The preprint manuscript gmd-2022-11 “Improved representation of plant physiology in the JULES-vn5.6 land surface model: Photosynthesis, stomatal conductance and thermal acclimation” presents the implementation and parametrisation of three additional photosynthesis and stomatal conductance formulations (Farquhar, Medlyn and “AcKK” in addition to original Collatz formulation) into the JULES LSM. Parametrisations are mostly based on large datasets or pre-existing peer-reviewed work. The four model configurations were analysed on site level with data from 17 EC sites as well as globally using drivers from the WFDEI meteorological dataset. An additional RCP8.5 scenario simulation was also performed and evaluated.

I found this manuscript to be an interesting read and I do not have any major revision requests. The results are (mostly) clearly presented, properly discussed and the conclusions are accurately drawn. While the presented model modifications are not new, they provide new insights in model behaviour and expand the JULES model functionality. Additionally, the new model configurations also provide a possibility for the model to be run in an “ensemble” mode (similar in style to CMIP) that would allow e.g., improved detection of model weaknesses and uncertainties related to processes and variable estimates. One thing that could be discussed is the effect of static PFT’s to the simulations and results.

Minor revision requests:

L167               The last reference Mercado et al. – ‘et al’ is in italics and missing point.

L181               I believe this should be GC3.1 instead of GC.1.

L200-202        So soil water stress is a linear response function (as in Best et al. 2011), but are the parameters fixed throughout the different biomes/soil types? I think it is reasonable to exclude possible β-function modifications (such as sigmoidal response or β^s formulations) from these experiments, but I would like a clarification to the manuscript, i.e., just stating that the parameters are fixed or that they depend on the soil type is ok.

L239               You have here an undefined parameter θ and I think you mean α.

L247               Just a suggestion for the last ‘large’ division – add normal brackets around both numerator and divider and add ‘-1’ exponent to the latter and write them as a product. If you like the original formulation more just use that.

L310-312        Although, fitting g₁ to g_s is not the most complicated of tasks, at least some more information should be given about this. I think you are using nonlinear regression, but did you estimate the goodness of the fit as in Lin et al. (2015) or use a separate validation set to control for overfitting? Regardless, the dataset seems heterogenous enough to prevent overfitting by itself.

L346               Is the mean seasonal cycle produced over entire measurement period or do you have one average cycle for each year? With former you are essentially smoothing the timeseries and reducing year-to-year variability which makes sense from a “climatic” perspective, but you might miss interannual variability. Although, since you are not calibrating/optimising model parameters this is not really an issue. Additionally, could you add time-intervals (which years of measurements you have used) to Table S1?

L385               You should explain what conservative here means (flux preserving) or give a reference.

L393-396        Just to clarify if I understood this correctly, you generate a separate spin-up for each model configurations and the difference of differences is taking “Collatz” as the baseline configuration for site simulations?

L398               You have a different realization of a process, not a complete lack of one, so I would suggest rephrasing this sentence.

L419               “Jmax:Vcmax” ratio has “:” in the index (or is this deliberate as it appears like this elsewhere in the paper as well).

L468               Because you have RMSE in the panel headings, I was initially reading this as “negative values are improvements” (since RMSE would be smaller). If you want to “fix” this, change the order in EQ 12 to “new – old”.

L481               Figure 2 headings are now telling a different story. Here you state that comparisons are to the original, but in the caption it is between the different configurations (based on headings you are not showing differences between Fq.Jac and AcKK.Med as well as Fq.Med and Clz.Jac.). Please change either image or text to what you want to show.

L612               It is a bit difficult to properly interpret the RMSE values from S8 and S9 so I wanted to ask, how clearly defined are these areas? In many places the differences seem quite minimal (and I don’t think there’s a good way to improve the images).

L706               I believe there was no competition among species or changing PFT zones/percentages (this came to mind with the mention of boreal region here)? This is one thing you might want to mention or speculate on somewhere in the discussion (maybe in or after the last paragraph under this heading).

L712               Missing space after bracket.

L890               You use two types of doi formats (long and short), and I would suggest sticking with just one.

Fig. S2            is not referenced in the manuscript or supplementary material.

see enclosed file

Comment on gmd-2022-11

Referee comment on "Improved representation of plant physiology in the JULES-vn5.6 land surface model: Photosynthesis, stomatal conductance, and thermal acclimation" by Rebecca J. Oliver et al., Geosci. Model Dev. Discuss.

The preprint manuscript gmd-2022-11 “Improved representation of plant physiology in the JULES-vn5.6 land surface model: Photosynthesis, stomatal conductance, and thermal acclimation” focused on improving the representation of photosynthesis and stomatal conductance parameterization within the JULES land surface model. This work is interesting and I think the paper should be published as it brings new and useful information to the scientific literature. 

Minor revision:

Line 202-206: The multi-layer canopy module can deal with the distribution of radiation and energy in vegetation very well. However, there is a hypothesis here that the parameterization of vegetation canopy needs accurate characterization, which is recommended to be described in a little detail.

Line 239: what is θ?

Line 241: it is suggested to show the Q10 functions here or in the supplementary.

Line 249 and 296: the unit of the gas constant R, 8.314 J·mol-1K-1, suggest using the same expression.

Line 411-424 and 452-462: How long is the simulation period for these sites? From Figure 1, different schemes have significant differences in the simulation performance of GPP and EF at different sites. Is it due to insufficient accumulation, especially for EBF/BET-tr and NET?

Line 522: Could you please show the results of GPP, LE, and H from the original JULES model? Then add the difference between each new configuration to the original one.

Line 599-601: The tropical forests appeared not only in Amazon and central Africa but also in Southeast Asia, including South China, Indo-China, Malay Peninsula, and regions to their south. It seems that AcKK.Med is not the only one who shows the best in the tropical forest.

Line 753-759: The reviewer agreed the understanding of tropical forests is still lacking. The complexity of the canopy process is not well handled in the current model, which is also one of the problems restricting the JULES model and other ESMs. It is suggested that the author consider adding some discussion from this aspect.

Line 856: It is suggested that authors could share data and scripts in a similar open-source way. MOSRS registration takes too long to examine and verify, so there is no way to review it more deeply.

Table S1: What is the period corresponding to the observation data of each station? And it is suggested to show the prescribed LAI value of each station.

After Figure S12: What is the purpose of this document? Restrepo-Coupe, N, 2013.