hypotheses and glossary

[lizzieinvancouver]

See also issue #25 ... List of all hypotheses (team?): @FrederikBaumgarten @AileneKane @kavs-P -- @FrederikBaumgarten will schedule meeting ASAP/make it happen ... they will consult with @alanaroseo

Please just work on here, or add new tex file to repo!

[AileneKane]

@rdmanzanedo Nice! I think they both look good and will be interested to hear what others think! i like the one with internal vs external as it aligns with the text, but agree dividing the different hypotheses gets tricky sometimes...Writing the legend for this figure will take a bit of time i think!

[alanaroseo]

I like both too. I think the it is hard to put all into internal external categories (e.g. species specific differences), maybe a layout like the first option but with symbols for internal/external so that some can have both?

On Fri, 8 Mar 2024 at 6:30 AM, AileneKane @.***> wrote:

@rdmanzanedo https://github.com/rdmanzanedo Nice! I think they both look good and will be interested to hear what others think! i like the one with internal vs external as it aligns with the text, but agree dividing the different hypotheses gets tricky sometimes...

— Reply to this email directly, view it on GitHub https://github.com/lizzieinvancouver/grephon/issues/29#issuecomment-1985795157, or unsubscribe https://github.com/notifications/unsubscribe-auth/AILTTQFAWIQQMIPV3JJ2ZOLYXHDRNAVCNFSM6AAAAAA4RFK5XKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSOBVG44TKMJVG4 . You are receiving this because you were mentioned.Message ID: @.***>

[lizzieinvancouver]

@rdmanzanedo Amazing! Thanks for doing this and thanks to @AileneKane @alanaroseo and @FrederikBaumgarten for all the help.

I have been trying to re-write the paper around external, internal and species-differences (a complicating factor for both external and internal) ... and interested to hear thoughts. For this figure it would mean 'population and species specific' could become 'population-level limits to growth' or such ... and then we'd have to move species outside (and divide up the hypotheses in that box).

[rdmanzanedo]

Great, if the text is going to be more in line with internal/external, then yeah, I agree with Ailene and Alana that the internal/external would be better!

[lizzieinvancouver]

@rdmanzanedo Thanks for this! I think it looks good for now and seems on me to get the text together for us all to look at. I made good progress over the weekend ... but did not finish, and now feeling a little underwater again. Hoping to resurface in a few days to early next week! Thanks everyone for your patience.

[lizzieinvancouver]

Parking a few things I would like to review ....

• [ ] If 5 is really the right total for species-specific responses. I get only FOUR when I look at my code (end of whathappened.R) and while michelot2012 cuny2012 etzold2021 all make sense, mckown2016 is a paper about one species. This number seems wrong.
• [ ] We also should have a an explanation of the difference in total papers versus how many total show up in the hypothesis figure ... I think some don't have hypotheses and we should explain this in the caption. Also ... some papers get multiple classifications -- we should add clear methods to supp.
• [ ] I also get a different number (I get 10 papers, not three and one is soolanayakanahally2013 which seems like a population limit paper only to me personally) for "Longer growing season != more growth" and I want to make sure we represent this correctly in terms of what we mean ... I mostly classify it under an internal limit. Is that correct?

Figure changes ... someday to consider

• [ ] The ONE paper we have on shift in allocation is mckown2016 which is not about shifting to reproduction, it's about shifting to roots. Maybe we could change: reproduction -> reproduction/roots/etc. to be clear about paper and cover what we discuss in the paper?

[FrederikBaumgarten]

Hi Lizzie Here is brief method description: Hypothesis were gathered by reviewing all publications that entered the meta-analysis for clearly stated hypothesis or analog research questions that were specifically addressed in the results (e.g. hypothesis testing). After extraction, each hypothesis was assigned to a broader main hypothesis (or hypothesis cluster) displayed in Fig. X. Studies addressing multiple hypothesis could end up in several hypothesis clusters. XX articles were discarded as they didn't provided a clear enough formulation of the knowledge gap they aimed to address.

All papers minus papers with assigned hyp = discarded paper not addressing properly

@kavs-P and @AileneKane: Do you agree? Did I forget anything? @AileneKane Can you help with the specific number issues?

Best, F

[AileneKane]

@lizzieinvancouver @rdmanzanedo @kavs-P Following up on Lizzie's parked items above:

• [ ] If 5 is really the right total for species-specific responses.....I agree with @lizzieinvancouver that 5 is wrong. I believe this this should be 4 (this is what I get 4 when I run "hypothesest.R"). In our last round of grouping hypotheses to get to a lower number, we proposed shifting from "species specific responses" to "species characteristics." I think we felt this would accommodate mckown2016, which is a paper about one species, as you say. McKown2016 was also coded as "shift in allocation toward roots" so we could just keep it there and remove it from "species characteristics"- I would agree witht that. Would you @FrederikBaumgarten (the one who coded this paper)?

[AileneKane]

@lizzieinvancouver For your item

• [ ] We also should have a an explanation of the difference in total papers versus how many total show up in the hypothesis figure ... you are correct that some were coded for multiple hypotheses. I propose that we have a table in the supp with all papers and their hypotheses to clarify this. Something like this table: https://github.com/lizzieinvancouver/grephon/blob/main/analyses/output/hyp_summarytab2.csv but with the wording fully aligned with the figure, and with the papers referenced. I can make this table and add to the supp if you like!

[AileneKane]

for "Longer growing season != more growth" I got 11 papers ... (see https://github.com/lizzieinvancouver/grephon/blob/main/analyses/output/hyp_summarytab2.csv ) @FrederikBaumgarten classified soolanayakanahally2013 under multiple different hypotheses, including this one. I defer to he and @lizzieinvancouver unless you want me to take another look and offer a 3rd opinion!

[AileneKane]

@lizzieinvancouver Thank you for keeping track of all the hypothesis assignments in this issue! I think those updates are the root of the discrepancies between the hypotheses figure and table. I have not updated the hypothesis table but I believe you have, based on all the notes and decisions in this issue. If you'd like me to go back and check that all papers are coded correctly there please let me know.

[lizzieinvancouver]

I talked with @FrederikBaumgarten on 2 May about some queries, mostly my concern that longer growing season != more growth seems to me like a larger category under which some of the other categories could fall under (e.g., population-specific responses) ... after a long discussion my plan is to use the other info in the table to try to re-classify into hypotheses that fit with the figure we actually have. So I am re-checking all of longer growing season != more growth AND earlier-longer-more growth (some of this relates to dealing with concerns from @cchambe12 on the last review, which I broadly agree with) and probably will just do a general check and re-assign.

Columns ... @AileneKane @kavs-P @FrederikBaumgarten

• Hypothesis (description): Explicit description of the hypothesis the paper used; not always filled out it if clearly fit into a category for Name already? But actually there are 22 unique entries for Hypothesis (description) and 49 for Name so I think I must be wrong. Could someone explain the difference in these columns? Were they perhaps supposed to be different, but in the end served the same purpose -- jot down notes on what the paper had for a hypothesis?
• The following FOUR columns: group group_final hypothesis_for_fig wording_figure are then larger and larger groups of the previous columns .. this seems true from quick glance through coding.

[lizzieinvancouver]

In the paper, we use earlier-longer-more growth to focus mostly on the idea that more time means more growth. Based on what is in the table, that seems right for about half the papers assigned this, but I was not sure so I checked the following and believe they fit (though I think people are pretty vague on this generally):

• camarrero 2022 (first paragraph covers this) -- FB
• cufar2014 (they touch on this only briefly in intro) -- no initials
• grossiord2022 -- KP (in file, "changes in the source (i.e. net photosynthesis and total C uptake) and sink activity (i.e. height and diameter increment) would be associated with shifts in leaf senescence and growing season length")
• silvestro2023 (seems good) -- KP
• wheeler2016 (they mention this has been shown before in intro) -- no initials

I started skimming this one and am not sure it counts, they talk mostly about WARMER seasons. @AileneKane Can you check? Again, I only skimmed...

• Stridbeck2022 -- AKE (file includes "At treeline temperature-growth relationships are remain positive during the summer")

[lizzieinvancouver]

@AileneKane For Richardson2020, can you fit in one of these categories?

"Carbon fertilization"                               "effect of growth rate not equal to growth duration"
 [4] "Higher temp = more growth (temp limitation)"        "internal constraints (including pop, photo)"        "Longer growing season != more growth"              
 [7] "Longer growing season = more growth"                "More temp = more drought (drought limitation)"      "shift in allocation"                               
[10] "shift of whole pheno sequence"                      "species-specific responses"

And update on:

@AileneKane Can you check? Again, I only skimmed... Stridbeck2022 -- AKE (file includes "At treeline temperature-growth relationships are remain positive during the summer")

I just figured out that earlier-longer-more growth is divided into two categories and Stridbeck ends up in `Higher temp = more growth (temp limitation)' so we are all set. Please ignore that query!

[lizzieinvancouver]

After my preliminary edits, I ended up with these categories (the numbers are not be right; they are the RAW numbers before Ailene's code makes some changes):

• No category: 3 (most get fixed in another script)
• Carbon fertilization: 2 (Finzi2020 and Oddi 2022)
• effect of growth rate not equal to growth duration: 6
• Higher temp = more growth (temp limitation): 9 -- this falls under earlier-longer-more growth
• internal constraints (including pop, photo): 6
• Longer growing season != more growth: 1 (Richardson2020 so we will hopefully re-assign)
• Longer growing season = more growth: 10 (and these relate to more time more growth)
• More temp = more drought (drought limitation): 8
• shift in allocation: 1
• shift of whole pheno sequence: 6 (longer growing season != more growth)
• species-specific responses: 3

@FrederikBaumgarten Can you confirm what hypothesis we should have for zhu2021 and for chen1998? I think one is More temp = more drought (drought limitation) and the other really is no hypothesis. Can you confirm? Including which is which? It was not clear above.

[lizzieinvancouver]

@lizzieinvancouver Thank you for keeping track of all the hypothesis assignments in this issue! I think those updates are the root of the discrepancies between the hypotheses figure and table. I have not updated the hypothesis table but I believe you have, based on all the notes and decisions in this issue. If you'd like me to go back and check that all papers are coded correctly there please let me know.

@AileneKane I thought I had, but today as I went through things again, I was not so sure.

1. I would appreciate if you could check that things seem coded correctly, I would much appreciate it. I could also use your help to ...
2. review the hypothesis code in general. It's in hypothesest.R (I updated this today to create a new column that I think matches the paper and thus figure better) and then end of whathappened.R and ...
3. also build a table we can add to the supp ....

Sorry, I realize this is a pain, but it has become a really nice part of the paper and I want to make sure we are doing it correctly. Let me know if you have questions/issues and thank you!

[AileneKane]

@lizzieinvancouver after reviewing Richardson, i think it is addressing 3 different hypotheses listed (the paper actually sets out to test 4 different possibilities but 2 of them fit into temperature limitation i think and "Longer growing season != more growth" could be replaced with shift in whole pheno sequence. I think "species specific responses" is tested as well, because they compare species and note differences across evergreen- deciduous groups. Its ok to have multiple hypotheses for it right? I updated hypothesis.R to -removed "Longer growing season != more growth" since you no longer want that.
-replaced it with "shift in whole pheno sequence" -add "species specific responses" (There are now 3 rows for Richardson

• [ ] Please let me know if I should undo or change any of these in some way

[AileneKane]

@lizzieinvancouver and @FrederikBaumgarten i also checked chen 1998, which seems to be motivated the "missing carbon sink"- I think we could call this one CO2 fertilization as they say "this sink might be found in undisturbed terrestrial ecosystems in which carbon sequestration has increased as a result of increasing atmospheric CO2 concentration (i.e. CO2 fertilization) and nutrients in soils.." I have updated the "hypotheses.R" file to make this change.

• [ ] Please let me know if I should revert to no hypothesis for chen1998

[AileneKane]

@lizzieinvancouver the other way that i checked hypotheses is to make a table here that includes our decisions about "authors_looked_external" and "authors_looked_endogenous" with the hypotheses we designated to see if they align (i.e., that for hypotheses that we deem "external" there is "yes" in the "authors_looked_external" and that for hypotheses we deem "internal" there is a "yes" in the "authors_looked_endogenous" column). They mostly do, but I found a few that don't (which doesn't necessarily mean our selected hypothesis is wrong, but i thought it was worth double checking these). The ones that do not align are chen1998 (which had no hypothesis, but was changed to c fertilization above), cufar2014, eckes-shephard2020, finzi2020,moser2019, oddi2022, silvestro2023,

i will plan to check these, and also get started on the table. Let me know if there are other specific papers you'd like me to check.

[lizzieinvancouver]

i also checked chen 1998, which seems to be motivated the "missing carbon sink"- I think we could call this one CO2 fertilization as they say "this sink might be found in undisturbed terrestrial ecosystems in which carbon sequestration has increased as a result of increasing atmospheric CO2 concentration (i.e. CO2 fertilization) and nutrients in soils.."

I am fine with this change personally (FYI, I probably will go back and look at the three CO2 fertilization papers to understand how that is really a mechanism for longer seasons leading to more growth ... they seem sort of unrelated to me). I chatted with @rdmanzanedo and we plan to put these in just the caption.

They mostly do, but I found a few that don't (which doesn't necessarily mean our selected hypothesis is wrong, but i thought it was worth double checking these). The ones that do not align are chen1998 (which currently has no hypothesis), cufar2014, eckes-shephard2020, finzi2020,moser2019, oddi2022, silvestro2023,

@AileneKane Thanks for all your work on this. I think one problem is those external and internal rules come from way before we re-organized to fit all mechanisms into external and internal. So  cufar2014, eckes-shephard2020, oddi2022, silvestro2023 all have no to both columns, but if they had a hypothesis it really should fit into one or the other column based on our current definitions, but back then our definitions were:

• Did they look for evidence that external factors, like resources or temperature, limits the length of the growing season (or growing season metric) or limits growth?
• Did they look for evidence that developmental, biophysical or provenance constraints (endogenous factors) limit the length of the growing season or limits growth?

I think any that mention TIME implicit or explicit as the factor will probably be entered as no no and that is okay to me. It was hard to tell if the others are wrong as the column names seem the same (authorslooked_externalfactors repeated twice?)

Which was all done before issue #22 when I reorganized into the current versions of these terms.

Finally, I suggest we have richardson be 'shift in whole pheno sequence' as the species groupings seem more about functional types and that's not how we present it in our paper (or how the other papers working on species look at it -- they look specifically at species).

[AileneKane]

@lizzieinvancouver ok! I removed species-specific responses from richardson and kept in shift in whole pheno sequence.

[AileneKane]

and i corrected the column names in the checkhyp table- sorry about that!

[lizzieinvancouver]

and i corrected the column names in the checkhyp table- sorry about that!

So quick! Thanks @AileneKane ... and I agree with you that finzi2020,moser2019 seem a little weird!

Let me know when you think the numbers are good to go for @rdmanzanedo to work on a new figure and thank you again!

[rdmanzanedo]

Notes from chatting between Rubén-Lizzie on final update of the figure:

There has been some neutral drift in the categories- We need to merge them and join with the last version (Comment March 8).

• Add back shift in whole phenosequence (separate from Longer GS = more growth.
• Add a new label for longer = faster growth rate
• Populations and species specific remain together but dash between them to distinguish numbers.

[AileneKane]

@lizzieinvancouver in checking Finzi, i think they also tested hypotheses related to internal drivers (external factors- carbon fertilization and temperature limitation). E.g., they say "Integrating site-based, long-term, data sets (e.g., provides the opportunity to separate the contributions of C accumulation into internal drivers of ecosystem development (i.e., regrowth and structural and compositional changes following disturbance, nutrient cycling) and global change drivers (e.g., temperature, CO2, invasive insects, atmospheric deposition, land use). " and their results include analyses of changes in forest species composition and biomass allocation to aboveground vs different below ground/soil categories.

So, I added hypotheses for Finzi that also included "species-specific responses" and "shifting allocation" to this one to the "hypothesest.R" file. Please let me know if you disagree.

[AileneKane]

@lizzieinvancouver moser hypotheses look correct

[AileneKane]

@lizzieinvancouver for now i put the code to set up the hypotheses table with references in the bottom of "whathappened.R" and saved a .csv of the file, that should be pretty easy to read in a make into a table in the manuscript. Currently just has 3 columns: hypothesis, number of studies, and list of refs. please let me know if you want anything different/additional.

Perhaps you could add to the sweave file? Or If you'd like me to add it to the Rnw file please let me know if you'd like in in the supplement or manuscript.

[lizzieinvancouver]

@lizzieinvancouver in checking Finzi, i think they also tested hypotheses related to internal drivers (external factors- carbon fertilization and temperature limitation). E.g., they say "Integrating site-based, long-term, data sets (e.g., provides the opportunity to separate the contributions of C accumulation into internal drivers of ecosystem development (i.e., regrowth and structural and compositional changes following disturbance, nutrient cycling) and global change drivers (e.g., temperature, CO2, invasive insects, atmospheric deposition, land use). " and their results include analyses of changes in forest species composition and biomass allocation to aboveground vs different below ground/soil categories.

So, I added hypotheses for Finzi that also included "species-specific responses" and "shifting allocation" to this one to the "hypothesest.R" file. Please let me know if you disagree.

@AileneKane Thanks for working on this! It's hard to know whether to add these hypotheses from Finzi without having looked at other papers, but from what you sent I feel like they are talking about separating things perhaps not related to longer seasons ("internal drivers of ecosystem development (i.e., regrowth and structural and compositional changes following disturbance, nutrient cycling)") and then separately "global change drivers" (including longer seasons). Right now our hypotheses of "species-specific responses" and "shifting allocation" are specific to reasons you do not see increases with longer seasons and this feels pretty different to me. I would personally not add these based on these sentences, but you have read the full paper so will trust you if you think they mean 'we hypothesize longer seasons do not lead to more growth because some species do not grow more with longer seasons' and 'we hypothesize longer seasons do not lead to more growth because they shift allocation to roots or other places.'

[lizzieinvancouver]

Perhaps you could add to the sweave file? Or If you'd like me to add it to the Rnw file please let me know if you'd like in in the supplement or manuscript.

@AileneKane Amazing -- thank you! I source whathappened.R in the supp (and main text but I think the table should go in the supp) so if you could just wrap the dataframe in a quick xtable and add a quick table to the supp I could finalize it. But no worries if you do not have time as you have done a TON already.

[AileneKane]

@lizzieinvancouver I pasted a table at the end of the supp. I couldn't get the file to compile (even before I added the table) so wasn't able to check the formatting- will likely require some tweaking of column widths, etc. If you want me to work on that, let me know!

[lizzieinvancouver]

@AileneKane Thank you for the amazing work to add the references into a table! I was able to fix the error in the table (it showed up if you just paste it into the end of whathappened.R) and fixed some citation errors, but I am still getting the following bibtex errors I cannot track down:

Database file #1: ..//bibtex/grephonbib.bib
Warning--I didn't find a database entry for "."
Warning--I didn't find a database entry for "zohner2023effecteffect"
Warning--I didn't find a database entry for "sebastian-azcona2020"

And the output looks odd to me. Any ideas on how to fix these problems would be most welcome!

[AileneKane]

@lizzieinvancouver I've updated the code to resolve those warnings. please let me know if it still doesn't work for you.

[AileneKane]

@lizzieinvancouver I removed the new Finzi hypotheses i had added lastweek because they were not as specific as you stated above. i believe the paper counts are ready to go

[lizzieinvancouver]

@AileneKane Thank you! The table is working (needed longtable in the preamble), but can you check the methods text that I have in the supp:

We returned to the papers to also assess them for which hypotheses they addressed. For this, we reviewed papers for stated hypotheses and/or analogous research questions that were specifically addressed in the results (e.g. hypothesis testing). After this we grouped each hypothesis to a broader main hypothesis (or hypothesis cluster) displayed in Fig. \ref{fig:hypotheses}. We extracted all hypotheses/questions from each paper, resulting in many papers had more than one hypothesis (\Sexpr{nrow(morethanonehyp)} of \Sexpr{papernum} papers).

@rdmanzanedo think categories for the figure are (in whatever order makes sense, feel free to shorten the phrasing a little as needed):

• Earlier/Longer growing season = more growth (more time to grow): 10
• Higher GS temperatures = more growth (release of temp. limitation): 8
• Earlier/longer season = more risks: 0
• More temp = more drought (drought limitation): 10
• shift of whole pheno sequence: 6
• effect of growth rate not equal to growth duration: 5
• internal constraints (including pop, photo): 5
• species-specific responses: 3
• shift in allocation: 1

After looking at the citations I think we should keep internal constraints and species-specific responses separate because one of the citations does not include populations so does fit with species to me.

We still need to:

• Add back shift in whole phenosequence (separate from Longer GS = more growth).
• Separate out current `earlier/longer = more growth' to be two separate ones above.

Thank you both @AileneKane and @rdmanzanedo for your help!

[rdmanzanedo]

got it! on my to-do list. Thanks indeed both for all the meticulous checking! 😄

[AileneKane]

@lizzieinvancouver the methods text looks good, except for one typo toward the end. I suggest replacing "resulting in many papers had more than one hypothesis" with "resulting in many papers addressing multiple hypotheses"

[rdmanzanedo]

@lizzieinvancouver Ok, great, so added these changes and had to rearrange a bit to make it make more sense in my brain (but happy to discuss quickly). So, I slightly rearranged because I think it made more sense that external>deltaGSL are the ones that have increase in the width of the curve and external>more.temp are the ones that do not change the width of the curve, does that make sense to all?

then the category 'internal constrains' we dind't had in previous versions, so I put it in internal and I assuming we are interpreting this as similar to shift in allocation but without a shift in allocation (something physiological preventing response to changing conditions). Is that correct?

The new figure:

for the caption, the width of the shaded areas in the #studies is proportional to the number of studies, and the last column is the 'expected growth response' for each of the mechanisms.

It could be possible to argue that the external/internal separation is a bit artificial (since the resposnes to deltaGSL or incr.temp are similarly driven by internal constrains or limitations), but I do see the value of separating these, maybe the word internal/external are a bit too broad. Maybe we can explain in the text somewhere that we meant external as those mostly driven by changes in environmental conditions and internal those most affected by physiological limitations. ??

[lizzieinvancouver]

@rdmanzanedo Thank for you this! It looks great and I think your choices make sense. Could you change the first column to be # of papers (not studies) as that is how we counted these? Could you also send me the SVG so I can have it on the repo?

[lizzieinvancouver]

@rdmanzanedo I think we explain external and internal in the main text, but I have edited the caption now to read:

Climate change may alter growing season (GS) length, which can then affect growth through diverse pathways. We review hypotheses for these pathways showing the number of papers (from a review of papers studying growth $\times$ growing season length) that mentioned each hypothesis (width of the shaded areas of left column is proportional to the number of papers with the number also given, right column shows the expected growth response for each hypothesis). We group hypotheses as focused on mechanisms moderated by the environment (`external') versus those focused on internal physiological constraints, which span both source (photosynthesis-limited) and sink limitation. For more details, see Supplement.

[lizzieinvancouver]

@AileneKane Thanks! I made your edits to the methods. If you have any comments on the caption for figure let me know.

[rdmanzanedo]

@rdmanzanedo Thank for you this! It looks great and I think your choices make sense. Could you change the first column to be # of papers (not studies) as that is how we counted these? Could you also send me the SVG so I can have it on the repo?

Sure, sent it with the small change ;)

[lizzieinvancouver]

@rdmanzanedo Thank you! And thanks to @AileneKane for all your help.

[lizzieinvancouver]

Just submitted so I think this is all set -- thank you again!