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 12 / 24 Did pre-existing places marked by monuments increase the likelihood that later people would build there? Because people respond to cultural places, we expected that monument builders would respond to prior locations and influence future perceptions. To our surprise, the results of the collocation analysis in Table 2 show that very few monuments of the successively occurring monument type are collocated with the chronologically preceding monument type. Indeed, when comparing the two most common monument types, we see that a large proportion of trilith monuments are significantly isolated from HCT monuments (HCT are on ridgelines, triliths on low terraces). Technological and labor requirements for monument construction We considered the overall labor contributions to different monuments over time and what, if any of it, had to be simulta- neous labor (episodic construction). Within each monument type (Fig 7, right) and within each of two categories of tech- nological requirement (Fig 7, left), we found no significant difference in total stone volume depending on the size of the largest stone (< 0.75 m or > 0.75 m, paired blue and yellow distributions). However, if all monuments are considered together (Fig 7, center), those with one or more large stones (> 0.75 m) had significantly higher (F(1,289) = 5.383, p = 0.021) total stone volumes than those with only small stones (< 0.75 m). A chi-squared test of independence determined a significant association between the size of the largest stone and the

Fig 4. a. Variable importance from the bootstrap aggregating (Left); 4b. Partial dependence plots, with deciles for data points, show the rela- tionship between environmental variables and the predicted probability of monument occurrence (Right). https://doi.org/10.1371/journal.pone.0323544.g004