The network analysis conducted in this study indicates that the seasonal mobility of current herders exhibits a distinct pattern characterized by a preferential connection between pairs of closer winter and summer clusters, leading to the generation of a modular connectivity pattern at the supra-regional scale. The network comprises at least two large modules, one of them includes the southern sector of the transhumant herders' distribution, encompassing the winter areas of Agrio and the middle course of Neuquén (Neuquén south), as well as the southern summer areas located in the Andes region, while the other includes clusters of the winter areas from Neuquén north, Curi Leuvú, and Barrancas, along with the geographically adjacent summer areas of Andes north, Domuyo, and Tromen. This pattern suggests that geographical proximity plays a significant role in shaping the observed modularity within the bipartite network of current herders. In other words, the modular structure seems to arise from the preferential movement of herds towards specific summer clusters that are also spatially close, except for those in the Añelo region, which represents the most distant area from the summer camps.
The conclusions drawn from our findings differ from those previously postulated for the crianceros of the area, which were based on a methodological approach that described the network of connections between pairs of winter and summer camps used by each herder to assess population resilience (Easdale et al. 2016). The authors of that study suggested that the use of space by current transhumant herders at a regional scale is highly decentralized. However, this conclusion is based on networks where each winter and summer camp node belongs to a single household, resulting in the reconstruction of numerous dyads and triads (Easdale et al. 2016). In contrast, our work defines nodes as geographic-ecological clusters, encompassing multiple households of herders, while the links represent the number of families moving between them. By grouping these camps based on environmental and ecological characteristics, we were able to identify mobility patterns that would not be discernible when analyzing the mobility of each household. Our definition of weighted bipartite networks, considering the number of herders moving between clusters, enabled us to apply statistical mechanics of complex networks to explore properties that cannot be examined in simple structures such as dyads and triads, including the preferential or modular use of space (Barabasi 2016; Newman 2018).
The pattern of modularity found in the bipartite network of summer and winter clusters of transhumant herders can be linked, beyond the importance of the underlying geographic space, to socio-historical aspects. The recurrence in the occupation and use of summer clusters by the same group of individuals belonging to households tied by kinship relations or other social relations, such as neighborhood, which also share the same winter area could account for this pattern to some extent. In this sense, several authors have documented the intergenerational transfer of summer fields (Bendini et al. 2002; Gonzalez Coll 2008) along with the transmission of traditional ecological knowledge (Berkes et al. 2000) linked to the transhumant way of life. Furthermore, an ethnographic study performed in the Tromen area showed that the main reasons for choosing specific summer areas indicated by the crianceros is that their ancestors went there (D'Abramo et al. 2021b). In other words, the selection of summer camps is not grounded on explicit knowledge about the available resources in each area (in the sense of omniscience proposed by models of ecological theory; Matsumura et al. 2010), but rather derived from the transmission of information across generations. For example, in the Tromen region, we found that a third of the households in this summer area correspond to relatives settled in the winter area from the Curi Leuvú basin (D'Abramo et al. 2021b). Moreover, activities involved in livestock management in the summer area require cooperation between neighbors or relatives for some tasks, such as herding assistance and temporary care of animals (Preda et al. 2022). The distances between winter and summer clusters, with the majority ranging from 7 to 100 km, may also contribute to conforming to a modular structure. Considering that the movement between patches is not cost-free (Matsumura et al. 2010), factors such as resource availability and water for the herders and animals would favor the selection of camps at a short distance from the winter areas, thus reinforcing the network modularity.
The framework developed here based on modern herder's mobility is suitable for interpreting the δ18O values corresponding to human remains recovered from the studied archaeological sites regarding the seasonal mobility of past hunter-gatherers. Our results show that the distances traveled between winter and summer clusters by current herders are similar to those reported for seasonal or logistical movements among hunter-gatherers (Kelly 2013). In this regard, if hunter-gatherers made seasonal movements from Curi Leuvú to summer areas, it is expected that the δ18O values of human remains would match the δ18O values of water in the areas near the Tromen massif and Domuyo volcano. Indeed, we found that between 30% and 50% of the δ18O values of human remains, chronologically ranging from the Middle to the Late Holocene (5000 − 200 years BP), matched this expectation. On the contrary, the δ18O values of the remaining individuals are consistent with two possible scenarios: a) a significant proportion of the individuals primarily remained in the Curi Leuvú basin, or b) they seasonally used the nearby summer areas of the Andes north, whose δ18O water values are indistinguishable from the winter areas. In this sense, other isotopes indicative of mobility, such as strontium (⁸⁷Sr/ ⁸⁶Sr) and obsidian, could contribute to distinguishing between these two alternatives and, therefore, the potentially utilized areas during the summer season.
In a previous study, Fernandez et al. (2021) showed that Cretaceous-Miocene volcanic deposits from the Andes mountains differ from Pliocene-Quaternary volcanic deposits of Domuyo volcano and Tromen massif, as well as from Jurassic-Cretaceous sedimentary rocks from the Curi Leuvú basin. Their preliminary results indicate that the human values from Hermanos Lazcano and Aquihueco burial sites correspond to the Jurassic-Cretaceous of Curi Leuvú and the Pliocene-Quaternary areas of Tromen and Domuyo. In this regard, although ⁸⁷Sr/ ⁸⁶Sr values do not allow differentiation between Curi Leuvú and the summer areas of Tromen-Domuyo, they enable distinguishing between the use of the summer areas of the Andes north and Tromen-Domuyo. Therefore, combining oxygen and strontium stable isotopes helps refine the range of seasonal mobility of prehistoric groups in the Curi Leuvú area. Accordingly, Fernández and colleagues (2018) suggest that obsidian artifacts found in the Tromen area originate from locally available rocks, particularly from Cerro Tilhue –the southernmost volcano in the Tromen massif, 12 km east of the Curi Leuvú river–, with no obsidians from the Andes north region, such as those from Laguna del Maule –one of the most important sources of raw materials in the region, at the border between Chile and Argentina, approximately 95 km north of the studied human burial sites– and the upper course of Varvarco (Barberena et al. 2019; Rindel et al. 2020). All these results support the preferential use of the summer areas of the Tromen massif and Domuyo volcano by groups settled in the Curi Leuvú basin, and the permanence of a large proportion of the group in the winter area throughout the year. This interpretation is consistent with the expectations derived from the analysis of the mobility of current herders and is compatible with a logistical mobility during summer season, with foraging movements of individuals or small groups from the winter area to the nearest summer areas, in agreement with previous studies (Barberena et al. 2017; Rughini et al. 2020; Bernal et al. 2020).
To summarize, our study applied a complex network approach to examine the current mobility patterns of transhumant herders, enabling the characterization of specific properties related to seasonal space utilization. Moreover, the use of a geographic-ecological criterion to define nodes, the definition of the network as a weighted bipartite, and the exploration of general properties of the network allowed us to derive a reference frame for mobility patterns and seasonal space use in the past in North Neuquén. This frame of reference can guide the study of diverse types of archaeological evidence, such as isotopes, lithic raw materials, rock motifs, and pottery styles, enabling informed interpretations of prehistoric mobility. Future studies can expand our approach beyond bipartite network models, which focus on modeling a subset of spatial usage, specifically individuals who move between summer and winter camps. Therefore, it is necessary to explore alternative approaches incorporating information on individuals who occupy the same area annually (microhabitats) without seasonal movements to summer camps. By exploring alternative models and methodologies, we could generate a more comprehensive framework for diverse archaeological evidence. Such advances will serve as a foundation for enhancing our comprehension of the prehistoric complementary use of summer and winter areas in the northern region of Neuquén.