Hacker News Reader

Comment by crygin | Hacker News Reader

I worked on this with Minze Stuiver -- the goal was to eliminate reliance on man-made artifacts of potentially ambiguous date. About the project: http://depts.washington.edu/qil/aboutqil.html

Datasets: http://depts.washington.edu/qil/datasets/

View on HN

>"14C, as stored in tree rings, is being measured to accurately determine the atmospheric 14C levels of the last 10000+ years. This is essential in getting precise dates for historic and prehistoric materials."

Can you explain more how this removes dependence on historical dates? Tree rings need to be calibrated too since more than one can form per year or years can be skipped based on the weather (too cold or too little precipitation -> growth stops).

I’m not sure I understand the question. Dendrochronology is independent of historical records, because you can start with a freshly-cut tree or core sample and work backwards to date an older sample. This is highly accurate and does not depend on records of when a tree was cut down. Yes, it is possible for rings to be ambiguous, but literally every measurement technique has error.

You then use the samples dated with dendrochronology to calibrate your radioisotope dating.

>"you can start with a freshly-cut tree or core sample and work backwards to date an older sample. This is highly accurate and does not depend on records of when a tree was cut down."

Assuming that the matching of rings is done perfectly (from my looking it up this is not trivial, where is the best place to find the raw data for one of the reference sequences btw?), I don't see how you could trust this without matching some rings to wood from known artifacts to "check it".

I mean can't you get a decade-long drought and so miss 10 years? And here is a source reporting that missing rings become more frequent as you look back further (it grows to 6% missing at 150 years): https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4885872/

Of course that all depends on the species, etc so maybe it can be avoiding by only selecting the "right" samples. Can you share a good source where I can see how one of the reference sequences has been arrived at from scratch?

Reply on news.ycombinator.com