the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Peatland-VU-NUCOM (PVN 1.0): Using dynamic PFTs to model peatland vegetation, CH4 and CO2 emissions
Tanya J. R. Lippmann
Monique M. P. D. Heijmans
Ype van der Velde
Han Dolman
Dimmie M. D. Hendriks
Ko van Huissteden
Abstract. Despite covering only 3 % of the planet’s land surface, peatlands store 30 % of the planet’s terrestrial carbon. The potential to both emit and drawdown CO2 and CH4, means that peatlands have a complex and multifaceted relationship with the global climate system. The net GHG emissions from peatlands depends on many factors but primarily vegetation composition, ground water level and drainage, land management, and soil temperature. Many peatland models use surface water levels to estimate CH4 exchange, neglecting to consider the efficiency of CH4 transported to the atmosphere by vegetation.
To assess the impact of vegetation on the GHG fluxes of peatlands, we have developed a new model, Peatland-VU-NUCOM (PVN). The new PVN model has been built from two parent models, the Peatland-VU and NUCOM-BOG models. To represent dynamic vegetation, we have introduced plant functional types and competition, adapted from the NUCOM-BOG model, into the Peatland-VU model. The PVN model includes plant competition, CH4 diffusion, ebullition, root, shoot, litter, exudate production, below-ground decomposition, and above-ground moss development, under changing water levels and climatic conditions. PVN is a site-specific peatland CH4 and CO2 emissions model, able to reproduce vegetation dynamics.
Here, we present the PVN model structure and explore the model’s sensitivity to environmental input data and the intro- duction of the new vegetation-competition schemes. We evaluate the model against observed chamber data collected at two peatland sites in the Netherlands to show that the model is able to reproduce realistic plant biomass fractions, and daily CH4 and CO2 fluxes. We find that this plot-scale model is flexible and robust and suitable to be used to simulate vegetation dynamics and emissions of other peatland sites.
- Preprint
(12001 KB) -
Supplement
(9730 KB) - BibTeX
- EndNote
Tanya J. R. Lippmann et al.
Status: final response (author comments only)
-
RC1: 'Comment on gmd-2023-48', Anonymous Referee #1, 28 Apr 2023
Peatland-VU-NUCOM (PVN 1.0): Using dynamic PFTs to model peatland vegetation, CH4 and CO2 emissions
The manuscript by Tanya Lippmann et al. provides a comprehensive description of the implementation of vegetation dynamics into the Peatland-VU model, based on the vegetation dynamics used in NUCOM-BOG. The manuscript documents the representation of vegetation dynamics that is used, and compares the new model with observations of greenhouse gas fluxes, observed cover fractions, and with model simulations with the unaltered Peatland-VU model.
I appreciate the detailed documentation of the equations that the model is based on, and I consider this well suited for GMD. In a few occasions, the rationale of these equations could be explained better – these are mentioned in the remarks below. Also, the simulation setup and driving data (frequency, timestep) ought to be described in more detail (see below). The results section is relatively straight-forward and provides first and foremost an overview of the basic behaviour for two Dutch sites, but I think that this is suitable for a paper that first and foremost documents a new model.
I consider this manuscript suitable for publication in GMD once these omissions have been corrected. I provide a list of detailed remarks below, which I hope can help the authors when revising the manuscript.
Major remarks
L 29: “Between 2005 and 2008, …” Please check this sentence – it is unclear which of the two studies the citation originates from, and the natural emissions that contributed 50% to the total CH4 budget are not all originating from peatlands, so the relevance is not clear here.
L 122: Please explain how the initialization of the carbon pools is done. Is this based on observations, or does the model require a spinup of some sort?
L 137-152, Eqs. 2-3: It would be nice if you could explain the competition for light a bit further. The equations for plant height (Eqs. 2-3) originate from applications to trees (with D being the stem diameter). Do you apply these equations to the grass vegetation from your sites? It would be good to assess how accurate these equations are for that purpose. Also, the description seems to indicate that all leaf area is assumed to be accessible for light if the PFT is highest, whereas the grassland vegetation you have at the sites will have leaf area distributed between the soil and the top of the plants (meaning that only part of the highest plant will receive light without shading by others, and part of the leaves may well be shaded). It would be good to see the rationale explained a bit further here.
L 159: The “harvest scheme” should probably be explained further: It is described only later in the manuscript that the model is made to describe dynamics of two grassland sites, and “harvesting” will hence be a grass cut. Maybe it would be good to introduce the sites and the typical land use of these earlier in the manuscript. The concept of a “harvest height” makes good sense once one understands that these are grasslands.
L 165, Eq. 7: I cannot understand the denominator in this equation. CB*SLA would result in a leaf area index, but it is subsequently divided by a long term including amongst others the growth rate of LAI. Please explain what this equation represents.
Section 2.5 or 2.7: I would welcome more details on the description of the simulation setup. How many years are the simulations (30, I think), what is the temporal frequency of the input data (daily?) and the time step of the model? Is there a spinup, or how is the model initialization done? What is the vertical resolution/how many soil layers are simulated? Also, this would be a good place to introduce the different simulations that are undertaken: “Standard” simulations with the observed input data, and simulations to derive model sensitivities (Table 3).
Minor remarks
- Overall, the text is nicely written, but it would be good to check the placement of commas throughout the manuscript
- L 46: replace “decompose” with “decomposes”
- L 90: replace “systems”
- L 97-98: Check formatting of citations
- L 99: Check consistency: write both gases as chemical formula (CO2 and CH4) or write out both
- L 113: Explain what is meant here with “plant dynamics”. Plant growth?
- L 121: Check sentence – one verb too much
- L 171: PG should probably be explained here – its definition follows only in Eq. 11.
- Eq. 9: Is the range of 1-P representing all PFTs or only the moss PFTs? Please clarify.
- Eq. 11: FPAR should probably be explained here – it follows only in L. 220
- Eq. 17 and L. 220: Please check the symbols here. I guess that AI*PAR is the total incident PAR, and not only AI? And if FPAR is the fraction, its unit should probably not be J m-2 d-1?
- L 225: Add “is” between “a” and “the ratio”
- L 228: Check the unit here – earlier, you defined PAR fluxes per day, now per second?
- L 231: Remove one of the occurrences of “where”
- L 254: FSP appears to describe some sort of empirical function of the timing of root exudates. It would be good to provide the function, or plot the activity as a function of DoY
- L 264: The equation seems to indicate a linear scaling with T(t). Does this give problems when T<0°C?
- L 277: Explain the SphagnumCarex here – is this Sphagnum peat, or is Carex somehow part of it?
- L 321: Check “photosynthetically active radiation”
- L 339 ff: Please provide information on the frequency of the measurements. How long was the time in-between the closures of the chambers?
- L 343/344, 357: It is not clear what “daily hourly average” or “daily monthly temperature” refers to. Are these daily or hourly / daily or monthly averages?
- L 365: remove “on”
- L 369: Sentence is a bit repetitive (“to outperform the Peatland-VU model”) – please check
- Section 3.1: Before presenting the results of the sensitivity work, the authors might consider presenting the basic behaviour of the model, which is now only presented in Sections 3.3 and 3.4. It is hard to relate to the sensitivities without a good understanding of the overall seasonal and interannual variations.
- L 376: Explain what is meant here with “belowground CO2 emissions”. Emissions from belowground pools to the atmosphere? And does this mean that the uptake of CO2 by plants is not accounted for in the results?
- L 381 and other occurrences: Replace “warming air temperatures” with “warming”, “rising air temperatures” or “increasing air temperatures”
- L 382: Put the reference to Figure 4 in brackets
- L 393: What does the “(5)” refer to? Figure 5?
- Fig. 2 and 3: Clarify what the individual points are presenting here. Annual means for each of the 30 years? Also, make sure to add a time unit to the CH4 and CO2 fluxes on the axis (I guess g CH4/y)
- L 422ff: The results present a number of additional simulations that have not been introduced before. It would be good to add the meaning of these in the methods section. In particular the PVN_FPAR_CONST is not intuitive to understand.
- Fig. 4: What do the lines display in these figures, are these means for all years? Also, it is unclear what the “x0.15” in the harvests indicates. Is this the cutting height?
- L 459: Add “and” after CO2.
- L 490: These are not histograms, as they do not display a frequency distribution. You can call them bar graphs or similar (but it is not necessary to describe the graph type in the text)
- L 536: remove apostrophe after “series”
- L 572: Clarify what data frequency the R2 values are based on. I guess this is based on the daily data? There would be too few years to compute statistics based on the annual numbers, I suppose.
- L 580: add “of” after “role”
- L 622: write out “AC”
- L 635: replace “limits” with “range” (or similar)
- L 642: explain the parameters used here (MethanePType, LeafRespirationCoeff, BiomassSenescence) to clarify what they describe and how they impact the simulation results
- L 660: The last sentence seems to be unfinished – it would be nice and relevant to compare your work with the studies mentioned here.
- L 666: Explain the “order” – are these ordered by relative decomposition rates?
- L 676: replace “large” with “thick”
- L 709: clarify which uncertainty is referred to here. Uncertainty in the simulated CO2 fluxes?
- L 726: I appreciate the comparison with the PEATBOG model. Maybe you can add more here than just the net annual GHG emissions? As this model also has dynamic PFTs, how does this influence their simulations? And what do you expect when comparing the Mer Bleue site to your sites?
Citation: https://doi.org/10.5194/gmd-2023-48-RC1 -
AC1: 'Reply on RC1', Tanya Lippmann, 23 May 2023
Dear Reviewer,
Thank you for taking the time to review this manuscript. We have used this feedback to correct and better explain several equations. In the revised manuscript, we will better explain model technicalities such as the model spin up, soil profile and the calibration process. Please find our responses to your comments below.
Peatland-VU-NUCOM (PVN 1.0): Using dynamic PFTs to model peatland vegetation, CH4 and CO2 emissions
The manuscript by Tanya Lippmann et al. provides a comprehensive description of the implementation of vegetation dynamics into the Peatland-VU model, based on the vegetation dynamics used in NUCOM-BOG. The manuscript documents the representation of vegetation dynamics that is used, and compares the new model with observations of greenhouse gas fluxes, observed cover fractions, and with model simulations with the unaltered Peatland-VU model.
I appreciate the detailed documentation of the equations that the model is based on, and I consider this well suited for GMD. In a few occasions, the rationale of these equations could be explained better – these are mentioned in the remarks below. Also, the simulation setup and driving data (frequency, timestep) ought to be described in more detail (see below). The results section is relatively straight-forward and provides first and foremost an overview of the basic behaviour for two Dutch sites, but I think that this is suitable for a paper that first and foremost documents a new model.
I consider this manuscript suitable for publication in GMD once these omissions have been corrected. I provide a list of detailed remarks below, which I hope can help the authors when revising the manuscript.
Response: Thank you taking the time to review our work. We will expand on the rationale used when choosing equations, as well as the input data, and simulation setup. Please see our responses below.
Major remarks
L 29: “Between 2005 and 2008, …” Please check this sentence – it is unclear which of the two studies the citation originates from, and the natural emissions that contributed 50% to the total CH4 budget are not all originating from peatlands, so the relevance is not clear here.
Response: We will clarify and correct this.
L 122: Please explain how the initialization of the carbon pools is done. Is this based on observations, or does the model require a spinup of some sort?
Response: We will create a new subsection within the methodology to explain the model spin-up and initialization of carbon pools.
L 137-152, Eqs. 2-3: It would be nice if you could explain the competition for light a bit further. The equations for plant height (Eqs. 2-3) originate from applications to trees (with D being the stem diameter). Do you apply these equations to the grass vegetation from your sites? It would be good to assess how accurate these equations are for that purpose. Also, the description seems to indicate that all leaf area is assumed to be accessible for light if the PFT is highest, whereas the grassland vegetation you have at the sites will have leaf area distributed between the soil and the top of the plants (meaning that only part of the highest plant will receive light without shading by others, and part of the leaves may well be shaded). It would be good to see the rationale explained a bit further here.
Response: A similar point was also raised by the second reviewer. The purpose for calculating height in the model is to sort the plants in descending height order to replicate shading. We understand that this relationship was not intended to be used for grass, sedges etc and whilst we do not have daily height measurements, heights estimated using this relationship are within the range of in situ observed heights and heights recorded in the literature (Chapin et al., 1996; Kattge et al., 2020). We are aware of one other allometric height relationship that has previously been used for grass, sedge and tree species within the JULES model (Harper et al., 2018). The JULES relationship is dependent on several additional parameters, and its use would introduce new sources of uncertainty into the model. Therefore, we opted for a simple allometric relationship (Eq 2 and Eq 3 of the manuscript). We will compare and evaluate the results of the height relationship with the height relationship used in JULES model, and if this impacts the sorting of PFT height, we will incorporate the allometric relationship from the JULES model into the PVN model.
It is an interesting point to consider that plants in the ‘top canopy’ may be partially overlapped/shaded by plants in lower canopies. However, we do not have observations to verify to what extent this is happening in these systems, or how this might be represented within a process-based model.
L 159: The “harvest scheme” should probably be explained further: It is described only later in the manuscript that the model is made to describe dynamics of two grassland sites, and “harvesting” will hence be a grass cut. Maybe it would be good to introduce the sites and the typical land use of these earlier in the manuscript. The concept of a “harvest height” makes good sense once one understands that these are grasslands.
Response: We agree with you that the introduction of the harvest scheme can be explained in more detail. We will give it a separate section.
L 165, Eq. 7: I cannot understand the denominator in this equation. CB*SLA would result in a leaf area index, but it is subsequently divided by a long term including amongst others the growth rate of LAI. Please explain what this equation represents.
Response: The second reviewer raised a similar comment. We agree that this equation should be solved differently. In equation 7, the denominator relates water availability with plant growth but this is better placed in a separate equation and does not belong here. We will amend the manuscript so that the denominator becomes its own equation that relates changes in plant growth and biomass to changes in water availability. We will explain the motivation, meaning and terms of the new equation in the revised manuscript.
Section 2.5 or 2.7: I would welcome more details on the description of the simulation setup. How many years are the simulations (30, I think), what is the temporal frequency of the input data (daily?) and the time step of the model? Is there a spinup, or how is the model initialization done? What is the vertical resolution/how many soil layers are simulated? Also, this would be a good place to introduce the different simulations that are undertaken: “Standard” simulations with the observed input data, and simulations to derive model sensitivities (Table 3).
Response: As we proposed in an earlier response, we will create a new subsection to explain the model spin-up where we will also define the timestep, length of the simulations, and the soil profile. We will define the ‘standard’ simulations in the Model Calibration section and refer back to these simulations when describing the testing of the model in section 2.5.
Minor remarks
We would like to thank the reviewer for making the minor remarks. We will amend the revised manuscript accordingly.
Sincerely,
Tanya J.R. Lippmann et al.
References
Chapin, F. S., Bret-Harte, M. S., Hobbie, S. E. and Zhong, H.: Plant functional types as predictors of transient responses of arctic vegetation to global change, J. Veg. Sci., 7(3), 347–358, doi:10.2307/3236278, 1996.
Harper, A. B., Wiltshire, A. J., Cox, P. M., Friedlingstein, P., Jones, C. D., Mercado, L. M., Sitch, S., Williams, K. and Duran-Rojas, C.: Vegetation distribution and terrestrial carbon cycle in a carbon cycle configuration of JULES4.6 with new plant functional types, Geosci. Model Dev., 11(7), 2857–2873, doi:10.5194/gmd-11-2857-2018, 2018.
Kattge, J., Bönisch, G., Díaz, S., Lavorel, S., Prentice, I. C., Leadley, P., Tautenhahn, S., Werner, G. D. A., Aakala, T., Abedi, M., Acosta, A. T. R., Adamidis, G. C., Adamson, K., Aiba, M., Albert, C. H., Alcántara, J. M., Alcázar C, C., Aleixo, I., Ali, H., Amiaud, B., Ammer, C., Amoroso, M. M., Anand, M., Anderson, C., Anten, N., Antos, J., Apgaua, D. M. G., Ashman, T. L., Asmara, D. H., Asner, G. P., Aspinwall, M., Atkin, O., Aubin, I., Baastrup-Spohr, L., Bahalkeh, K., Bahn, M., Baker, T., Baker, W. J., Bakker, J. P., Baldocchi, D., Baltzer, J., Banerjee, A., Baranger, A., Barlow, J., Barneche, D. R., Baruch, Z., Bastianelli, D., Battles, J., Bauerle, W., Bauters, M., Bazzato, E., Beckmann, M., Beeckman, H., Beierkuhnlein, C., Bekker, R., Belfry, G., Belluau, M., Beloiu, M., Benavides, R., Benomar, L., Berdugo-Lattke, M. L., Berenguer, E., Bergamin, R., Bergmann, J., Bergmann Carlucci, M., Berner, L., Bernhardt-Römermann, M., Bigler, C., Bjorkman, A. D., Blackman, C., Blanco, C., Blonder, B., Blumenthal, D., Bocanegra-González, K. T., Boeckx, P., Bohlman, S., Böhning-Gaese, K., Boisvert-Marsh, L., Bond, W., Bond-Lamberty, B., Boom, A., Boonman, C. C. F., Bordin, K., Boughton, E. H., Boukili, V., Bowman, D. M. J. S., Bravo, S., Brendel, M. R., Broadley, M. R., Brown, K. A., Bruelheide, H., Brumnich, F., Bruun, H. H., Bruy, D., Buchanan, S. W., Bucher, S. F., Buchmann, N., Buitenwerf, R., Bunker, D. E., et al.: TRY plant trait database – enhanced coverage and open access, Glob. Chang. Biol., 26(1), 119–188, doi:10.1111/gcb.14904, 2020.
Citation: https://doi.org/10.5194/gmd-2023-48-AC1
-
RC2: 'Comment on gmd-2023-48', Anonymous Referee #2, 28 Apr 2023
The manuscript by Lippmann et al. describes the new Peatland-VU-NUCOM (PVN) model. The new model has been assembled from two parent models, the NUCOM-BOG model and the Peatland-VU model. The specific aim of the new model has been to better represent vegetation dynamics and implement different plant functional types (PFTs) to improve simulated fluxes of CO2 and CH4. The authors then use the model implementation to simulate CH2 and CH4 fluxes and vegetation dynamics in two wetland sites. A substantial sensitivity analysis is also carried out, which I appreciate.
The authors’ objective is sound, and the model does include a few new and novel improvements which could be a welcome addition to the scientific literature. Mainly, the inclusion of mosses and reeds (typha) is interesting. The model also includes different formulations of CH4 dynamics compared to the widely used scheme by Wania et al., which is interesting.
The text is generally well-written with clear language. The model description mostly provides good insight into the framework and theory of the model. However, I believe that the text would benefit largely from a major restructuring. Much of the description of vegetation dynamics is scattered over large parts of the text instead of following a clear line of presentation. An example of such a structure would be to start with the plant carbon source (photosynthesis), then move on to tissue turnover (mostly referred to as senescence in the text), and allocation of each PFT. Then follow up with competition and the general path/column vegetation dynamics. The implications of the vegetation dynamics for the carbon fluxes could then be described. Variables or functions included in equations sometimes lack descriptions in the text or are not mentioned. I understand that these formulations might be described in other sources, but they should be properly referred to so the reader can easily look them up.
Lastly, and most importantly, some of the assumptions and equations described in the manuscript are erroneous and/or poorly described. I will describe these in detail below, however, due to the severity of these errors I feel that the theoretical framework is not yet ready for publication. I thus recommend the editor reject the manuscript in its current form but encourage a re-submission once these errors have been sorted out.
I have further divided my comments into three sections. One deal with the model description, the second deals with major comments on the other parts of the manuscript, and finally technical notes.
Comments on the model description
The allometric equations used to partition the carbon are developed and parameterised for trees. The D in the model is the diameter of the tree and BD is the wood density of the tree. It is difficult – or even impossible – to transfer these allometric relationships to grasses, and even more so to mosses. The authors better motivate how this adaptation could be done and why these are valid assumptions. I believe that this assumption needs to be completely revised and filled with scientifically more sound parameterisations.
FPAR, which is calculated in eq. 4, is usually calculated only as the Lambert–Beer law (i.e., e-k * LAI). This could then be integrated over the canopy. I believe that is the intention with the latter formulation of CB * SLA in eq. 4. I do not understand why the plant's total biomass (CB) should be included in this calculation. A more reasonable approach would be to use LAI or leaf biomass * SLA. The authors need to motivate why total plant biomass is a good variable to include in the calculation of the fraction of absorbed light.
I am a bit confused by the formulations in equations 5 and 6. Does the senescence of leaf material happen daily? This sounds like a biomass turnover to me, which should be subtracted from GPP (photosynthesis) to obtain NPP. In that way, the turnover is double counted when SM is substituted into equation 5. Furthermore, NPP is a rate. This means that in eq. 6 the variable SM would also be a rate where RS will represent the partitioning of new biomass into the aboveground carbon pool. Since eq. 5 (and other functions) uses the dx/dt approach to denote change, SM in eq. 6 erroneously appears to be a state variable.
LAI is commonly defined as just leaf mass * SLA (mass/area). The addition of a light extinction coefficient and water growth factor with the change in LAI makes no sense to me and is not explained further in the text. LAI is a state variable in the model and its temporal development is determined by the development of the leaf mass, sometimes modified by a phenology factor. If the authors make a new formulation for LAI it should be clearly derived in the text.
A dynamic thickness of a moss layer is an interesting feature of the model and could potentially be of interest to other models and the vegetation modelling community. I appreciate the authors' work on this subject but have a few comments that I believe will strengthen the manuscript. First, I believe that this subject is not very well known in the broader modelling community and the theory could thus be expanded somewhat, perhaps even moved to its own section. Secondly, I am not sure why Gmax is used instead of NPP, please provide a justification for this in the text. The parameter DBD might potentially be important, and it would be good with a sentence or reference to how this value is constrained by data. Lastly, I cannot find any other mention of moss thickness in the manuscript. A sentence or two about how this is used in the would be good. There are several places where this could improve the model, but it is not stated.
Please provide the equation for TG as well. It is not obvious to me what temperature parameters are congruent to WL. Lastly, please consider reformatting the equation into the commonly used grouping by curly braces.
Eq 13. Seems to have a few sign errors. The production of CH4 should of course be positive while plant transport, CH4 oxidation, and ebullition are losses of CH4 from the layer and should have a negative sign but are now positive. The diffusive flux could be either negative (more CH4 leaves the layer than what is added from lower layers) or positive, its sign is thus determined by what direction the flow of CH4 has and the sign convention used. Furthermore, diffusive fluxes generally need to be solved numerically in computer programs. Please include a sentence on how this flux is calculated.
The formulation of plant transport of CH4 described in Eq 14 and 15 contains at least four PFT-specific parameters. This would yield a large parameter space for calibration. Please described how these parameters are constrained by data or theory.
Eq 16 seems to have the wrong sign for the integral of BCO2.
The photosynthesis algorithm used is derived from (Haxeltine et al., 1996) which is a general light-use efficiency model. The potential photosynthesis (AP) described in eq. 17 has however been modified with the addition of a temperature function and instantaneous photosynthesis (AI). While I can see the value of an additional temperature modifier, it is difficult to understand what the AI parameter adds. The modification is also not derived or referenced in the text. Please also provide a reference for the temperature function.
The description of root exudates in eq 23-26 seems mostly ok. However, I think the same issue with double counting of the turnover is present here. It is an interesting feature with the spring addition of exudates. The manuscript would however benefit from describing the theory behind this more as well as how the parameters are constrained and how the data for this is obtained. In Eq 26 the growth of RM seems unnecessary to describe the addition of new root biomass.
Finally, I would recommend the authors add mass balance to the code to make sure that carbon is not lost through double counting.
Major comments
Note that LPJ-WhyMe by Wania et al. (2010) and LPJ-GUESS (Smith et al., 2014) are not the same model. Especially the version used by Chaudhary et al. (2020) and earlier work has an advanced peat accumulation scheme, dynamic vegetation and CH4 dynamics. The statements on lines 55-58 need to acknowledge this.
Please provide a table of all the model inputs and their units that are required. I have a bit of difficulty understanding what might be simulated dynamically within the model and what is prescribed through input. This could be added to the supplementary material.
Under section 2.2 it is stated that the PFTs have bioclimatic limits. Why include such limits in a site-based model where the PFTs are prescribed? Furthermore, none of the included PFTs is strictly evergreen. Please consider changing this language to say ‘lifeform’ or similar.
I believe that Tables 1 and 2 could be merged. This would provide a better overview of the parameters for each PFT and the reader would not need to know in which table to look for e.g., units. Also, please ensure that parameter names are uniformly used in the text. For instance, Gmax is referred to as both its parameter name, ‘maximum growth’, and Gmax in the equation.
Why start the model description with how the model is initiated? This could be done later after the description of the equations. Please also include a section on model spinup including what forcing data that is used and what steady-state condition you have (or the number of years that you run the spinup).
Please move the description of the harvest scheme from section 2.2.2 (competition among PFTs) into its own section. Also, the statement on line 161 that biomass is uniformly distributed is wrong since you use the allometric equations by Smith et al. (2001).
For the soil model in section 2.3.2 it would be interesting with either a conceptual figure or a table describing how the pools are related to each other. Also, the k-values for each pool should be provided.
The model calibration is well described and follows standard protocols. However, I would like a few sentences on what data was used for calibration. Was the dataset split up into a calibration and a test set or were all data used for model calibration? Also, what parameters were used in the calibration and how were these selected?
The model sensitivity test is also well-designed and performed. This scheme should be kept for future testing.
Technical comments
General
There is sometimes the use of parenthesis when referring to the opposite. For example lines 139-140 “… and increase (decrease) of foliage of taller PFTs …” This is a bit confusing to me, please consider using other expressions such as ‘vice versa’ or similar.
The placement of some references feels odd, for example on line 59. Please review this.
On lines 101-103 the model time-step and simulation duration seems a bit confused. Please consider a different formulation that clarifies this.
In Figure 1, are the different background pictures intended to represent different types of peatlands? In such a case, please add a description/statement of this and how it is reflected in the flowchart.
Please consider using the capital letter delta in differential equations.
Please consider changing the heading of section 2.2.1. It mostly describes soil carbon pools. If these are inherited from a parent model, please also provide the reference for this.
The RD function mentioned on line 204 is not described in detail. Please reference the source of this function
Specific
18 Please add a reference for the first statement.
36 GHG has not been used before in the text, please spell out
58 Table Fig. S1? Please clarify what you refer to here, table/figure
74 plot-scale ecosystem competition model
75 process-based plot-scale peatland model
116 Is threshold the correct word here? Should it not be range?
120 Exudates seems to appear twice in this sentence
123 DBD has not been used before in the text, please spell out
124 The number of what? I assume soil layers but the word is missing
128-129 This introduction of subscripts is done later on as well. One is enough.
236 I believe the word formation should be used instead of storage.
246 I believe NPP is intended instead of NEE.
265-269 This text should be moved to the acknowledgement or data statement.
References
Chaudhary, N., Westermann, S., Lamba, S., Shurpali, N., Sannel, A. B. K., Schurgers, G., Miller, P. A., & Smith, B. (2020). Modelling past and future peatland carbon dynamics across the pan-Arctic. Glob Chang Biol, 26(7), 4119-4133. https://doi.org/10.1111/gcb.15099
Haxeltine, A., Prentice, I. C., & Creswell, I. D. (1996). A coupled carbon and water flux model to predict vegetation structure. Journal of Vegetation Science, 7(5), 651-666. https://doi.org/10.2307/3236377
Smith, B., Prentice, I. C., & Sykes, M. T. (2001). Representation of vegetation dynamics in the modelling of terrestrial ecosystems: comparing two contrasting approaches within European climate space. Global Ecology and Biogeography, 10(6), 621-637. https://doi.org/10.1046/j.1466-822X.2001.t01-1-00256.x
Smith, B., Wårlind, D., Arneth, A., Hickler, T., Leadley, P., Siltberg, J., & Zaehle, S. (2014). Implications of incorporating N cycling and N limitations on primary production in an individual-based dynamic vegetation model. Biogeosciences, 11(7), 2027-2054. https://doi.org/10.5194/bg-11-2027-2014
Wania, R., Ross, I., & Prentice, I. C. (2010). Implementation and evaluation of a new methane model within a dynamic global vegetation model: LPJ-WHyMe v1.3.1. Geoscientific Model Development, 3(2), 565-584. https://doi.org/10.5194/gmd-3-565-2010
Citation: https://doi.org/10.5194/gmd-2023-48-RC2 -
AC2: 'Reply on RC2', Tanya Lippmann, 23 May 2023
Dear Reviewer,
Thank you for taking the time to review this manuscript. We are appreciative of your useful suggestions and comments. We have used this feedback to correct or better explain several equations. In the revised manuscript, we clearly introduce key peatland processes. We will also improve the clarity of the manuscript by merging Table 1 and Table 2 and providing a schematic of the different SOM pools. Please find our responses to your comments below.
Original comment:
The manuscript by Lippmann et al. describes the new Peatland-VU-NUCOM (PVN) model. The new model has been assembled from two parent models, the NUCOM-BOG model and the Peatland-VU model. The specific aim of the new model has been to better represent vegetation dynamics and implement different plant functional types (PFTs) to improve simulated fluxes of CO2 and CH4. The authors then use the model implementation to simulate CH2 and CH4 fluxes and vegetation dynamics in two wetland sites. A substantial sensitivity analysis is also carried out, which I appreciate.
The authors’ objective is sound, and the model does include a few new and novel improvements which could be a welcome addition to the scientific literature. Mainly, the inclusion of mosses and reeds (typha) is interesting. The model also includes different formulations of CH4 dynamics compared to the widely used scheme by Wania et al., which is interesting.
The text is generally well-written with clear language. The model description mostly provides good insight into the framework and theory of the model. However, I believe that the text would benefit largely from a major restructuring. Much of the description of vegetation dynamics is scattered over large parts of the text instead of following a clear line of presentation. An example of such a structure would be to start with the plant carbon source (photosynthesis), then move on to tissue turnover (mostly referred to as senescence in the text), and allocation of each PFT. Then follow up with competition and the general path/column vegetation dynamics. The implications of the vegetation dynamics for the carbon fluxes could then be described. Variables or functions included in equations sometimes lack descriptions in the text or are not mentioned. I understand that these formulations might be described in other sources, but they should be properly referred to so the reader can easily look them up.
Lastly, and most importantly, some of the assumptions and equations described in the manuscript are erroneous and/or poorly described. I will describe these in detail below, however, due to the severity of these errors I feel that the theoretical framework is not yet ready for publication. I thus recommend the editor reject the manuscript in its current form but encourage a re-submission once these errors have been sorted out.
I have further divided my comments into three sections. One deal with the model description, the second deals with major comments on the other parts of the manuscript, and finally technical notes.
Response: We thank the reviewer for their constructive comments and particularly, for their close revision of the equations, terms, and units. We understand the rationale to begin with the carbon source (photosynthesis), and move to plant turn over, methane dynamics, and finally to vegetation competition. We will revise the sequence of the methods section accordingly. Please see our replies below.
Comments on the model description
Original comment:
The allometric equations used to partition the carbon are developed and parameterised for trees. The D in the model is the diameter of the tree and BD is the wood density of the tree. It is difficult – or even impossible – to transfer these allometric relationships to grasses, and even more so to mosses. The authors better motivate how this adaptation could be done and why these are valid assumptions. I believe that this assumption needs to be completely revised and filled with scientifically more sound parameterisations.
Response: Thank you for raising this point and providing an interesting discussion. A similar point was also raised by reviewer 1. The purpose for calculating height in the model is to sort PFTs in descending height order to estimate shading. Please note that this height relationship is not used for moss species because moss species are always considered to be the lowest plants, at surface level. We will amend the manuscript to make it clear that this relationship is not used for moss PFTs. We have revised the allometric height equation and can confirm that there was an error in the writing of constants, k1, k2, and k3. Equation 2 should read:
H = k2 * D ^ (K3)
where, k2 and k3 are equal to 40 and 0.85, respectively (Smith et al., 2001).
We understand that this relationship was not initially intended to be used for grass, sedges etc and whilst we do not have daily height measurements, heights estimated using this relationship are within the range of in situ observed heights and heights recorded in the literature (Chapin et al., 1996; Kattge et al., 2020). We are aware of one other allometric height relationship that has previously been used for grass, sedge and tree species within the JULES model (Harper et al., 2018). The JULES relationship is dependent on several additional parameters, and its use would introduce new sources of uncertainty into the model. Therefore, we opted for a simple allometric relationship (Eq 2 and Eq 3 of the manuscript). We will compare and evaluate the results of the height relationship with the height relationship used in JULES model, and if this impacts the sorting of PFT height, we will incorporate the allometric relationship from the JULES model into the PVN model.
Original comment: FPAR, which is calculated in eq. 4, is usually calculated only as the Lambert–Beer law (i.e., e-k * LAI). This could then be integrated over the canopy. I believe that is the intention with the latter formulation of CB * SLA in eq. 4. I do not understand why the plant's total biomass (CB) should be included in this calculation. A more reasonable approach would be to use LAI or leaf biomass * SLA. The authors need to motivate why total plant biomass is a good variable to include in the calculation of the fraction of absorbed light.
Response: We apologise that these parentheses were placed erroneously and thank you for picking up on this. This equation should read:
FPAR = 1 - e^(-LEC * CB * SLA)
We will amend this equation in the final manuscript. This approach is an application of Lambert-Beer's law taken from (Heijmans et al., 2008), which is aligned with previous applications of the relationship (Huisman and Olff, 1998; Prentice et al., 1993).
Original comment: I am a bit confused by the formulations in equations 5 and 6. Does the senescence of leaf material happen daily? This sounds like a biomass turnover to me, which should be subtracted from GPP (photosynthesis) to obtain NPP. In that way, the turnover is double counted when SM is substituted into equation 5. Furthermore, NPP is a rate. This means that in eq. 6 the variable SM would also be a rate where RS will represent the partitioning of new biomass into the aboveground carbon pool. Since eq. 5 (and other functions) uses the dx/dt approach to denote change, SM in eq. 6 erroneously appears to be a state variable.
Response: We agree with you that the units of NPP are Kg C day-1. We will correct this in the revised manuscript. This equation is also missing a term. Equation 5 should read:
d(CB)/dt= SM – BS * CB
where the units of SM and NPP (Eq 6) are both Kg C m-2 day-1.
We believe this removes the indication of double accounting.
Original comment: LAI is commonly defined as just leaf mass * SLA (mass/area). The addition of a light extinction coefficient and water growth factor with the change in LAI makes no sense to me and is not explained further in the text. LAI is a state variable in the model and its temporal development is determined by the development of the leaf mass, sometimes modified by a phenology factor. If the authors make a new formulation for LAI it should be clearly derived in the text.
Response: The first reviewer raised a similar comment. We agree that this equation should be solved differently. In equation 7, the denominator relates water availability with plant growth. This is better placed in a separate equation and does not belong in equation 7. We will amend the manuscript so that the denominator becomes its own equation that relates changes in plant growth and biomass to changes in water availability. We will explain the motivation, meaning and terms of the new equation in the revised manuscript.
Original comment: A dynamic thickness of a moss layer is an interesting feature of the model and could potentially be of interest to other models and the vegetation modelling community. I appreciate the authors' work on this subject but have a few comments that I believe will strengthen the manuscript. First, I believe that this subject is not very well known in the broader modelling community and the theory could thus be expanded somewhat, perhaps even moved to its own section. Secondly, I am not sure why Gmax is used instead of NPP, please provide a justification for this in the text. The parameter DBD might potentially be important, and it would be good with a sentence or reference to how this value is constrained by data. Lastly, I cannot find any other mention of moss thickness in the manuscript. A sentence or two about how this is used in the would be good. There are several places where this could improve the model, but it is not stated.
Response: Thank you for your interest in this feature. Gmax is used in the calculation of potential growth by the NUCOM model (Heijmans et al., 2008). We will provide a more detailed introduction to the dynamic thickness of a moss layer in the revised manuscript. We will explain how dry bulk density is constrained within the model. Unfortunately, the thickness of the moss layer is not yet used by the model. A next step is to recalculate surface height and also the soil properties (DBD, pH, OM content of top soil layer(s).) However, such an undertaking is beyond the aims of this study (introducing PFTs into the model) but we hope that we have laid the foundations to allow a dynamic relationship between moss thickness, soil profile, and surface height in future model versions.
Original comment: Please provide the equation for TG as well. It is not obvious to me what temperature parameters are congruent to WL. Lastly, please consider reformatting the equation into the commonly used grouping by curly braces.
Response: We will amend the manuscript to include the equation for TG.
Original comment: Eq 13. Seems to have a few sign errors. The production of CH4 should of course be positive while plant transport, CH4 oxidation, and ebullition are losses of CH4 from the layer and should have a negative sign but are now positive. The diffusive flux could be either negative (more CH4 leaves the layer than what is added from lower layers) or positive, its sign is thus determined by what direction the flow of CH4 has and the sign convention used. Furthermore, diffusive fluxes generally need to be solved numerically in computer programs. Please include a sentence on how this flux is calculated.
Response: We agree and will correct the +/- signs in this equation.
Original comment: The formulation of plant transport of CH4 described in Eq 14 and 15 contains at least four PFT-specific parameters. This would yield a large parameter space for calibration. Please described how these parameters are constrained by data or theory.
Response: LAI is calculated by a model function, represented using equation 17. Plant parameters, vP (MethanePType_PFT) and PlOx (MethanePlantOx_PFT) are described in the PFT table with references. We will provide an intext reference to this. Whilst, we provide equations for the calculation of root mass, and root distribution, we will amend the manuscript to include the equation that defines root density. In section 2.4.1 PFT attributes, we will amend the manuscript to provide more detail on how these parameters are constrained by data and theory. We highlight the consequences of parameter uncertainty in the discussion (4.1 Sources of uncertainty). The introduction of PFTs allowed several Peatland-VU parameters that were previously calibratable to become observation-informed parameters, thereby reducing the parameter space. We will list these in section 2.4.1 PFT attributes and also include reference to this in the discussion (section 4.1.1 Input parameters).
Original comment: Eq 16 seems to have the wrong sign for the integral of BCO2.
Response: We agree and will correct this in the revised manuscript.
Original comment: The photosynthesis algorithm used is derived from (Haxeltine et al., 1996) which is a general light-use efficiency model. The potential photosynthesis (AP) described in eq. 17 has however been modified with the addition of a temperature function and instantaneous photosynthesis (AI). While I can see the value of an additional temperature modifier, it is difficult to understand what the AI parameter adds. The modification is also not derived or referenced in the text. Please also provide a reference for the temperature function.
Response: We appreciate this opportunity to reflect on our application of the carbon-flux model. We appreciate the reviewer’s close assessment of these equations and apologise for any confusion. ‘PAR’ is an erroneous term in this equation. Equation 17 should now read:
AP = FPAR * sigma * phi * AI *FG
Furthermore, RE in equation 19 is erroneous and needs to be replaced by,
a* Amax,
where ‘a’ is constant (also in Eq. 18), 0.08, taken from Haxeltine et al. and the LHS of Eq 20 should read, RT. We will amend these equations and their descriptions in the revised manuscript.
AI is defined in equation 19 and originally taken from equation 1 (Haxeltine et al., 1996). In the original Haxeltine et al. paper, a carbon-flux model is used to calculate NPP for specific plant types. Monthly potential photosynthesis (similar to daily potential photosynthesis or eq 17, in our manuscript) is represented by Apin Eq. 3 of (Haxeltine et al., 1996)). Actual monthly photosynthesis (Eq 5 in Haxeltine et al) is calculated using monthly potential gross photosynthesis (Eq 3 in Haxeltine et al) and environmental temperature scalar, phiT. Instead of writing monthly potential gross photosynthesis (eq3 in Haxeltine et al) as a separate equation, we have used the RHS of eq3 in Haxeltine et al. Therefore, our equation for daily potential photosynthesis (Eq 17) is equivalent to the original equation for actual monthly photosynthesis (Eq 5 in Haxeltine et al). Whereas, the original model converts between instantaneous photosynthesis and monthly GPP, we convert instantaneous photosynthesis to daily GPP.
Original comment: The description of root exudates in eq 23-26 seems mostly ok. However, I think the same issue with double counting of the turnover is present here. It is an interesting feature with the spring addition of exudates. The manuscript would however benefit from describing the theory behind this more as well as how the parameters are constrained and how the data for this is obtained. In Eq 26 the growth of RM seems unnecessary to describe the addition of new root biomass.
Response: Thank you for highlighting this equation. We are happy to amend the manuscript to elaborate on the role of root exudates. These equations were taken from the Peatland-VU model (van Huissteden et al., 2006). As we mentioned in a previous response, we will amend section 2.4.1, PFT attributes, to provide more detail on how these parameters are constrained by data and theory.
Equation 26 should read:
Rd = (1-RS) * NPP
Where, RM is omitted. We believe this removes the indication of double accounting.
Original comment: Finally, I would recommend the authors add mass balance to the code to make sure that carbon is not lost through double counting.
Response: We think this is a good idea and whilst formulating the model, we checked the carbon mass balance by combining the results of all carbon pools and stores. Unfortunately, this check was not included within the model code. We are working on incorporating this into the model code in future versions.
Major comments
Original comment: Note that LPJ-WhyMe by Wania et al. (2010) and LPJ-GUESS (Smith et al., 2014) are not the same model. Especially the version used by Chaudhary et al. (2020) and earlier work has an advanced peat accumulation scheme, dynamic vegetation and CH4 dynamics. The statements on lines 55-58 need to acknowledge this.
Response: We are aware of these differences. Table S1 lists these models separately with their appropriate references. We will make sure that the introduction clearly differentiates these models.
Original comment: Please provide a table of all the model inputs and their units that are required. I have a bit of difficulty understanding what might be simulated dynamically within the model and what is prescribed through input. This could be added to the supplementary material.
Response: We are happy to do this.
Original comment: Under section 2.2 it is stated that the PFTs have bioclimatic limits. Why include such limits in a site-based model where the PFTs are prescribed? Furthermore, none of the included PFTs is strictly evergreen. Please consider changing this language to say ‘lifeform’ or similar.
Response: These bioclimatic limits do not determine geographic extent but are used by the photosynthesis function (adapted from (Haxeltine et al., 1996)), and the potential growth function (adapted from (Heijmans et al., 2008)). It is unclear what the reviewer means when they say these PFTs are not strictly evergreen. It is unclear whether the reviewer means that some PFTs can be either deciduous or evergreen or does the reviewer mean that the evergreen PFTs do litter some (but not all) leaves throughout the year. In the case of the former, PFTs can be re-defined as evergreen or deciduous in the PFT input parameter file. In the case of the latter, it is typical for evergreens to regularly lose old leaves. We will amend the text to mention that evergreen PFTs lose (old) leaves, described in Eq 5.
Original comment: I believe that Tables 1 and 2 could be merged. This would provide a better overview of the parameters for each PFT and the reader would not need to know in which table to look for e.g., units. Also, please ensure that parameter names are uniformly used in the text. For instance, Gmax is referred to as both its parameter name, ‘maximum growth’, and Gmax in the equation.
Response: We will merge Table 1 and Table 2 into a single table. We will make sure to use parameter names in text and will also reference the parameter name in the descriptions of equations, where relevant.
Original comment: Why start the model description with how the model is initiated? This could be done later after the description of the equations. Please also include a section on model spinup including what forcing data that is used and what steady-state condition you have (or the number of years that you run the spinup).
Response: We are happy to re-order the methods section. We will to include more details regarding the model spin-up, the forcing data and the length of the simulations.
Original comment: Please move the description of the harvest scheme from section 2.2.2 (competition among PFTs) into its own section. Also, the statement on line 161 that biomass is uniformly distributed is wrong since you use the allometric equations by Smith et al. (2001).
Response: We are happy to move the harvest scheme into its own section. What we meant by ‘biomass is uniformly distributed’ is that the aboveground living biomass is not separated into separate organs. We will amend the manuscript to make this clearer.
Original comment: For the soil model in section 2.3.2 it would be interesting with either a conceptual figure or a table describing how the pools are related to each other. Also, the k-values for each pool should be provided.
Response: The k values are site specific input parameters that will be included in the table of site parameters in the supplementary material, on your suggestion. We are happy to describe the SOM pools using a table or schematic, possibly in the supplementary material.
Original comment: The model calibration is well described and follows standard protocols. However, I would like a few sentences on what data was used for calibration. Was the dataset split up into a calibration and a test set or were all data used for model calibration? Also, what parameters were used in the calibration and how were these selected?
Response: We will provide these details in the revised manuscript.
Original comment: The model sensitivity test is also well-designed and performed. This scheme should be kept for future testing.
Response: Thank you for this feedback.
We hope that these responses have resolved your concerns.
Sincerely,
Tanya J.R. Lippmann et al.
References
Chapin, F. S., Bret-Harte, M. S., Hobbie, S. E. and Zhong, H.: Plant functional types as predictors of transient responses of arctic vegetation to global change, J. Veg. Sci., 7(3), 347–358, doi:10.2307/3236278, 1996.
Harper, A. B., Wiltshire, A. J., Cox, P. M., Friedlingstein, P., Jones, C. D., Mercado, L. M., Sitch, S., Williams, K. and Duran-Rojas, C.: Vegetation distribution and terrestrial carbon cycle in a carbon cycle configuration of JULES4.6 with new plant functional types, Geosci. Model Dev., 11(7), 2857–2873, doi:10.5194/gmd-11-2857-2018, 2018.
Haxeltine, A., Prentice, I. C. and Creswell, I. D.: A coupled carbon and water flux model to predict vegetation structure, J. Veg. Sci., 7(5), 651–666, doi:10.2307/3236377, 1996.
Heijmans, M. M. P. D., Mauquoy, D., Van Geel, B. and Berendse, F.: Long-term effects of climate change on vegetation and carbon dynamics in peat bogs, J. Veg. Sci., 19(3), 307–320, doi:10.3170/2008-8-18368, 2008.
Huisman, J. and Olff, M.: Competition and facilitation in multispecies plant-herbivore systems of productive environments, Ecol. Lett., 1(1), 25–29, doi:10.1046/j.1461-0248.1998.00015.x, 1998.
van Huissteden, J., van den Bos, R. and Marticorena Alvarez, I.: Modelling the effect of water-table management on CO2 and CH4 fluxes from peat soils, Netherlands J. Geosci. - Geol. en Mijnb., 85(1), 3–18, doi:10.1017/S0016774600021399, 2006.
Kattge, J., Bönisch, G., Díaz, S., Lavorel, S., Prentice, I. C., Leadley, P., Tautenhahn, S., Werner, G. D. A., Aakala, T., Abedi, M., Acosta, A. T. R., Adamidis, G. C., Adamson, K., Aiba, M., Albert, C. H., Alcántara, J. M., Alcázar C, C., Aleixo, I., Ali, H., Amiaud, B., Ammer, C., Amoroso, M. M., Anand, M., Anderson, C., Anten, N., Antos, J., Apgaua, D. M. G., Ashman, T. L., Asmara, D. H., Asner, G. P., Aspinwall, M., Atkin, O., Aubin, I., Baastrup-Spohr, L., Bahalkeh, K., Bahn, M., Baker, T., Baker, W. J., Bakker, J. P., Baldocchi, D., Baltzer, J., Banerjee, A., Baranger, A., Barlow, J., Barneche, D. R., Baruch, Z., Bastianelli, D., Battles, J., Bauerle, W., Bauters, M., Bazzato, E., Beckmann, M., Beeckman, H., Beierkuhnlein, C., Bekker, R., Belfry, G., Belluau, M., Beloiu, M., Benavides, R., Benomar, L., Berdugo-Lattke, M. L., Berenguer, E., Bergamin, R., Bergmann, J., Bergmann Carlucci, M., Berner, L., Bernhardt-Römermann, M., Bigler, C., Bjorkman, A. D., Blackman, C., Blanco, C., Blonder, B., Blumenthal, D., Bocanegra-González, K. T., Boeckx, P., Bohlman, S., Böhning-Gaese, K., Boisvert-Marsh, L., Bond, W., Bond-Lamberty, B., Boom, A., Boonman, C. C. F., Bordin, K., Boughton, E. H., Boukili, V., Bowman, D. M. J. S., Bravo, S., Brendel, M. R., Broadley, M. R., Brown, K. A., Bruelheide, H., Brumnich, F., Bruun, H. H., Bruy, D., Buchanan, S. W., Bucher, S. F., Buchmann, N., Buitenwerf, R., Bunker, D. E., et al.: TRY plant trait database – enhanced coverage and open access, Glob. Chang. Biol., 26(1), 119–188, doi:10.1111/gcb.14904, 2020.
Prentice, I. C., Sykes, M. T. and Cramer, W.: A simulation model for the transient effects of climate change on forest landscapes, Ecol. Modell., 65(1–2), 51–70, doi:10.1016/0304-3800(93)90126-D, 1993.
Smith, B., Prentice, I. C. and Sykes, M. T.: Representation of vegetation dynamics in the modelling of terrestrial ecosystems: comparing two contrasting approaches within European climate space, Glob. Ecol. Biogeogr., 10(6), 621–637, doi:10.1046/j.1466-822x.2001.t01-1-00256.x, 2001.
Citation: https://doi.org/10.5194/gmd-2023-48-AC2
-
AC2: 'Reply on RC2', Tanya Lippmann, 23 May 2023
-
RC3: 'Comment on gmd-2023-48', Andrew Baird, 11 May 2023
Review of Lippmann et al.: ‘Peatland-VU-NUCOM (PVN 1.0): Using dynamic PFTs to model peatland vegetation, CH4 and CO2 emissions’ submitted to Geoscientific Model Development.
Overview
This paper reports on the coupling of a peatland vegetation dynamics model (NUCOM-BOG) with a model (Peatland-VU) that simulates CO2 and CH4 dynamics in peat soils. The paper seems suitable for the journal. The model description is reasonably clear (but see later comment). The sensitivity analysis and the assessment of model mechanisms are clearly presented, as are the model-data comparisons. Nevertheless, I do have some reservations about the paper in its current form and recommend that it undergoes reasonably substantive revision before being considered for publication. My concerns are articulated below.
Model purpose and model complexity
The authors start by saying that CO2 and CH4 fluxes to and from peatlands are strongly mediated by plants and that it is necessary to include key plant processes in peatland CO2-e flux models. Later they suggest that they weren’t expecting their new model to outperform the existing Peatland-VU model. For example, on lines 84-85 they say: “All three models (NUCOM, PeatlandVU, and PVN) depend heavily on calibration using (often limited) observational data and for this reason, we do not expect to reproduce observed CH4 and CO2 more accurately.” (see also lines 367-370). This leads the reader to question the purpose of the new model. The implication in places is that it can help guide restoration efforts, but an arguably over-complicated and over-parameterised model is perhaps not what is needed by wetland managers. The authors do discuss model equifinality and suggest the new model is less prone to this problem than Peatland-VU because it contains fewer optimisable (i.e., non-measurable) parameters. However, the new model contains many parameters/inputs, and measurements for all of these are unlikely to be available for many sites, so the problem of model equifinality must surely remain. If that’s the case, I would also question how much the new model can provide insights into the effect of plants on CO2 and CH4 fluxes in peatlands beyond what is already known through numerous observational and experimental studies. In short, I’d like to seek a stronger justification for the new model. In writing such a justification the authors may wish to comment on the importance of plants and plant dynamics relative to water table position, which we know is a first-order control on annual emissions of CO2 and CH4 from peat soils.
Model equations and explanations
I found the subsections in which the model equations are presented quite difficult to follow in a few places. As someone who is not familiar with the parent models, it would have been useful to have had more written information on the equations and their derivation or even some simple plots showing the form of the equations. Several equations also seem to contain errors – some are not dimensionally homogenous and the signs of the terms in others don’t make sense (to me, at least). I have indicated some examples in a marked-up copy of the paper supplied separately. I recommend the authors check the equations carefully to ensure they are rendered correctly and are dimensionally balanced.
Peat accumulation and water table inputs
Some aspects of the new model didn’t really make sense to me. It is not clear how net peat accumulation or loss is accounted for in the model. Can the designated soil layers expand and contract as the size of their SOM pools varies? Also, I am confused about why only mosses can apparently contribute to peat formation as suggested in Figure 1 and noted on line 236: “Non-moss PFTs do not contribute to the storage of the peat.” If the new model is to be applied over timescales of many decades, then changes in peat thickness and mass should surely be considered. Also, peat properties such as permeability are important at these timescales because they feed back into a peatland’s hydrological functioning and the position of the water table which, as noted above, is a first-order control on CO2 and CH4 fluxes.
The authors note that site water tables were provided by the Netherlands Hydrological Instrument. This doesn’t seem to be an accurate way of obtaining water tables. It would have been better to have measured water tables directly or to have used a site-specific water-table model.
I have made more comments (mostly minor) on a pdf of the paper and this is posted separately for the authors and the editor.
I operate a policy of open reviewing and ask that my name be revealed to the authors.
Andy Baird,
Chair of Wetland Science, University of Leeds, UK;
11th May 2023.
-
AC3: 'Reply on RC3', Tanya Lippmann, 23 May 2023
Dear Andrew,
Thank you for taking the time to review this manuscript. We have used this feedback to more clearly outline the motivations for developing the model and how this model may be used for wetland management. We’ve reflected on the model’s parameter space and we will carefully check all equations in the revised manuscript.
Overview
Original comment: This paper reports on the coupling of a peatland vegetation dynamics model (NUCOM-BOG) with a model (Peatland-VU) that simulates CO2 and CH4 dynamics in peat soils. The paper seems suitable for the journal. The model description is reasonably clear (but see later comment). The sensitivity analysis and the assessment of model mechanisms are clearly presented, as are the model-data comparisons. Nevertheless, I do have some reservations about the paper in its current form and recommend that it undergoes reasonably substantive revision before being considered for publication. My concerns are articulated below.
Response: We thank the reviewer for their constructive comments. Please find our replies to your comments below.
Model purpose and model complexity
Original comment: The authors start by saying that CO2 and CH4 fluxes to and from peatlands are strongly mediated by plants and that it is necessary to include key plant processes in peatland CO2-e flux models. Later they suggest that they weren’t expecting their new model to outperform the existing Peatland-VU model. For example, on lines 84-85 they say: “All three models (NUCOM, PeatlandVU, and PVN) depend heavily on calibration using (often limited) observational data and for this reason, we do not expect to reproduce observed CH4 and CO2 more accurately.” (see also lines 367-370). This leads the reader to question the purpose of the new model. The implication in places is that it can help guide restoration efforts, but an arguably over-complicated and over-parameterised model is perhaps not what is needed by wetland managers. The authors do discuss model equifinality and suggest the new model is less prone to this problem than Peatland-VU because it contains fewer optimisable (i.e., non-measurable) parameters. However, the new model contains many parameters/inputs, and measurements for all of these are unlikely to be available for many sites, so the problem of model equifinality must surely remain. If that’s the case, I would also question how much the new model can provide insights into the effect of plants on CO2 and CH4 fluxes in peatlands beyond what is already known through numerous observational and experimental studies. In short, I’d like to seek a stronger justification for the new model. In writing such a justification the authors may wish to comment on the importance of plants and plant dynamics relative to water table position, which we know is a first-order control on annual emissions of CO2 and CH4 from peat soils.
Response:
The model is capable of estimating the greenhouse gas balance in response to environmental changes (changes in temperature or radiation or water levels) or new management decisions (changes in harvest regime or vegetation management) for peatland sites. Therefore, the model can serve wetland management by estimating changes in the greenhouse gas balance of peatland sites in response to management decisions. We will highlight this in the revised manuscript. In the development of the model, we made steps to minimise the calibratable parameter space. The introduction of PFTs allowed several Peatland-VU parameters that were previously calibratable to become observation-informed parameters, whilst introducing few new parameters, thereby the net result is a reduction in the parameter space. We will list the observation-informed PFT parameters that were calibratable model parameters in the Peatland-VU model (in section 2.4.1 PFT attributes) and also refer back to this in the discussion (section 4.1.1 Input parameters).
While the effects of groundwater table on peatland GHG emissions are extensively described (Evans et al., 2021), the impacts of plant type and plant community composition on GHG emissions are less understood (Malmer et al., 2003). With this model we open up the possibility to explore the combined effects of groundwater and plant species composition on GHG emissions. Plant functional types have been found to explain uncertainties in GHG emissions from wetlands in response to warming in a meta-analysis of wetlands exposed to warming (Bao et al., 2023). Changes in vegetation composition have been observed in long running water table manipulation experiments (Peltoniemi et al., 2009; Strack et al., 2006). Generally, sedges and mosses establish during wetter conditions and shrubs and trees develop during dryer conditions. Enhanced Sphagnum growth during warming experiments have been found to outcompete shrubs (Dorrepaal et al., 2006). Changes in vegetation have been accompanied by changes in fungal and bacterial biomass (Jaatinen et al., 2008) as well as decreases in methanogenic and methanotrophic community diversity (Yrjälä et al., 2011). Following changes in plant community composition, changes to CO2 (NPP) have been observed, further impacting root exudation (Ballantyne et al., 2014). Vegetation composition change directly impact litter inputs (Malmer et al., 2005), impacting the quality and quantity of fresh SOM contributions. Peat mineralization rates were observed to decline as readily decomposable material is already mineralized (Davidson and Janssens, 2006; Dorrepaal et al., 2009). Differences in vegetation composition within the same site and with the same water levels have been observed to lead to differences in CH4 fluxes (Bubier, 2016; Jackowicz-Korczyński et al., 2010). To understand how peatland GHG emissions respond to environmental change, the representation of plant-environmental feedbacks is critical. We will include this text, justifying the development of the new model, in the introduction of the revised manuscript.
Model equations and explanations
Original comment: I found the subsections in which the model equations are presented quite difficult to follow in a few places. As someone who is not familiar with the parent models, it would have been useful to have had more written information on the equations and their derivation or even some simple plots showing the form of the equations. Several equations also seem to contain errors – some are not dimensionally homogenous and the signs of the terms in others don’t make sense (to me, at least). I have indicated some examples in a marked-up copy of the paper supplied separately. I recommend the authors check the equations carefully to ensure they are rendered correctly and are dimensionally balanced.
Response: We will add more detailed descriptions to several of the equations. We will carefully re-check all equations for errors. This point was also raised by reviewer 1 and 2.
Peat accumulation and water table inputs
Original comment: Some aspects of the new model didn’t really make sense to me. It is not clear how net peat accumulation or loss is accounted for in the model. Can the designated soil layers expand and contract as the size of their SOM pools varies? Also, I am confused about why only mosses can apparently contribute to peat formation as suggested in Figure 1 and noted on line 236: “Non-moss PFTs do not contribute to the storage of the peat.” If the new model is to be applied over timescales of many decades, then changes in peat thickness and mass should surely be considered. Also, peat properties such as permeability are important at these timescales because they feed back into a peatland’s hydrological functioning and the position of the water table which, as noted above, is a first-order control on CO2 and CH4 fluxes.
Response: We will amend the manuscript to specify that there is no feedback between peat loss/accumulation and the soil layers (such as dry bulk density, or recalculation of the pF curve). This is a feature that we plan to incorporate in future versions of the model. This will be considerable work and was beyond the aims of this study. The introduction of calculating the thickness of the living moss layer was one first step towards doing this. Vascular plants have significantly faster decomposition rates than bryophyte plants in both undisturbed and harvested fens (Graf and Rochefort, 2009). Vascular plants are more likely to contribute to SOM pools such as labile organic matter, exudates, microbial biomass, litter, and dead roots and therefore, we have divided their OM products between these belowground pools. We will amend the manuscript to include this explanation.
Original comment: The authors note that site water tables were provided by the Netherlands Hydrological Instrument. This doesn’t seem to be an accurate way of obtaining water tables. It would have been better to have measured water tables directly or to have used a site-specific water-table model.
Response: We agree with the reviewer that the use of in situ water table measurements would be ideal. Unfortunately, such measurements are unavailable. The NHI is a reasonably high spatial resolution water level product (250mx250m). One aim of developing the PVN model is to eventually develop a model of all Dutch peatlands in conjunction with the NHI product. For this reason, we’ve used the NHI product in this application of the model.
Original comment: I have made more comments (mostly minor) on a pdf of the paper and this is posted separately for the authors and the editor.
Response: We have read the comments in the attached pdf and we hope to have the opportunity to incorporate the feedback into a revised manuscript.
We hope that these responses have resolved your concerns.
Sincerely,
Tanya J.R. Lippmann et al.
References
Ballantyne, D. M., Hribljan, J. A., Pypker, T. G. and Chimner, R. A.: Long-term water table manipulations alter peatland gaseous carbon fluxes in Northern Michigan, Wetl. Ecol. Manag., 22(1), 35–47, doi:10.1007/s11273-013-9320-8, 2014.
Bao, T., Jia, G. and Xu, X.: Weakening greenhouse gas sink of pristine wetlands under warming, Nat. Clim. Chang., 13(May), doi:10.1038/s41558-023-01637-0, 2023.
Bubier, J. L.: The Relationship of Vegetation to Methane Emission and Hydrochemical Gradients in Northern Peatlands, J. Ecol., 83(3), 403–420 [online] Available from: http://www.jstor.org/stable/2261594, 2016.
Davidson, E. A. and Janssens, I. A.: Temperature sensitivity of soil carbon decomposition and feedbacks to climate change, Nature, 440(7081), 165–173, doi:10.1038/nature04514, 2006.
Dorrepaal, E., Aerts, R., Cornelissen, J. H. C., Van Logtestijn, R. S. P. and Callaghan, T. V.: Sphagnum modifies climate-change impacts on subarctic vascular bog plants, Funct. Ecol., 20(1), 31–41, doi:10.1111/j.1365-2435.2006.01076.x, 2006.
Dorrepaal, E., Toet, S., Van Logtestijn, R. S. P., Swart, E., Van De Weg, M. J., Callaghan, T. V. and Aerts, R.: Carbon respiration from subsurface peat accelerated by climate warming in the subarctic, Nature, 460(7255), 616–619, doi:10.1038/nature08216, 2009.
Evans, C. D., Peacock, M., Baird, A. J., Artz, R. R. E., Burden, A., Callaghan, N., Chapman, P. J., Cooper, H. M., Coyle, M., Craig, E., Cumming, A., Dixon, S., Gauci, V., Grayson, R. P., Helfter, C., Heppell, C. M., Holden, J., Jones, D. L., Kaduk, J., Levy, P., Matthews, R., McNamara, N. P., Misselbrook, T., Oakley, S., Page, S. E., Rayment, M., Ridley, L. M., Stanley, K. M., Williamson, J. L., Worrall, F. and Morrison, R.: Overriding water table control on managed peatland greenhouse gas emissions, Nature, 593(7860), 548–552, doi:10.1038/s41586-021-03523-1, 2021.
Graf, M. and Rochefort, L.: Examining the peat-accumulating potential of fen vegetation in the context of fen restoration of harvested peatlands, Ecoscience, 16(2), 158–166, doi:10.2980/16-2-3128, 2009.
Jaatinen, K., Laiho, R., Vuorenmaa, A., Del Castillo, U., Minkkinen, K., Pennanen, T., Penttilä, T. and Fritze, H.: Responses of aerobic microbial communities and soil respiration to water-level drawdown in a northern boreal fen, Environ. Microbiol., 10(2), 339–353, doi:10.1111/j.1462-2920.2007.01455.x, 2008.
Jackowicz-Korczyński, M., Christensen, T. R., Bäckstrand, K., Crill, P., Friborg, T., Mastepanov, M. and Ström, L.: Annual cycle of methane emission from a subarctic peatland, J. Geophys. Res. Biogeosciences, 115(2), 1–10, doi:10.1029/2008JG000913, 2010.
Malmer, N., Albinsson, C., Svensson, B. M. and Wallén, B.: Interferences between Sphagnum and vascular plants: Effects on plant community structure and peat formation, Oikos, 100(3), 469–482, doi:10.1034/j.1600-0706.2003.12170.x, 2003.
Malmer, N., Johansson, T., Olsrud, M. and Christensen, T. R.: Vegetation, climatic changes and net carbon sequestration in a North-Scandinavian subarctic mire over 30 years, Glob. Chang. Biol., 11(11), 1895–1909, doi:10.1111/j.1365-2486.2005.01042.x, 2005.
Peltoniemi, K., Fritze, H. and Laiho, R.: Response of fungal and actinobacterial communities to water-level drawdown in boreal peatland sites, Soil Biol. Biochem., 41(9), 1902–1914, doi:10.1016/j.soilbio.2009.06.018, 2009.
Strack, M., Waller, M. F. and Waddington, J. M.: Sedge succession and peatland methane dynamics: A potential feedback to climate change, Ecosystems, 9(2), 278–287, doi:10.1007/s10021-005-0070-1, 2006.
Yrjälä, K., Tuomivirta, T., Juottonen, H., Putkinen, A., Lappi, K., Tuittila, E. S., Penttilä, T., Minkkinen, K., Laine, J., Peltoniemi, K. and Fritze, H.: CH4 production and oxidation processes in a boreal fen ecosystem after long-term water table drawdown, Glob. Chang. Biol., 17(3), 1311–1320, doi:10.1111/j.1365-2486.2010.02290.x, 2011.
Citation: https://doi.org/10.5194/gmd-2023-48-AC3
-
AC3: 'Reply on RC3', Tanya Lippmann, 23 May 2023
Tanya J. R. Lippmann et al.
Data sets
The git repository of Peatland-VU-NUCOM (PVN 1.0), hosted by bitbucket. Tanya J. R. Lippmann and Jacobus (Ko) van Huissteden https://bitbucket.org/tlippmann/pvn_public/src/master/
Model code and software
Peatland-VU-NUCOM (PVN 1.0): A peatland GHG emissions model using dynamic plant functional types Tanya J. R. Lippmann https://doi.org/10.5281/zenodo.7701698
Tanya J. R. Lippmann et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 342 | 85 | 19 | 446 | 33 | 3 | 5 |
- HTML: 342
- PDF: 85
- XML: 19
- Total: 446
- Supplement: 33
- BibTeX: 3
- EndNote: 5
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1