Modeling nitrous oxide emissions from agricultural soil incubation experiments using CoupModel
- 1Department of Agroecology, iClimate, Aarhus University, Tjele, Denmark
- 2Department of Physical Geography and Ecosystem Science, Lund University, Lund, Sweden
- 3Department of Sustainable Development, Environmental Science and Engineering, KTH Royal Institute of Technology, Stockholm, Sweden
- 1Department of Agroecology, iClimate, Aarhus University, Tjele, Denmark
- 2Department of Physical Geography and Ecosystem Science, Lund University, Lund, Sweden
- 3Department of Sustainable Development, Environmental Science and Engineering, KTH Royal Institute of Technology, Stockholm, Sweden
Abstract. Efforts to develop effective climate mitigation strategies for agriculture require methods to estimate nitrous oxide (N2O) emissions from soil. Process-based biogeochemical models have been used for such estimations but were mainly tested with field-scale measurements. In this study, results from a short-term (43-day) factorial incubation experiment were used to investigate the ability of a process-oriented model (CoupModel) to estimate N2O and carbon fluxes, and soil mineral nitrogen (N) dynamics. This study identified the sensitivities of model parameters when estimating three output variables using a global sensitivity analysis approach. Our results suggested that important parameters regarding N2O flux estimates were linked to the decomposability of soil organic matter (e.g. organic C pool sizes) and the denitrification process (e.g. Michaelis constant and denitrifier respiratory rates). The model was able to simulate low-magnitude daily and cumulative N2O fluxes with model errors (MEs) close to zero, but tended to underestimate N2O fluxes as observed daily values increased over 0.1 g N m-2 day-1. Besides, the response of N2O emissions to soil moisture was not well reflected in the model, probably related to the indirect involvement of soil moisture response function in the denitrification process. We also evaluated ancillary variables regarding N cycling, which indicates that more frequent measurements and additional types of observed data such as soil oxygen content and the microbial sources of emitted N2O are required to further evaluate model performance and biases. The current description of the N cycling process in the model may not consistently represent the temporal scale of nitrification and denitrification processes behind N2O emissions. The major challenges for calibration are associated with high sensitivities of denitrification parameters to initial soil moisture abiotic conditions and residue amendment. For the development of process-based models, we suggest there is a need to address soil heterogeneity, and to revisit current subroutines of moisture response functions.
- Preprint
(2221 KB) -
Supplement
(1510 KB) - BibTeX
- EndNote
Jie Zhang et al.
Status: final response (author comments only)
-
RC1: 'Comment on bg-2022-56', Anonymous Referee #1, 09 May 2022
The manuscript is well written and provides an interesting review of challenges to biogeochemical modeling of N2O fluxes. It presents an incubation experiment paired with modeling to better resolve drivers of error in N2O modeling. However, the most interesting discovery from that effort, which is the contribution of biases in NO3- and NH4+ towards N2O flux biases, is touched on fairly superficially and should be delved into in much more detail. Subsequently, the paper discusses many potential drivers of model error and challenges in experimentation to better identify and address contributions to this error. However, the study conducted does not help address these shortcomings much at all. Hence, in my opinion this article is of limited value as an original research paper and is in fact a mix of limited original research and interesting review. I urge the researchers to push towards work to unravel these meaningful issues they coherently discuss here.
Other comments:
How were the parameter ranges derived? It's insufficient to just describe them as "with realistic ranges" or according to model defaults. The ranges are important to model sensitivity and calibration equifinality issues.
Too much is shown in the figure 5 subplots for interpretation. This data needs to represented in a better manner.
In table 1, why is the rRMSE so much different between the single treatment and multi treatment for NH4+?
You describe a pattern of better model fit as the simulations progress with time. This sounds like a model initialization issue. Did you make any attempts to spin-up the model?
Do you have ideas of what caused the second flux peak? Was it the residue decomposition? Something else?
Isn't seeing ranges of calibrated parameters oscilating heavily across treatments a sign that the calibration is largely fitting noise?
-
CC1: 'Comment on bg-2022-56', Lorenzo Brilli, 10 May 2022
The manuscript provide results from a short-term (43-day) factorial incubation experiment to investigate the ability of a process-oriented model (CoupModel) to estimate N2O and carbon fluxes, and soil mineral nitrogen (N) dynamics. The manuscript is well written, it fluently flows and the whole structure is coherent with the adopted approach. To my opinion, all three objectives indicated by authors at the end of the introduction were satisfying investigated. This would make the paper suitable to be published.
However, many similar works were developed and published through years, reporting similar issues and conclusions. This make the novelty of the paper very poor, despite the large work done. To overcome this huge limitation, I suggest authors to be more proactive at presenting solutions on how to solve the detected issues under current modelling limitations. Within the text, in fact, only general suggestions to cope with these issues are reported (i.e., revisit basic model assumptions and equations, increasing high-quality measurement data, etc.), but none proper solution (new equations to implement and their description, description of further steps in soil incubation experiments, previous chemical analysis to do, etc.) and related changes in final results were reported. I understand that this is not the primarily objective declared within the paper, but since an exponential number of modelling works were published in the last 30 years, a step forward to indicate how to overcome these limitations would be done.
-
RC2: 'Comment on bg-2022-56', Anonymous Referee #2, 13 May 2022
Zhang et al. present a study where results from a short term incubation study was used to assess the ability of a process oriented model to simulate N2O emissions from soil with the addition of different crop residues and nitrate levels. The paper is generally well written and the topic within the scope of Biogeosciences and presents an interesting discussion and review about challenges of accurately simulating soil N2O fluxes. Therefore, I think, that the manuscript should be valuable for other researchers trying to model N2O emissions from soils and could be potentially published in Biogeosciences. However, I agree with the comment by Lorenzo Brilli, that the approach is not really novel and that there is the need to focus more on solutions rather than discussing the limitations.
In addition, I am not really sure to what extent results from a short term incubation, with sieved and repacked soil cores and limited measurements can be used to calibrate and quantify the uncertainty of a process based model used for simulating N cycling and N2O emissions under field conditions. The conditions used in the incubation (sieved, repacked cores, constant temperature and soil moisture) are not typically found in the field and highest N2O emissions are often associated with wetting and drying cycles. Moreover, sieving the soil will result in the destruction of soil aggregates and lead to increased SOM mineralization. I think that theses points need to be better highlighted in the paper and their implications for modelling N2O emissions under field conditions discussed.
LN 25 ff: “For the development of process-based models, we suggest there is a need to address soil heterogeneity, and to revisit current subroutines of moisture response functions.”
Soil heterogeneity was very much reduced in this experimental set up by sieving and re-packing the soil. Can you comment what this implies for field measurements?
Ln 108, to what size was the soil sieved?
Jie Zhang et al.
Jie Zhang et al.
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
256 | 83 | 12 | 351 | 24 | 5 | 6 |
- HTML: 256
- PDF: 83
- XML: 12
- Total: 351
- Supplement: 24
- BibTeX: 5
- EndNote: 6
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1