Fig 1.
Maps of the Pacific Islands (A), Samoa (B), and the Falefa Valley (C).
Table 1.
Archaeological feature descriptions and uses summarized from [53].
Fig 2.
Overview of sampling and survey locations (A), and lidar identified archaeological features (B) in the Falefa Valley.
Fig 3.
Relative density of walls (A) and ditches (B) recorded digitally.
The boundary of the feature distributions toward the inland extent of the valley is tied to the boundary (dashed line) of available ground return lidar data and does not denote the absence of these features. Density was calculated using the line density tool in ArcMap. Cell size is 200 m and density was calculated using a 100 m search radius.
Fig 4.
Relative density of mounds recorded digitally.
The boundary of the feature distribution toward the inland extent of the valley is tied to the boundary (dashed line) of available ground return lidar data and does not denote the absence of these features. Density was calculated using the line density tool in ArcMap Cell size is 200 m and density was calculated using a 100 m search radius.
Fig 5.
Position of the largest earthen mound in Falefa Valley.
The basal dimensions of this feature make it one of the largest earthen features in Sāmoa by that measure. Several additional mounds are visible as well. The location of the western valley profile is indicated by the arrow in the hillshade image and the dot-bounded line in the inset of Falefa Valley. The location of the large mound is indicated by the rectangle in the inset of Falefa Valley.
Fig 6.
Similar feature arrangements in the Falefa Valley.
Walled rectilinear areas often enclose mounds (black) in the southwestern (A), central (B), and (C) eastern portions of the valley, although less often found in the latter. Background is a shaded relief map generated using the lidar dataset with yellow color indicating a higher elevation than green.
Fig 7.
Abutting patterns of walls in the Falefa Valley.
Single walls (Red) typically do not cross large walls (Black), abutting them instead, indicating a relative construction chronology. Some single walls also abut each other. Along the length of some large walls, sections may be identified as small walls and this is often the result of stone removal such that the section is less wide or tall and not typical of a large wall. Background in a shaded relief map generated from the lidar dataset (z = 1).
Fig 8.
North and east excavation profiles of large wall, feature 1156.
Basic layer descriptions are below profile image.
Table 2.
Modelled ages and calibrated radiocarbon ages for archaeological events in the Falefa Valley.
Fig 9.
Tempo of wall construction across the Falefa Valley.
MCMC results from the Bayesian models were analyzed using ArchaeoPhases [81] in R [R Core 82]. Note the increase in wall construction in the 14th century and later.
Fig 10.
Results of various analyses on D-section cores.
Depths are marked along the x axes. Red lines indicate the depths and results for radiocarbon analyses of samples Wk-51320, -51321, and -51322. Bottom graph from analysis of Falevao 1 core. All other results from Falevao 19 core. Missing values for K and TOC/TN result from no samples measured at those depths. Graph of grain size percents: green, solid line is sand, black, dotted line is silt, blue, dashed line is clay.
Fig 11.
pH and %Base Saturation of Falefa Valley soils.
The three samples on the left side of each graph are from the Lalomauga Alluvium in the northeastern portion of the valley. All other samples are from the Mulifanua Volcanics.
Fig 12.
Components of the territoriality and niche construction hypotheses to explain portions of the archaeological record in the Falefa Valley.
The diagrams indicate components that can be theoretically evaluated (linked evolutionary/ecological theory), those that are empirical predictions (modelled or hypothesized outcomes) presented here, or components that are current observations.
Fig 13.
Relative digital surface model of the southeastern Falefa Valley.
Elevevation shading is relative to river in black. Reticulated ditches with variable upstream-downstream relationships are shown. Rectilinear light gray areas are mounds.