This is the review for the manuscript “Implications of climate and litter quality for simulations of litterbag decomposition at high latitudes” submitted by Aas et al. In it they present their work of using the MIMICS+ model to simulate two different litter bag measurement campaigns. Their results affirmed that the model is capable of projecting litter decomposition in high latitudes.

Right at the start I will admit that it was really difficult for me to figure out which recommendation give for this manuscript. Overall I think it is a well written paper that while ultimately detailing a baseline test, still manages to present thought-out, clear reasoning for its necessity. Additionally, the experimental setup is explained well and the results analysis is thorough. If it wasn’t a central flaw, while my recommendation would have been returned major revisions, almost all my comments would have been easily addressed.

However, there is a central problem with the current version of the manuscript. For the modelling comparison here, time series data from the Canadian Intersite Decomposition Experiment (CIDET) and the Vestland Climage Grid (VCG) which was buried in 10 cm. The issue here is that the CIDET litter bags were not buried, they were left on the surface. For example, in the CIDET manual detailed explanation of the setup on page 12 is explicit about that while the visual presentation of the experiment setup on page 13 makes it evident as well. It is also explained as such in the Trofymow paper referenced in the manuscript. This is not meant to be snarky towards the error made here, but rather to establish that in writing this, I did go through the material to confirm that this is how it was. Although I will admit that I remain terrified that it will be revealed that I had fundamentally misunderstood something here.

Now, I don’t think that this has major bearing on the results and most of the conclusions would remain that same. This does, though, require redoing the simulations and analysis to be redone for the CIDET sites. Which brings us to me being conflicted as I am leaning towards recommending major revisions for the manuscript, but depending on the model, the work required to address this correction might require more time than major revisions would allow for.

With all that written, my recommendation will still remain returned for major revisions. But if the editor and/or authors come to the conclusion that they need more time to properly correct the manuscript, then I still hope that the manuscript will be resubmitted after a rejection that allows for the necessary time period.

Below are a few line-to-line recommendations about the manuscript:

Line 25: Bradford et al., 2016; Joly et al., 2023

It is slightly distracting that the same two articles are quoted on two subsequent lines directly above/below each other. Only the first reference is required or split the references between the sentences.

Line 33: “…litter are buried…”

Kin of related to the larger criticism given to the manuscript before, but there are plenty of litter bag experiments that are not buried.

Line 36: “…and methods using standard litter…”

I would recommend first introducing the issue before stating that there have been suggestion on how to overcome it. In the current form it is a bit confusing as my initial concern was that something had been removed before this part.

Line 70: “2. Methods”

I’m putting this comment here as I couldn’t figure out where exactly to put in the text. Your description of the two experiments is missing central information regarding the properties of the bags. Not just their dimensions, but the size of the holes they have, are they single or two layer, etc.

This isn’t necessary information just for accuracy of description, but a huge challenge in comparing information between different litter bag experiments is that the physical differences in the bags do affect the development of the decomposition timeline, especially in multi-annual experiments.

Line 80: “…well-established, relatively long-term experiment (hypothesis 1)”

The litter bags in CIDET are collected every year, so you also have access to the decomposition state after the first years. So why are you only doing using the CIDET data to compare the long-term performance?

This is especially pertinent because just looking at the results in Figure 2, I’m not initially convinced that the MIMICS+ model would perform as well in both time frame here as the litter bag experiment sees the largest drop of the mass during the first year before levelling off, something that doesn’t appear to happen in most of the model runs. Although would need to see the measurement comparison there to be certain.

Line 97: 2.3 Description of the soil decomposition model, MIMICS+

First, I don’t think that comma there between model and MIMICS+ is necessary, is it?

Second, as far as I could tell neither in the model explanation or in appendix B is there any mention how the model was calibrated? Doesn’t need to be an exhaustive explanation, just a few lines about the datasets and methods used.

Line 194: “This gave initial decomposer…”

Just to confirm, these initial values are for the date when the litter bags were added? Not steady state approximation?

Line 202: “…we assume minor contributions from…”

Do you mean here in the model context? As this is a very assumption to make for multi-year litter bag experiments.

Line 306: “Unpublished results from…”

I will admit being puzzled by this reference as I don’t understand how you can have unpublished results from a published paper? Also if you are referencing unpublished results here, they need to still accessible somewhere.

Line 319: “…leaching processes likely also contributed to the observed mass loss.”

This part would benefit from a slight reordering. First explain what the model representation of leaching does and does not include before from there shifting to explaining why it might be contributing here.

Additionally, and this is a little bit tricky, I don’t think the leaching being a major factor in the discrepancy is a convincing argument based even on the results presented here. For that to be the case, the leaching needs to be insufficient during the first 6-12 months before kicking in correctly? This isn’t mean as a discouraging comment, but rather that the rapid decomposition during the first year of almost any decomposition experiment is a well-known challenge for soil carbon decomposition models and it simply appears to be the case also for MIMICS+.

Line 323: “Their experiment suggested…”

But this is in a bit of a contradiction with the VCG results where the initial decomposition drop lasts for longer than the first couple of weeks, isn’t it?

Line 403: “Litterbag mesh sizes…”

As already noted, the mesh sizes for the different experiment should have been established in the methods section instead of just mentioned all the way here.

Review of “Implications of climate and litter quality for simulations of litterbag decomposition at high latitudes.”

In this manuscript, the authors test the effectiveness of the MIMICS+ model in representing litter decomposition over time in northern latitudes by using the model to replicate both a short term litter decomposition study in Norway and a long-term decomposition study in Canada. The authors find that MIMICS+ generally represents longer-term litter decomposition well, and performs better over longer timescales when climate variables hold more control over decomposition outcomes than litter quality variables. The effects of litter quality on modeled deposition were smaller than the authors predicted based on experimental data. Further, different representations of microclimate variables using observational and model-derived temperature and moisture in MIMICS+ did not improve model fidelity to observed decomposition rates. Based on these results, the authors reveal several research needs that could advance model representations of decomposition, including work examining litter C:N and microbial interactions, divergent influences of soil moisture on decomposition, and model representations of leaching and freezing. 

Overall, this is a well-written article that presents an interesting and useful dataset, especially with respect to improving model-data integration. The work is an important advance that highlights steps to be taken by both modelers and empiricists that will improve our understanding and representations of C dynamics across scales. My comments on this manuscript are relatively minor as I think that the authors did a nice job conducting and presenting their study.  Further, I appreciated that the manuscript was accessible from multiple disciplines--both in modeling and empirical arenas--which is a crucial step in advancing ecosystem to global scale research. 

My comments primarily involve adding a bit more detail regarding the similarities and differences between the field and model experiments that could help the authors and reader more completely assess potential reasons empirical results were divergent from those modeled. As an extension of this, I also wondered how mycorrhizal differences may have played into the ability to replicate field experiments in the model, especially as one of the advances of MIMICS+ is the additional of mycorrhizal types, and CIDET mycorrhizae could be quite different than VCG mycorrhizae given the forest/grassland distinction. Included below are my line-specific comments, which give some specific locations where I think these overarching topics could be addressed. On the whole, though, this is a nicely done experiment and clearly written manuscript. 

Line-by-line comments

Introduction

Line 56: Should ‘of’ be ‘and’ at the start of the line?

Line 60: Just something I’m curious about (although I don’t know that you need to include this in the manuscript specifically)—I’m familiar with testing of MIMICS in various sites but has MIMICS+ been tested primarily in northern ecosystems to this point or more broadly?   

Methods

Line 111: It’s stated previously what rMMK stands for, but the abbreviation isn’t included when reverse Michaelis Menten Kinetics is introduced. It would be good to include the abbreviation rMMK in line 100.

Section 2.3/2.4: Given that CIDET includes many forested sites and the VCG sites are grasslands, did you make any adjustments to the model to reflect potential differences in mycorrhizal types between these ecosystems and could mycorrhizal parameters contribute to any of the differences in the ability of the model to match observations at the Canadian versus Norway sites? From the methods section, it seems like C allocated to mycorrhizae is determined by CLM data--does that generate differences in the portion of AM and EcM for each site?

Figure 1 caption: In the 2^nd line of the figure caption, ‘of’ after (DEF) should be ‘or.’

Line 197: Is the selected model layer a bit deeper than the depths of the empirical litter bags? For VCG the temperature logger is 5 cm deep—is there a reason the authors opted to use the model layer below that and could this decision alter the degree of consistency between the model and the empirical litterbag results? The area of C added also seems larger than the size of empirical litterbags but maybe this is a feature of model resolution?

Results

Line 221: Here it’s noted that the sites with highest MAT have some of the largest mass loss during the model experiments indicated in Fig. 2, but it’s somewhat hard to tell that these are the sites with the warmest MAT given the data displayed in Fig. 2. Opting to show the June-September temperature in the figure makes it seem like the MAR and HID sites are the warmest, so when PMC and SHL were introduced in line 221, I did a double take. Since the results text is focused on MAT rather than summer season temperature, it might be clearer if MAT is used in the figure as well.  

Figure 3: I don’t recall the authors making a note of this in their methods, but why did the authors decide to normalize values using hemlock mass loss values rather than plotting simulated cumulative mass loss for each litter type, including hemlock?

Figure 6: Should BGC in the top legend be MOD instead?

Discussion:

Line 306: Should the Althuizen et al. reference include “in prep” or “in press”? 

Line 310: Possibly Rocci et al.’s recent publication in SBB would of interest to the authors here? 

Rocci, K. S., Cleveland, C. C., Eastman, B. A., Georgiou, K., Grandy, A. S., Hartman, M. D., ... & Wieder, W. R. (2024). Aligning theoretical and empirical representations of soil carbon-to-nitrogen stoichiometry with process-based terrestrial biogeochemistry models. Soil Biology and Biochemistry, 189, 109272.

Line 404: Due to mesh sizes, could litter at VGC be interacting with mycorrhizal fungi, as well, and therefore would mycorrhizal representation in the model be even more important to replicating field results? Further, do you know if mesh size is the same for CIDET litter bags? I might expect that the underestimation of mass loss by MIMICS+ has more to do with these fauna interactions than microbial stoichiometry (as discussed in the previous section) given numerous studies that show microbial C:N:P to be fairly constrained.

Line 408: Soil nutrient availability, microbial communities, and fauna also change with soil depth so differences in depth representation could result in a number of differences affecting decomposition results. It might be worth bringing up a couple of these other actors as well in this paragraph as the authors see fit.