the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 12 Aug 2026
Warming-induced agricultural nitrous oxide emissions can offset mitigation benefits of fertiliser management
Abstract. Agriculture is the dominant source of anthropogenic nitrous oxide (N2O), a potent greenhouse gas with a high global warming potential. In Switzerland, substantial changes in fertiliser use alongside climatic conditions have occurred over the past four decades, yet the relative contributions of management practices versus environmental change to long-term N2O emission trends remain incompletely understood. Here, we applied the biogeochemical model DayCent at 1 km resolution across Switzerland, integrating spatially explicit datasets on climate, soil properties and agricultural management, to quantify N2O emissions for the period 1981–2020 from croplands and grasslands and attribute emission changes to management versus climate drivers. Simulated national N2O emissions from agricultural soils declined by roughly 5 % from 4.0 to 3.8 kt N yr⁻¹ between the 1980s and 2010s, primarily due to a 25 % reduction in N fertiliser use. Our attribution simulations suggested that under real climate conditions, such a decrease in fertiliser N inputs (–25 %) over the period studied lowered emissions by 15.2 % in croplands and by 12.0 % in grasslands (including permanent meadows and pastures, and high alpine summer pastures). However, compared to a control scenario (in the absence of climate change), rising temperatures over the past 40 years offset these gains, increasing emissions by 5.7 % in croplands and 13.6 % in grasslands. These results show that warming-induced N2O emissions partially negate mitigation from improved fertiliser management, highlighting the need for integrated agricultural N2O mitigation strategies that are resilient to future warming.
- Preprint
(4559 KB) - Metadata XML
-
Supplement
(3232 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2026-4445', Anonymous Referee #1, 03 Sep 2026
- AC1: 'Reply on RC1', Jize Jiang, 05 Oct 2026
-
RC2: 'Comment on egusphere-2026-4445', Anonymous Referee #2, 22 Sep 2026
Dear Authors and Editor,
This manuscript targets a relevant and important topic and provides valuable insights into N2O emissions dynamics in agriculture of relevance for GHG emissions monitoring and reporting as well as policy.
The paper is well-written. The introduction is clear an motivates the study well. The methods section is likewise clearly structured and the discussion brings up important uncertainties in the modelling approach.
However, I have a number of comments that the authors need to address before I can recommend it for publication, as outlined in the sections below.
Major comments:
1) Clarify the use of emission factors (EFs)
The authors use emission factors as one of the key output variables in their analysis. This is in the introduction motivated by the use of EFs in Tier 1 IPCC methodology for GHG reporting. However
- The definition of EFs in this study deviates from the one used in IPCC guidelines (see detailed comment line 192). Please clarify why this alternative definition was what this means for the comparison between your derived EFs and the IPCC default ones.
- Clarify the intention of using EFs as a primary output variable for analysis. Is the goal to develop alternative EFs for use in GHG reporting or to facilitate comparisons?
2) Clarify development of background emissions
Background emissions of N2O were derived from a counterfactual scenario where climate drivers where the same but fertiliser and manure N inputs were set to zero. It is however not clearly described how e.g. crop rotations and harvests are handled in this counterfactual. And won’t these “background emissions” at least partly represent legacy emissions from SOM built up in the spin-up? See also my detailed comment on lines 186-190.
3) Revise statistical treatment of Monte Carlo results
Monte Carlo simulations were used by varying a set of influential model parameters around assumed distributions. From this a 95% confidence interval (CI) around the main results were calculated. However, it is not possible to derive confidence intervals from Monte Carlo simulations as the number of Monte Carlo runs will impact the width of the CI. So, increasing your Monte Carlo from 200 iterations to, say 400, would increase your perceived confidence. See e.g. von Brömssen & Röös (2020) and my detailed comment on lines 242-243.
Detailed comments by line:
46: “Switzerland” appears a bit suddenly here. It comes in more naturally at line 60 so can be dropped here.
49: Clarify that the 1% EF implies that 1% of N inputs is released as N2O-N.
51-53: You write that “EFs are typically fixed in space and time” but in the same sentence that they still “partly represent” climatic conditions which seem like a contradiction. I am assuming that you refer to the disaggregation of default EFs for some climate conditions, but please clarify.
73-74: Clarify “agricultural information” to show the detail of management variables included in the model
108-110: Please clarify the extent of this “evaluation” of the model to Swiss data and how this feeds into your work. Has this resulted in a Swiss calibration of model parameters that you use? How many crops have been evaluated and how representative are these to the total Swiss cropland? Does model evaluation cover the climatic, topographic and soil gradient in Switzerland?
128-131: How was data for the 24 agri-climatic zones resampled to the 1x1km grid?
147-148: Clarify if the “fixed dates” were assumed to be the same across Switzerland or varied by agri-climatic zone
186-190: In the “nofert” scenario I am assuming that net mineralisation of SOM will be higher than in the scenarios with continued fertilisation due to mineralisation of SOM built up from manure and crop C inputs in spin-up and the lower inputs of C in scenario with no manure application and reduced crop growth. So, is it really “background” N2O emissions you observe or rather a combination of that and legacy N2O emissions due to previous agricultural management? Or did you use a separate spin-up for the “nofert” scenario that did not include manure and fertilisers during spin-up?
192: This definition of EF diverges from the definition of Tier 1 EF (i.e. EF1) in IPCC Guidelines for greenhouse gas inventories, which in addition to N in mineral and organic fertilisers also include N in crop residues and SOM mineralisation in the denominator. Please clarify and motivate why an alternative definition was used here as you relate your work to the default IPCC EFs in your introduction.
235-236: Why does your default value differ from the calibrated mean?
242-243: Am I correct to interpret this as that you calculated a confidence interval around the mean (or median) based on the Monte Carlo simulations? If so, that is not correct as the number of Monte Carlo runs performed (i.e. the number of “samples”) will impact the width of the confidence interval. See e.g. von Brömssen & Röös (2020).
von Brömssen, C., Röös, E. Why statistical testing and confidence intervals should not be used in comparative life cycle assessments based on Monte Carlo simulations. Int J Life Cycle Assess 25, 2101–2105 (2020). https://doi.org/10.1007/s11367-020-01827-4
252-253: Consider the wording “despite declining fertiliser inputs” which indicate that reduced fertiliser inputs are expected to reduce EFs, but that’s not the case right?
271: See comment above.
360-361: What does “spatial increases” mean?
422: Clarify if the 25% reduction refers to total national use or per hectare application rate.
424: Change “in line with” to “along with”
503: Missing “have” in “can have”
505-506: Clarify what you mean by “substantial shifts”
531-532: You do not show that climate change “renders mitigation measures less effective” but rather that climate-induced N2O emission increases offsets some of the reduction from mitigation.
536-537: Would be interesting if you could elaborate on what “site-specific mitigation strategies” might be. What strategies could be suitable in one place but not another due to pedoclimatic conditions?
553-554: Related to my comment on lines 531-532 this is not really a trade-off but rather that some of the gains from mitigation are offset by climate-induced emissions increases resulting in lower absolute emissions reductions. However, from your results I can’t see that they show that this compromises the effectiveness of mitigation measures (here reduced fertiliser N inputs), but rather that it’s a battle up-hill.
Citation: https://doi.org/10.5194/egusphere-2026-4445-RC2 - AC2: 'Reply on RC2', Jize Jiang, 05 Oct 2026
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 238 | 140 | 55 | 433 | 76 | 41 | 50 |
- HTML: 238
- PDF: 140
- XML: 55
- Total: 433
- Supplement: 76
- BibTeX: 41
- EndNote: 50
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Dear Editor,
The manuscript, “Warming-induced agricultural nitrous oxide emissions can offset mitigation benefits of fertiliser management” by Jiang et al., is interesting and addresses a scientifically important question. However, I would recommend major revisions before acceptance.
The topic is timely and important. The manuscript addresses a strong research question with potentially important implications for agricultural climate mitigation. The combination of high-resolution, process-based explicit modelling with a valuable national-scale dataset/model framework is a major strength. The manuscript is also transparent about several limitations of the modelling framework.
The scientific results are presented clearly and concisely.
However, I have substantial concerns regarding the representation of agricultural management dates and fertiliser timing, land-use change, and biological nitrogen fixation (BNF) in the scenario study.
Major comments
1. Representation of management dates and fertiliser timing
The authors state that they use static/fixed management calendars rather than dynamically varying management dates between years. For a 40-year climate-change analysis, this could be important because warming may affect planting and harvest dates, growing-season duration, mowing and grazing dates, and plant N demand, among other processes.
If fertilisation occurs on a fixed calendar date despite a progressively earlier growing season, the model may create an artificial mismatch between N fertilisation and plant N demand. This mismatch could itself affect simulated N₂O emissions. Similarly, keeping mowing dates fixed could affect plant residue inputs and SOM mineralisation.
The conclusion that warming increases N₂O emissions may therefore partly reflect the interaction between warming and fixed management schedules. I suggest that the authors quantify the effect of climate-responsive management dates for fertilisation, mowing, and other relevant management activities, in order to demonstrate whether the main conclusions remain robust when management responds to climate change.
2. Land-use change and fertiliser inputs (Lines 200–201)
The authors state that “total fertiliser inputs also depend on land use areas which changed land use over the past 40 years, especially for croplands.” How have the authors addressed this issue? Were land-use areas kept static, as in Figure S1?
If land-use areas were kept static, please quantify the impact of this assumption on the study outcomes. In particular, how does it affect the partitioning of national total fertiliser inputs among different crops?
3. Biological nitrogen fixation (Lines 440–441)
The authors state that “BNF became an increasingly important N input in unfertilised grasslands, potentially contributing to an increase in modelled background emissions under warming conditions.” Is temporal variation in the area undergoing BNF included in the DayCent simulations? If so, please show the variation in BNF area over time and report the simulated contribution of BNF to total N inputs for both cropland and grassland. This would help determine whether the increasing grassland background N₂O emissions shown in Fig. 1b are driven, at least partly, by changes in BNF.
4. Synthetic versus organic fertiliser N inputs
I think Fig. 1c and the description of fertiliser N inputs would benefit from separating synthetic and organic N inputs for both cropland and grassland. I suggest showing separate time series for synthetic and organic fertiliser N inputs for cropland and grassland, for example by adding these results to Fig. S6. This would clarify whether the reported decline in fertiliser N inputs is primarily attributable to changes in synthetic fertiliser, manure/organic fertiliser, or both.
5. Suggested modification of the title
I suggest modifying the title to: “Climate warming-induced agricultural nitrous oxide emissions can offset mitigation benefits of fertiliser reduction”
This would clarify that the study specifically concerns climate warming and reductions in fertiliser N inputs.
Minor comments
6. Please clarify in the caption whether “Fertilizer N inputs” in Fig. 1c includes both synthetic and organic fertilisers, as well as manure deposited during grazing.
7. How is fertilisation distributed among crops in the model setup? Are synthetic and organic fertilisation data available separately for individual crops? Please specify this in the Methods section.
8. Please specify whether the nofert scenario also sets grazing-derived N inputs to zero. This could be clarified in the text, for example around Lines 186–189.