Undefined codes

[rudeboybert]

Here is the codebook of codes from data/metadata/codes_metadata.csv

Code	Description
A	alternate POM (point of measurement, different than 1.3m)
B	stem broken above 1.3m
C	stem dead above 1.3m
DS	dead, stem standing
DC	dead, stem fallen
DT	dead, only tag found
DN	dead, no stem or tag found
F	inside fence (deer exclosure)
I	stem irregular where measured
J	bent stem
L	leaning stem
M	multiple stems
main	main stem
P	prostrate stem (growing parallel to ground)
RE	dendroband needs replacing or fixing
S	secondary stem
X	stem broken below 1.3m

Of the 1715 total dendroband measurements, 184 have codes D (incomplete version of existing dead code), G, Q that aren't defined:

codes	n
D	91
G	55
Q	21
I;Q	9
B;L;D	3
Q;D	3
G;D	1
I;B;D	1

@teixeirak How should we resolve these missing codes?

[teixeirak]

Hmmm... @gonzalezeb , @mcgregorian1 , @camerondow35 , do any of you know what "G" or "Q" codes mean? I don't know!

[mcgregorian1]

I believe G was an older code for "ground" as in, stem is on ground. But I could be mistaken.

Q I knew at one point but it's escaping me.

[teixeirak]

Thanks, @mcgregorian1. So if "G" is dead fallen (DC), would "D" mean dead standing (DS)? In which case,
D--> DS
G--> DC

[mcgregorian1]

That makes sense to me. I can't be 100% sure though that "G" didn't refer to just ground irrespective of being live or dead (but I'm ~90% sure). Those codes were used before we did the 2018 census, and I think we only knew what they were because we asked @gonzalezeb when comparing with the 2013 / most recent dendroband data at the time.

[teixeirak]

So these are all old codes, prior to your revamp, @mcgregorian1? In that case it predates @camerondow35 (you can ignore this), and hopefully @gonzalezeb remembers.

[mcgregorian1]

Yep, that's what I remember!

…

On Tue, 22 Jun 2021, 16:36 Kristina Anderson-Teixeira, < ***@***.***> wrote: So these are all old codes, prior to your revamp, @mcgregorian1 <https://github.com/mcgregorian1>? In that case it predates @camerondow35 <https://github.com/camerondow35> (you can ignore this), and hopefully @gonzalezeb <https://github.com/gonzalezeb> remembers. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#42 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AJNRBEMZTAG2Z5CUPQOPME3TUDX5VANCNFSM47ELQMEQ> .

[gonzalezeb]

"Q" was an original code used for 'stem broken above 1.3 m". By original I mean used in the CTFS census protocol but also in the master dendroband protocols. HOWEVER, at SCBI, "B" was also used for "stem broken above 1.3 m" (see codes for the 2013 recensus here).
I noticed that trees that have "Q" (tags 112467, 171247, 201261, 202050) do so from 2011 to 2018 but then the code is not in years 2019 and 2020, I think that need to be corrected (keep the code in all years if true). I also noticed that tag '171247' disappears in year 2016 and onward because it died in 2015, but not all trees that died in a given year are "removed" from the following years, that also need to be checked to be consistent.

For 'G", I am pretty sure the code was used when the tree was only identified to genus level but in the dendro data all trees with G are identified to species, so definitely Ian must be right. BUT, if G is for 'ground" then all trees with G any year need to be reviewed (for example tree 121301 has G in 2012 but not in 2020).

Definitely G is not DC (dead fallen).

For all trees with a "D" I would find them in the tree mortality census and assign the dead code given there. Of course, that would work for dendro-trees dead after 2014. For dendro-trees dead in 2013 use the recensus survey. For years 2010-2012 maybe one can assume D-->DS