Spatial modeling of archaeological site locations based on summed probability distributions and hot-spot analyses: A case study from the Three Kingdoms Period, Korea

Archaeologists typically use radiocarbon ages to date human activities on archaeological sites. However, radiocarbon ages can also serve as independent proxies for human demographic patterns through space and time. Spatial modeling of archaeological data often involves taking diachronic datasets and developing synchronic views of the distribution of cultural patterns. Here, we evaluate spatial dynamics of human activity areas assayed by radiocarbon ages in the Baekje Kingdom of Korea before and after the invasion of Koguryeo, which is historically documented as occurring in AD 475. The statistical techniques used in this research apply spatial autocorrelations on archaeological sites with weighted attributes from summed probability distributions (SPDs) to evaluate regional scale diachronic changes from radiocarbon datasets. Ripley’s K analysis shows an increasing tendency toward clustering of weighted SPD attributes from sites after AD 325. The Getis-Ord Gi* statistic shows the changing hot spots of Baekje settlement shift prior and concurrent to the establishment of a military zone in northern Gyeonggi Province, South Korea. The settlement systems realign from clusters in the north and south of the province to clusters in the southeast with avoidance of the north.