Fig 1.
The location of the Noceto site.
(A) Digital terrain model of northern Italy illustrating the areal distribution of the Terramare culture along the Po Plain. (B) Simplified geomorphological sketch of the region illustrating the distribution of the Terramare settlements (dots) and the position of the Noceto settlement (star) (modified from [5]). (C) DTM of the Noceto area illustrating the position of the Vasca Votiva (star) and the depression related to the quarry (arrow), which was the location of the Bronze Age settlement. Key for (B): 1, pre-Quaternary formations; 2, Alpine Pleistocene glacial deposits; 3, Pleistocene deposits at the foot of the Apennine; 4, Pre-Holocene terraces; 5, alluvial plain.
Fig 2.
The lower tank at Noceto under excavation.
Fig 3.
Details of the basal network of beams from the lower tank at Noceto.
Fig 4.
The upper tank at Noceto during the excavation.
Fig 5.
Overhead photogrammetric plan image of the upper tank at Noceto.
Fig 6.
The fill of the upper tank at Noceto showing the laminated deposits lying upon the beams placed at the base of the structure.
Fig 7.
Some of the Bronze Age vessels and wooden items, deposited in the upper tank at Noceto (in situ).
Fig 8.
Selection of archaeological material from the lower tank dating to the BM2B phase (from [3, 9]).
Pots 1, 3, 4, 5 and 7 are from SU 172 alfa (this SU is associated to the 14C age UGAMS-29349 in Table 4 below); pots 2, 6, 8 are from SU 542.
Fig 9.
Selection of archaeological material from the upper tank dating to the BM3 phase (from [10]).
Pots 1 and 2 are from SU 87a; pots 3 and 9 are from SUs 81, 81c, 81d-e; pots 10 and 14 are from SU 66base; pot 15 is from SU 5003 (this SU is associated to the 14C age Poz-19307 in Table 4 below); pots 16 and 17 are from SU 5002c-e.
Table 1.
Find locations of the samples from the two superposed Noceto tanks and the relation with the two phases identified by the dendrochronological analysis.
Table 2.
Descriptions of the 14 trees represented by the 25 oak samples and the 2 elm trees represented by 3 samples.
Fig 10.
Location of the samples submitted to the dendrochronological analysis.
(A) lower structure, (B) upper structure: a) sample of board, b) sample of post, c) sample of beam, d) beam, e) scattered board inside the lower structure, f) standing board, g) post, h) post hole.
Table 3.
The relative year (RY) dates of the Noceto (NOC) tree-ring samples cut for 14C dating from NOC-12A and 14A with Groningen (GrM) lab codes, δ13C values and measured 14C ages.
Table 4.
14C dates run on other samples from the Noceto upper and lower tanks.
Fig 11.
Comparison of the tree-ring widths of the Noceto oak chronology composed of the 9 long sequences with the Noceto oak chronology comprising the 5 short sequences.
Supporting statistics are t = 7.42, r = 0.58, tr = 0.67, all with p < 0.01, with an overlap of 113 years.
Fig 12.
The combined Noceto tree-ring width chronology comprising 14 constituent elements.
Fig 13.
Bargraph showing the crossdated placements of the timbers from the two phases identified among the Noceto samples.
Fig 14.
The tree-ring width patterns for the NOC-12 and 14 samples showing where the segments of NOC-12 (green bars) and NOC-14 (gray bars) were taken for 14C dating.
Fig 15.
OxCal wiggle-match of the 11 dated NOC elements against the IntCal20 14C calibration curve.
A. 14C dates (or weighted average 14C values) for NOC dated elements (each 5 tree-rings/years) shown at 68.3% probability (Y axis) against the IntCal20 calibration curve (68.3% probability band) [29]. The date centered RY1162, indicated with arrow, is a minor possible outlier (outlier probability ~12%). B. The fit probabilities for the latest 14C dated element: RY1202. Data from OxCal [26–28] version 4.4.2.
Fig 16.
χ2 fit function for the NOC wiggle-match against the IntCal20 14C calibration curve [29] expressed in terms of the placement of RY1202 (the last 14C dated element).
Fig 17.
Bayesian chronological model combining (i) the wiggle-match date for lower tank felling-construction, (ii) 14C dates on short-lived sample(s) associated with the use of the Noceto lower tank, (ii) a 14C on wood of the upper tank and the wiggle-match date for the upper tank felling-construction, and then (iii) 14C dates on short-lived samples associated with the use of the upper tank.
The OxCal General Outlier model is applied to the 14C dates on short(er)-lived samples, the OxCal Charcoal Outlier model is applied to the date on a wood sample (unknown what tree rings dated and we assume some in-built age likely) [27]. A. All data with UGAMS-29350 a clear outlier (too recent, indicated by arrow)–perhaps a sample that should belong with the upper tank use period? B. The model re-run after re-assigning UGAMS-29350 to the upper tank period of use. The results are very similar. Data from OxCal [26, 27] version 4.4.2 using IntCal20 [29] with curve resolution set at 1 year. The lines under each distribution indicate the modelled 68.3% and 95.4% hpd ranges respectively.
Fig 18.
The results of an OxCal Interval query applied to the Noceto upper tank use phases in Fig 17A and 17B to estimate the period of time after construction through end of use of the upper tank and placement of the upper tank use period versus the 14C calibration curve.
A. The use period from the Fig 17A model with all data. B. The use period from the Fig 17B model with UGAMS-29350 re-assigned. C. The five 14C dates from the upper tank use period (Fig 17B model) shown as placed against the calibration curve. The hollow distributions indicate the unmodelled calibrated calendar age probabilities; the solid distributions indicate the modelled calendar age probabilities. The area labelled as 1 indicates the most likely range for these data and a shorter period of use for the upper tank; the area labelled as 2 is the alternative placement due the wiggle in the calibration curve.