Where to start reading about CDendro & CooRecorder?

Last update Nov 8 2013.
Start out by getting some practical experience of CDendro and also to provide you with some new unanswered questions.
Under the heading
"Using CDendro" read and go through the exercises within the following sections: Please set up your CDendro system and perform the exercises as shown in these sections!

Then have a sincere look at the section:

If you are a somewhat experienced dendrochronologist, then continue with this:
  • Basics: Detrending, Normalization, Discrimination and Autocorrelation

    That section might also look basic to a reasonably experienced dendrochronologist, but nevertheless the basics of this section is what I get most first questions on from dendrochronologists, like "How do you take care of autocorrelation?", "How do you detrend?". When doing practical crossdating with CDendro, I never think about these problems as they are taken care of automatically, nota bene for the crossdating and building of "collections" of crossdated samples - not for creating any sort of climate proxy! A less experienced user might skip this section for a first reading and continue with what follows.

A reader who is new to dendrochronology should read the section
  • Basics: Where is a reference curve?" telling not only about my own experience of getting referential data but also about what is now freely available for the Scandinavian region. There is also a mentioning of the current free availability of some mean value curves of oak data from NW-Europe. Here I also want to "push" for our Wiki at www.cybis.se/wiki containing both reference texts and also ring width data from the Scandinavian region.
Other fundamental sections under the heading
Using CDendro are: Then read When you have become more used to the program you should absolutely read the section Then it might also be time for the section
  • Methodology: What is a good TTest value? (In short: You need an overlap >80 and  a T-value above 6 to be fairly confident with a crossdating, though as low as T=5.5 is probably the correct match.)

How to start out with CooRecorder?

CooRecorder is described in a separate section with its own "left-bar contents". To learn the program I would recommend you to start from the beginning "CooRecorder basics" and then continue with the following sections one after the other. Please note that it is very easy to switch between CooRecorder and CDendro by using the respective program icons on the menu bar. If  you consider to start measuring several radii from the same tree, you should also read the CDendro section
  • Using CDendro: Multi-radii collections. Then you should carefully consider how you name each radius file from the same tree so that CDendro can understand this.
When you have gone through all this, you are probably quite able to dig into and hopefully understand the rest of this documentation. You are always welcome to write to me so that we can clear out any questions.

Our research project:

Finally Petra and I hope that you will take some time to look at our research project that now has gone on for ten years. We want to prove or disprove whether there are "invented years" in our time line within the first millennium AD. Years ago this was a big issue in Germany with TV programs and newspaper articles. Since then the open discussion has faded. Our investigation - after crunching a terrible lot of data - clearly shows

  • that a European oak chronology (about 1500 years long), containing samples of archaeologically Roman origin, has a significant match (T > 6.5) against an absolute supra-long pine curve from northern Scandinavia but 218 years later than expected!
Is it really possible to match curves from that distant regions and between species? An absolute Danish oak curve covering AD 200-1986 also shows a significant match, however this time at the expected position. We also can demonstrate that other absolute long European oak curves match in the same way towards the north Scandinavian pine. The correlation coefficient is indeed low, but the match covers such a long overlap that the T-value goes high enough to indicate a true match, and there is no competing match against the more than 7000 years long reference.

Now if this is right, what would it mean? Most probably that Roman time is dated conventionally too old by 218 years, quite a big issue for our historians to tackle! Unfortunately it would also mean that the 14C calibration curve is corrupt within the first millennium BC.

We want to see more dendrochronological research about this matter. So we hope that you and your colleagues will start discussing and check the validity of our results. The case is described here: On the continuity of the current European tree ring oak chronology

Lars-Åke Larsson



Copyright © 2017, Cybis Elektronik & Data AB, www.cybis.se