the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Coordination of rooting, xylem, and stoma strategies explains the response of conifer forest stands to multi-year drought in the Southern Sierra Nevada of California
Polly Buotte
Roger Bales
Bradley Christoffersen
Rosie A. Fisher
Michael Goulden
Ryan Knox
Lara Kueppers
Jacquelyn Shuman
Chonggang Xu
Charles D. Koven
Abstract. Extreme droughts are a major determinant of ecosystem disturbance, which impact plant communities and feed back to climate change through changes in plant functioning. However, the complex relationships between above- and belowground plant hydraulic traits, and their role in governing plant responses to drought, are not fully understood. In this study, we use a plant hydraulics model, FATES-Hydro, to investigate ecosystem responses to the 2012–2015 California drought, in comparison with observations, for a site in the southern Sierra Nevada that experienced widespread tree mortality during this drought.
We conduct a sensitivity analysis to explore how different plant water sourcing and hydraulic strategies lead to differential responses during normal and drought conditions.
The analysis shows that:
1) deep roots that sustain productivity through the dry season are needed for the model to capture observed seasonal cycles of ET and GPP in normal years, and that deep-rooted strategies are nonetheless subject to large reductions in ET and GPP when the deep soil reservoir is depleted during extreme droughts, in agreement with observations.
2) risky stomatal strategies lead to greater productivity during normal years as compared to safer stomatal control, but lead to high risk of xylem embolism during the 2012–2015 drought.
3) for a given stand density, the stomatal and xylem traits have a stronger impact on plant water status than on ecosystem level fluxes.
Our study reveals the importance of resolving plant water sourcing strategies in order to represent drought impacts on plants, and consequent feedbacks, in models.
- Preprint
(1309 KB) - Metadata XML
-
Supplement
(1611 KB) - BibTeX
- EndNote
Junyan Ding et al.
Status: final response (author comments only)
-
RC1: 'Comment on bg-2023-16', Anonymous Referee #1, 16 Mar 2023
Ding et al present an interesting study using the FATES model at the Soaproot site in the southern Sierra Nevada Mountains, USA, which is dominated by ponderosa pine trees. The authors explore parameter space for root depth, hydraulic, and stomatal parameters in their experimental design. The authors initialize FATES with the observed demography and turn off growth and mortality to observe how changes in climate over a major drought period impact simulated physiology, soil moisture, and water and carbon fluxes. The model is forced with 4km resolution MACA climate. Model predicted ET and GPP are compared to flux tower observed LH and tower modeled GPP. The exploration of hydraulic parameter space and rooting depths is a really interesting and important set of model experiments to perform for hydraulically enabled vegetation models and their application to terrestrial ecosystem processes. However, I have several major methodological concerns that I hope the authors can address
Method clarity: Some aspects of the experimental design are not clear. For example, how was the soil moisture initialized or spun up? Did the authors test the sensitivity of their conclusions to this method? What is the vertical resolution of the new multi soil layer model? Given that soil water is fundamental to these experiments, I think these are important details. I also wonder why the authors forced the model with MACA instead of the flux tower met.
Assessment of model performance: I see this as a contextualized OSSE experiment. However, I do think that the authors should do a little more than calculate TMSE relative to GPP/ET. Perhaps use some of the standard ILAMB metrics in addition to RMSE like inter annual variability, monthly variability and phase shifts in annual cycles?
The authors come off as defensive about the model predicted leaf water potentials, understandably because they are a physical (there are basically no trees that allow for LWP lower than -4 MPa, the -10 MPa cited in the text is for California chaparral and the correct citation is Tyree 1997 not Vesala 2017). I do appreciate that the experiments are designed to test relative sensitivities of physiological diagnostics to model parameters and that ecosystem dynamics are turned off, but with some of the parameterizations all the trees would be VERY dead before the drought started. I think with such ridiculous LWP values, the authors are going to lose the confidence of a large portion of their audience that has a physiology but not a modeling background, so I hope they will try to make some modifications to their experiments. The fact that the LWPs are dropping so low suggests that the authors might want to reconsider the vulnerability curve parameterization and parameter space that they explore for their experiments. Another option is to use the simulated water potentials to tell us more about the system (for example, to screen what parameter combinations are physiologically impossible at the site). It seems like these trees must have really deep roots to exist at this site with is in agreement with the conclusions of Goulden and Bales 2019. Why not direct the discussion in this direction instead?
Minor line specific comments
There are grammatical problems throughout the text which could use further proofreading (tense problems etc)
Almost everywhere water potential units are written as ‘Mpa’ when they should be ‘MPa’. Similarly, the authors should be consistent with capitalization/abbreviation of ‘Fig.’, ‘fig’, ‘Figure’ throughout the text. Id also like the authors to denote the denominator as either 1/x or x-1 rather than using both in the text
Instead of using kLWP as one of the parameters, can the authors choose a different letter, this is easily confused with conductance (k)
All of the figures would benefit from increased font size.
L359-361 I am having trouble understanding what the authors mean here, can they clarify?
Make sure to write out all abbreviations in figures in the captions
Citation: https://doi.org/10.5194/bg-2023-16-RC1 -
AC1: 'Reply on RC1', Junyan Ding, 21 Apr 2023
The comment was uploaded in the form of a supplement: https://bg.copernicus.org/preprints/bg-2023-16/bg-2023-16-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Junyan Ding, 21 Apr 2023
-
RC2: 'Comment on bg-2023-16: "Coordination of rooting, xylem, and stomatal strategies explains the response of conifer forest stands to multi-year drought in the Southern Sierra Nevada of California" by Ding et al.', Anonymous Referee #2, 29 Mar 2023
This manuscript presents a new model of plant hydraulics in the framework FATES, named FATES-Hydro, and kind of explores its sensitivity on some parameters for the ponderosa pine forest US-CZ2.
The manuscript is a mixed bag of several elements, which leaves the reader alone on most of it. It might be that it presents the model development of FATES-Hydro (but this is not clear), it does a very informal sensitivity analysis, and it assesses model behaviour during a very long drought without the pretence of realism.
Was the newly model developed for the study? The text reads (somehow) as if it should have been presented in Fisher et al. (2015) and Koven et al. (2020), but they are rather for FATES itself. So I guess the model development is presented here for the first time. If not, the reference is missing. If yes, it is presented very badly:
- It changes notation all the time, for example using LWP or Psi_l for water leaf potential.
- It uses unusual notation such as Se for saturation (called standardized relative water content in the manuscript).
- It uses strange definitions such as "e_i is the saturation vapor pressure (Pa) inside the leaf at a given vegetation temperature when An=0", which might be true but it is not explained (sounds strange anyway, why saturation should depend on An).
- There are different parts of the model that are not connected in the manuscript. For example, how is the formulation of Vesala et al. (2017) connected to the rest such as Ball-Berry?
- Also in Vesala et al. (2017), what is k_LWP in the Kelvin equation? It is not given in Vesala et al.
- It is not explained how Psi_l is calculated.
- Are there several stem sections?
- It is not mentioned how root water uptake is calculated. I seem to have guessed at one point in the result section that it might be proportional to root length density.
- What is "we have sequentially solved the Richards' equation for each individual soil layer"? How is this working?
There are also weird choices like using Ball-Berry while calculating leaf water potential. There are a number of good papers that discuss this such as Anderegg et al. (PLoS One 2017, 10.1371/journal.pone.0185481) and references therein.
It would also have been interesting to know why the factor beta_t (why t?) is applied on Vcmax when using actual leaf water potential. What is the physiological mechanism that reduces Vcmax during the day following leaf water potential?
However, I was wondering why the model was developed in the first place given that the current manuscript uses CLM5 in FATES, which already includes plant hydraulics.
There are also different developments of plant hydraulics in the literature such as all the work about the models SurEau and FETCH, Janott et al. (Plant and Soil 2011, 10.1007/s11104-010-0639-0), Huang et al. (New Phytolo 2017, 10.1111/nph.14273), to name just a few that could have been considered or discussed.For me, the really interesting part would have been the interactions between soil water and cohorts. It sounds like that the model has one tile per cohort so there is no interaction between cohorts. This is not explained in the manuscript let alone explored or discussed.
The manuscript presents further an ad hoc sensitivity analysis. It is also not really explained so I have to guess. Presenting sensitivity as change of model output due to change of a parameter within limits depends, of course, on the chosen limits. "the effective rooting depth, above which 95% of root biomass stays, varies from 1m to 8m", which is such a large range that, of course, everything will depend on it. I need no hydraulic model to know that.
The explanation why the authors chose to change only P50_gs and not ags is wrong (p11, l297ff; p13, l354ff). P50_gs tells "only" at what potential close the stomata, eventually. It is actually ags that determines the strategy, i.e. isohydric vs anisohydric behaviour (I think to remember that this is explained in one of the Sperry papers).I was wondering at the results section if it would not have been better to adapt the model to the site first. Now it looks like that the model cannot reproduce GPP with any of the hydraulic strategies.
Also, the Psi_l values are so strange that one should ask oneself it there is not something wrong with the model.
Are the curves of Fig. 1 realistic? How do they compare to measurements? Are there PLC (percent loss of conductance) curves for ponderosa pine?
Why are the K_max different for the safe and efficient strategies?
How comes that theta/theta_sat increase at the bottom of the soil column in Fig. 5? Is there no drainage (Fig. 6 suggests otherwise) or is theta_sat very different at the bottom of the soil column?I would have lots of further comments but I think the manuscript should be brought into shape first.
Just as a last comment on the conclusions: it is very obvious that deeper rooting plants are less influenced by droughts, and that risky and safer xylem is more and less vulnerable to drought but it also has more and less GPP, by definition. These obvious results should not be the main conclusions. And the statement "that deep-rooted pines with risky stomata have the highest GPP but also the highest drought mortality risk" is misleading because shallow-rooted pines with risky stomata actually have the highest drought mortality risk.Citation: https://doi.org/10.5194/bg-2023-16-RC2 -
AC2: 'Reply on RC2', Junyan Ding, 21 Apr 2023
The comment was uploaded in the form of a supplement: https://bg.copernicus.org/preprints/bg-2023-16/bg-2023-16-AC2-supplement.pdf
Junyan Ding et al.
Junyan Ding et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 331 | 132 | 21 | 484 | 39 | 9 | 12 |
- HTML: 331
- PDF: 132
- XML: 21
- Total: 484
- Supplement: 39
- BibTeX: 9
- EndNote: 12
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
