The Proceedings of the Eighth International Conference on Creationism (2018)

dendrochronologists (such as the edge-wise juxtaposition of hand- written skeleton plots*, or the superimposition of graphed series over a light table). Computerized dendrochronology should not be conflated with the likes of computer climate change models and their complex, interactive, potentially assumption-based algorithms. The metrics used in computerized dendrochronology are in no sense “wooly”: They are quite simple [see cybis.se for details], straightforwardly verifiable, and entirely practical. Had the computerized statistical methods failed to provide distinctively high t-values for valid crossmatches that had previously always been identified visually, and/or yielded nonsensical high t-values for trees known to have lived at great geographic distances or at different times, computerized dendrochronology would never have “caught on”, let alone earned the widespread usage it now enjoys among dendrochronologists. As a matter of fact, we are actually getting close to the point where machine can replace man in dendrochronology! Larsson and Larsson (2018) come close to saying this, as they discuss the semi- automatic chronology-building process [that is detailed in the next section under: How Individual… ]. They comment, “The iterative procedure can be interactive (incremental) which means that CDendro suggests a candidate to add and the dendrochonologist manually accepts of rejects each sample after visual inspection of the match. But the procedure can also be fully automatical, which only makes sense with a relevant parameter setup (see above). If the criteria for acceptance are sufficiently high, the dendrochonologist most likely would accept the match also in incremental mode, that means we do not think that the dendrochronologist will make better decisions than CDendro .” (p. 7; Emphasis added). There is no concern about the somewhat different t-values, obtained for the same crossmatch, by the various older dendrochronological software programs. Consider an occasional difference in t-values of up to 2.0. While there could have been a potentially-valid controversy surrounding the significance of 6.5 and 8.5 for a crossmatch, in pre-CDendro software, there is no practical difference in, for example, the virtual certainty implied by, say, a crossmatch with a t-value of 12.6 or a value of 14.6. TREE-RING CROSSMATCHING VALIDATED, WITH A SYNOPSIS OF MAJOR PROCEDURES In this work, use of words such as “correct” or “valid” crossmatches refers to compellingly-visual and/or compellingly- statistical matches of trees, at their proper overlap, that are known to have lived at the same time in close proximity. It also refers to the crossmatches conventionally regarded as valid, for the long chronologies, based on visual and/or statistical criteria that is comparable to those found between compellingly-matched known- contemporaneous (living) trees. Permit an example, of a correct crossmatch, involving currently or recently-living trees. We have a long-lived tree, designated as the reference, which was cut down right after the 1963 growing season. Another long-lived tree, which grew nearby, was cored right after the 2017 growing season. The series may not crossmatch strongly, if at all, but if they do, it is expected that the distinctively- high-t crossmatch point will be at the bark-ward (youngest) ring of the second tree offset -54 years relative to the bark-ward (youngest) ring of the reference tree, and at no other position . That is exactly what we find. 1. The “Floor” and the “Ceiling” for Recognizing Correct Crossmatches Traditionally, dendrochronologists have been using a rough cutoff t≥3.5 before accepting a crossmatch as potentially valid (e. g, as in Figure 5). This may well serve as the minimum credible working value (“floor”), but a virtual-certainty (“ceiling”) value is also needed for a prospective crossmatch, especially when one is dealing with prehistoric timbers for which little or no independent context (e. g, historical or archeological data) exists to back up Woodmorappe ◀ Tree-ring chronology shortening via disturbances ▶ 2018 ICC 654 Figure 2. The Disturbance-Clustering Hypothesis. Actually-contemporaneous trees [the constituents of 1, 2, 3, and 4], now disturbance-induced to crossmatch only within their respective disturbance-imposed clusters (1)-(4), get staggered and assembled into an artificially-old long tree-ring chronology. “Bridge” series that connect the clusters may occur at sites of short overlaps [(1)-(2) junction], but more likely occur at zones where there are only “few6matchers”(*) and/or where (as illustrated) the series are relatively weakly “attached” to the master chronology [(2)-(3) and (3)-(4) junctions]. Both of the latter are suggestively identified in the actual TRN/FIN (Figure 7).