Category talk:Trees Working Group
- 1 initiated discussion on detrending -- JEG (talk) 17:35, 17 October 2016 (PDT)
First off, this page looks great - thanks to anyone who has contributed.
Re: detrending, this reminds me that in the rest of LinkedEarth, we encourage users to distinguished between native observations and any processed form thereof. Because detrending may result in a loss of information, I would encourage the group to describe:
- Guidelines to store raw data, for each measurement type
- Guidelines to store metadata about detrending method (e.g. method name, software implementation, parameter values, etc).
I've put in some basics for what I think is needed, but isn't this a problem shared by all proxies? These are fairly standard mathematical techniques that I'd assume are used in other proxies too?
- Actually, they are not. Tree rings are somewhat unique in the amount of mathematical processing required to isolate the climatic component of the growth curve. In many cases it would be desirable to reprocess old TRW data with different detrending techniques depending on th goal, but that is often not feasible if the original data are not archived, or if the detrending process is irreversible.
Re: initiated discussion on detrending -- EK Sutherland (talk) 13:48, 17 March 2017 (PDT)
- I will note that the value of TRW data lies not only in its use for climatology and detrending is not necessarily desirable. TRW data can be applied to other dendrochronological disciplines where climate is considered ‘noise.’ For example, annual area increment (AAI) calculated from TRW series might be used to evaluate growth before and after an event, such as an insect outbreak. The AAIs would be different derived series to be archived.