidence of human occupation from the end of Prehistory (Cots, 2003). Finally, the third part of the analysis was devoted to achieving the T = 0.001 / (6 * (Exp((-3.5) * (Abs((Tan((([S] * 3.1416) / 180))) + 0.05))))) second objective of the research: to examine the relationship between the paths and the sites and to verify whether there are changes according where to the chronology of the settlements. To achieve this, the distance from T = Time in minutes to walk 1 km. each settlement to the nearest LPC route was calculated. The develop­ S = Slope in percentage. ment of GIS applications makes it possible to perform different calcu­ Finally, the third function employed is the Nunes algorithm lations for distance (see Bevan, 2011; Herzog, 2014 or Verhagen, 2018; expressed in the following formula (Gassiot Ballbè et al., 2020): among others). Some studies consider distance according to geometric T = S*1,2532 + 9,1806.where criteria using density analyses or Thyssen Polygons. This kind of appli­ In the first part of the analysis, the proposals put forward by cation is problematic outside of flat zones. In steep terrain, such as that Güimil-Fariña and Parcero-Oubiña (2015) to analyse the reasoning of this case study, there is no reason to measure human mobility in behind the routes of historical paths are followed. Specifically, the Euclidean terms without considering topography due to the fact that it is research of these authors is devoted to understanding the rationality literally impossible to cross the area by walking in a straight line. For behind the route taken by a Roman road. In the present study, a similar this reason, in the present study, distance is assimilated to a cost surface procedure was performed. The route of the cabanera was compared with which produces a certain friction depending on given geographical an LCP created with some of the most commonly used algorithms in criteria. As was the case in the first part of the analysis, slope is the basis archaeological research, taking only biomechanical considerations into of this friction. For this part of the analysis, isotropic algorithms were account. This is a null hypothesis and no greater coincidence was ex­ taken into consideration in order to calculate a friction surface expressed pected. Rather, the goal was to compare the routes, to explore how a in travel time (e.g., Bevan, 2011; Kantner, 2012, as well as other cases transhumant route deals with physical constraints such as slope, altitude discussed in Herzog, 2014). In this part of the analysis, it must be borne or water courses and to attempt to create a GIS model in order to in mind that the roads or mobility axes tend to run along the bottom of perform LCP analysis with results similar to real livestock routes. To the valleys while the archaeological sites are located on the slopes. In perform the LCP paths, the ArcMap command called Path Distance, other words, in practically all cases, pastoral settlements are located at a available in the ArcToolbox (White, 2015), was used, which calculates higher altitude than the nearest possible axis of circulation. Therefore, an accumulative cost surface from one point in an isotropic way, taking the positive effect of using an anisotropic cost algorithm in terms of into account negative values (downhills). The workflow followed in this displacement cost is minimal. Considering positive and negative slope is step provides the 25m DEM and a table of slope values and their asso­ ciated costs as well as a raster with possible obstacles to pedestrian movement. These costs change depending on the algorithm used. Table 2 Therefore, three different tables are provided for each function. The 25m Historical and archaeological evidences of human settlement surrounding the DEM was used in order to facilitate the calculation over a large area and study zone. Points used to start and end the LCP crossing the Park. INVARQUE is the patrimonial inventory of Catalonia and is cited in Web and cartographical to reduce the bias in the DEM due to the surface of modern roads which references (Departament de Cultura, 2022). The other references are also cited did not exist in the past. The next step was to model the layout of in “References) (Bolos i Hurtado, 2012; Ordeig, 1993; Cots, 2003). different possible routes using the Cost Path tool. The second part of the analysis consisted of performing LCP routes to Cost factor Ascribed cost cross the PNAESM taking into account the observations in the previous Slope-based pedestrian cost (Llobera and MSw = 2.635 + 17.73 Sw+42.37 analysis. At this point, the 5m DEM was used for the following reasons: Sluckin, 2007) Sw2 − 21.43 Sw3+14.93 Sw4 Extra cost to avert river beds (depending on 3X the area is smaller; the presence of modern roads is negligible; and in the sector) order to take into consideration smaller paths or passageways unno­ Extra cost to avert mountain ridges (wit a 4x ticeable in a larger DEM. In this part, the main methodological goal was buffer of 10 m) (Carrer and Angelucci, to modify the usual workflow of LCP analysis to adapt it to geophysical 2017) Extra cost to avert slopes above 60% 3X patterns of mountainous zones and to obtain an output route that deals Extra cost to block crossing slopes above 10X with physical constraints in a similar way to the ethnographically 100% studied route. In this part, the main issue in any LCP analysis was Extra cost to block crossing lakes 10X 6 D. Garcia Casas and E. Gassiot Ballbè Quaternary International xxx (xxxx) xxx useful in calculations aimed at obtaining LCPs as the specific route may Thus, there are no differences in the altitudinal distribution of Medieval vary depending on whether negative slope is considered in the outward and Contemporary sites. and return. However, this step is not intended to determine a specific To perform the first LCP model, it was considered necessary to add an route from an LCP crossing the study area to the archaeological site. The extra cost for the lakes and water masses to avoid the LCP passing aim is to obtain a magnitude of the distance in terms of human mobility through them. For this reason, an extra cost of x10 to cross lakes or water between the two entities (LCP and settlement). In this step, the 5m DEM masses was added. The results of LCP01 are shown in Fig. 6, in which it was used in order to obtain a more detailed result. is possible to observe differences depending on the cost function per­ Since the aim was to obtain plausible magnitudes and not exact re­ formed. The outcome of the LCP performed with the Tobler cost function sults, the obstacles were simplified by only considering lakes as a barrier (LCP01TO) is quite different to the route of the cabanera, coinciding at to movement. The procedure followed consisted of: 1) Calculating the its beginning with a slight variation going to the east before coinciding slope in percentage from a DEM distributed by the ICGC with a pixel again. However, the most important divergence occurs when arriving at width of 5m (Institut Cartogràfic i Geològic de Catalunya, 2022); 2) Alta Ribagorça. It does not go to Pont de Suert and continue to the north, Producing a friction surface with each one of the 3 algorithms (Llobera, following the course of the Noguera Ribagorçana river avoiding the Boí Nunes and Tobler) 3) Performing a cost surface using the Cost Distance Valley. Arriving close to Espitau de Vielha, the route of LCP01TO turns command. The procedure generated three continuous maps of to the east and reaches the PNAESM via the Canau de Rius mountain displacement in minutes from the LCPs obtained to the rest of the ana­ pass. Next, it crosses the Valarties Valley and arrives in the town of lysed area (the surface area of the PNAESM) (Fig. 12). Next, the value of Arties. The route does not match again with the cabanera until the final the pixel corresponding to each settlement was extracted in the three stretch in the Val d’Aran. It should be noted that both routes cross the maps (results shown in Table 5). Finally, a statistical test was performed PNAESM, albeit via different valleys. The LCP performed with the Nunes in order to check whether the distance of the archaeological sites to the algorithm is even more different, matching with LCP01TO along the nearest LCP depends on their chronology. course of the Noguera Ribagorçana river and following the same route. However, it only coincides with the cabanera in the previously 4. Results mentioned final stretch. The third function, the Nunes algorithm (see Gassiot Ballbè et al., 2020), shows a great coincidence with the first one 4.1. Modelling transhumant movements in the upper zone. However, in the lowlands it has a different layout. On the other hand, the LCP created with the Llobera algorithm The study assumes that proximity to the optimal route is an impor­ (LCP01LLO) shows a greater degree of coincidence, particularly in the tant factor in the location of an archaeological site. This premise was Pyrenees, the mountain zone on which the analysis is focused. Despite verified by comparing the amount of terrain accessible from a route in some differences ascending through Aragon, the LCP01LLO model time intervals with the number of sites at the same distance. Specifically, crosses the Noguera Ribagorçana in the town of Pont de Suert, follows a a cost terrain surface was performed from a section of the cabanera route similar path through the Boí Valley and crosses the PNAESM area crossing the PNAESM to the surrounding areas which was surveyed through similar valleys. There are two differences which are of partic­ during archaeological fieldwork, the results of which can be observed in ular interest in terms of this study. The first divergence occurs after Fig. 4a and 4b and Table 3. In the first test, we used the Llobera and crossing the Caldes mountain pass with the route going straight up a Sluckin algorithm. As a result, 33% of the archaeological sites are near vertical zone (highlighted in Fig. 6). From personal experience and located less than 15 min away from the transhumant route despite the observations during fieldwork in the area, passing through this route is fact that this threshold only includes 11% of the terrain evaluated. On difficult and dangerous without climbing equipment. The second dif­ the contrary, only 5 out of 91 sites (5%) are located more than 150 min’ ference is a detour of the Cabanera arriving at Vall d’Aran and the walk from the transhumant route in contrast to 33.5% of the terrain Garona watershed. There is no apparent geophysical reason for taking (Fig. 5). Therefore, these observations suggest a preference for locating this detour. the archaeological sites close to the transhumant route. In order to understand the first divergence mentioned, attention was The second test was used to verify whether altitude is a variable focused on one of the methodological goals of this paper: the need to which can explain the variability of the spatial distribution. In other adapt the cost functions to the features of high mountain zones and to words, to check whether there are differences in the altitudinal distri­ explore the effects of features of mountain topography on the routes of bution of the sites depending on their chronology. To perform the test, transhumant pathways, the most important of which is, in our model­ the absence of normality in the elevation values was verified with the ling, the slope gradient. In the last LCP model, at least one unrealistic Shapiro-Wilk test (P-value 0.142). Therefore, the non-parametric Krus­ stretch in the output route was shown. It can categorically be stated that kal Wallis test was used to check the distribution of altitude sites ac­ slopes of more than 125% are always dangerous for human movement cording their chronology as independent value. The outcome (P-value without special equipment. As far as the movement of shepherds with 0.915) confirms no significant differences between the two samples. their flocks is concerned, the following aspects can be stated: a) Fig. 4. a (Left). Map of the PNAESM zone selected for the test, transhumant pathway (Cabanera) and nearby archaeological sites. Fig. 4 b (Right). The same map with the cost distance in minutes reclassified on intervals. 7 D. Garcia Casas and E. Gassiot Ballbè Quaternary International xxx (xxxx) xxx Table 3 First test results. Amount of land and number of archaeological sites at each cost distance interval. NORMALITY TESTS LLobera Tobler Nunes W p-value W p-value W p-value Distance to nowadays transhumant route 0.87954 0.0008346* 0.87473 0.000628* 0.88795 0.001386* Distance to LCP 0.95687 0.1605 0.96001 0.2022 0.95356 0.1257 SIGNIFICANCE TESTS LLobera Tobler Nunes T p-value T p-value T p-value Distance to LCP (T Test) − 2.7655 0.009405* − 2.6741 0.01171* − 2.8369 0.007957* Chi-square Chi-square Chi-square Distance to nowadays transhumant route (KW-Test). 0.00092879 0.9757 0.014861 0.903 0.0083591 0.9272 walk through lakes and large rivers. However, if the route of the cab­ anera is examined, it crosses some large rivers and streams. It is neces­ sary to bear in mind that what is being modelled is the movement of shepherds with their flocks numbering hundreds of livestock. Even crossing a river (in this case Noguera Ribagorçana) with small boats seems impractical. It is known that there were bridges to cross the river, but the great majority were built in the 19th or 20th centuries. However, on the route there is a village (Pont de Suert) the name of which suggests the presence of a bridge from at least Medieval times. Indeed, topo­ nymical and historical studies (Coromines, 1970) suggest the existence of the town and the bridge from at least the 1000 AD. Therefore, in subsequent models the existence of a bridge in this village was taken into consideration. The beds of the main rivers (the Segre and its affluents Noguera Ribagorçana and Noguera Pallaresa) were added as a constraint. Partially following the methodology of another case study (Güimil-Fariña and Parcero-Oubiña, 2015), an extra cost of x3 was added to a buffer area of 5 pixels (25 m) around the courses of the rivers, with the exception of Pont de Suert, where the cost was not altered. A relatively low degree of difficulty was attributed due to the fact that the Fig. 5. Graphic comparing the number of sites and the amount of terrain sur­ existence of fords or other bridges that have now disappeared cannot be rounding the transhumant pathway. discarded. Finally, it was taken into account that the route of the cab­ anera crosses the main rivers at different points. Therefore, they cannot Shepherds do not usually use climbing equipment in their movements; be modelled as insurmountable obstacles. In this case the extra cost is b) There is a threshold at which movement becomes dangerous and not high, meaning that the movement along rivers is not completely should be avoided. The main question is: where is this threshold? The hindered. Finally, a further constraint (ridges) was added to the new perception of a “steep slope” can be subjective and imprecise, especially model (LCP04). Mountain paths usually avoid ridges, due to their in terms of performing GIS analysis. To respond to this question, the exposure to high winds and lightning. To create the ridges in the model, slope of the cabanera was examined via the analysis of its altitudinal another study in ancient mountain pastoralism (Carrer and Angelucci, profile (the results are shown in Fig. 8). More than half of the terrain 2017) was followed. A hydrographic analysis was performed using an crossed by the cabanera has slopes of less than 15%. Indeed, slopes inverse DEM to execute the Flow Direction and Flow Accumulation com­ steeper than 60% are extremely rare, accounting for less than 5% of the mands. Then, the ridges were extracted creating a buffer of 10 m around pixels in the layout. It should be pointed out that the pixels are imprecise them and an extra cost of 3x was added. As with the case of the rivers, due to their size (625 m2). Extreme values in the graphic can be related the goal was to avoid LCP routes crossing this type of terrain, albeit with these pixel inaccuracies. To improve the analysis, the methodology without creating impassable barriers (if necessary, a shepherd and proposed by Güimil-Fariña and Parcero-Oubiña was followed, adapting his/her flocks can cross a ridge or a river). it to our study case. An extra cost of x3 was assigned to slopes steeper The outcome (LCP04, Fig. 10) now shows a remarkable similarity than 60% (in an attempt to avoid them and only use them if absolutely with the route of the cabanera. The pathway avoids crossing zones with necessary) and x10 to slopes of more than 100% (to definitively avoid steep slopes although there is still a divergence when the routes descend passing through terrain with this gradient). According to the first test from the PNAESM to the Val d’Aran and cross the Garona river at results, the analysis was performed with the Llobera cost function (see another point. Here, the cabanera makes a detour, apparently with no Fig. 9). relation to geophysical constraints. This may have been caused due to The results (LCP02) are not sufficiently satisfactory to understand the existence of a bridge or ford, rights of access or other sociohistorical the route of the cabanera. The outcome shows a divergence even higher reasons. At this point, it is considered that the LCP04 model is useful for than that of LCP01LLO in comparison with the original route of the modelling ancient shepherds’ paths between different villages crossing cabanera. Altitude could be another possible factor in understanding the archaeological zone of study (the PNAESM). why the cabanera runs where it does. Fig. 7 shows that the maximum altitude is 2564 m.a.s.l. Thus, in the next test an extra cost factor of x4 was added for land above 2350 m, providing exactly the same outcome 4.2. Making the least Cost Paths (LCP03) as LCP02. Therefore, altitude can be discarded as a movement factor to model in our case study. At this point of the analysis, a method was defined to obtain LCP At this point, another possible constraint for mobility in the study routes with a route similar to well-documented ethnohistorical trans­ area was examined. Common sense suggests that it is not possible to humant routes passing through mountain zones. To summarise, the method employs the Llobera anisotropic algorithm giving an extra cost 8 D. Garcia Casas and E. Gassiot Ballbè Quaternary International xxx (xxxx) xxx Fig. 6. Comparison of the Transhumant Route layout, and the LCP’s obtained using Tobler, LLobera and Nunes algorithms. Fig. 7. Histogram of the slope transhumant route. of x3 to rivers, x10 to lakes, x4 to ridges x3 to the slopes above 60% and Fig. 8. Topographical section of the transhumant route selected for the study x10 to the slopes above 100% (Table 4). The next step was to plot the shown in Fig. 1b and Fig. 2. The 2564 m threshold is highlighted. LCPs between the aforementioned points surrounding the PNAESM. The goal was to produce a hypothetical mobility map of shepherds in the past 4.3. Site locations, pathways and mobility based on current data regarding transhumant routes which could be of assistance in evaluating the location of archaeological sites (the results Having addressed the question of how to model ancient pastoral can be observed in Fig. 11). As can be seen, the routes of the three mobility in a mountain zone, the second part of this study is devoted to transhumant paths (cabaneras) inside the PNAESM show a high degree exploring spatial differences in the distance from the archaeological sites of coincidence with the LCP routes obtained. On the other hand, it to the LCP outcomes. In this part of the analysis the methodology should be noted that the LCP also coincides with present-day pedestrian employed for modelling geographical relationships is different as the pathways. It should also be noted that the LCPs performed are similar to spatial phenomena to explore them are not the same. First of all, the goal those modelled in previous studies (Gassiot Ballbè et al., 2020; Garcia was to understand the geographical factors which can influence the Casas, 2018). layout of a transhumant path. Then, another spatial (and social) phe­ nomenon was addressed: the distance from a possible livestock pathway to the locations of certain archaeological sites. The goal was not to address how much time it takes to move from one place to another. 9 D. Garcia Casas and E. Gassiot Ballbè Quaternary International xxx (xxxx) xxx Fig. 9. LCP02 and the layout of transhumant route (cabanera). Rather, the purpose was to obtain a plausible magnitude regarding the chronology were statistically significant. To achieve this, first of all, the distance from a spatial entity (the LCP) to two samples of archaeological Shapiro-Wilk test was performed (VanPool and Leonard, 2011) on each sites with different chronologies in terms of human mobility. of the three series of frequencies according to the algorithm. The out­ Table 5 shows the distance in minutes from the LCP pathways to the comes (shown in Table 5) demonstrate that the frequencies in each of archaeological sites. Firstly, attention was paid to the distance to the LCP the three series (distance according to the Llobera, Nunes and Tobler routes. The time differences depending on the algorithm used are, in algorithms) are distributed normally. Therefore, the T-test was used to general, quite significant if the sites are located at a short distance from perform the analysis with the results showing a high degree of signifi­ the LCP (less than 20 min). At higher time thresholds the distance can cance (see Table 6) in the outcomes provided by the three algorithms. increase by more than 20 min depending on the algorithm. If the results The distances in time provided by each algorithm are different, although are examined, it can be seen that these differences do not follow a linear in all three cases there are significant statistically differences depending pattern either for short or long distances. They can be attributed to how on the period. Therefore, it can be stated that differences in the distance the algorithms deal with different thresholds of slope. It is interesting to to an LCP route are related to the chronology of the archaeological site. note that the Llobera algorithm provides, in all cases, higher distance However, if the distance from each settlement to present-day trans­ values than the other algorithms. In general, the Tobler algorithm pro­ humant routes (cabaneras) is taken into consideration, there are no vides lower values, although, in some cases, the results of the Nunes significant differences (see Table 6). The distribution was not normal algorithm are similar. However, if the means are taken into account, the and therefore the Kruskal-Wallis Test was used. The values for distance results are clearer, showing considerable differences depending on in minutes to the archaeological sites are high even in the medieval or chronology. In this regard, it is useful for the purposes of this research, contemporary settlements, with the exception of two cases in each which focuses on modelling and obtains possible magnitudes rather than chronology (SO-007 for the Contemporary period and NA-026 for the reconstructing the exact time that a shepherd took in the past. The next Late Antiquity/Medieval period). This could be explained by the layouts step was to check if the differences in the frequencies depending on of present-day transhumant routes which only cross some areas of the 10 D. Garcia Casas and E. Gassiot Ballbè Quaternary International xxx (xxxx) xxx Fig. 10. LCP04 and the layout of transhumant route (cabanera). PNAESM. For this reason, LCP analysis proved useful in studying ancient Valley with the PNAESM in the Central Pyrenees. In brief, the maximum mobility. Even if the reconstruction of the routes is not the goal, GIS slope thresholds of this route have been determined, along with the role models make it possible to approach possible axes of circulation. of rivers, water masses and ridges as factors affecting the mobility of Moreover, the differences in distance to the archaeological sites provide livestock and the role played by a village as a node. However, more than information concerning historical changes in landscape use. its explanatory factor, in our opinion, the main achievement of this part of the research has been to obtain a methodology to obtain viable LCPs 5. Discussion as regular circulation routes. As pointed out in the introduction, the direct application of some of the procedures and algorithms for The research presented here has been successful in understanding obtaining LCPs in archaeology risk obtaining routes that are unfeasible mobility linked to the seasonal movements of livestock between the for human circulation. In many cases, the outputs prioritise crossing plains of the Ebro Valley and the Central Pyrenees and how this provided extremely steep slopes rather than travelling long distances to get structure to the territory in the high mountain landscapes of the Pyr­ around them, and, although in some cases it is physically possible to enees in the past. It has also made it possible to verify the existence of a avoid them, these are routes with a high risk of falling without climbing historical change in the organisation of the territory and settlement equipment. Thus, they are unlikely to constitute frequently used patterns in relation to the possible routes crossing the PNAESM that communication routes. arose between the end of the Late Middle Ages and the last two centuries The analysis raised has made it possible, via the observation of a before the present. Below, these achievements will be discussed known route, to obtain quantifiable and reproducible factors in GIS following the order in which they were presented. operations. The resulting LCP returns a path very similar to the trans­ The analysis of a present-day transhumance route has enabled an humant route used as an ethnographic reference. The same factors were understanding of the topographical and geophysical factors that influ­ then applied to LCP analyses of mobility between permanent settlements ence the route of a current transhumance pathway linking the Ebro and other points of settlement surrounding the PNAESM, obtaining 11 D. Garcia Casas and E. Gassiot Ballbè Quaternary International xxx (xxxx) xxx Fig. 11. LCP’s between the selected points surrounding the PNAESM and transhumant pathways. Permanent settlements and other evidences of permanent occu­ pation attested in the region where PNAESM is located (see Table 2). possible paths to cross the study area where the archaeological sites methodology is non-reconstructive: the aim is not to reconstruct the analysed are located. These are potential paths that could have been exact route, but to understand factors which provide information about used in the past and for which no documentation has been preserved, a pastoral movements in mountain zones, leading to the creation of normal fact given that until recent times the mention of the specific plausible possible pathways. On the other hand, other studies on LCPs in livestock routes is anecdotal and extremely imprecise. It is understood archaeology (Kantner, 2012) suggest that comparing the resulting that these possible routes are suitable for livestock movements similar to routes with the distribution of archaeological sites along them can those recorded today or in recent centuries. In our research a present- suggest that the possible pathway was indeed used. As explained above, day transhumant route was used as a basis to infer possible pathways the first group of sites in the Medieval era fulfil this condition. in the past. This approach implicitly assumes that the subjects who are Furthermore, it could be of interest to compare the outcome of the moving and subjective human behaviour have not changed significantly. network of pathways in this research with a previous study in the same It can be argued that pastoral movement can change depending on the area focused on the accessibility of Neolithic archaeological sites to different size of flocks, the animal species concerned, if there are preg­ grazing areas and possible roads (Gassiot Ballbè et al., 2020). In that nant animals in the flock, etc. These factors can affect the ability to cross paper, LCP pathways were performed using the Nunes algorithm adding areas of rough terrain and therefore affect the decision-making process extra cost only to water masses such as lakes. Despite the differences in when choosing one route or another. These arguments are true and must methodology, the degree of divergence in the routes obtained is low. In be taken into account. However, as has been argued above, the general, the routes cross the same valleys and mountain pass with only 12 D. Garcia Casas and E. Gassiot Ballbè Quaternary International xxx (xxxx) xxx Fig. 12. National Park of Aigüestortes and Estany de Sant Maurici (PNAESM). Map with LCP routes and cost distance to archaeological sites. two exceptions. Furthermore, it can be suggested that using a indicators suggest a significant increase in livestock activity in the study well-documented route as the basis of the analysis makes the perfor­ area (Gassiot Ballbè and Pèlachs Mañosa, 2017). The morphology of the mances strong. settlements also changed. A new type of settlement appeared with The third part of the research was devoted to analysing the spatial several huts and enclosures in the same location. In other publications relationships between possible communication routes and two groups of (Garcia Casas and Gassiot Ballbè, 2022), these have been called archaeological settlements from different time periods. The results “grouped settlements”. They seem to be settlements in which different indicate a historical change in the position of seasonal summer pastoral groups of shepherds with several flocks came together in the same place settlements between the Middle Ages and the 18th-20th centuries. These to spend at least part of the summer. Such settlements suggest a way of interpretations complement our knowledge of the archaeological record organising animal husbandry which was quite different to that of the of the PNAESM and the evolution of seasonal pastoralism from its pre­ transhumant shepherds of the previous centuries. In other publications, historic origins to the present day. The beginning of seasonal pasto­ two sub-groups of these settlement types have been mentioned accord­ ralism in the Central Pyrenees seems to be situated at the end of the ing to their chronology, along with certain differences in the internal Neolithic (Gassiot Ballbè et al., 2017). However, over the next two morphology of the architectural structures (Garcia Casas and Gassiot millennia, rather than a progressive evolution, discontinuities can be Ballbè, 2022). The first group (AA-024, AA-026, AA-043, ESP-024, observed in the social practices. NA-084, VB-085, VB-098, NA-026, ESP-015) is dated from the Late Another remarkable change took place in Late Antiquity. The num­ Roman Empire to the Late Antiquity (approximately between the 2nd ber of known archaeological sites increased enormously compared to the and 5th centuries) (Gassiot Ballbè, 2016). There are also differences in previous period and paleoenvironmental indicators indicate a general spatial positioning as the first group is concentrated in the northern part clearing of the forest to create pastureland. Both types of archaeological of the PNAESM in the Val d’Aran and the locations are clearly close to an 13 D. Garcia Casas and E. Gassiot Ballbè Quaternary International xxx (xxxx) xxx Table 4 seem to be located less than an hour from an LCP to cross the park, with Cost factors used to make the LCP’ the different locations with population evi­ only five exceptions (those located in the aforementioned valleys of dence surrounding the study zone. MSw = energetic expenditure (Kj/m), Sw = Llacs and Casesnoves and one settlement in the Valley of Cabanes, which terrain gradient. Based on (Parcero-Oubiña et al., 2019). is located far away from the rest). In these three valleys there are no Location Type First Notes References potential circulation routes. In this regard, its location pattern is similar insight to the Neolithic settlements mentioned above, close to optimal routes Ares Toponymy, that could have been the road network for livestock circulation in this mountain pass sector of the Pyrenees (Gassiot Ballbè et al., 2020). Arties Archaeological. III - IV Cots 2003, At the end of the Middle Ages another major change took place in the Roman stela and Century d. INVARQUE Roman necropolis C forms of occupation of the territory and the social organisation of Espitau Documentary. 1180 Bolós i pastoralism. The grouped settlements were abandoned, with the de Chapel, crossroad Hurtado 2012 archaeological sites formed over the following centuries presenting a Vielha different morphology and internal organisation of space. Modern and, Espot Documentary. 819 Ordeig (1993) especially, contemporary settlements are much more similar to the Church and fortification ethnographic image of traditional transhumance (Garcia Casas and Espui Documentary. 819 Bolos i Gassiot Ballbè, 2016). A previous study has suggested a significant Church and Hurtado discontinuity of pastoral practices from the end of the Middle Ages and fortification 2012; Ordeig the first centuries of the Modern Era (15th-17th centuries) (Garcia Casas (1993) and Gasiot Ballbè, 2022). This has been observed in changes to the Llessui Documentary. 852 Bolos i Church and Hurtado 2012 morphology of pastoral structures. In the present research, the discon­ settlement tinuity in space in the same chronology was observed. Both studies Senet Archaeological. III-IV INVARQUE provide information regarding significant changes in the Pyrenean Insights of Roman Century d. Landscape. The majority of the sample of the contemporary settlements Villa. C Suert Archaeological, From Evidences of INVARQUE, analysed is located more than an hour away from the optimal routes to different Bronze different type Cots, 2003 cross the PNAESM, an interpretation that suggests a significant change evidences. Age and chronology in the social organisation of movement and the role of communication (megalithic routes. structures, How can this be explained in historical terms? A difference was necropolis, houses …) observed in the distance to the archaeological sites from routes crossing Vall de Documentary. 1064 Vall de Boí Bolós and the PNAESM, showing a change in the social organisation of the land­ Boí Different churches includes the Hurtado, scape. What were the general historical mutations in the Pyrenean so­ and medieval medieval towns 2012, Ordeig cieties which led to this change? At this point, with the current state of towns and (1993) Vielha Documentary and Medieval. INVARQUE research on this issue, it is difficult to provide definitive explanations. It archaeological. Unclear. would be necessary to collect more archaeological data and cross them Medieval town with documentary sources. However, a hypothesis can be suggested based on our historical knowledge of socio-historical changes in the Modern Era in rural zones of north-western Catalonia (Ros, 2016). Be­ Table 5 tween the 16th and 17th centuries, written documentation shows Distance in minutes from each archaeological site to the nearest LCP and nearest numerous conflicts over grazing rights between local communities and cabanera (transhumant route). The table shows the results using each one of the feudal lords in the Pallars region (Bringue, 1996). The outcomes of these three algorythms LLO (Llobera), TO (Tobler) and N (Nunes). MED: Medieval conflicts were quite different depending on the village and the valley, sites. CONT: Contemporary sites. but in cases such as the Valls d’Aneu (in the north-east of the study area), Cost distance Total area % Cases Archaeological % Cases the communal councils gained the power to regulate all elements related (min) (ha) sites to seasonal livestock farming (Padilla Lapuente, 1999). In addition to 0–15 774.34 11.66 30 32.97 the ownership of many grazing areas converted into communal terri­ 15–30 410.75 6.18 14 15.38 tory, they gained the right to regulate the maximum number of cattle 30–45 394.27 5.94 8 8.79 that could graze, the timing of herds moving in and out, the entry of 45–60 425.46 6.41 11 12.09 foreign cattle and the areas where they could graze. The municipal 60–75 425.19 6.40 13 14.29 75–90 387.59 5.84 6 6.59 councils also controlled the payments and paid the corresponding taxes. 90–105 407.81 6.14 2 2.20 Disputes over grazing rights also took place between municipalities. 105–120 414.73 6.24 0 0.00 Numerous court rulings between the 15th and 17th centuries defined the 120–135 414.89 6.25 2 2.20 territories in which a community could send livestock to graze, be it its 135–150 361.39 5.44 0 0.00 own livestock or foreign livestock to which this right was leased. This division of the territory would give rise to the pletas, a traditional term to LCP route. Less than 20 min away, the second group (NA-091, NA-076, refer a mountain area with a refuge for the shepherds and grasslands to NA-090, SO-006, VB-001A, VB-014, VB-017sup, VB-019, VB-088) pre­ feed the flocks (Violant i Simorra, 2001). sents chronologies in the Middle Ages (between the 11th and 14th These delimitations may have had an effect on the location of the centuries). Their locations are more variable as they are spread over livestock settlements where huts or shelters for shepherds and enclo­ several areas of the PNAESM, some of them in the valleys of Llacs and sures were built. The decision-making was different in comparison with Casesnoves. However, not all of the archaeological sites dating from Late earlier times. It is known that in the Middle Ages the mountain grass­ Antiquity or the Medieval period correspond to this type of settlement. lands were owned by feudal lords and, especially, abbeys, as was the Settlements can also be found which consist of one or two huts with one case with Lavaix and Ovarra (Bolós and Hurtado, 2012). However, very or two enclosures (VB-098 and NA-026) or rock shelters (ESP-016 and little is known about the specific forms of pastoralism, who the herders VB-014). A description of the settlements can be found in the supple­ were and the technical work processes they employed. The documen­ mentary material. All of the Late Antique and Medieval settlement types tation reflects ownership and grazing rights, but not who carried out the seasonal herding or their ways of life (Garcia Casas et al., 2022). 14 D. Garcia Casas and E. Gassiot Ballbè Quaternary International xxx (xxxx) xxx Table 6 Outcomes of the statistical analysis of the results showed in Table 4. (Distance in minutes depending if the site is medieval o contemporary). Code Chronology LCP_LLO LCP_N LCP_TO CAB_LLO CAB_N CAB_T0 NA-091 MED 4.7 4.4 3.4 70.9 58.4 47.2 NA-076 MED 8.9 7.5 7.3 174.2 139.9 117.6 NA-084 MED 20.4 17.9 14.5 173.0 124.7 127.3 NA-079 MED 5.9 5.4 4.4 112.9 94.6 75.8 AA-026N MED 43.4 33.5 30.9 362.0 279.8 311.0 NA-090 MED 32.3 26.6 27.6 180.4 130.7 138.0 AA-024A MED 30.7 21.7 21.6 356.9 287.5 303.8 NA-026 MED 0.3 0.3 0.2 0.7 0.5 0.4 AA-043 MED 162.6 116.6 126.2 392.1 325.8 274.5 ESP-015 MED 11.5 10.1 8.0 314.6 265.9 224.6 ESP-024 MED 43.9 37.3 30.9 283.3 238.1 205.1 ESP-016 MED 1.7 1.5 1.2 302.5 260.8 209.5 VB-014 MED 4.2 3.4 3.1 171.2 143.9 133.5 VB-001A MED 59.9 53.8 41.2 235.1 199.6 171.2 VB-017 sup MED 87.2 75.8 59.5 260.8 219.3 189.3 VB-019 MED 151.9 113.6 114.4 225.8 178.1 188.8 VB-085 MED 123.9 103.1 87.4 299.9 249.5 218.3 VB-098 MED 172.9 132.7 127.8 246.9 197.3 201.9 SO-006 MED 86.5 75.7 58.3 90.3 77.0 58.8 VB-088A MED 2.1 2.0 1.5 271.6 234.3 185.5 NA-102 CONT 4.2 3.6 2.8 54.1 44.9 35.4 NA-054 CONT 31.3 24.0 22.2 64.5 52.3 44.8 AA-060 CONT 186.7 149.7 142.9 350.8 291.4 246.4 TC-009 CONT 26.4 18.8 25.2 354.8 280.1 344.3 VB-017inf CONT 75.0 65.9 50.9 248.0 209.3 180.9 VB-018sup CONT 144.9 108.1 109.9 218.8 172.3 184.0 VB-110 CONT 116.9 88.4 84.4 128.2 91.8 112.4 VB-111 CONT 166.4 125.9 123.5 240.5 190.5 197.6 VB-112 CONT 166.1 126.4 123.2 240.2 191.0 197.3 VB-086 CONT 124.9 105.4 87.7 302.5 252.6 217.7 VB-096 CONT 197.0 146.3 151.0 281.1 223.3 225.0 VB-004 CONT 167.0 144.6 113.3 336.8 269.7 243.0 SO-005 CONT 89.1 76.3 60.1 92.7 78.9 59.9 SO–008B CONT 39.0 34.5 26.2 42.2 36.1 27.1 SO-007A CONT 7.2 6.8 4.8 12.8 11.4 7.9 TDG-01 CONT 151.5 124.5 99.5 374.5 297.2 272.0 TDG-02 CONT 194.5 164.1 130.8 418.4 337.4 302.7 Moreover, in high mountain areas, the boundaries of each property are regard, after comparing our results with other studies in the same area vague and imprecise. It is possible that the feudal lords who owned the (Gassiot Ballbè et al., 2020) and in other high mountain zones of Europe mountain areas and their rights of use exercised little control over where (Carrer, 2013), it can be stated that there are no statistical differences shepherds settled in the high-altitude pastures. On the contrary, written regarding the location of the archaeological sites over time. It can be documentation of disputes between local communities in the following suggested that shepherds’ criteria for choosing a settlement location centuries shows that the high mountain areas were completely divided according to altitude or slope remain the same in different chronologies. up in modern times. Thus, in modern and contemporary times, in the Therefore, explanations must be sought in sociohistorical reasons and, decision-making process to build a new settlement, the priority would be more specifically, in changes in territorial organisation for the differ­ to settle on the clearly-defined plots of land on which a community had ences in the geographic location of the archaeological sites. It would be grazing rights and not near roads. Another reason suggested by a pre­ interesting to test this hypothesis in other high mountain areas where vious study (Gassiot Ballbè et al., 2020) could be the great availability of archaeological studies of livestock settlements have been carried out grass to feed animals near the site. This may be an explanatory hy­ over large time scales. pothesis that should be corroborated in future studies. However, beyond the specific historical causes, this research has 6. Conclusion demonstrated the existence of differences in the position of settlements according to chronology. These interpretations add to other research The study of past societies through the spatial dispersion of archae­ that warns that many of the elements of mountain landscapes attributed ological remains has been revealed as a successful trend which has by “common sense” to the Middle Ages have a modern origin. In this contributed new data and interpretative models. These studies have case, the spatial layout adds to these observations. Other research, both been applied in different environments as well as in prehistoric and in the same area and in other mountain areas, has already warned of the historic periods. Our research has focused on mobility related to sea­ danger of making direct analogies between modern transhumance and sonal pastoralism, analysing pastoral pathways as an archaeological the different forms of pastoralism that have been practised in the long element which can provide an insight into the location of different sites. sequence of time since the end of Prehistory (Garcia Casas, 2013; Garcia It is also interesting to note the differences between medieval sites and Casas and Gassiot Ballbè, 2022). Furthermore, the results of this study transhumant settlements. The mountain landscape, in modern and make it possible to affirm that the location of archaeological sites contemporary times, was, and still is, divided into plots and their cor­ relating to mountain livestock rearing does not depend merely on the responding grazing rights. On the contrary, the location of medieval biophysical elements of the landscape (such as altitudinal level, slope, sites near to axes of mobility demonstrates a different organisation of water sources, etc.) or on environmental determinisms. The upland territories. decision-making process regarding where to establish a pastoral settle­ These results complement other archaeological research in the ment was also influenced by social factors that changed over time. In this PNAESM which rejects a supposed continuity between prehistoric, 15 D. Garcia Casas and E. Gassiot Ballbè Quaternary International xxx (xxxx) xxx ancient and medieval livestock and traditional transhumance docu­ Arqueológico Del Territorio: Actas Del V Simposio Internacional de Arqueología de Mérida. CSIC, Mérida, pp. 383–394. mented by ethnographers in the early 20th century. Additionally, Bolós, J., Hurtado, V., 2012. Atles dels comtats de Pallars i Ribagorça. Dalmau, combined with other archaeological and paleoenvironmental studies, Barcelona. they help to gain an understanding of the long history of human occu­ Bringue, J.M., 1996. Comunitats i bens comunals al Pallars Sobira, segles XV- XVIII. pation of high-mountain Pyrenean landscapes with its different phases Butlleti de la Societat Catalana d’Estudis Historics 7, 115–128. Burri, S., Py, V., Cesarini, R., 2015. Moving up and down throughout the seasons: winter and discontinuities. In the present study, LCP analysis was employed. In and summer grazing between provence and southern Alps (France) AD 1100-1500. future, it could be interesting to evaluate the applicability of other In: Costello, E., Svensson, E. (Eds.), Historical Archaeologies of Transhumance across methods, such as focal mobility networks (Llobera et al., 2011) or cor­ Europe: [21st Annual Meeting of the European Association of Archaeologist ‘“Historical Transhumance in Europe”’ Glasgow, 2-5 September 2015], EAA ridors of movement (Pinto and Keitt, 2009), to this kind of studies. As a Monograph. Routledge, Glasgow, United Kingdom, pp. 135–154. final remark, we believe that the methodology employed here may also Carrer, F., 2013. An ethnoarchaeological inductive model for predicting archaeological be of interest if applied in other zones. Firstly, as a methodological tool site location: a case-study of pastoral settlement patterns in the Val di Fiemme and Val di Sole (Trentino, Italian Alps). J. Anthropol. Archaeol. 32, 54–62. https://doi. in zones with rough topography to analyse mobility, to obtain plausible org/10.1016/j.jaa.2012.10.001. potential routes and to explore the relationship with archaeological Carrer, F., Angelucci, D., 2017. Continuity and discontinuity in the history of upland sites. And finally, not only in other upland zones but also in pastoral landscapes: the case study of val molinac and val poré (val di Sole, trentino, eastern Italian Alps). Landsc. Res. 43, 862–877. https://doi.org/10.1080/ mid-mountain areas and plains where there is evidence of ancient 01426397.2017.1390078. transhumance, the results can improve our knowledge of the origins and Della Casa, P., 2013. Valleys, pastures and passes: new research issues from the Swiss development of seasonal pastoralism as well as the complementarity Central Alps. Preistoria Alp. 47, 17–30. Chapman, B., 1979. Transhumance and megalithic tombs in Iberia. Antiquity 53, between different altitudinal ranges and the role of these activities in the 150–151. https://doi.org/10.1017/S0003598X00109214. formation of present-day rural landscapes. Coromines, J., 1970. Estudis de toponímia catalana. Barcino, Barcelona. Costello, E., Svensson, E., 2018. Historical Archaeologies of Transhumance across Europe, Historical Archaeologies of Transhumance across Europe. https://doi.org/ Author contributions 10.4324/9781351213394. Cots, P., 2003. Los primeros pobladores: un paisaje en transformación. In: Esteban, A. David Garcia Casas: Conceptualization, Methodology, Data cura­ (Ed.), La Humanización de Las Altas Cuencas de La Garona y Las Nogueras (4.500 AC tion, Writing- Original draft preparation, Writing – review & editing, – 1955 DC). Organismo Autónomo Parques Nacionales, Madrid, pp. 43–143. Le Couedic, M., 2010. Les pratiques pastorales d’altitude dans une perspective Visualization. Ermengol Gassiot Ballbè: Direction of archaeological ethnoarchéologique. Cabanes, troupeaux et territoires pastoraux pyrénéens dans la research at PNAESM, field work and sites recording. Supervision, Re­ longue durée. Unpublished PhD dissertation, Université de Tours. sources, Writing – review & editing, Visualization. Le Couedic, M., Calastrenc, C., Rendu, C.H., Allée, P.H., 2016. A la recherche des sites. In: Rendu, Ch, Calastrenc, C., Le Couédic, M., Berdoy, A. (Eds.), Estives d’Ossau, 7000 ans de pastoralisme dans les Pyrénées. Le Pas d’oiseau, Toulouse, pp. 61–84. Funding de Cultura, Departament, 2022. Inventari del Patrimoni Arqueológic de Catalunya. http://invarque.cultura.gencat.cat/ (Accesed 14 June 2022). Departament d’Acció Climàtica, Alimentació i Agenda Rural, 2012. Camins Ramaders This research was supported by the Galician Agency of Innovation Classificats. https://agricultura.gencat.cat/ca/serveis/cartografia-sig/bases-cartogra (GAIN), part of the Government of Galicia that funds the postdoctoral fiques/ramaderia/camins-ramaders/. contract of David Garcia Casas (Reference number IN606B-2019/006) Departament de Medi Ambient y Agència Catalana de l’Aigua, 2015. Xarxa Hidrográfica. http://aca.gencat.cat/. (Accessed 9 June 2022). and the Spanish Ministry of Science and Innovation who founds the Fairen, S., Berrocal, M., Lopez-Romero, E., Walid, S., 2006. Las vías pecuarias como project “Las primeras comunidades agropastorales de la vertiente sur del elementos arqueológicos del paisaje. In: Grau, I. (Ed.), La Aplicación de Los SIG En Pirineo central: economia y paisaje (5600-4500 calANE)” leaded by Arqueología Del Paisaje. Servicio de Publicaciones de la Universidad de Alicante, Alicante, pp. 55–68. Ermengol Gasiot Ballbè (Reference number PID2020-115205 GB-I00 ). Fonte, J., Parcero-Oubiña, C., Costa-García, J., 2017. A GIS-based analysis of the rationale behind Roman roads. The case of the so-called via XVII (NW Iberian Peninsula). Mediterranean Archaeology and Archaeometry 17, 163–189. https:// Declaration of competing interest doi.org/10.5281/zenodo.1005562. Galán, E., Ruiz-Galvéz, M., 2001. Rutas ganaderas, transterminancia y caminos antiguos: The authors declare that they have no known competing financial el caso del Occidente peninsular entre el Calcolítico y la Edad del Hierro. In: Gómez- interests or personal relationships that could have appeared to influence Pantoja, J. (Ed.), Los Rebaños de Gerión: Pastores y Trashumancia En Iberia Antigua y Medieval. Casa de Velázquez, Madrid, pp. 263–278. the work reported in this paper. Galop, D., 2016. Évolutions paléo-environnementales en vallée d’Ossau, du Néolithique à l’Époque contemporain. In: Rendu, Ch, Calastrenc, C., Le Couédic, M., Berdoy, A. Acknowledgements (Eds.), Estives d’Ossau, 7000 ans de pastoralisme dans les Pyrénées. Le Pas d’Oiseau, Toulouse, pp. 161–165. Garcia Casas, D., 2013. Aproximación al poblamiento de las zonas de alta montaña The data used in this research come from the collective investigation pirenaicas desde la arqueologia y la etnografia. SAGVNTVM. Papeles del Laboratorio that the Grup d’Arqueologia de l’Alta Muntanya carried out in the de Arqueología de Valencia 45, 221–239. https://doi.org/10.7203/ SAGVNTVM.45.2006. PNAESM; we thank all its members and all the participants in the Garcia Casas, D., 2018. Arqueologia d’un territori d’alta muntanya del Pirineu Central. fieldwork on which this study is based. The authors are also grateful to Persones, ramats i prats al llarg de la història al Parc Nacional d’Aigüestortes i Ignasi Ros for provide ethnographic cartography about transhumant Estany de Sant Maurici. Universitat Autònoma de Barcelona. Unpublished PhD dissertation. Available online. https://www.educacion.gob.es/teseo/mostrarRef.do? routes. Finally, we would like to thanks Cesar Parcero Oubiña for his ref=1719432. help and advices to model GIS based pathways. Garcia Casas, D., Gassiot Ballbè, E., 2016. Los últimos siglos de ocupaciones pastoriles. In: Gassiot Ballbé, E. (Ed.), Montañas Humanizadas. Arqueología Del Pastoralismo In: El Parque Nacional d’Aigüestortes i Estany de Sant Maurici. Organismo Appendix A. Supplementary data Autónomo de Parques Nacionales, Madrid, pp. 219–237. Garcia Casas, D., Gassiot Ballbè, E., 2022. The archaeology of pastoralism in the central Supplementary data to this article can be found online at https://doi. Pyrenees: a diachronic analysis of livestock structures in Aigüestortes i Estany de Sant Maurici national park. J. Mediterr. Archaeol. 35, 5–31. https://doi.org/ org/10.1016/j.quaint.2023.07.007. 10.1558/jma.23767. Garcia Casas, D., Rodríguez-Antón, D., Costa Badia, X., Gassiot Ballbè, E., 2022. References Arqueología del paisaje en la alta montaña. Una primera aproximación al estudio de las ocupaciones ganaderas de época medieval en el Parque Nacional de Aigüestortes i Estany de Sant Maurici (Pirineo Occidental de Cataluña). Cuadernos de Adell, J., Riu Barrera, E., 1994. Formes d’urbanisme, implantació d’esglèsies i feudalisme Arqueología de la Universidad de Navarra 30 (2), 211–238. a les valls altes del Comtat de Pallars. In: Cultures i Medi, de La Prehistòria a l’edat Gassiot Ballbè, E., 2016. Montañas Humanizadas. Arqueología del pastoralismo en el Mitjana (Homenatge Al Prof. Dr. Jean Guilaine). Institut d’Estudis Ceretans, Parque Nacional d’Aigüestortes i Estany de Sant Maurici. Organismo Autónomo Puigcerdà, pp. 625–629. Parques Nacionales, Madrid. Bevan, A., 2011. Computational models for understanding movement and territory. In: Mayoral, V., Celestino, S. (Eds.), Tecnologías de Información Geográfica y Análisis 16 D. Garcia Casas and E. Gassiot Ballbè Quaternary International xxx (xxxx) xxx Gassiot Ballbè, E., Garcia Casas, D., 2014. Ovelles i pastures. Arqueologia dels darrers Padilla Lapuente, J.I., 1999. L’Esperit d’Aneu: Llibre Dels Costums I Ordinacions. Consell segles de ramaderia a l’alta muntanya. Afers: Fulls de recerca i pensament 78, Cultural de les Valls d’Aneu, Esterri d’Aneu. 451–470. Parcero-Oubiña, C., Güimil-Fariña, A., Fonte, J., Costa-García, J.M., 2019. In: Gassiot Ballbè, E., Pèlachs Mañosa, A., 2017. The husbandry occupation of the western Verhagen, P., Joyce, J., Groenhuijzen, M.R. (Eds.), Footprints and Cartwheels on a Pyrenees of Catalonia in Roman times and the beginning of the Middle Ages. Treballs Pixel Road: on the Applicability of GIS for the Modelling of Ancient (Roman) Routes arqueol. 21, 287–306. https://doi.org/10.5565/rev/tda.71. BT - Finding the Limits of the Limes: Modelling Demography, Economy and Gassiot Ballbè, E., Rodríguez Antón, D., Pélachs Mañosa, A., Pérez Obiol, R., Julià Transport on the Edge of the Roman Empire. Springer International Publishing, Brugués, R., Bal-Serin, M.-C., Mazzucco, N., 2014. La alta montaña durante la Cham, pp. 291–311. https://doi.org/10.1007/978-3-030-04576-0_14. Prehistoria: 10 años de investigación en el Pirineo catalán Occidental. Trab. Prehist. Pinto, N., Keitt, T.H., 2009. Beyond the least cost-path: evaluating corridor redundancy 71 (2), 261–281. using a graph theoretic approach. Landsc. Ecol. 24, 253–266. https://doi.org/ Gassiot Ballbè, E., Clemente Conte, I., Mazzucco, N., Garcia Casas, D., Obea Gómez, L., 10.1007/s10980-008-9303-y. Rodríguez Antón, D., 2016. Surface surveying in high mountain areas, is it possible? Polla, S., Verhagen, P., 2014. Computational approaches to the study of movement in Some methodological considerations. Quat. Int. 402, 35–45. https://doi.org/ archaeology, theory, practice and interpretation of factors and effects of long-term 10.1016/j.quaint.2015.09.103. landscape formation and transformation. Degruyter. Gassiot Ballbè, E., Mazzuco, N., Clemente Conte, I., Rodríguez Antón, D., Obea Reimer, P.J., Austin, W.E.N., Bard, E., Bayliss, A., Blackwell, P.G., Bronk Ramsey, C., Gómez, L., Quesada Carrasco, M., Díaz, 2017. The beginning of high mountain Butzin, M., Cheng, H., Edwards, R.L., Friedrich, M., Grootes, P.M., Guilderson, T.P., occupations in the Pyrenees. Human settlements and mobility from 18,000 cal BC to Hajdas, I., Heaton, T.J., Hogg, A.G., Hughen, K.A., Kromer, B., Manning, S.W., 2000 cal BC. In: Catalán, J., Ninot, J., Aniz, M. (Eds.), High Mountain Conservation Muscheler, R., Palmer, J.G., Pearson, C., van der Plicht, J., Reimer, R.W., Richards, in a Changing World. Springer International Publishing, 75–105. D.A., Scott, E.M., Southon, J.R., Turney, C.S.M., Wacker, L., Adolphi, F., Büntgen, U., Gassiot Ballbè, E., Garcia Casas, D., Nunes Alonso, J., Salvador Baigues, G., 2020. Capano, M., Fahrni, S.M., Fogtmann-Schulz, A., Friedrich, R., Köhler, P., Kudsk, S., Modelización de territorios ganaderos en la alta montaña al final del Neolítico: una Miyake, F., Olsen, J., Reinig, F., Sakamoto, M., Sookdeo, A., Talamo,S., 2020. The integración de análisis espacial e información etnográfica. Trab. Prehist. 77, 48–66. IntCal20 northern hemisphere radiocarbon Age calibration curve (0–55 cal kBP). https://doi.org/10.3989/tp.2020.12246. Radiocarbon 62, 725–757. doi:10.1017/RDC.2020.41. Gietl, R., Doneus, M., Fera, M., 2007. Cost distance analysis in an alpine environment: Rendu, C.H., 2006. Transhumance : prélude à l’histoire d’un mot voyageur. In: Lafont, P. comparison of different cost surface modules. In: Proceedings of the 35th Y. (Ed.), Transhumance et estivage en Occident des origines aux enjeux actuels. 26es International Conference on Computer Applications and Quantitative Methods in Journées Internationales d’Histoire de l’abbaye de Flaran, 9,10, 11 septembre 2004. Archaeology (CAA). Layers of Perception, Berlin, Germany. April 2–6, 2007. Berlin. Presses universitaires du Mirail, Toulouse, pp. 7–29. Güimil-Fariña, A., Parcero-Oubiña, C., 2015. “Dotting the joins”: a non-reconstructive Rendu, C., Calastrenc, C., Le Couédic, M. Le, Berdoy, A., 2016. Estives d’Ossau, 7000 ans use of Least Cost Paths to approach ancient roads. The case of the Roman roads in the de pastoralisme dans les Pyrénées. Le Pas d’oiseau, Toulouse. NW Iberian Peninsula. J. Archaeol. Sci. 54, 31–44. https://doi.org/10.1016/j. Ros, I., 2016. Notes per a una etnografia de la vida pastoral a cavall dels S. XVI-XVII, a la jas.2014.11.030. ciutat de Lleida. Unpublished paper presented at I Congres de Transhumancia i Herzog, I., 2014. A review of case studies in archaeological least-cost analysis. camins ramaders de Catalunya, Lleida, Spain. Archeologia e Calcolatori 25, 233–239. Saintenoy, Th, Mendez-Quiros, P., 2021. Movimiento e imperialismo en los valles de Herzog, I., 2017. Reconstructing pre-industrial long distance roads in a hilly region in Arica (Andes 18◦ S). Lat. Am. Antiq. 32, 1–19. https://doi.org/10.1017/ Germany, based on historical and archaeological data. Studies in Digital Heritage 1, laq.2020.100. 642–660. https://doi.org/10.14434/sdh.v1i2.23283. Saintenoy, T., Ajata, R., Romero Guevara, A.L., Sepulveda, M., 2017. Arqueología del Ingold, T., Vergunst, J.L., 2008. Ways of Walking: Ethnography and Practice on Foot, territorio aldeano prehispanico tardío en los altos de arica: aportes de la Anthropological Studies of Creativity and Perception. Ashgate, Aldershot. fotointerpretación satelital para el estudio regional de la cuenca alta de azapa. Estud. Institut Cartogràfic i Geològic de Catalunya, 2022. Model d’elevacions del terreny de 5x5 Atacameños 54, 85–110. m. https://www.icgc.cat/Descarregues/Elevacions/Model-d-elevacions-del-terreny Snead, J.E., Erickson, C.L., Darling, J.A., 2009. Making human space: the archaeology of -de-5x5-m (Accesed 9 June 2022). trails, paths and roads. In: Snead, J.E., Erickson, C.L., Darling, J.A. (Eds.), Jimeno, A., 2001. El modelo de trashumancia aplicado a la cultura de Cogotas I. In: Ruiz- Landscapes of Movement: Trails, Paths, and Roads in Anthropological Perspective. Gálvez, M. (Ed.), La Edad Del Bronce: ¿primera Edad de Oro de España? Crítica, University of Pennsylvania, Museum of Archaeology, and Anthropology, pp. 139–180. Philadelphia, pp. 1–19. Kantner, J., 2012. Realism, reality and routes. In: White, D., Surface-Evans, S. (Eds.), Tobler, W., 1993. Three presentations on geographical analysis and modelling. Technical Least Cost Analysis of Social Landscapes: Archaeological Case Studies. The Report 93, 1–25. Univeristy of Utah Press, Salt Lake City, pp. 225–238. Union, European, 2022. Copernicus Land Monitoring Service. European Environment Laborda, V., Villalba-Mouco, P., Lanau, M., Gisbert, M.S., Domingo, R., Montes, L., 2018. Agency (EEA). https://land.copernicus.eu/ (Accesed 6 June 2022). El Puerto Bajo de Góriz (Parque Nacional de Ordesa y Monte Perdido). Ocupación y VanPool, T.L., Leonard, R.D., 2011. Quantitative Analysis in Archaeology. John Wiley & explotación de un paisaje de alta montaña desde la prehistoria hasta el siglo XX, 26. Sons, Chichester. Bolskan: Revista de arqueología del Instituto de Estudios Altoaragoneses, pp. 9–30. Verhagen, P., 2018. In: Siart, C., Forbriger, M., Bubenzer, O. (Eds.), Spatial Analysis in Lanen, R.J. van, Kosian, M.C., Groenewoudt, B.J., Spek, T., Jansma, E., 2015. Best travel Archaeology: Moving into New Territories BT - Digital Geoarchaeology: New options: modelling Roman and early-medieval routes in The Netherlands using a Techniques for Interdisciplinary Human-Environmental Research. Springer multi-proxy approach. J. Archaeol. Sci.: Report 3, 144–159. https://doi.org/ International Publishing, Cham, pp. 11–25. https://doi.org/10.1007/978-3-319- 10.1016/j.jasrep.2015.05.024. 25316-9_2. Llobera, M., 2020. Memory at your feet: modelling the agency of past trails. J. Archaeol. Vilà Valentí, J., 1950. Una encuesta sobre la trashumancia en Cataluña. Pirineos 17–18, Sci.: Report 29, 102–177. https://doi.org/10.1016/j.jasrep.2019.102177. 405–445. Llobera, M., Sluckin, T.J., 2007. Zigzagging: theoretical insights on climbing strategies. Violant i Simorra, R., 2001. La Vida Pastoral Al Pallars. Garsineu, Tremp. J. Theor. Biol. 249, 206–217. https://doi.org/10.1016/j.jtbi.2007.07.020. Walsh, K., Court-Picon, M., Bealieu, J., Guiter, F., Mocci, F., Richer, S., Sinet, R., Llobera, M., Fábrega-Álvarez, P., Parcero Oubiña, C., 2011. Order in movement: a GIS Talon, B., Tzortzis, S., 2014. A historical ecology of the ecrins (southern French approach to accessibility. J. Archaeol. Sci. 38, 843–851. https://doi.org/10.1016/j. Alps): archaeology and palaeoecology of the mesolithic to the Medieval period. jas.2010.11.006. Quat. Int. 353, 52–73. Ordeig, R., 1993. Les dotalies de les esglésies de Catalunya (Segles IX-XII) Volum 2. White, D., 2015. The basics of least cost analysis for archaeological applications. Advan. Estudis històrics, Vic. Archaeol. Pract. 3, 407–414. Orengo, H.A., Palet, J.M., Ejarque, A., Miras, Y., Riera, S., 2014. Shifting occupation White, D., Surface-Evans, S., 2012. Least Cost Analysis of Social Landscapes: dynamics in the Madriu–Perafita–Claror valleys (Andorra) from the early Neolithic Archaeological Case Studies, Least Cost Analysis of Social Landscapes: to the Chalcolithic: the onset of high mountain cultural landscapes. Quat. Int. 353, Archaeological Case Studies. University of Utah Press, Utah. https://doi.org/ 140–152. https://doi.org/10.1016/j.quaint.2014.01.035. 10.7183/2326-3768.3.4.407. Web and cartographical references 17