Meeting with Ben notes - vvandermeersch/climategrowthshifts GitHub Wiki

26 August 2026 (Victor and Ben and Lizzie)

  • I didn't take many notes but I had some ideas during the meeting that I wrote down here here

6 August 2026 (Victor and Ben and Lizzie!)

  • Ben is exited about the work! He even seemed to not be embarrassed if we send it to Science first. Away we goooo!
  • What is the relative influence of phi VPD and WP -- basically, can we make some plots to better interpret the 0.2 effect ...
  • We still need the climate betas from a model without shocks
  • We need to run 1900 onward (or 1920 I guess, but Lizzie is excited for 1900)
  • Speaking of going further back in time, Ben is excited to see the shocks in the 1600s...that's another project!
  • We need to do TWO important sensitivities analyses before running forward:
  1. what's the size of the VPD x WP on linear(ish) climate and shocks? Run on a subsample of data (1960 onward, subsample the uber-common species)
  2. what's the result if you subsample to the 2010 stands?
  • Forecasting: Let's start with IPSL/CNRM/GFDL and do just the overlapping historical period (if it works, we can do 245 moderate warming only) using this this dataset
  • Alternative projections (Daymet/Livneh based)... I think it goes with this paper
  • Also, this paper on 2015 being weird in the PNW: https://iopscience.iop.org/article/10.1088/1748-9326/aa8fde
  • Divide out the middle rockies
  • Reviewers: Bill Anderegg (Ben is in COI with everyone else who has touched a western tree ring)

27 April 2026 (Victor and Ben and Lizzie!)

19 March 2026 (Victor and Ben! Lizzie arrived late)

20th Century

Should we layer on wildfire data and beetle data? (Frame from ecological perspective rather than an climatological reconstruction -- unified framework for endogenous and exogenous drivers of tree growth.)

Can the VS (Vaganov Shaskin) model get shocks?

25 February 2026 (Victor, Lizzie, and Ben!)

  • East coast project: XBASE forced with climate... circularity?
  • last progresses on shock model (on 7 species!): adding non-concordance looks good (to me), and trees stands with high non-concordance really seem less concordant... look here for some preliminary results and ugly figures
  • How to choose a sensible time window to group shocks (e.g. decade: 1950-59, 1960-69, etc.)? And is it problematic if a known drought period gets split across two windows (for example 1987–1992), potentially diluting its effect compared to something like 2010–2019?
  • Next: what's our plan? Model is running, but we can already come up with some ideas we want to test (+ figures)... see here

Lizzie asks, are trees growing more or less?

  • Western US ... Ben is not sure, he thinks the last 20 years have not been great
  • Eastern US: Trees are growing faster and more maybe due to CO2 fertilization and summer precip (see article, 'Age class, longevity and growth rate relationships: protracted growth increases in old trees in the eastern United States')
  • Check FIA data thinks Lizzie

5 February 2026 (Victor, Lizzie, and Ben!)

  • ACC attribution? Still forgot to ask ...
  • Ben suggested maybe ask Mukund Rao to join as tree ring and Bayesian expert

21 January 2026 (Victor, Lizzie, and Ben!)

8 December 2025 (Victor, Lizzie, and Ben!)

24 November 2025 (Victor, Lizzie, and Ben!)

7 October 2025 (Victor, Lizzie, and Ben!)

Discussion points:

  • Look at GP for a stand (click here to be SAD :disappointed_relieved:), see how poorly (our) climate predicts declines, discuss! Any other metrics?
  • What are his predictions for soil moisture x VPD? See some food for thought here
  • If time allows: Has anyone looked at climate responses across trees across US West from tree rings? If not, why not?

Notes after our discussion

Best quote from the meeting, "a lot of ecologists, like, hate tree rings."

22 September 2025 (Victor, Lizzie, and Ben!)

Incomplete notes that Lizzie tossed together after.

We discussed some multi-species and other approaches

Possible future projections (though Ben can also look around later, once we agree on model and if we need VPD etc.)

Ben is totally excited to extend to carbon, "it's the carbon that matters" -- yeah!

Questions we could ask:

  • Differential response to VPD and SM (Ben suggested, but goes with our interaction model)
  • Where do tree grow more or less over last decades?
  • Think about uncertainty stuff (snore)

After the meeting ...

  • Victor and Lizzie came up with a bunch of to-do items
  • Check the literature, especially stuff by Evans and also by Williams

7 August 2025 (Victor, Lizzie, and Ben!)

What to do about GSL? (see also issue)

  • Lizzie suggested we just drop all those annoying Mediterranean and desert southwest climates (not a great idea)
  • Maybe stick with annual GDD and summer soil moisture if we cannot find better soon, they's okay large-scale integrators perhaps for this spatial scale?
  • See St George and Ault paper
  • Exploratory analysis?
  • Maybe work on this with flux data instead, see Ameriflux data and X-Base

Climatic representativeness (see also issue 47)

Other datasets for climate that go before 1980?

Other precip variables to try instead of soil moisture?

  • winter precip
  • summer soil moisture at different depths, so try 0-1 m and 0-2 m; these should integrate spring precip so try these before expanding the window

11 July 2025 (Victor's incomplete notes, Lizzie not here)

Regarding 2002 bump:

  • US drought monitor
  • maybe 2m is too deep for soil moisture, try shallower? but 2m should better capture winter precipitation (which is what matters for trees in Western NA)
  • another possibility: simply look at seasonal precipitation variables? Even if it should be less related to tree growth (what directly matters is soil moisture), I should look at patterns in other predictors (without even fitting a model)
  • remember that WLDAS soil moisture is really just a model
  • seems like it's a large-scale pattern according to stand GPs, usually we believe it's drought on such scale but could be really interesting to find another large-scale pattern (e.g. big frost event!)

Regarding $\beta_\text{VPD}$, Angio and Gymno:

  • could be the explanation I wrote in issue #42 (Gymno close stomate before, so stop to grow before Angio)
  • but could also be that we're sampling Gymno and Angio on different sites, Gymno more dry sites? Make maps of distribution, and maybe look at sites where both occurs?

Regarding interactions VPDxSM:

  • VPD kind of already includes a soil moisture effect (so kind of already includes the interaction?)... Moisture in the air is related to moisture in the soil, and colder air is related to higher soil moisture... That's why VPD is interesting, because it's to some extent integrative
  • but still make sense to include SM and VPD, since SM should reflect the water recharge (do we say this in English?) from winter (especially up to 2m...)

Regarding legacies:

  • IMHO legaciesfrom a physiological point of view should not be positive (positive effect is related to less competition, so it's more a community-level thing that a ecophysiological legacy, e.g. damage) $\rightarrow$ Ben agrees, and adds that ITRDB typically samples more isolated trees, so competition should matter less
  • in the literature, a lot of discussion about cold vs. hot droughts $\rightarrow$ today droughts are happening in a much warmer world, and it would be super interesting to see what it implies for legacies

Regarding GSL:

  • maybe try first with higher elevation site, where we know that snowpack melt has a very strong signal?

23 May 2025 (Lizzie's incomplete notes)

  • Bristle cone pine (Pinus longaeva) data — check how the model fairs on these data (where species is more sensitive to growth …)
  • Check east vs west sites for soils moisture (make plots of model predictions for growth ~ moisture across sites, comparing east and west (west is much wetter)
  • Make plots of model predictions for growth ~ GDD across sites, organized by elevation.
  • Maybe there is a signal of soil moisture being too high? (Though Victor shows it does not show up in full model plots as missing.)
  • Try VPD instead of soil moisture? (Lizzie likes this idea)
  • Ben made lots of suggestions on how to model soil moisture (what parts of WLDAS to use etc.) where I did not take notes but Victor did.
  • Part of challenge is the sites that we're starting with....
  • Discussing whether to move before 1980 with different data ... Ben suggested there is other data and maybe worth going back since these spatial precisions (e.g., 800 m) may be `false precision' and leading us astray, BUT for now he thinks 40 years of tree growth and these data is good.

9 April 2025 notes