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

of all the indices for every single year. A narrow channel between the upper, mean, and lower curves indicates that the indices agree well, with each other, for that year: A wide channel indicates that they are quite discordant. Not surprisingly, the master chronology for the perturbed series (Table 1) exhibits wide channels, except for the experimentally narrowed rings. However, my examination of the TRN and FIN master chronologies shows that wide channels do occur naturally (albeit sporadically and on a few-decades scale), notably at 5076-5086 BC in the combined TRN/FIN master chronology (not shown). 3. Individual Series, in Various Combinations, and Their Role as “Bridges” Between “Clusters” The ability of convincing but fortuitous <100 year and especially <50 year OVLcrossmatches to arise is unremarkable: It is called the “segment-length curse” in dendrochronology. Yet convincing false crossmatches are not limited to short overlaps, and I have identified a number of them. For example, there is one with considerable overlap (149 years), of a pair of trees, in FIN, that had ostensibly lived thousands of years apart (Figure 3 and 4 screenshots), and this pair-crossmatch satisfies all the gateway statistics, including the block test. Interestingly, both the P2Yr and P2YrL are nearly the same as the T-statistic from COFECHA--6.8, the latter of which, as noted earlier, is suspected of sometimes giving inflated t-values. Additionally, the gleichlaufigkeit*, though not usually important in the acceptance of crossmatches as valid, is above the minimally- informative 60%. The success at identifying credible but false pair crossmatches inspired me to expand the research to ensembles consisting of more than two falsely-crossmatched series and/or assortments of correctly- and incorrectly-crossmatched series. The results (simulated “bridges”) are summarized in Table 2, and discussed later. Of course, dendrochronologists would not have accepted the mistaken crossmatch, shown in Figure 3 and 4, as valid, because of the elementary fact that the two involved series each match the master chronology better, as placed and dated, than they do each other in the shown pairing-crossmatch. However, were it not for the existence of TRN/FIN chronologies, the crossmatch shown in Figure 3 and 4 would likely be accepted as valid. Let us look more closely at the existent/nonexistent concept raised in the previous paragraph in order to appreciate its importance in the Disturbance-Clustering hypothesis—specifically to the generation of “bridge” series. Initially, most if not all trees growing in the first millennium after the Flood (3000 BC to 2000 BC) likely crossmatched with each other according to the usual common climatic signal. After the “clusters” formed as specified, some individual trees likely escaped this process, and thus continued crossmatching with the few remaining unaffected climate-controlled trees, just as before. However, in the case of time intervals in which virtually all series were disturbed, the few undisturbed “survivor” series no longer had a chronology to match towards. They had thus been “released” from their “obligation” of “fitting-in” with a chronology. That is, they could now freely crossmatch towards each other (as illustrated in Figure 3 and 4), or to become part of crossmatching ensembles as illustrated in Table 2. Some of these ensembles encountered particular spots of satisfactorily crossmatching with some “cluster” of disturbed series. Depending upon the location of the crossmatching point, they either became “padding” that merely added to the local sample depth, or they became the crucial “bridges” that connected two “clusters” together. To pursue the analogy with the cardboard crossword puzzle, imagine someone taking scissors and cutting the pieces, thereby destroying their edge-matching characteristics. Thus, any and all previous fits of the puzzle pieces are now irrelevant: The pieces now fit in accordance with how they had been cut [disturbed: Table1]. Any stray pieces that had escaped the cutting no longer have to “fit-in” with the better-fitting pieces: They can now freely fit amongst themselves [Table 2] and to fit with the newly-cut edges of the new-fitting mutilated puzzle pieces [as “padding”], even connecting [as “bridges”] these mutilated assemblies together, eventually forming a long chronology. One of the challenges in constructing the false ensembles (Table 2) had involved the choosing of series that would not incur the “warning low-outlier effect*” upon their inclusion. An even bigger challenge was to avoid series with conspicuously-poor pair- crossmatching tendencies, in the matrix, against other series. This extended to the avoidance of obviously bifurcated* collections in the matrix. All of series involved, in each matrix summarized in Table 2, had to crossmatch at r≥0.3, at OVL≥50 years, in order to be included. To keep this challenge in perspective, note that many series in the respective matrices consisting of all series of TRN, FIN, or TRN/FIN combination, are below r=0.3, and quite a few of them are flagged by CDendro (as “Bad Dating”) by falling below r=0.2. As for bifurcated collections, these, although preferentially avoided, are not necessarily indicative of false assemblages. They can occur naturally whenever two sets of trees grew at significant distances from each other. For instance, the combining of TRN and FIN, done in order to complement the intervals of low sample depth in each (Larsson and Larsson 2018), is actually a double- edged sword. The geographic distances between TRN and FIN cause a bifurcated collection: TRN series have many high pair- crossmatches amongst themselves, and the same is true of FIN series amongst themselves, but few individual TRN series and individual FIN series pair-crossmatch strongly with each other. Furthermore, quite a few TRN series glaringly “stand out” in the matrix owing to their row-after-row asterisk-marked “Bad Dating” against many FIN series. Bifurcated collections can also result amongst the constituents of P2Aut. Consider “seed” FILKOM6750 (-289 34), which, incidentally, forms Subchron 9 (Figure 6). The 19 series involved, when displayed in the matrix, include series FIL6236X and PIT5494, which form almost-opposite tendencies of crossmatching strongly with one set of series but not the remainder of them. Let us now focus on the 11 false master chronologies (potential “bridges”) in Table 2. The ensembles include those with large (>150 year) OVL’s with their respective master chronologies. Sample depths equal or greater to 5 are readily achieved. The largest master chronology effect observed, despite the fact that they all are small collections, was an improvement of 4.8 (last entry). The first three listed ensembles have constituents that all crossmatch individually, within each respective ensemble, at P2YrsL≥6.0. Although my Woodmorappe ◀ Tree-ring chronology shortening via disturbances ▶ 2018 ICC 664