A predictive model for Palaeolithic sites: a case study of Monforte de Lemos basin, NW Iberian Peninsula

Journal of Archaeological Science: Reports 49 (2023) 104012 Available online 22 April 2023 2352-409X/© 2023 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). A predictive model for Palaeolithic sites: A case study of Monforte de Lemos basin, NW Iberian Peninsula Mikel Díaz-Rodríguez a , b , c , d , * , Ram´ on F´ abregas-Valcarce a , d , Augusto P´ erez-Alberti e a Grupo de Estudos para a Prehistoria do Noroeste Ib´ erico, Arqueoloxía, Antigüidade e Territorio (GEPN-AAT), Dpto. de Historia, Universidade de Santiago de Compostela, Praza da Universidade, n◦1, 15782 Santiago de Compostela, Spain b Department of Archaeology and Heritage Studies. Aarhus University, Moesgård All´ e 20, 8270 Højbjerg, Denmark c BIOCHANGE – Center for Biodiversity Dynamics in a Changing World, Aarhus University, Aarhus, Denmark d CISPAC – Centro de Investigaci´ on Interuniversitario das Paisaxes Atl´ anticas, Edificio Font´ an, Cidade da Cultura de Galicia, Monte Gai´ as, s/n, 15707 Santiago de Compostela, Spain e Departamento de Edafoloxía e Química Agrícola. Facultade de Bioloxía, Universidade de Santiago Compostela, Spain ARTICLE INFO Keywords: Predictive model Palaeolithic Settlement Patterns Spatial analysis Iberian Peninsula Europe ABSTRACT Although a theoretical model for the settlement patterns of Galician Palaeolithic has been proposed in the last decades, it has not been statistically tested. The present paper aims to check whether this previous theoretical model can be verified statistically. For this purpose, a methodology based on the creation of a predictive model has been used in which the main environmental variables were analysed and their suitability for predicting the location of Palaeolithic sites statistically verified. The predictive model shows that the most accurate variables are elevation, slope, cost to potential hydrology, the cost to wetland areas, and visual prominence. The results demonstrated that the theoretical model was fulfilled in some of the variables previously proposed. Thus, we have shown the usefulness of this approach to test hypotheses and the results obtained open new possibilities of analysis in the study of the Palaeolithic sites in NW Iberia. 1. Introduction The study of settlement patterns provides a large amount of information to understand past societies and it has generated considerable interest in recent years. Currently, some approximations allow us to reconstruct past territories and analyse environmental variables that could be determining the settlement patterns in hunter-gatherer societies during the Palaeolithic period (e.g. Turrero et al., 2013; Burke et al., 2014; 2017; 2021b; García Moreno and Fano Martínez, 2014; Ludwig et al., 2018; Wren and Burke, 2019). This study is based on the reconstruction of past societies based on the analysis of those environmental variables that could determine the choice of places of occupation of hunter-gatherer societies during the Palaeolithic period. For this, there are a series of tools such as Geographic Information Systems (GIS) and spatial statistics, the combination of which facilitates the creation of a predictive model, therefore allowing the analysis and quantification of different variables related to the choice of a specific location by the societies of the past. However, this type of approach based on spatial analysis is practically non-existent in the Northwest of the Iberian Peninsula (de Lombera Hermida et al., 2015; Díaz Rodríguez, 2017; Díaz Rodríguez and Carrero Pazos, 2019; Díaz-Rodríguez et al., 2021; Díaz- Rodríguez and F´ abregas-Valcarce, 2022), except for some approximations applied to other chronologies and based on the application of locational patterns, quantitative modelling and predictive modelling (Llobera, 2015; Rodríguez Rell´ an and F´ abregas Valcarce, 2015; Carrero- Pazos, 2018; Rodríguez-Rell´ an and F´ abregas Valcarce, 2019; Carrero- Pazos et al., 2020). The main objective of this paper is to study the settlement patterns of the Palaeolithic sites from the Monforte de Lemos basin (NW Iberia). We have decided to choose this zone because it is an area that has been intensively studied in the last two decades and it is one of the districts with the highest density of archaeological sites of this chronology in that part of Iberia. As mentioned above, there are some previous approaches based on the application of GIS, but a study based on the application of predictive modelling has never been carried out in this region. We believe that it is a good opportunity to apply this methodology since it * Corresponding author at: Facultade de Xeografía e Historia, Dpto. de Historia, Universidade de Santiago de Compostela, Praza da Universidade, n ◦1, 15782 Santiago de Compostela, Spain. E-mail address: [email protected] (M. Díaz-Rodríguez). Contents lists available at ScienceDirect Journal of Archaeological Science: Reports journal homepage: www.elsevier.com/locate/jasrep https://doi.org/10.1016/j.jasrep.2023.104012 Received 7 November 2022; Received in revised form 7 April 2023; Accepted 12 April 2023 Journal of Archaeological Science: Reports 49 (2023) 104012 2 can provide relevant results. In the present work we hope to be able to identify some of the predictive variables of the occupation of these sites. However, it is very likely that the resulting variables do not match the predictive variables identified in other regions, since each area has its own characteristics shaping the patterns of occupation. In this case we have an inland basin as opposed to a coastal area (García Moreno, 2013; Garate et al., 2020; Parow-Souchon et al., 2021), in a mountainous region (Leloch et al., 2022). Also, other studies carried out on a larger scale Fig. 1. (a) Location of the region studied (in red). (b) Study area with the archaeological sites (red dots) (c) Geomorphological and geological surface of the study area (modified from de Lombera Hermida et al., 2015). (d) Longitudinal profile of the Monforte de Lemos basin (modified from de Lombera Hermida et al., 2015). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) M. Díaz-Rodríguez et al. Journal of Archaeological Science: Reports 49 (2023) 104012 3 are bound to yield different results (Burke et al., 2017; 2021a; Wren and Burke, 2019; Jochim, 2022). In other words, in coastal areas the altitude of the archaeological sites will be lower than in the Monforte de Lemos basin; archaeological sites, moreover, will be nearer to the shore and to the hydrological resources (García Moreno, 2010). In the present study, settlement patterns are analysed based on the creation of a predictive model. That model has been obtained using the different variables that had been previously proposed for the theoretical model established by different researchers. Predictive modelling can be defined as a technique that foresees the location of archaeological sites in a region (Kholer and Parker, 1986). A basic premise of predictive modelling is that human spatial behaviour is to a large extent predictable (Verhagen, 2018). For this reason, we believe that it is possible to identify those environmental variables that could be behind the decision of hunter-gatherer societies when choosing a given location for settlement in that region. We think that some variables, previously established in the theoretical model, would be important when choosing the place of settlement. Identifying these variables allows us to understand the hunter-gatherer ways of interacting with the territory. 2. Regional setting The study area analysed in this paper is the Monforte de Lemos basin, which is located south of the province of Lugo (Galicia, Spain) in the Northwest of the Iberian Peninsula (Fig. 1a). This area is placed east of the Mi˜ no river and north of its main tributary, the Sil river (Fig. 1b) and between two areas with high altitude, how are the Chantada’s Surface at the west (600 m.a.s.l.) and O Courel Mountain Range at the east (1600 m.a.s.l.). Another important river course of the basin is the Cabe River, the main fluvial course of the basin. In the Monforte de Lemos basin the transition of endorheic to exoreic drainage and the development of transcontinental drainage to the Atlantic Ocean is similar to the one proposed for the genesis of the Douro and the Tejo/Tajo (Tagus) rivers, involving as main mechanism an overspill induced by a major climatic change of increasing humidity by middle Pliocene (Cunha and P´ erez- Alberti, 2022). The surface of the basin corresponds to the whole of the council of Monforte and to part of those that border it, such as O Savi˜ nao, Pant´ on, Sober, Pobra do Broll´ on and B´ oveda. The Sil and tributaries mainly run cross the basin area, which mainly consists of Palaeozoic metamorphic rocks with minor granites, that are intensely faulted with WNW-ESE direction. After an episode of neotectonics and a subsequent fluvial reorganization, Pleistocene sediments linked to paleochannels, and alluvial fans covered the margins of silts and tertiary clays in a lake environment. These quaternary deposits, arranged in a sequence of flat surfaces, are identified as fluvial terraces, glacis, and pediments (Ameijenda Iglesias, 2011). The Monforte de Lemos basin presents certain peculiarities that make it impossible to reconstruct a relative sequence of topographic levels that provides a referential framework for the lithic assemblages located on its surfaces, as has been obtained in some of the main fluvial basins of the Iberian Peninsula (Santonja and P´ erez-Gonz´ alez, 2000–2001; Cano et al., 1997). One of the problems is the sparse character of the Quaternary fluvial terrace levels (Middle Cabe), that are asymmetrically distributed in the valley, being concentrated in the northern sector (Fig. 1d). Another problem consists of the large extension of pre-Quaternary surfaces that makes it possible for several lithic dispersions ascribed to different technological Modes to be found on the same surface. Finally, there is also an important incidence of the morphogenetic processes in the deposits of the Monforte de Lemos basin (Ameijenda Iglesias et al., 2010). In some previous studies, a preliminary sequence of different levels of erosion was established that brought together the surfaces of the Monforte basin (Ameijenda Iglesias, 2011). A geomorphological analysis has also been carried out, which has allowed the identification of the different geomorphological units of the basin in greater detail (de Lombera Hermida et al., 2015). These works have allowed us to verify that the opening of the basin brought with it the progressive disarticulation of the existing fluvial network in relation to a post-alpine tectonic reactivation and linked to the existence of tropical climatic conditions. The Mi˜ no and Sil rivers were “expelled” to the west (Mi˜ no) and south (Sil) progressively fitting into the terrain favoured by antecedent processes (P´ erez Alberti et al., 1993). The result of the interaction between tectonic dynamics and the progressive change in fluvial dynamics has resulted in a set of staggered levels in the territory (Fig. 1c). Among them, due to their archaeological importance, we must highlight the levels associated with Paleo Mi˜ no, the waters that flowed through where the basin is today during a phase prior to the embedding of the Mi˜ no-Sil system, and on the other hand, the Cabe River, that would drain the basin later. The recent studies had make possible the identification of several evolutionary stages in the basin area (Cunha and P´ erez-Alberti, 2022): (1) During the Paleogene and Miocene, endorheic sedimentation occurred. (2) During the latemost Miocene to Zanclean, probably ca. 9.7 to 3.7 Ma, the climax of lithosphere compression created most of the modern relief; sedimentation only occurred locally at piedmonts, as heterometric alluvial fans. (3) The transition of endorheic to exorheic drainage to the Atlantic Ocean could be similar to the one proposed for the genesis of the Douro and the Tejo/Tajo (Tagus) rivers, involving as main mechanism an overspill induced by a major climatic change of increasing humidity by middle Pliocene. 4) Probably during the last 1.8 Ma, the stage of fluvial incision has been responsible for the development of terrace staircases, valley entrenchment and captures. The relationship of the sites with the different geomorphological surfaces of the basin allows us to observe that a large number of archaeological remains are found mainly on the alluvial fans identified in the south of the basin (n =19), on Palaeozoic or Tertiary substrata with little development of the Quaternary landfills (n =34), the fluvial/ alluvial surfaces of the Paleo Mi˜ no (n =12), the fluvial formations related to the Middle section of the Cabe River (n =6) and a few are located in the current alluvial plain (n =5). The first Palaeolithic reference to the Monforte basin was a quartzite handaxe found in the middle XX century at Vilaescura (Cano Pan and V´ azquez Varela, 1991). The systematic investigation of this area began in 2006 in the frame of two concatenated research projects that were carried out and developed between 2006 and 2010, making possible the location of more than a hundred archaeological sites ascribed to the Palaeolithic and revealing a long human occupation during that period in this region (de Lombera Hermida et al., 2008, 2006; F´ abregas Valcarce et al., 2010, 2009, 2008, 2007; Rodríguez ´ Alvarez et al., 2008). Along the survey campaigns, a total of 16 km 2 was reviewed, representing 9.14 % of the total extension of the Monforte de Lemos basin (Fig. 1b), a common coverage when surveying large regions (Díez Martín, 2000). Through Landsat images it has been observed that a dense vegetation cover or forest accounts for 37.23 % of the surface, pasture areas represent 29.4 %, while open land or land dedicated to agriculture accounts for 6.4 % and it is on the latter that the works have focused (Miller et al., 2010). Along with the surveying, archaeological excavations were undertaken at some sites (de Lombera-Hermida et al., 2011; Rodríguez ´ Alvarez et al., 2014). After the discovery of lithic materials in the area, two research projects were proposed, as mentioned above. During the execution of both projects, some 104 archaeological sites from the Palaeolithic period were discovered. In the study area, there are few works framed in the Palaeolithic and carried out from spatial analysis and GIS. We can highlight a paper that discusses large-scale territorial mobility and attempts to define the routes of entry to the Northwest of the Iberian Peninsula (de Lombera- Hermida et al., 2011). Other articles analyse isolated variables such as visibility or mobility in the Valverde site, framed in the Upper Palaeolithic (Rodríguez ´ Alvarez et al., 2008; de Lombera Hermida et al., 2012). Another work has been carried out that deals with the analysis of M. Díaz-Rodríguez et al. Journal of Archaeological Science: Reports 49 (2023) 104012 4 settlement patterns through the study of variables such as visibility, altitude, aspect, slope, proximity to river courses and least-cost paths (de Lombera Hermida et al., 2015). It is a preliminary study based on the descriptive analysis of the mentioned variables, without applying analyses based on spatial statistics that could provide more accurate and precise information. Finally, a study analysing the occupation patterns of Lower Palaeolithic sites in the Monforte de Lemos basin through the application of descriptive statistical methods was carried out (Díaz Rodríguez et al., 2021). 3. Material and methods 3.1. Data acquisition and software In the present study we made use of the information and data obtained from the two research projects previously mentioned. We have not established a chronological differentiation among the archaeological sites, and we have decided to create a single model that encompasses them all. We lack chronological data since we have a reduced number of sites and that prevents a statistical treatment of a minimum quality. During the execution of these projects and with the objective of identifying and documenting archaeological sites during the survey, the study area was divided into four theoretically defined sectors, considering their geomorphological characteristics: the northern sector (where the main Plio-quaternary formations are concentrated), the eastern and southern sectors (in which the alluvial fans are found) and finally the central sector of the sub-basin. The latter could not be thoroughly surveyed because the present urban nucleus of Monforte de Lemos is located there. The constructions of this nucleus prevented the review of large areas of land likely to contain material remains. But despite this, some finds of lithic material were reported during building works and probably many others have been destroyed in that way (de Lombera Hermida et al., 2015). The archaeological surveys were carried out in several campaigns from 2006 to 2010. These focused on arable land and land clearance areas where soil visibility was good. These tasks were nearly impossible to undertake in densely vegetated areas that, alas, make up a large part of the basin. The survey tasks were carried out by groups of between 5 and 9 people who inspected land plots smaller than 1 ha, with surveyors keeping a distance of 3–7 m from each other, walking along transects. In the larger plots, the separation between transects was less than 15 m. The plot was used as a registration unit, identifying in it the number of artefacts. The cadastral registry was used to calculate the extension of the plots and georeferencing them on the surveying maps. The coordinates were taken at the centre point of the dispersions using a Trimble GEO Explorer 2005 (GeoXT) GPS with submeter precision. In addition to the surveys, archaeological excavations were carried out in those places where the density of artefacts could indicate the existence of sites in which the stratigraphic context was preserved (de Lombera Hermida et al., 2015). The total number of artefacts recovered in the Monforte de Lemos sites amounts to 3522, although a large part of them come from the Valverde excavations (n =2037, objects recovered in excavation) (Fig. 2a). The distribution of the number of artefacts per site can be observed in the Fig. 2b. There is great homogeneity in terms of the technological and rolling characteristics of the recovered lithic pieces. Although more than 100 sites were located, in this study, we will only use 76. These are those archaeological sites for which we have more accurate information based on the stratigraphic position of the pieces at the time of collection or the bearing level of the lithic industry. In some cases, we decided to exclude some points because these are very close to each other; otherwise, they would cause an overrepresentation within the sample, since there would be two sites in the same cell of the raster map. For these reasons, we have included sites that meet the criteria mentioned above and this includes some isolated findings considered. Based on the morpho-technical analysis of the lithic assemblages, 21 sites were assigned to Mode 2, 17 to Mode 3, 9 to Mode 4 and 29 present scarce and non-diagnostic lithic assemblages (defined as indeterminate), many of them correspond to isolated findings (1–4 artefacts) and very dispersed (F´ abregas Valcarce et al., 2007, 2008, 2009, 2010, 2011; Rodríguez ´ Alvarez et al., 2008; Rodríguez ´ Alvarez et al., 2014; de Lombera-Hermida et al., 2011; de Lombera Hermida et al., 2012). The treatment of spatial data has been carried out with different software that allows GIS analysis. The coordinate system used is EPSG: 25,829 (ETRS89 / UTM zone 29 N). GRASS GIS has been used in versions 6.4.3, 7.0.2 and 7.0.4 (Grass Development Team, 2020). Quantum GIS (versions 2.8.1 and 2.10.1) (QGIS.org, 2021) and SAGA GIS (version 2.2.1) (Conrad et al., 2015) have also been used. The latest GIS software used has been ArcGIS 10.3 (USC license) (Esri, 2011). It is the only one that does not share the GNU-GPI license. Finally, to carry out the different analytical approaches, R version 4.0.5 was used, with the R Studio graphical interface (R Core Team, 2021) and the packages required to run the analysis (Table 1). The Digital Elevation Model (DEM) has been used as a base map to elaborate the different locational variables analysed in this paper. This DEM has been obtained from the National Centre for Geographic Information (CNIG) and has a resolution of 25 m. It is cartography that collects the information obtained from the photogrammetric and LiDAR flights of the National Plan for Aerial Orthophotography (PNOA) Fig. 2. (a) Violin plot that shows the number of artefacts per site (red dots). (b) Violin plot that shows the number of artefacts per site (red dots) except for Valverde site. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) M. Díaz-Rodríguez et al. Journal of Archaeological Science: Reports 49 (2023) 104012 5 (https://centrodedescargas.cnig.es/CentroDescargas/index.jsp). Also, the Geologic Map has been used and obtained from the Spanish Government’s online repository (L´ opez Olmedo et al., 2022). 3.2. Spatial distribution of sites The first step in carrying out the spatial analysis of the archaeological sites of the Monforte de Lemos basin consisted of checking whether Complete Spatial Randomness (CSR) exists. For this, a sample of random points was created, with the same number of points as the archaeological sample, with the objective to compare both samples. The next step was to observe whether the distribution of both data sets belongs to the same population or not, which would indicate that there are no differences between the two samples and therefore we could not reject the CSR. For this, we have chosen the variable UTM X, corresponding to the x-coordinate of each point, and we have compared it between both samples. The Shapiro-Wilk test was used to test normality, and the Kolmogorov-Smirnov (K-S) test was used to check if both samples belong to the same population. Similarly, Ripley’s K functions and their L and G variants were used (Bivand et al., 2013). The homogeneous and inhomogeneous K, L and G functions were calculated (Baddeley et al., 2015) using a confidence interval based on Monte Carlo simulations (n =99). 3.3. Definition of covariates Intending to elaborate the statistical analysis, we have selected some covariates to establish the theoretical model based on previous works carried out in the study area and other similar areas. In the next lines, we will define the covariates used (Table 2 and Fig. 3). The process of obtaining each covariate is explained in greater detail in the SI file. Altitude (ALT) can be defined as the elevation calculated based on some reference datum. Usually, it is the sea level (if it refers to the Table 1 Synthesis of R packages used, authors and application details. Package Author/s Application Details dismo Hijmans et al., (2017) Methods for species distribution modelling. geostatsp Brown, (2015) Geostatistical Modelling with Likelihood and Bayes. GGally Schloerke et al., (2021) This package is a plotting system based on the grammar of graphics. ggplot2 Wickham et al., (2020) Package for creating graphics. maps Becker et al., (2021) Display of maps. maptools Bivand et al. (2020a) Set of tools for manipulating geographic data. MASS Ripley et al., (2020) Functions and datasets to support Venables and Ripley. patchwork Pedersen, (2022) Package for combining multiple plots. plyr Wickham, (2020) Set of tools that solves problems relates with applying or combining data. raster Hijmans et al., (2020) Reading, writing, manipulating, analysing, and modelling of spatial data. readxl Wickham et al., (2019a) Package for read excel files. rgdal Bivand et al. (2020b) Provides bindings to the “GDAL” and “PROJ” library. rgeos Bivand et al. (2020c) Package for topology operations on geometries. sp Pebesma et al. (2020) Classes and methods for spatial data. spatstat Baddeley et al. (2020) Toolbox for analysing Spatial Point Patterns. tidyverse Wickham et al. (2019) Data representations and API design. vioplot Adler and Kelly (2022) This package allows extensive customisation of violin plots. Table 2 Conditioning type, variables, acronym, description and ID number. Conditioning type Variables Acronym Description ID Number Abiotic Altitude ALT Elevation at a given point, in meters, above sea level. It is calculated from a DEM. v1 Abiotic Topographic Prominence Index TPI100 Calculation of the TPI that consists of comparing the elevation of each of the cells of the DEM with the average of the surrounding elevations. Calculated for 100 m radii. v15 Abiotic Topographic Prominence Index TPI500 Calculation of the TPI that consists of comparing the elevation of each of the cells of the DEM with the average of the surrounding elevations. Calculated for 500 m radii. v16 Abiotic Topographic Prominence Index TPI1000 Calculation of the TPI that consists of comparing the elevation of each of the cells of the DEM with the average of the surrounding elevations. Calculated for 1000 m radii. v17 Abiotic Slope SLO Slope of the ground at a given point. It is calculated from the DEM. v13 Abiotic Aspect ASP Aspect of the ground at a given point. It is calculated from the DEM. v12 Abiotic Cost to potential hydrology HYDROC Distance, in time, at a given point to the potential hydrology. It is calculated from the DEM. v5 Abiotic Euclidean distance to potential hydrology HYDROE Distance, in meters, at a given point to the potential hydrology. It is calculated from the DEM. v8 Abiotic Cost to wetland areas WET Water accumulation in a conjoin of points. It is measured in displacement cost time. It is calculated from the DEM. v2 Abiotic Cost to potential geology GEOLC Distance, in time, at a given point to the potential geology. It is calculated from the DEM. v4 Abiotic Euclidean distance to potential geology GEOLE Distance, in meters, at a given point to the potential geology. It is calculated from the DEM. v7 (continued on next page) M. Díaz-Rodríguez et al. Journal of Archaeological Science: Reports 49 (2023) 104012 6 absolute altitude) or, for example, the bottom of a valley (if it is relative altitude). This covariate is mentioned in the specific bibliography since it is believed that archaeological sites are located at high points of the landscape (de Lombera Hermida et al., 2015; F´ abregas Valcarce et al., 2010; Ramil Rego, 1989/1990, p. 194). Palaeolithic sites have been considered landmarks in the landscape. We would be facing reference points that would stand out from the surrounding terrain and would be visible or observable from a certain distance (F´ abregas Valcarce and de Lombera Hermida, 2010). To model this covariate, the topographic prominence has been calculated, which can be obtained using different methodological approaches. In this case, we use the Topographic Prominence Index (TPI). Following the recommendations of specialized researchers (Nakoinz and Knitter, 2016) and our previous experience (Díaz Rodríguez and Carrero Pazos, 2019), the TPI was calculated for 3 different radii (100, 500 and 1000 m) (TPI100, TPI500 and TPI1000). The slope (SLO) can be defined as the maximum degree of elevation variation at a given position. It is obtained from the DEM. This is a covariate that has been taken into account in some previous works on the Galician Palaeolithic (de Lombera Hermida et al., 2015, p. 280) or other Iberian regions like the Cantabrian (García Moreno, 2010), Asturian (Fern´ andez Fern´ andez, 2010) and in the Sierra de Atapuerca (Marcos S´ aiz, 2006). Aspect (ASP) has been defined as an important covariate to find the location of archaeological sites (de Lombera Hermida et al., 2015, p. 289). In some studies, it has been described that the majority of sites are oriented toward the second and third quadrants (Ramil Rego and Ramil Soneira, 1996). This covariate has been obtained from the DEM. The relation between the palaeolithic sites and the river courses has been defended in previous studies (F´ abregas Valcarce and de Lombera Hermida, 2010; Ramil Rego, 1989/1990, p. 193; Villar Quinteiro, 1996). The currently hydrological map presents actualisms because of human anthropization and the passage of time in the landscape. To achieve a more approximate image of what existed in the past, we have decided to create our hydrographic network based on the DEM and use a methodology that has been applied in previous works (García García, 2015). The proximity of the sites to the potential hydrology was calculated from all the points of the study area to the close water course. It has been measured in the distance (HYDROE) and displacement cost time (HYDROC). Although, in previous studies, it has been taken into account the proximity of wetland areas and the visual control over these areas (Criado Boado et al., 1991; L´ opez Cordeiro, 2002; 2015; de Lombera Hermida et al., 2015). The wetland areas could be defined as the accumulation of water in a conjoin of points. For this, the SAGA GIS software has been used, and more specifically the Topographic Wetness Index (TWI), which indicates the topographic humidity index in each of the cells of the map used. Once this map was obtained and given that we were interested in those areas in which this humidity index is higher, we proceeded to calculate the quartiles. In this way, we are left with the cells of the map with values above the third quartile. On the other hand, we must bear in mind that the calculation of the TWI is going to attribute a very high value to the cells in which a river coincides, but we are not interested in those cells since they would be falsifying the data. So, we subtracted the hydrological map, previously created, from the TWI polygon map. Thus, we manage to stay with those higher humidity values in which the rivers are not found. Then, we calculated the displacement cost time from every point of the study area to the close wetland areas divided into points (WET). For the Palaeolithic hunter-gatherers, it was important to have raw materials such as quartzite or quartz in the vicinity. These resources could be obtained from river courses, in the form of pebbles located on river beaches, or collecting raw materials from the veins. We have defined that as potential geology which has been mentioned in previous studies (Ramil Rego, 1989/1990; Villar Quinteiro, 1996; L´ opez Cordeiro, 2002; 2015; F´ abregas Valcarce and de Lombera Hermida, 2010; de Lombera Hermida et al., 2012). We also considered the potential geology to carry out the analyses. For this purpose, we have used the data obtained from the Mining Geological Institute (IGME). Those areas that could contain raw materials of interest to Palaeolithic huntergatherers have been selected and divided into points at established radii. Subsequently, the cost of moving, in time (GEOLC) and distance (GEOLE), from the rest of the cells in the study area to the closest potential geology points has been calculated. Within the biotic conditioning, we used the variable cost to potential hunting areas. We employed the Central Place Foraging Prey Choice (CPFPC). This model was proposed by M. Cannon (2003) and is based on the theory of foraging. It was used by Marín Arroyo to study the patterns of mobility and control of the territory in eastern Cantabrian. For that purpose, Cannon’s model was applied to deer hunting and goats as the most representative species of the Magdalenian diet (Marín Arroyo, 2008; 2009). In the present study, we have used the potential areas of hunting goats and obtained a covariate based on the cost, in time, from any point of the study area to these potential hunting areas (CPFPCG). In order to find these, we have calculated the 1.2-hour isochrones from the sites and we have used the slope map to select, within the isochrones, those cells with slopes greater than 30◦. Other conditioning types have also been considered. One of them is visibility, which has been contemplated to define the occupation of archaeological sites in previous works (L´ opez Cordeiro, 2002; 2004; 2015; Rodríguez ´ Alvarez et al., 2008; F´ abregas Valcarce and de Lombera Hermida, 2010; de Lombera-Hermida et al. 2011; de Lombera Hermida et al., 2015). In this case, we have used the analysis of visual prominence. This consists of calculating the points visible from each of the Table 2 (continued) Conditioning type Variables Acronym Description ID Number Biotic Cost to potential goat hunting areas CPFPCG Distance, in time, at a given point to the potential goat hunting areas. It is calculated from the DEM, based in slope and CPFPC model. v3 Other Visual Prominence VISPR Visible number of points, in each cell, from any of the points selected. It is calculated from the DEM. v14 Other Cost to Least cost path LCPC Potential Least cost paths between given points. It is calculated from the DEM and considering the slope and hydrology. v6 Other Index of potential direct insolation DIRINS Calculation of potential incoming direct insolation. It is obtained from the DEM. v10 Other Index of potential diffuse insolation DIFINS Calculation of potential incoming diffuse insolation. It is obtained from the DEM. v9 Other Index of potential total insolation TOTINS Calculation of potential incoming total insolation. It is obtained from the DEM. v11 Other Wind Exposition Index WIND Calculation of Wind Exposition Index. It is obtained from the DEM. v18 M. Díaz-Rodríguez et al. Journal of Archaeological Science: Reports 49 (2023) 104012 7 given points. For this, an observer height of 1.75 m has been used. Once the calculation is done, a raster file is obtained where the cells have a value that shows the number of cells visible from each cell (VISPR). The calculation of potential least cost paths (LCP) will be measured as the relationship that exists between the sites and the movement through the surrounding landscape. We could define it as the transit route generated between two points, depending on various factors and corresponding to the lower cost in energy or time. The natural transit routes or the so-called royal roads were treated in the bibliography. Establishing these routes as one of the variables that mark the pattern of location of the Palaeolithic sites in the Galician region (Ramil Rego and Ramil Soneira, 1996, p. 125; F´ abregas Valcarce and de Lombera Hermida, 2010, p. 267; de Lombera Hermida et al., 2015, p. 289; L´ opez Cordeiro, 2015, p. 301; Díaz Rodríguez, 2017). This consists of carrying out a calculation of transit routes in a specific area, without considering the archaeological sites. For this, it is necessary to have a starting and arrival point. In response to this, we were inspired by a previous work that used a methodology based on the calculation of optimal routes between all points of the landscape, called FETE (From Everywhere to Everywhere) (White and Barber, 2012). In this work, an analysis was employed that uses all the points of a grid as starting points and at the same time as arrival points, in such a way that the calculation allows representing the territory covering everything and creating the least cost path. However, that requires computer equipment that is powerful enough to run those analyses. Therefore, we have decided to adapt this model following the methodology used in another study (Rodríguez Rell´ an and F´ abregas Valcarce, 2015). It is a simplification that has been previously described in more detail (Díaz Rodríguez, 2017; Díaz- Rodríguez et al., 2021), but briefly, it consists in dividing the border of the study area into points at a certain established radius between them and calculating the LCP between them using a point as a starting point and the rest as stopping points and repeating the analysis with all the points (LCPC). In some previous studies, it has been mentioned that the archaeological sites of the Galician Palaeolithic could be located on slopes facing west to better take advantage of the calorific value of the sun’s rays (Ramil Rego and Ramil Soneira, 1996, p. 125). In addition, it seems logical to think that insolation could have played an important role in the occupation of an archaeological site. However, it has not been openly considered in the bibliography referring to this area, but it has Fig. 3. Covariates analysed in the present study. (a) ALT covariate. (b) TPI100 covariate. (c) TPI500 covariate. (d) TPI1000 covariate. (e) SLO covariate. (f) ASP covariate. (g) HYDROC covariate. (h) HYDROE covariate. (i) WET covariate. (j) GEOLC covariate. (k) GEOLE covariate. (l) CPFPCG covariate. (m) VISPR covariate. (n) LCPC covariate. (o) DIRINS covariate. (p) DIFINS covariate. (q) TOTINS covariate. (r) WIND covariate. M. Díaz-Rodríguez et al. Journal of Archaeological Science: Reports 49 (2023) 104012 8 been considered in other regions of the Iberian Peninsula. Some studies have analysed its influence when choosing a place of occupation, as occurs in the area of the As´ on River Valley, in Cantabria (García Moreno, 2008; 2015). In our study, the insolation has been obtained from SAGA GIS, through the Potential Incoming Solar Radiation module (Conrad et al., 2015) and calculated in three different ways (DIRINS, DIFINS and TOTINS). The last of the variables used in the analysis is that of the prevailing winds (WIND). Shelter from prevailing winds may be another variable to take into consideration when hunter-gatherer societies choose their places of occupation (Villar Quinteiro, 1996; García Moreno, 2010; de Lombera Hermida et al., 2015, p. 290). As to quantify this variable, it has been obtained the wind exposition index through the Wind Effect Index from the SAGA GIS module (B¨ ohner and Antoni´ c, 2009; Conrad et al., 2015). Table 3 Results for the statistical tests applied. Shapiro-Wilk Test (sites) Shapiro-Wilk Test (random sites) K-S Test (sites vs random sites) W p-value W p-value D p-value 0.87411 2.188 e-06 0.95329 0.007074 0.36807 7.229 e-05 Fig. 4. K, L and G Functions (a-f). M. Díaz-Rodríguez et al. Journal of Archaeological Science: Reports 49 (2023) 104012 9 3.4. Predictive model The predictive model can be considered one of the first tools in the GIS applications to archaeology (Church et al., 2000; Vermeulen, 2001). It is a method that allows the prediction of the value or the probability of presence of a dependent covariate in a place using one or more independent covariates. Predictive models are defined as tools to project known patterns or relationships into unknown times or places (Warren and Asch, 2000). Also, can be defined as a technique that tries to predict the location of archaeological sites or materials in a region (Kholer and Parker, 1986). The use of GIS in archaeology arose in North America due to the need to manage large extensions of land, difficult to manage with conventional survey methods. The objective was to catalogue, inventory and protect the largest possible number of sites. Faced with this dichotomy, predictive models were developed that would allow delimiting the areas that were more likely to contain archaeological sites. The first applications of the predictive model were based on the intersection of environmental and, to a lesser extent, cultural variables that were measured in a Boolean way, considering the absence or presence of certain conditions that favoured human occupation in a specific place. That is if in a specific place human habitation could not exceed 800 m of altitude, this duality was established, distinguishing those sites with lower altitudes as suitable (a value of 1 was given) and those with higher altitudes as unsuitable (a value of 0 was given). The passage of time and the evolution of GIS allowed to carry out predictive models more complex, treating the study variables quantitatively, in such a way that they could be given more importance to some over others. This is based on the weighted value method. In this way, it seeks to predict the location of archaeological sites, and if the patterns of occupation of the human societies of the past responded to a series of conditions. In which some had greater importance than others in response to the different circumstances affecting those societies. Predictive models can be divided into three categories (Nakoinz and Knitter, 2016). First are the point density approximations, which do not consider the localization preference of sites (Ben-Said, 2021; Bevan, 2020; Bevan et al., 2013; Bivand et al., 2013). Second are inductive approximations, which are based on known points that have locational characteristics that can be extrapolated to the whole population (Carrer, 2013; Deeben et al., 1997; Verhagen and Whitley, 2012). Finally we find deductive approaches, which seek to answer the question of why sites are in certain places (Kamermans and Rensink, 1999; Kvamme, 2005; Verhagen, 2007). Based on Conolly and Lake (2006), we can establish that to carry out a model prediction, a series of phases must be followed. First of all, it is necessary to collect the data, the following is the statistical analysis of these data, later the application of the model and finally its validation is carried out (Duncan and Beckman, 2000, p. 36; Warren and Asch, 2000, p. 13). In addition, it is based on the premise that it is possible to differentiate between areas of the landscape with evidence of occupation (sites) and landscape areas without such evidence (non-sites) in the function of one or more landscape attributes. In the first phase, the location of the sites and “non-sites” through an arbitrary sampling program. Regardless Fig. 5. Pearson’s correlation test for the different covariates analysed. Table 4 Multivariate regression model results. Coefficients Estimate Std. Error Z value Pr (>|z|) (Intercept) 7.2294176 2.6165387 2.763 0.00573 ** ALT −0.0167014 0.0073815 −2.263 0.02366 * SLO −0.3456405 0.1244542 −2.777 0.00548 ** VISPR 0.0458370 0.0204803 2.238 0.02521 * WET −0.0017355 0.0011814 −1.469 0.14183 HYDROC 0.0007585 0.0004113 1.844 0.06514 . Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘’ 1 M. Díaz-Rodríguez et al. Journal of Archaeological Science: Reports 49 (2023) 104012 16 Acknowledgements This study is part of the research projects Din´ amicas poblacionales y tecnol´ ogicas durante el Pleistoceno final-Holoceno de las Sierras orientales del Noroeste ib´ erico (R&D Projects of the Ministry of Science, PID2019- 107480 GB-I00), Ocupaciones humanas durante el Pleistoceno en la cuenca media del Mi˜ no (HUM2007-63662/HIST) and Poblamiento durante el Pleistoceno Medio/Holoceno en las comarcas orientales de Galicia (HAR2010-21786/HIST). M. Díaz-Rodríguez is part of the CLIOARCH project which has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement 817564); and the NeanderEDGE project which has received funding from the Independent Research Fund Denmark (case number 9062-00027B). We are very grateful to the two anonymous reviewers and the editor, whose comments have contributed significantly and helped improve the initial version of this paper. The analyses of this work have been carried out in the statistical environment R (R Core Team, 2021). Formatting of funding sources. M. Díaz-Rodríguez is granted by Margarita Salas Fellowship from the European Union - NextGeneration EU by the Programme for the requalification, international mobility, and attraction of talent through the University of Santiago de Compostela (Spain) and the Ministry of Universities (Spain). Data accessibility. The data used in this paper and the code executed in R are available at https://github.com/mikeldiazrodriguez/patterns_monforte and archived on zenodo (https://doi.org/10.5281/zenodo.7808861). Appendix A. Supplementary material Supplementary data to this article can be found online at https://doi. org/10.1016/j.jasrep.2023.104012. References Adler, D., Kelly, T.S., 2022. vioplot: violin plot. R package version (4). Ameijenda Iglesias, A., de Lombera Hermida, A., P´ erez Alberti, A., Rodríguez ´ Alvarez, X. P., F´ abregas Valcarce, R., 2010. Geomorphological and geoarchaeological evolution of the Monforte de Lemos basin (Galicia, Spain). Erosion phases and postdepositional processes in NW Iberia. Estudos do Quatern´ ario/Quaternary Studies 6, 5–22. Ameijenda Iglesias, A., 2011. Geomorphology and Relative Chronology of the Human Occupations during the Pleistocene at the Basin of Monforte de Lemos (Lugo, Galicia). In: Lombera Hermida, A. de, F´ abregas Valcarce, R. (Eds.), To the West of Spanish Cantabria. The Paleolithic Settlement of Galicia. British archaeological Reports - International Series; 2283, Oxford, pp. 81–92. Baddeley, A., Rubak, E., Turner, R., 2015. Spatial point patterns: methodology and applications with R. Chapman and Hall/CRC Press. Baddeley, A., Turner, R., Rubak, E., 2020. spatstat: Spatial Point Pattern Analysis. Model- Fitting, Simulation, Tests. Becker, R.A., Wilks, A.R., Brownrigg, R., Minka, T.P., Deckmyn, A., 2021. maps: Draw Geographical Maps. Ben-Said, M., 2021. Spatial point-pattern analysis as a powerful tool in identifying pattern-process relationships in plant ecology: an updated review. Ecol. Process. 10, 1–23. https://doi.org/10.1186/s13717-021-00314-4. Bevan, A., 2020. Spatial point patterns and processes. In: Gillings, M., Hazigüzeller, P., Lock, G. (Eds.), Archaeological Spatial Analysis. A Methodological Guide. Routledge. Taylor & Francis Group, London and New Yotk, pp. 60–76. Bevan, A., Crema, E.R., Li, X., Palmisano, A., 2013. Intensities, Interactions, and Uncertainties: Some New Approaches to Archaeological Distributions. In: Bevan, A., Lake, M. (Eds.), Computational Approaches to Archaeological Spaces. pp. 27–52. Bivand, R., Lewin-Koh, N., Pebesma, E., Archer, E., Baddeley, A., Bearman, N., Bibiko, H.-J., Brey, S., Callahan, J., Carrillo, G., Dray, S., Forrest, D., Friendly, M., Giraudoux, P., Golicher, D., G´ omez Rubio, V., Hausmann, P., Hufthammer, K.O., Jagger, T., Johnson, K., Lewis, M., Luque, S., MacQueen, D., Niccolai, A., Pebesma, E., Prepi˜ n´ an Lamigueiro, O., Plunkett, E., Rubak, E., Short, T., Snow, G., Stabler, B., Stokely, M., Turner, R., 2020a. maptools: Tools for Handling Spatial Objects. Bivand, R., Rundel, C., Pebesma, E., Stuetz, R., Hufthammer, K.O., Giraudoux, P., Davis, M., Santilli, S., 2020c. rgeos: Interface to Geometry Engine - Open Source (’GEOS’). Bivand, R., Keitt, T., Rowlingson, B., Pebesma, E., Sumner, M., Hijmans, R., Rouault, E., Warmerdam, F., Ooms, J., Rundel, C., 2020b. rgdal: Bindings for the “Geospatial”. Data Abstraction Library. Bivand, R.S., Pebesma, E., G´ omez-Rubio, V., 2013. Applied Spatial Data Analysis with R. Springer, New York. B¨ ohner, J., Antoni´ c, O., 2009. Land-Surface Parameters Specific to Topo-Climatology. In: Hengl, T., Reuter, H.I. (Eds.), Geomorphometry: Concepts, Software, Applications. Elsevier, Amsterdam (The Netherlands), pp. 195–226. doi:10.1016/S0166-2481(08) 00008-1. Brown, P.E., 2015. Model-Based Geostatistics the Easy Way. J. Stat. Softw. 63, 1–24. https://doi.org/10.18637/jss.v063.i12. Burke, A., Levavasseur, G., James, P.M.A., Guiducci, D., Izquierdo, M.A., Bourgeon, L., Kageyama, M., Ramstein, G., Vrac, M., 2014. Exploring the impact of climate variability during the Last Glacial Maximum on the pattern of human occupation of Iberia. J. Hum. Evol. 73, 35–46. https://doi.org/10.1016/j.jhevol.2014.06.003. Burke, A., Kageyama, M., Latombe, G., Fasel, M., Vrac, M., Ramstein, G., James, P.M.A., 2017. Risky business: The impact of climate and climate variability on human population dynamics in Western Europe during the Last Glacial Maximum. Quat. Sci. Rev. 164, 217–229. https://doi.org/10.1016/j.quascirev.2017.04.001. Burke, A., Alcaraz-Casta˜ no, M., Cardineau, C.A., Bisson, M., Cuartero, F., Maillo- Fern´ andez, J.M., Fuertes-Prieto, N., Yravedra Sainz de los Terreros, J., 2021a. Archaeological Survey in Guadalajara: Human Occupation in Central Spain during the Late Pleistocene. J. F. Archaeol. 46, 1–20. https://doi.org/10.1080/ 00934690.2020.1838829. Burke, A., Peros, M.C., Wren, C.D., Pausata, F.S.R., Riel-Salvatore, J., Moine, O., Vernal, A. de, Kageyama, M., Boisard, S., 2021b. The archaeology of climate change: The case for cultural diversity. Proceedings of the National Academy of Sciences of the United States of America 118, 1–10. doi:10.1073/pnas.2108537118. Cannon, M.D., 2003. A model of central place forager prey choice and an application to faunal remains from the Mimbres Valley, New Mexico. J. Anthropol. Archaeol. 22, 1–25. https://doi.org/10.1016/S0278-4165(03)00002-3. Cano, J., Fumanal, M.P., Ferrer, C., Usera, J., Bl´ azquez, A.M., Olmo, J., 1997. Evoluci´ on de la costa meridional de Galicia durante el Cuaternario superior. In: Rodríguez Vidal, J. (Ed.), Cuaternario Ib´ erico. Huelva, pp. 33–46. Cano Pan, J.A., V´ azquez Varela, J.M., 1991. O Paleolítico Inferior e Medio. In: Galicia. Historia. Tomo I - Prehistoria e Historia Antiga. A Coru˜ na, pp. 45–85. Carrer, F., 2013. An ethnoarchaeological inductive model for predicting archaeological 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. org/10.1016/j.jaa.2012.10.001. Carrero-Pazos, M., 2018. Density, intensity and clustering patterns in the spatial distribution of Galician megaliths (NW Iberian Peninsula). Archaeol. Anthropol. Sci. https://doi.org/10.1007/s12520-018-0662-2. Carrero-Pazos, M., Bustelo-Abuín, J., Barbeito-Pose, V., Rodríguez-Rell´ an, C., 2020. Locational preferences and spatial arrangement in the barrow landscape of Serra do Barbanza (North-western Iberia). J. Archaeol. Sci. Rep. 31, 102351 https://doi.org/ 10.1016/j.jasrep.2020.102351. Cascalheira, J., Gonçalves, C., Maio, D., 2022. The spatial patterning of Middle Palaeolithic human settlement in westernmost Iberia. J. Quat. Sci. 37, 291–299. https://doi.org/10.1002/jqs.3286. Church, T., Brandon, J., Burget, G.R., 2000. GIS applications in archaeology: Method in search of Theory. In: Wescott, K.L., Brandon, J. (Eds.), PrActicAl ApplicAtions of GIS for ArchAeoloGist. A Predictive Modeling Kit. Taylor & Francis, Londres, pp. 135–155. Conolly, J., Lake, M.W., 2006. Geographical Information Systems in Archaeology. Cambridge University Press, Cambridge. Conrad, O., Bechtel, B., Bock, M., Dietrich, H., Fischer, E., Gerlitz, L., Wehberg, J., Wichmann, V., B¨ ohner, J., 2015. System for Automated Geoscientific Analyses (SAGA) v.2.1.4. Geoscientific Model Development 1991–2007. doi:10.5194/gmd-8- 1991-2015. Criado Boado, F., Bonilla Rodríguez, A., Cerqueiro Landín, D., Díaz V´ azquez, M., Gonz´ alez M´ endez, M., Infante Roura, F., M´ endez Fern´ andez, F., Penedo Romero, R., Rodríguez Puentes, E., Vaquero Lastres, J., 1991. Arqueología del paisaje, el ´ area Bocelo-Furelos entre los tiempos paleolíticos y medievales (campa˜ nas de 1987 1988 y 1989). Direcci´ on Xeral do Patrimonio Cultural, Santiago de Compostela (A Coru˜ na). Cunha, P.P., P´ erez-Alberti, A., 2022. Discussion of the transition of endorheic to exoreic drainage of Cenozoic Basins in Galicia (NW of Iberia), the development of the ancestral transverse drainage to the Atlantic and the later stage of fluvial incision. In: 10th International Conference on Geomorphology, Coimbra, Portugal, 12–16 Sep 2022, ICG2022-322. Coimbra (Portugal). doi:10.5194/icg2022-322. Lombera Hermida, A. de, Vaquero Rodríguez, M., Alonso Fern´ andez, S., Rodríguez ´ Alvarez, X.P., Lazu´ en Fern´ andez, T., F´ abregas Valcarce, R., 2008. A cunca media no Mi˜ no dentro das ocupaci´ ons paleolíticas do NW peninsular. In: M´ endez Quintas, E. (Ed.), Homenaxe a Xos´ e María ´ Alvarez Bl´ azquez II. Estudos Sobre Paleolítico. Asociaci´ on Cultural Instituto de Estudos Mi˜ noranos, Gondomar (Pontevedra), pp. 321–346. de Lombera Hermida, A., Rodríguez ´ Alvarez, X.P., Lazu´ en Fern´ andez, T., F´ abregas Valcarce, R., 2006. El Paleolítico Inferior y Medio en el interior de Galicia. La Depresi´ on de Monforte de Lemos. CROA (Boletín da Asociaci´ on de Amigos do Museo de Viladonga) 16, 31–42. de Lombera Hermida, A., Rodríguez ´ Alvarez, X.P., Rabu˜ nal Gayo, X., Ameijenda Iglesias, A., Martínez G´ omez, F., Soares Remeseiro, M., P´ erez Alberti, A., F´ abregas Valcarce, R., 2012. El yacimiento de Valverde (Monforte de Lemos, Lugo, Galicia, Espa˜ na) y las primeras evidencias de poblamiento en el pleniglacial del NO de la Península Ib´ erica. Espacio, Tiempo y Forma. Serie I, Prehisoria y Arqueología. 5, 363–382. https://doi.org/10.5944/etf. de Lombera Hermida, A., Díaz Rodríguez, M., P´ erez Alberti, A., Ameijenda Iglesias, A., Rodríguez ´ Alvarez, X.P., F´ abregas Valcarce, R., 2015. Evoluci´ on de los patrones de asentamiento de los yacimientos paleolíticos de la Depresi´ on de Monforte de Lemos (Lugo, Galicia). S´ emata: Ciencias Sociais e. Humanidades 27, 267–297. M. Díaz-Rodríguez et al. Journal of Archaeological Science: Reports 49 (2023) 104012 17 Lombera-Hermida, A. de, Rodríguez ´ Alvarez, X.P., F´ abregas Valcarce, R., Lazu´ en Fern´ andez, T., 2011. The Paleolithic settlement of the Monforte Basin (Lugo , Galicia). In: Lombera Hermida, A. de, F´ abregas Valcarce, R. (Eds.), To the West of Spanish Cantabria. The Paleolithic Settlement of Galicia. British archaeological Reports - International Series; 2283, pp. 95–112. Deeben, J., Hallewas, D.P., Maarleveld, T.J., 1997. Predictive Modelling in Archaeological Heritage Management of the Netherlands: the Indicative Map of Archaeological Values (2nd Generation). In: Willems, W.J.H., Kars, H., Hallewas, D. P. (Eds.), Berichten van de Rijksdienst Voor Het Oudheidkundig Bodemonderzoek. pp. 76–118. Díaz Rodríguez, M., 2020. Tecnologías Geoespaciales aplicadas al estiduo del Paleolítico en el Noroeste Peninsular. Universidade de Santiago de Compostela. Díaz Rodríguez, M., Carrero Pazos, M., 2019. La Prominencia Topogr´ afica de los yacimientos en abrigo en las sierras septentrionales (Lugo, Galicia). Una aproximaci´ on de estudio con SIG. In: Los Tiempos Cambian, de La Piedra Al Teclado. X Jornadas de J´ ovenes En Investigaci´ on Arqueol´ ogica. Burgos, pp. 363–373. Díaz Rodríguez, M., F´ abregas Valcarce, R., P´ erez Alberti, A., 2021. Definiendo el patr´ on de ocupaci´ on del Paleolítico Inferior en la Depresi´ on de Monforte de Lemos (Lugo, Galicia). In: Domínguez-Solera, S.D. (Ed.), Cuando Empez´ abamos a Ser Nosotr@s: Curso Sobre El Paleolítico Inferior y Medio a Nivel Mundial. Volumen III (2020). UIMP Cuenca, IDEC, IDEA, El Provencio (Cuenca), pp. 97–104. Díaz Rodríguez, M., 2017. Patrones de asentamiento en el Paleolítico superior y el Epipaleolítico en el Noroeste Peninsular. C´ alculo de rutas ´ optimas y movilidad. In: V´ azquez Martínez, A., Cordeiro Macenlle, R., Carrero Pazos, M., Díaz Rodríguez, M., Rodríguez N´ ovoa, A.A., Vilas Est´ evez, B. (Eds.), (Re)Escribindo a Historia. Achegas Dos Novos Investigadores En Arqueoloxía e Ciencias Da Antigüidade. Santiago de Compostela (A Coru˜ na), pp. 25–46. Díaz-Rodríguez, M., F´ abregas-Valcarce, R., Lombera-Hermida, A. de, Rodríguez-´ Alvarez, X.P., Xos´ e,) &, Rodríguez-´ Alvarez, P., 2021. An´ alisis locacional del yacimiento Paleolítico de Cova Eir´ os (Triacastela, Lugo). Estudos do Quatern´ ario 21, 14–32. doi: https://doi.org/10.30893/eq.v0i20.199. Díaz-Rodríguez, M., F´ abregas-Valcarce, R., 2022. Evaluating the effectiveness of three spatial cluster analysis methods on Palaeolithic site location distributions in Galicia, NW Iberian Peninsula. J. Archaeol. Sci. Rep. 41, 103323 https://doi.org/10.1016/j. jasrep.2021.103323. Díez Martín, F., 2000. El Poblamiento paleolítico en los p´ aramos del Duero. Universidad de Valladolid. Duncan, R.B., Beckman, A., 2000. Site location in Pennsylvania and West Virginia. In: Wescott, K.L., Brandon, R.J. (Eds.), PrActicAl ApplicAtions of GIS for ArchAeoloGist. A Predictive Modeling Kit. Taylor & Francis, Londres, pp. 33–58. Esri, 2011. ArcGIS Desktop. F´ abregas Valcarce, R., de Lombera Hermida, A., 2010. El paleolítico superior en Galicia a la luz de las últimas investigaciones. In: El Paleolítico Superior Peninsular. Novedades Del Siglo XXI. Homenaje Al Profesor Javier Fortea. Monografíes SERP. Universitat de Barcelona, Barcelona, pp. 255–270. F´ abregas Valcarce, R., Lazu´ en Fern´ andez, T., de Lombera Hermida, A., Pe˜ na Alonso, J.A., P´ erez Alberti, A., Rodríguez ´ Alvarez, X.P., Rodríguez Rell´ an, C., Terradillos Bernal, M., 2007. Novos achados paleolíticos no interior de Galicia. A Depresi´ on de Monforte de Lemos e as súas industrias líticas. Gallaecia 26, 7–30. F´ abregas Valcarce, R., Alonso Fern´ andez, S., Lazu´ en Fern´ andez, T., de Lombera Hermida, A., P´ erez Alberti, A., Rodríguez ´ Alvarez, X.P., Rodríguez Rell´ an, C., Terradillos Bernal, M., Serna Gonz´ alez, M.R., Vaquero Rodríguez, M., 2008. Aportaci´ ons ´ o estudo da Prehistoria da cunca media do Mi˜ no. Os asentamentos en cova e ´ o aire libre. Gallaecia 27, 63–88. F´ abregas Valcarce, R., Alonso Fern´ andez, S., Ameijenda Iglesias, A., Grandal d’Anglade, A., Lazu´ en Fern´ andez, T., de Lombera Hermida, A., P´ erez Alberti, A., P´ erez Rama, M., Rodríguez ´ Alvarez, X.P., Rodríguez Rell´ an, C., Serna Gonz´ alez, M. R., Terradillos Bernal, M., Vaquero Rodríguez, M., 2009. Novos resultados das intervenci´ ons arqueol´ oxicas no sur lucense. Os xacementos paleolíticos da Depresi´ on de Monforte de Lemos (Monforte de Lemos), Cova Eir´ os (Triacastela) e Valdavara (Becerre´ a). Gallaecia 28, 9–32. F´ abregas Valcarce, R., Alonso Fern´ andez, S., Ameijenda Iglesias, A., Grandal d’Anglade, A., Lazu´ en Fern´ andez, T., de Lombera Hermida, A., P´ erez Alberti, A., P´ erez Rama, M., Rodríguez ´ Alvarez, X.P., Serna Gonz´ alez, M.R., Vaquero Rodríguez, M., 2010. Completando o mapa. Novas dataci´ ons absolutas para o Paleolítico e Mesolítico do interior galego. Gallaecia 29, 5–28. F´ abregas Valcarce, R., Alonso Fern´ andez, S., Ameijenda Iglesias, A., Fern´ andez Marchena, J., Grandal d’Anglade, A., Lazu´ en Fern´ andez, T., de Lombera Hermida, A., P´ erez Rama, M., Rodríguez ´ Alvarez, X.P., Vaquero Rodríguez, M., Veiga Moares, A., 2011. Variabilidade das ocupaci´ ons paleolíticas na cunca media do Mi˜ no. Gallaecia 30, 5–27. Fern´ andez Fern´ andez, J., 2010. Una aportaci´ on desde la arqueología del paisaje al conocimiento del primer poblamiento humano del Valle del Trubia. Universidad de Oviedo, Oviedo, Estudio geoarqueol´ ogico y an´ alisis SIG del territorio. Garate, D., Intxaurbe, I., Moreno-García, J., 2020. Establishing a predictive model for rock art surveying: The case of Palaeolithic caves in Northern Spain. J. Anthropol. Archaeol. 60, 101231 https://doi.org/10.1016/j.jaa.2020.101231. García García, M., 2015. Poblamiento y sociedad en el valle medio del Duero durante la edad de los metales (ca, 2800–450 cal. ANE). Un estudio de arqueología del paisaje. Universidad de Valladolid. García Moreno, A., 2008. Insolaci´ on y h´ abitat Paleolítico en el Valle del As´ on (Cantabria, Espa˜ na). Cuaternario y Geomorfología 22, 93–105. García Moreno, A., 2010. Patrones de asentamiento y ocupaci´ on del territorio en el Cant´ abrico Oriental al final del Pleistoceno: Una aproximaci´ on mediante SIG. Universidad de Cantabria. García Moreno, A., 2013. GIS-based methodology for Palaeolithic site location preferences analysis. A case study from Late Palaeolithic Cantabria (Northern Iberian Peninsula). J. Archaeol. Sci. 40, 217–226. https://doi.org/10.1016/j. jas.2012.08.023. García Moreno, A., 2015. The Incidence of Potential Insolation on Settlement Dynamics and Site Location Preferences: A Case Study From the Cantabrian Late Palaeolithic. J. Archaeol. Sci. Rep. 3, 90–99. https://doi.org/10.1016/j.jasrep.2015.05.020. García Moreno, A., Fano Martínez, M.´ A., 2014. Paleolithic sites beyond the archaeological deposits. In: García, A., García, J., Maximiano, A., Ríos-Garaizar, J. (Eds.), Debating Spatial Archaeology: Proceedings of the International Workshop on Landscape and Spatial Analysis in Archaeology (Santander, June 8th-9th, 2012). Santander, pp. 231–241. García-Moreno, A., 2013. To see or to be seen... is that the question? An evaluation of palaeolithic sites’ visual presence and their role in social organization. J. Anthropol. Archaeol. 32, 647–658. https://doi.org/10.1016/j.jaa.2013.03.003. Grass Development Team, 2020. Geographic Resources Analysis Support System (GRASS). Software. Hijmans, R.J., Phillips, S., Leathwick, J., Elith, J., 2017. dismo: Species Distribution Modeling. Hijmans, R.J., Etten, J. van, Sumner, M., Cheng, J., Baston, D., Bevan, A., Bivand, R., Busetto, L., Canty, M., Forrest, D., Ghosh, A., Golicher, D., Gray, J., Greenberg, J.A., Hiemstra, P., Hingee, K., Karney, C., Mattiuzzi, M., Mosher, S., Nowosad, J., Pebesma, E., Perpinan Lamigueiro, O., Racine, E.B., Rowlingson, B., Shortridge, A., Venables, B., Wueest, R., 2020. raster: Geographic Data Analysis and Modeling. Jochim, M.A., 2022. Dots on the Map: Issues in the Archaeological Analysis of Site Locations. J. Archaeol. Method Theory. https://doi.org/10.1007/s10816-022- 09580-8. Kamermans, H., Rensink, E., 1999. GIS in Palaeolithic Archaeology. A case study from the southern Netherlands. In: Archaeology in the Age of the Internet. CAA97. Computer Applications and Quantitative Methods in Archaeology. Proceedings of the 25th Anniversary Conference, University of Birmingham, April 1997. BAR International Series 750, Oxford, pp. 20–33. Kholer, T.A., Parker, S.C., 1986. Predictive Models for Archaeological Resource Location. Adv. Archeol. Method Theory 9, 397–452. Kvamme, K.L., 2005. There and Back Again: Revisiting Archaeological Locational Modeling. GIS and Archaeological Site Location Modeling. CRC Press. Leloch, M., Kot, M., Pavlenok, G., Szymczak, K., Khudjanazarov, M., Pavlenok, K., 2022. Tracing the Palaeolithic settlement patterns in the Western Tian Shan piedmont: an example of predictive GIS modelling use. J. Quat. Sci. 37, 527–542. https://doi.org/ 10.1002/jqs.3393. Llobera, M., 2015. Working the Digital: Some Thoughts from Landscape Archaeology. Learning from Archaeological Practice, Material Evidence, pp. 173–188. L´ opez Cordeiro, M. del M., 2002. Ocupaci´ on de espacios graníticos en la Sierra de Xistral: el ´ area arqueol´ ogica de Nord´ es (Valadouro, Lugo). Gallaecia 21, 61–86. L´ opez Cordeiro, M. del M., 2004. Resultados obtenidos en la construcci´ on de una “Nueva Arqueología” del paleolítico gallego. Cuadernos de Estudios Gallegos LI, 133–151. L´ opez Cordeiro, M. del M., 2015. Industrias Líticas y Ocupaci´ on del espacio en el Paleolítico de Galicia. Universidade de Santiago de Compostela. L´ opez Olmedo, F., Palacio Su´ arez, J.P., D´ avila Ruiz, M.D.T., Luís L´ opez, F., García Rojo, E., Martínez Cano, M., P´ erez Ruiz, J., García-Brazales G´ omez, R., Monz´ on Lara, P., 2022. Mapa Geol´ ogico Digital continuo E. 1: 50.000, Zona Ib´ erica (Zona-1700). In: GEODE. Mapa Geol´ ogico Digital Continuo de Espa˜ na. Ludwig, P., Shao, Y., Kehl, M., Weniger, G.C., 2018. The Last Glacial Maximum and Heinrich event I on the Iberian Peninsula: A regional climate modelling study for understanding human settlement patterns. Global Planet. Change 170, 34–47. https://doi.org/10.1016/j.gloplacha.2018.08.006. Marcos S´ aiz, F.J., 2006. La Sierra de Atapuerca y el Valle del Arlanz´ on. Patrones de asentamiento prehist´ oricos, Burgos. Marín Arroyo, A.B., 2008. Patrones de movilidad y control del territorio en el Cant´ abrico oriental durante el Tardiglaciar. Trab. Prehist. 65, 29–45. Marín Arroyo, A.B., 2009. The use of optimal foraging theory to estimate Late Glacial site catchment areas from a central place: The case of eastern Cantabria, Spain. J. Anthropol. Archaeol. 28, 27–36. https://doi.org/10.1016/j.jaa.2008.11.001. Miller, A., Lombera Hermida, A. de, Barton, C.M., F´ abregas Valcarce, R., 2010. A New View of the Paleolithic Settlement of the Galician Hinterland, Spain. In: Abstract of the SAA 75th Anniversary Meeting. p. 173. Nakoinz, O., Knitter, D., 2016. Modelling Human Behaviour in Landscapes. Quantitative Archaeology and Archaeological Modelling, Springer, Londres (Reino Unido). Parow-Souchon, H., Zickel, M., Manner, H., 2021. Upper Palaeolithic sites and where to find them: A predictive modelling approach to assess site expectancy in the Southern Levant. Quat. Int. https://doi.org/10.1016/j.quaint.2021.05.020. Pebesma, E., Bivand, R., Rowlingson, B., Gomez-Rubio, V., Hijmans, R., Sumner, M., MacQueen, D., Lemon, J., Lindgren, F., O’Brien, J., O’Rourke, J., 2020. sp: Classes and Methods for Spatial Data. Pedersen, T.L., 2022. patchwork: The Composer of Plots. P´ erez Alberti, A., Rodríguez Guiti´ an, M., Valcarcel Díaz, M., 1993. Las formas y dep´ ositos glaciares en las Sierras Orientales y Septentrionales de Galicia (NW Península Ib´ erica). In: P´ erez Alberti, A., Guiti´ an Rivera, L., Ramil Rego, P. (Eds.), La Evoluci´ on Del Paisaje En Las Monta˜ nas Del Entorno De Los Caminos Jacobeos. Cambios Ambientales y Actividad Humana, Xunta de Galicia, Santiago de Compostela (A Coru˜ na), pp. 61–90. QGIS.org, 2021. QGIS Geographic Information System. R Core Team, 2021. R: A language and environment for statistical computing. R Foundation for Statistical Computing. Ramil Rego, E., Ramil Soneira, J.M., 1996. El fin de los tiempos glaciares en Galicia. Magdaleniense y Epipaleolítico. In: F´ abregas Valcarce, R. (Ed.), Os Primeiros M. Díaz-Rodríguez et al. Journal of Archaeological Science: Reports 49 (2023) 104012 18 Poboadores de Galicia: O Paleolítico. Edici´ os do Castro, Sada (A Coru˜ na), pp. 117–146. Ramil Rego, E., 1989/1990. Habitabilidad cavernícola: elecci´ on de asentamientos. Brigantium 6, 191–197. Rey-Rodríguez, I., L´ opez-García, J.M., Benn` asar, M., Ba˜ nuls-Cardona, S., Blain, H.A., Blanco-Lapaz, ´ A., Rodríguez-´ Alvarez, X.P., de Lombera-Hermida, A., Díaz- Rodríguez, M., Ameijenda-Iglesias, A., Agustí, J., F´ abregas-Valcarce, R., 2016. Last Neanderthals and first Anatomically Modern Humans in the NW Iberian Peninsula: Climatic and environmental conditions inferred from the Cova Eir´ os small-vertebrate assemblage during MIS 3. Quat. Sci. Rev. 151, 185–197. https://doi.org/10.1016/j. quascirev.2016.08.030. Ripley, B., Venables, B., Bates, D.M., Hornik, K., Gebhardt, A., Firth, D., 2020. MASS: Support Functions and Datasets for Venables and Ripley’s MASS. Rodríguez ´ Alvarez, X.P., Lombera Hermida, A. de, F´ abregas Valcarce, R., 2014. Las ocupaciones paleolíticas en la Depresi´ on de Monforte de Lemos (Lugo , Galicia). In: Carbonell, E., Bermúdez de Castro, J.M., Arsuaga, J.L. (Eds.), Los Cazadores Recolectores Del Pleistoceno y Del Holoceno En Iberia y El Estrecho de Gibraltar. Estado Actual Del Conocimiento Del Registro Arqueol´ ogico. Universidad de Burgos, Burgos, pp. 26–31. Rodríguez ´ Alvarez, X.P., F´ abregas Valcarce, R., Lazu´ en Fern´ andez, T., Hermida, de, L., P´ erez Alberti, A., Pe˜ na, J.A., Rodríguez Rell´ an, C., Terradillos Bernal, M., Ameijenda Iglesias, A., Rodríguez ´ Alvarez, E., 2008. Nuevos yacimientos paleolíticos en la Depresi´ on de Monforte de Lemos (Lugo, Galicia, Espa˜ na). Cuaternario y Geomorfología 22, 71–92. https://doi.org/10.1016/S0034-7094(11)70016-2. Rodríguez Rell´ an, C., F´ abregas Valcarce, R., 2015. Arte rupestre galaica: unha achega dende a estatística espacial e os SIX. Semata. Ciencias Sociais e Humanidades 27, 9–34. Rodríguez-Rell´ an, C., F´ abregas Valcarce, R., 2019. Monuments on the move. Assessing megaliths’ interaction with the NW Iberian landscapes. In: Müller, J., Hinz, M., Wunderlich, M. (Eds.), Megaliths, Societies and Landscapes: Early Monumentality and Social Differentiation in Neolithic Europe. Universit¨ at Zu Kiel, Kiel (Germany). Kiel (Germany), pp. 621–639. https://doi.org/10.13140/RG.2.2.27215.38567. Santonja, M., P´ erez-Gonz´ alez, A., 2000–2001. El Paleolítico Inferior En El Interior De La Península Ib´ erica. Un Punto De Vista Desde La Geoarqueología. Zephyrus 53–54, 27–77. Schloerke, B., Cook, D., Larmarange, J., Briatte, F., Marbach, M., Thoen, E., Elberg, A., Toomet, O., Crowley, J., Hofmann, H., Wickham, H., 2021. GGally: Extension to “ggplot2.”. Turrero, P., Domínguez-Cuesta, M.J., Jim´ enez-S´ anchez, M., García-V´ azquez, E., 2013. The spatial distribution of Palaeolithic human settlements and its influence on palaeoecological studies: A case from Northern Iberia. J. Archaeol. Sci. 40, 4127–4138. https://doi.org/10.1016/j.jas.2013.06.003. Verhagen, P., 2007. Case Studies in Archaeological Predictive Modelling. Case Studies in Archaeological Predictive Modelling. https://doi.org/10.5117/9789087280079. Verhagen, P., 2018. Predictive Modeling. In: L´ opez Varela, S.L. (Ed.), The Encyclopedia of Archaeological Sciences. John Wiley & Sons. https://doi.org/10.1002/ 9781119188230.saseas0475. Verhagen, P., Whitley, T.G., 2012. Integrating Archaeological Theory and Predictive Modeling: A Live Report from the Scene. J. Archaeol. Method Theory. https://doi. org/10.1007/s10816-011-9102-7. Vermeulen, F., 2001. The potential of GIS in Landscape Archaeology. In: Slapsak, B. (Ed.), COST Action G2. On the Good Use of Geographical Information Systems in Archaeological Landscape Studies, European Communities, Luxemburgo, pp. 9–16. Viana-Soto, A., P´ erez-Alberti, A., 2019. Periglacial deposits as indicators of paleotemperatures. A case study in the Iberian Peninsula: The mountains of Galicia. In: Permafrost and Periglacial Processes. John Wiley and Sons Ltd, pp. 374–388. https://doi.org/10.1002/ppp.2026. Villar Quinteiro, R., 1996. El Paleolítico superior y el Epipaleolítico en Galicia. UniversidaDe De Santiago De Compostela. Warren, R.E., Asch, D.L., 2000. A predictive model of Archaeological Site Location in the Eastern Prairie Peninsula. In: Wescott, K.L., Brandon, J. (Eds.), PrActicAl ApplicAtions of GIS for ArchAeoloGist. A Predictive Modeling Kit. Taylor & Francis, Londres, pp. 5–32. White, D.A., Barber, S.B., 2012. Geospatial modeling of pedestrian transportation networks: A case study from precolumbian Oaxaca, Mexico. J. Archaeol. Sci. 39, 2684–2696. https://doi.org/10.1016/j.jas.2012.04.017. Wickham, H., Averick, M., Bryan, J., Chang, W., D’Agostino McGowan, L., François, R., Grolemund, G., Hayes, A., Henry, L., Hester, J., Kuhn, M., Lin Pedersen, T., Miller, E., Milton Bache, S., Müller, K., Ooms, J., Robinson, D., Piage Seidel, D., Spinu, V., Takahashi, K., Vaughan, D., Wilke, C., Woo, K., Yutani, H., 2019. Welcome to the tidyverse. Journal of Open Source Software 4, 1686. https://doi.org/10.21105/ joss.01686. Wickham, H., Bryan, J., Kalicinski, M., Valery, K., Leitienne, C., Colbert, B., Hoerl, D., Miller, E., 2019a. readxl: Read Excel Files. Wickham, H., Chang, W., Henry, L., Pedersen, T.L., Takahashi, K., Wilke, C., Woo, K., Yutani, H., Dunnington, D., 2020. ggplot2: Create Elegant Data Visualisations Using the Grammar of Graphics. Wickham, H., 2020. plyr: Tools for Splitting, Applying and Combining Data. Wren, C.D., Burke, A., 2019. Habitat suitability and the genetic structure of human populations during the Last Glacial Maximum (LGM) in Western Europe. PLoS One 14, 1–22. https://doi.org/10.1371/journal.pone.0217996. M. Díaz-Rodríguez et al.