Accuracy and Precision in Dendrochronologies Compared to Historical Events
Accuracy and Precision in Dendrochronologies Compared to Historical Events
Tree rings and dendrochronology are one of the most accurate and precise annual measuring systems available to both scientist and the layperson: it is fairly easy to count rings in a cross-section or coring of a tree. Once the problems of false rings and missing rings are identified it becomes possible to identify these in the chronologies, and thus correct them for such errors. This is well documented in Message 9, Dendrochronology Basics.
Note that carbon-14 (14C) measurements and age calculations based on 14C measurements are NOT discussed yet, as the focus is on the accuracy and precision of the tree ring chronologies.
As seen above the ecology and local climates in Ireland and Germany differ from each other and from the ecology and local climate for the Bristlecone pines.
This means that a direct comparison of tree ring widths would not be likely -- there is too much variation in the local (micro) climates for the overall patterns to be similar -- except where the sites are relatively close, and of course, for extreme incidents that affect the climate of the whole earth, such as the "year without a summer" (1816 CE) mentioned in Message 9, Additional Information on European Oaks.
The two Bristlecone pine chronologies correlated with such high accuracy and precision because they came from adjoining sites.
The two Oak chronologies correlated as accurately as they did because they shared a common basic European climate pattern, even thought they had individual microclimate differences.
However, there are several additional historical incidents of note that do show up in the tree chronologies:
quote:The extreme weather events of 535536 were the most severe and protracted short-term episodes of cooling in the Northern Hemisphere in the last 2,000 years. The event is thought to have been caused by an extensive atmospheric dust veil, possibly resulting from a large volcanic eruption in the tropics, or debris from space impacting the Earth. Its effects were widespread, causing unseasonal weather, crop failures, and famines worldwide.
Tree ring analysis by dendrochronologist Mike Baillie, of the Queen's University of Belfast, shows abnormally little growth in Irish oak in 536 ...
So there is consilience between history and the Irish oak chronology: 100% accuracy and precision at 1816 CE and 536 CE, from independent sources of information with the same values. Thus we can test dendrochronologies with historical events, and we can look at this aspect in greater detail here:
quote:Frost damage to the wood of mature trees is a rare phenomenon caused by temperatures well below freezing at some time during the growing season, when secondary wall thickening and Iignification of immature xylem cells in the annual ring is not yet complete. Freezing promotes extracellular ice formation and dehydration which result in crushing of the outermost zone of weaker cells, leaving a permanent, anatomically distinctive record in the wood . ...
Frost-damage zones have been produced in the annual rings of subalpine bristlecone pines (Pinus longaeva D. K. Bailey and P. aristata Engel.) at intervals of a few decades to a few hundred years for at least the past 4,000 yr. They are observed at localities ranging from California to Colorado, a distance of some 1,300 km. In the course of tree-ring chronology development, the presence and type of frost damage in dated annual rings from living trees [10-12] and sub-fossil wood [13,14] was routinely noted. ...
... Wexler's basic premise seems to be supported by Lamb's observation  of southward displacement of the sub-polar low-pressure zone in the North Atlantic sector in the first July following a great eruption, and continuing in some cases for 3 - 4 yr. ...
... Synoptic situations more typical of winter may be expected to occur in late spring and in early autumn. Such a scenario seems to have been followed in the frost-ring year of 1884, ...
We can see the evidence of frost-rings for 1817 (following the "year with no summer"), and for 1884, after the eruption of Krakatoa in 1883. There are several other notable events shown going back to 1601 CE, however there was no frost-ring for 1785 when one of the highest DVI's was recorded.
Such effects may not occur in all locations, due to weather patterns, and thus may not affect all three chronologies, but effects can still be found in many wide spread localities. Frost rings are reported in pines in Sweden for example. Similar correlations can be found for other volcanic eruptions, demonstrating that these events can affect the tree-ring growth in a way that can provide accurate information of the interaction of eruptions with climate:
This high consilience between these independent chronologies gives us high confidence in the accuracy and precision of the Irish oak chronology, and increases our (already high) confidence in the Bristlecone pine chronology.
Comparing European Oak and Bristlecone Pine Chronologies by 14C Levels
Comparing European Oak and Bristlecone Pine Chronologies by 14C Levels
As noted for the Bristlecone pines and European oaks, crossdating between many trees in a location is one method to check for errors. Another is to compare two or more independent chronologies from different locations, as was done with the two Bristlecone pine chronologies and the German oak chronologies. However the climate differences between Europe and the White mountains in California are too different to use dendrochronological methods to compare the oaks to the pines. We have indirect comparisons through the historical events that show up in these chronologies, but these are limited to the historic period.
Another comparison can be done, however, using 14C quantity measurements as alternate "ring" data: the generation of 14C changes year to year, so there is a pattern of 14C variation in the atmosphere similar to tree ring climate patterns, and these can be matched in a similar manner. Because the atmospheric levels are essentially the same around the globe (there is some difference between north and south hemispheres), we can compare the Bristlecone pines and European oaks by their 14C level patterns.
For an initial idea of the accuracy of the data and the amount of error involved we have this piece of evidence:
quote:In a recent series of papers studying annual tree rings, Miyake et al. (2012) reported on the existence of excursions in the radiocarbon record at A.D. 774 775, followed by a less intense event at A.D. 993 994 (Miyake et al., 2013). The A.D. 774 775 spike is observed as a change in Δ14C of ~12 15 in a 1 2 year period. Apart from the event at A.D. 993 994, there are no other reported excursions of this magnitude in the last several thousand years (Usoskin and Kovaltsov, 2012). The initial work of Miyake et al. (2012) on the A.D. 774 775 event was based on annual rings from Japanese cedar trees. The first event has been independently confirmed by other investigators on European oak trees (Usoskin et al., 2013), with a change in Δ14C of~15.In addition, Gόttler et al. (2013a, 2013b) report on a record from the Southern Hemisphere using Kauri wood from New Zealand. This Kauri record shows the same amplitude in Δ14C but with a small offset due to the Southern Hemisphere regional effect. Liu et al. (2014) have recently reported on a similar excursion in Δ14C determined from dated corals in the South China Sea.
In a recent paper, Liu et al. (2014) proposed that the 14C increase at A.D. 774 775 was caused by a cometary impact into the Earths atmosphere. ... They also cited Chinese historical records from A.D. 773 that described a major atmospheric disturbance at the time, including a significant dust event (e.g., Napier, 2001).
We have confirmed the A.D. 774 775 event in the 14C record at two additional locations, in the western United States and Russia. The amplitude of the event is very similar to previously reported results from Japan, Germany, and New Zealand. ... The fact that the 14C signal is observed in five very different locations with exactly the same amplitude is remarkable in itself. The exact cause of the event is unclear, although a number of mechanisms have been proposed, all of which require an extraterrestrial origin. ...
The consilience displayed with these independent dendrochronologies is remarkable: if any one of them were off by 1 year it would be an 0.13% error. Thus this consilience is good confirmation that the tree ring chronologies are both accurate and precise this far back in the chronological record.
quote:We will employ tree rings and carbon-14, but not in the way readers may be accustomed to seeing. We will not use carbon-14 to determine an age at all. We will simply measure how much carbon-14 is currently found in each tree ring. Carbon-14 decays with time, so if each tree ring represents one year of growth, we should see a steady decline in the carbon-14 content of each successive ring. Figure 5 shows tree-ring carbon-14 data from living trees extending back 4000 rings. ...
If additional confidence in this data is desired, it may be helpful to note that the amount of carbon-14 found in a timber from a tunnel in Jerusalem thought to have been built by Hezekiah is approximately the same as the amount found in tree ring number 2700, which places its ring-counting age where expected from Biblical records if each ring equals one year. Even better, consider the Dead Sea Scrolls the book of Isaiah in particular. ... The amount of carbon-14 in the Isaiah scrolls is equal to or less than the amount in tree ring number 2100, meaning carbon-14 confirms its before-Christ historicity.
This graph appears to start with year 2000 CE (rather than 1950). This adds 2050 BP (100 BCE) and 2650 BP (700 BCE) to the list of correlations of historical artifact to dendrochronological age by 14C content.
Then there is consilience with Egyptian history and the dating of various finds (artifacts), for example:
quote: ... Radiocarbon dating, which is a two-stage process involving isotope measurements and then calibration against similar measurements made on dendrochronologically dated wood, usually gives age ranges of 100 to 200 years for this period (95% probability range) and has previously been too imprecise to resolve these questions.
Here, we combine several classes of data to overcome these limitations in precision: measurements on archaeological samples that accurately reflect past fluctuations in radiocarbon activity, specific information on radiocarbon activity in the region of the Nile Valley, direct linkages between the dated samples and the historical chronology, and relative dating information from the historical chronology. Together, these enable us to match the patterns present in the radiocarbon dates with the details of the radiocarbon calibration record and, thus, to synchronize the scientific and historical dating methods. ...
... We have 128 dates from the NK, 43 from the MK, and 17 from the Old Kingdom (OK). The majority (~75%) of the measurements have calibrated age ranges that overlap with the conventional historical chronology, within the wide error limits that result from the calibration of individual dates.
The modeling of the data provides a chronology that extends from ~2650 to ~1100 B.C.E. ... (red lines added)
The results for the OK, although lower in resolution, also agree with the consensus chronology of Shaw (18) but have the resolution to contradict some suggested interpretations of the evidence, such as the astronomical hypothesis of Spence (24), which is substantially later, or the reevaluation of this hypothesis (25), which leads to a date that is earlier. The absence of astronomical observations in the papyrological record for the OK means that this data set provides one of the few absolute references for the positioning of this important period of Egyptian history (Fig. 1A).
Note that there are several other sample dates with similar correlation of 14C measurement to dendrochronology correlations, here it is the earliest/oldest set that is of interest as a measure of accuracy and precision. The dendrochronology correlation is shown as two lines in Fig 2 (+1σ and -1σ ) -- I added the red lines in the image for discussion:
The earliest/oldest dates in Fig 2 are shown at ~2660 BCE, with 7 samples placed together (with two more placed nearby). There are several possible matches for each of these samples, running from 2580 BCE to 2860 BCE -- due to the wiggle of the 14C amounts in that portion of the graph -- I get 5 possible matches for the lowest point with an average age of 2693 BCE, 8 possible matches for the next point with an average of 2660 BCE, 6 possible matches for the third point for an average of 2702 BCE, 12 possible matches for the fourth point for an average of 2733 BCE, 9 possible matches for the fifth point for an average of 2754 BCE, 6 possible matches for the sixth point for an average of 2750 BCE, 8 possible matches for the seventh point for an average of 2771 BCE, 8 possible matches for the eight point for an average of 2787 BCE, and 6 possible matches for the highest point for an average of 2788 BCE. Assuming these points all represent the same age, the overall average age is ~2740 BCE with σ of +/-88 years (2827 BCE to 2651 BCE).
Shaw's date for the tomb is 2660 BCE, so this falls inside the margin of error and thus is in close agreement with that dating.
Note that +/-88 years in over 4,700 years of tree ring chronology is an error of +/-1.9%. The error is partly due to the two stage process of using 14C data to convert to dendrochronological calendar age, but it is mostly due to the wiggle of the 14C levels that match these sample data points to several different times.
Note that this conversion to dendrochronological time does not depend on the calculation of 14C 'age' (which is a purely mathematical conversion of the measured amounts of 14C and 12C in the samples), but to comparing the measured 14C/12C ratios to ones found in the tree rings to find the best match to the tree rings. Using 14C levels to match chronologies introduces an error due to the number of different rings that match those levels inside the +/-1σ margins of error.
So we have another historical calibration date of 2660 BCE with 98% consilience between history and European oak chronology. This chronology extends back to 12,410 cal BP (before 1950), or 10,460 BCE, and ~40% of its length is consilient with documented historical events\artifacts.
For a final idea of the accuracy of the tree ring data and the amount of error involved in just the dendrochronologies, we have this:
quote:For inclusion in the calibration data set, dendrochronological dating and cross-checking of tree rings is required. ...
The Holocene part of the 14C calibration is based on several millennia-long tree-ring chronologies, providing an annual, absolute time frame within the possible error of the dendrochronology, which was rigorously tested by internal replication of many overlapping sections. Whenever possible, they were cross-checked with independently established chronologies of adjacent regions. The German and Irish oak chronologies were cross-dated until back into the 3rd millennium BC (Pilcher et al. 1984), and the German oak chronologies from the Main River, built independently in the Gottingen and Hohenheim tree-ring laboratories, cross-date back to 9147 cal BP (Spurk et al. 1998).
Due to periodic narrow rings caused by cockchafer beetles, some German oak samples were excluded from IntCal98. Analysis of these tree rings, with an understanding of the response of trees to the cockchafer damage, allowed some of these measurements to be re-instated in the chronology (Friedrich et al., this issue).
The relation between North American and European wood has been studied using bristlecone pine (BCP) and European oak (German oak and Irish oak), respectively. Discrepancies have become evident over the years, in particular when the German oak was corrected by a dendro-shift of 41 yr towards older ages (Kromer et al. 1996). Attempts were made to resolve the discrepancies by remeasuring BCP samples, measured earlier in Tucson (Linick et al. 1986). The University of Arizona Laboratory of Tree-Ring Research provided dendrochronologically dated bristlecone pine samples to Heidelberg (wood from around 4700 and 7600 cal BP), Groningen (around 7500 cal BP), Pretoria (around 4900 cal BP), and Seattle (around 7600 cal BP). The replicate measurements have a mean offset of 37 +/- 6 14C yr (n = 21) from the Tucson measurements. ... Because of this offset, the IntCal working group has decided not to include the BCP record in IntCal04.
Uncertainty in single-ring cal ages for dendrochronologically-dated wood is on the order of 1 yr for highly replicated and cross-checked chronologies and is therefore ignored in the analysis.
The Bristlecone Pine was not included in the calibration data because it was 37 years younger than the two oak chronologies at 7600 BP (before 1950). This indicates that the Bristlecone pine chronology is likely missing rings, especially in the more ancient rings where the number of samples is small and where the offset is noticed. Even so, this is an error of only 0.48% at 5650 BCE, which is still very high accuracy.
Two studies look at the accuracy and precision between these chronologies, first:
quote:High-precision measurement of dendrochronologically dated Irish oak at bi-decade/decade intervals has continued in the Belfast laboratory, extending the 14C data base from ca AD 1840 to 5210 BC. The dendrochronology is now considered absolute (see Belfast dendrochronology this conference) (Brown et al, 1986) and a continuous detailed curve is presented, showing the natural variations in the atmospheric concentration of 14C over >7000 years. Each data point has a precision of <2.5 , and some 4500 years have now been compared with Seattle, giving excellent agreement.
It has been shown above that it is now possible by combining Seattle and Belfast data to provide an internationally acceptable calibration curve within a 1σ envelope of ca ± 14 years, covering a time period of some 4500 years. The remaining Belfast curve from 25005210 BC would be valid using an error multiplier of 1.23 to give an average calibration band-width of <± 20 years.
Note that the "" symbol is "parts per thousand," so this is <0.25% error in the measured 14C levels in all the samples. At 5210 BCE (7160 BP) an error of 0.25% in ~7200 years would be a error of +/-18 years, very precise and accurate.
quote:The availability of absolutely (dendrochronologically) dated German oak has allowed the Belfast laboratory to extend its published high-precision 14C measurements of Irish oak (Pearson et al. 1986) by 2680 yr. The samples were selected at contiguous 20-yr intervals, following a precedent adopted and considered satisfactory in previous publications. All samples were measured for at least 200,000 counts within the 14C channel. The statistical counting error, together with the error on standards, backgrounds and applied corrections, give a realistic precision quoted on each sample of ± 2.5%o (± 20 yr). This error is considered high-precision for sample ages of 7000-8000 BP.
To help justify a claim to accuracy, and at the same time, help to determine a laboratory error multiplier, both replicate analysis and interlaboratory comparisons are necessary. We measured six contiguous samples to give an overall precision of ± 20 yr on each sample. They gave a mean age difference of ca. 13 yr, when compared to the same samples (some already duplicated) measured some 8-11 yr before (Table 1). This difference is considered reasonable, although just at the acceptable limit of statistical expectation. We compared recent replicate analysis of Irish oak samples from 5170-5090 BC to German oak, and the mean values differed by <10 absolute.
The mean difference between the Irish Oak Chronology and the German Oak Chronology was <10 years over ~8,000 years, based on matching 14C levels, or 0.13%. This compares to the 0.48% for the Bristlecone Pines.
The combined information from these chronologies now extends back to 7980 BCE, or 9930 BP (before 1950), slightly longer than the Bristlecone Pine chronology. The significant point though, is not the extension of the chronologies, but the consilience of the data and their ability to match historical dates with very small error.
So the data from all three dendrochronologies is consilient with 99.5% accuracy and precision and the German oak and Irish chronologies back to 7980 BCE with 99.9% accuracy and precision. This results in very high confidence for the accuracy and precision of the dendrochronologies.
Jull,A.J.T., Panyushkina,I.P., Lange,T.E., Kukarskih,V.V., Myglan,V.S., Clark,K.J., Salzer,M.W., Burr,G.S., and Leavitt,S.W., (2014), Excursions in the 14C record at A.D. 774 775 in tree rings from Russia and America, Geophys. Res. Lett., 2015 41,doi:10.1002/2014GL05 [2015, March 01] http://www.geo.arizona.edu/....arizona.edu/files/JullAGU.pdf
Reimer, P. J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J. W., Bertrand, C. J. H., Blackwell, P. G., Buck, C. E., Burr, G. S., Cutler, K. B., Damon, P. E., Edwards, R. L., Fairbanks, R. G., Friedrich, M., Guilderson, T. P., Hogg, A. G., Hughen, K. A., Kromer, B., McCormac, G., Manning, S., Ramsey, C. B., Reimer, R. W., Remmele, S., Southon, J. R., Stuiver, M., IntCal04 Terrestrial Radiocarbon Age Calibration, 0-26 CAL KYR BP, Radiocarbon, Vol 46, Nr 3, 2004, p 10291058 https://journals.uair.arizona.edu/...icle/download/4167/3592
In addition to the three absolute dendrochronologies discussed so far, there is a "floating" chronology of interest in measuring the age of the earth by counting annual tree ring layers: the German pine chronology. There are two aspects to this floating chronology: (1) the annual rings and the amount of time (age) they cover, and (2) the way it is tethered to the absolute German oak chronology.
quote:THE EARLY HOLOCENE-LATE GLACIAL GERMAN PINE CHRONOLOGY
Along the Danube and Rhine rivers, remnants of well-preserved pine trunks (Pines sylvestris) are frequently dredged from river gravels, along with oaks. When the first 14C dates of H. E. Suess (unpublished data) attributed a surprisingly old age to these pine trees, I started collecting both pines and oaks. This project led to the construction of an unbroken 1768-yr floating late Younger Dryas and early Holocene pine chronology, as well as a 405-yr Allerod pine series.
First, we have linked 185 subfossil pines to an unbroken floating record of 1605 tree rings. This sequence crosses the boundary between the early Holocene and the Late Glacial, a boundary recently detected in 13C and 2H records of pine tree-ring cellulose (Becker, Kromer & Trimborn 1991). We derived an absolute minimum age of 11,370 dendroyr BP for the beginning of the pine sequence by linking the end of the pine series with oak at the 8800 BP 14C oscillation, which is visible in both series (Kromer & Becker 1992).
Very recently, five additional pines covering 14C dates in the 8800 BP range have cross-matched with the end of the pine master chronology, extending the sequence to 1768 tree rings. This series must now overlap the beginning of the oak master near 8800 BP, which meanwhile is extended to 8021 BC. Indeed, a reasonable cross-match between the 8800 BP pine/oak masters is now visible. The overlap between both curves consists of 295 tree rings, but this important linkage is still tentative and must be confirmed by additional 14C measurements. However, this link extends the absolute German oak/pine dendrochronology by an additional 1550 yr, to 9494 BC. The calibration data beyond 7800 BC presented here are derived from this tentative zero point of 9494 BC.
Note that this matches the pine chronology to the oak chronology by wiggle-matching the 14C levels, a slightly less accurate method than by matching tree ring patterns. This method of tethering a floating chronology will be discussed in greater detail later in [msg=TBD].
Further study with additional samples not only corrected some of the oak chronology (the 41 year shift mentioned previously in Message 17) but improved the linkage to the pine chronology by matching tree ring patterns instead of using 14C wiggle matching.
quote:The combined oak and pine tree-ring chronologies of Hohenheim University are the backbone of the Holocene radiocarbon calibration for central Europe. Here, we present the revised Holocene oak chronology (HOC) and the Preboreal pine chronology (PPC) with respect to revisions, critical links, and extensions. ...
We have indicated the revisions and extensions of the combined oak and pine tree-ring chronology for central Europe constructed at Hohenheim University. This chronology forms the backbone of the Holocene 14C calibration. The Holocene oak chronology (HOC) has been strengthened by new trees starting at 10,429 BP (8480 BC). Oaks affected by cockchafer predation have been identified and removed from the chronology. The formerly floating Preboreal pine chronology (PPC) has been cross-matched dendrochronologically to the absolutely dated oak chronology. In addition, the 2 parts of the PPC were linked dendrochronologically. Including the 8-yr shift of the oak-pine link, the older part of the PPC (pre-11,250 BP) needs to be shifted 70 yr to older ages with respect to the published data (Spurk et al. 1998). The southern German part of the PPC now covers 2103 yr from 11,993 to 9891 BP (10,044 -- 7942 BC). Furthermore, the PPC was extended significantly by new pine chronologies from Avenches and Zόrich, Switzerland, and by the pine chronology from the Younger Dryas forest at Cottbus, eastern Germany. The absolutely dated tree-ring chronology now starts at 12,410 cal BP (10,461 BC). Therefore, the tree-ring-based 14C calibration now reaches back into the mid-Younger Dryas. ...
The Younger Dryas was basically a mini Ice Age that occurred just after the last major Ice Age. Note that the pine chronology is now anchored dendrochronologically rather than tethered by 14C wiggle pattern matching.
quote:The Younger Dryas stadial, also referred to as the Big Freeze, was a 1,300 (± 70) year period of cold climatic conditions and drought which occurred between approximately 12,800 and 11,500 years BP (between 10,800 and 9500 BC). The Younger Dryas stadial is thought to have been caused by the collapse of the North American ice sheets, although rival theories have been proposed.
The combined European Oak and German Pine anchored (absolute) dendrochronology now extends from the present day back to 10,461 BCE.
The earth is at least 12,477 years old (2017)
The minimum age for the earth is now at least 12,477 years old (2017), based on the very accurate and precise German pine dendrochronology anchored to the absolute German oak chronology and extending back to 10,461 BCE. This also means that there was no major catastrophic event that would have disturbed the growth of any of the overlapping trees -- ie no catastrophic flood occurred in this time.
This is significantly older than many YEC models (at least 6,000 years older, for those using Archbishop Usher's assumption filled calculations of a starting date of 4004 BCE), as this chronology extends to 10,461 BCE.
This also begins to be a problem for the type of "Gap Creationism" where the earth is old but life is young because trees are living things.
And this is still only the start of annual counting methods.
Friedrich, Michael et al, The 12,460-Year Hohenheim Oak and Pine Tree-Ring Chronology from Central Europe - a Unique Annual Record for Radiocarbon Calibration and Paleoenvironment Reconstructions, Radiocarbon, Volume 46, Nr 3, 2004, p 1111-1122 https://journals.uair.arizona.edu/...icle/download/4172/3597