South Arabia’s prehistoric monument landscape shows social resilience to climate change

PLOS One | https://doi.org/10.1371/journal.pone.0323544 May 28, 2025 16 / 24 than rely on the technologies of communication involving monuments with culturally specific construction and mnemon- ics[42]. Moreover, cloud forests and tall grasses on the escarpment limited visibility, thereby limiting the utility of a monu- ment. (Ancient houses, corrals, and hearths are visible, so preservation and sampling do not prevent archaeologists from detecting monuments). Large differences exist in the numbers of examples we recorded for each monument type. Prior experience in a compa- rable landscape (Hadramawt, southern Yemen) demonstrated that these numbers are not significantly explained by pres- ervation conditions of different landforms. Moreover, platforms are significantly associated with lower elevations and HCT in higher elevations [58], both captured in our viewshed boundary in this research so that our monument counts would be relatively unaffected by terrain differences. Therefore other factors—such as period, cultural conventions, available labor—account for differences in the numbers of monuments of each type. Results of collocation and environmental analyses suggest that historic place attachments alone cannot explain spatial distributions of different monument types. A shift from earlier platforms close to springs to HCT built further from springs coincides with major aridification. Perhaps desert people no longer clustered as close to reliable water (and there is no

evidence they adopted oasis farming in Dhofar). Instead, the shift reflects adaptive resilience through major subsistence change during an era when goats replaced cattle in the arid interior [100, p. 228–9]. Goats can go longer without water and browse a wider range of forage [101]. HCT are also located on elevated sites compatible with the steep terrain of goats. The northern face of a slope would have less vegetation, given that prevailing moisture comes as fog from the south, so HCT would have higher prominence on bare terrain, perhaps consonant with the territorial signaling archaeologists have suggested as their function [8,102]. HCTs and triliths occurring across a broader geographical range are consistent with populations unable to support clustered herds on the sparse graze available near springs and adapting through dispersal. This adjustment in pastoralism (while eschewing farming) raises the role of monuments in the adaptive social resilience of a dispersed, thinned, Middle Holocene Dhofari population responding to climate change. Harrower’s analysis of water management in Wadi Sana, Hadramawt, noted the contemporaneity of HCTs and water management in a drying land- scape, speculating that “death and ancestry-infused landscape markers maintained deep social importance” [102, p. 505]. To other archaeologists also, the appearance of Bronze Age tombs suggests that resilient people now forged social identities and mediated access to resources through kinship with memorialized dead [8,34]. Our analysis

scrutinizes and quantifies these purported relationships, explaining why monuments of different eras do not appear to cluster as historic attachment to place. People’s choice to build different styles in different places signals or reflects a changing social landscape in differ- ent eras. Collocation analysis and the significantly different local environments of different monument types support the chronological framework developed from a dated sample of monuments. While triliths extend beyond our study area and may mark the once-expanded range of a particular ethnic group [74,103, p.73-4], the distribution of boat graves is restricted to eastern Dhofar, with possible parallels elsewhere [39]. Anecdotally, our team noted very few instances in which early monuments were quarried in antiquity to erect later monuments. If people did respond to prior monument placement at all, it was to place some later monuments apart from their antecedents, not to rebuild close by. We suggest that the technology of monuments communicating culturally specific social identities was a persistent feature of pastoral groups, even as the messages that monuments communicated varied through time and space. Changes in the volume and size of stones used in construction point to variable labor investments. Containing larger stones, Neolithic platforms are the largest and earliest single-episode monuments. To assemble such a platform requires balancing uprights with interior chock stones until all are held in place by a rapidly accumulated rubble fill. The construc-

tion sequence suggests that simultaneous cooperation among multiple herding households provided labor for these significantly larger monuments. At minimum, the group size could be several families, with seven strong adults to lift the largest stones. At one Neolithic platform in Yemen, rare preservation conditions indicate gatherings of up to 5000 peo- ple [29,53,71]. Ethnographic case studies document episodic monument construction events that aim to showcase an