Resolving effects of leaf pigmentation changes and plant residue on the energy balance of winter wheat cultivation in the ORCHIDEE-CROP model

Yu, Ke; Su, Yang; Lauerwald, Ronny; Ciais, Philippe; Xi, Yi; Xu, Haoran; Zhang, Xianglin; Viovy, Nicolas; Pickering, Amie; Collard, Marie; Goll, Daniel S.

doi:10.5194/gmd-19-7217-2026

Articles | Volume 19, issue 15

https://doi.org/10.5194/gmd-19-7217-2026

© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/gmd-19-7217-2026

© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 19, issue 15

Development and technical paper

|

06 Aug 2026

Development and technical paper |

| 06 Aug 2026

Resolving effects of leaf pigmentation changes and plant residue on the energy balance of winter wheat cultivation in the ORCHIDEE-CROP model

Ke Yu, Yang Su, Ronny Lauerwald, Philippe Ciais, Yi Xi, Haoran Xu, Xianglin Zhang, Nicolas Viovy, Amie Pickering, Marie Collard, and Daniel S. Goll

Download

Final revised paper (published on 06 Aug 2026)
Supplement to the final revised paper
Preprint (discussion started on 02 Jul 2025)
Supplement to the preprint

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2025-1861', Anonymous Referee #1, 26 Aug 2025

This manuscript investigates the effects of leaf pigmentation changes and crop residues on the surface energy balance of winter wheat, by improving the ORCHIDEE-CROP model with dynamic albedo, soil evaporation, and surface roughness parameterizations. The study is clearly presented, methodologically sound, and provides valuable insights into biophysical processes often overlooked in land surface models. Some aspects, however, would benefit from refinement: The site-to-site variability in residue effects (on evaporation and roughness) deserves more explicit discussion of limitations. The treatment of uncertainties and sensitivity to input data and climate scenarios could be expanded. Figures are informative but some captions and explanations are too brief; adding interpretive detail would improve readability. Overall, the work is solid and novel. I recommend minor revision before acceptance.

Citation: https://doi.org/10.5194/egusphere-2025-1861-RC1
- AC2: 'Reply on RC1', Ke Yu, 06 Nov 2025
  
  Dear reviewer #1,
  Thank you so much for your time in reviewing our manuscript and for your positive feedback on our work. Following your suggestions, the revision of this manuscript is as follows:
  (1) Added a sensitivity analysis of key model parameters to analyze the variability of residue effects. The corresponding methodology and results of the sensitivity test were described in section 2.6 and section 3.5; (2) Expanded the discussion of model limitations and uncertainties in section 4.5.
  For more details, please find our responses to all of your comments in the response letter (Supplement)
  Sincerely,
  
  Ke Yu
  
  on behalf of all co-authors
  
  Citation: https://doi.org/10.5194/egusphere-2025-1861-AC2
RC2:
'Comment on egusphere-2025-1861', Anonymous Referee #2, 16 Sep 2025
Yu et al present the ORCHIDEE-CROP model that incorporates a time-varying surface albedo that considers foliar yellowing and crop residue. The effect of crop residue on surface roughness, surface temperature, and surface energy partitioning was also analyzed. Overall, the work is solid and addresses effects not considered in current models, while it may benefit from some clarification on methods and interpretations. I suggest publication of the manuscript after minor revision.
In several places across the manuscript (e.g., L36), the manuscript claims a significant impact on surface energy balance. However, in most scenarios, the impact seems small (less than a few W/m2) to me, and the uncertainty is at similar magnitudes as the difference. I think the magnitude of the differences can be better addressed.

L231, L242: It would be good to explain how the calibrations were performed in the supplementary.

L312: The equation number seems incorrect

L355: How was the αsurf model trained? Random forest is mentioned here, but Section 2.4.2 describes a direct fitting of the parameters in Eqs. 13 and 14.

Figure 4: It would be good to show the relative difference in the main text or supplementary to provide more context.
Citation: https://doi.org/10.5194/egusphere-2025-1861-RC2
- AC1: 'Reply on RC2', Ke Yu, 06 Nov 2025
  
  Dear reviewer#2:
  Thank you so much for your time in reviewing our manuscript and for providing your constructive comments and suggestions. The valuable comments helped us improve the paper. Following the comments, our main revision is as below:
  
  (1) Clarified the magnitude of impacts of foliar yellowing and residue covering on water and heat variables in the Abstract, and included the relative changes in targeted variables between the improved model and the initial version in the results sections; (2) Expanded the discussion of the uncertainty and spatiotemporal variability of residue impacts in sections 4.2 and 4.5; (3) Supplemented more details about the calibration of crop development and harvest timing in sections 2.2.5 and 2.3.
  
  Please find our detailed responses to all of your comments in the response letter (Supplement).
  Sincerely,
  
  Ke Yu
  
  on behalf of all co-authors
  
  Citation: https://doi.org/10.5194/egusphere-2025-1861-AC1

Peer review completion

AR – Author's response | RR – Referee report | ED – Editor decision | EF – Editorial file upload

AR by Ke Yu on behalf of the Authors (06 Nov 2025) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (26 Nov 2025) by Nathaniel Chaney

RR by Anonymous Referee #2 (03 Dec 2025)

RR by Anonymous Referee #3 (13 Dec 2025)

Suggestions for revision or reasons for rejection

This study proposes modifications to surface albedo representation in the ORCHIDEE-CROP model during leaf pigmentation changes and crop residue phases. The impacts of these changes on surface energy balance are analysed, and potential implications for Soil Water Content (SWC) under present-day and drier climate conditions are inferred.
While the modeling choices are generally reasonable, the proposed modifications do not convincingly demonstrate consistent improvements compared to the original ORCHIDEE-CROP formulation. In some cases, model performance appears to degrade, and the added complexity does not appear to translate into robust or statistically significant improvements. As currently presented, the manuscript would benefit from clearer methodological justification, more rigorous evaluation against the original model, and a more quantitative and transparent assessment of improvements and limitations.
I revised the manuscript up to Section 3.3 and recommend major revisions. Should the revised evaluation become more convincing, I would be happy to provide further feedback on the remaining sections.

Below, the authors may find some suggestions for methodological improvements. However, the authors are also encouraged to explore additional paths to improve model performance beyond the specific suggestions listed here.

Major comments:
L122-136: It is said that 2 sites use a different methodology for calculations of albedo. The authors should try this methodology on the other 8 sites, compare the results and make sure that it is a reasonable approximation – this information should be added in the SI. Also, it is mentioned that data are quality-controlled and gapfilled, and if I remember correctly these dataset come with quality flags for the quality of the measurement or for signaling gapfilling. Did the authors filtered the data based on this information? I wonder whether using only measurements flagged for good quality could improve the final performance.

L147-150: I disagree with this approach. Using rainfall from the driest year, while the other variables vary means that decoupled quantities are being used, which is not physically realistic. Meteorological fields should be allowed to change coherently, and then the effects that the authors find interesting should be isolated via statistical methods.

L162-167: It is true that LAI measurements are sparse at ICOS site. However, I suggest that the authors use the locally available information to correct the MODIS values and match the locally shown variability and range (after all, a max LAI=2.5 is quite low for wheat in my opinion)

L257-258: This sentence is unclear. Does it mean that sowing and maturity dates are determined via GDD and then a number of days after maturity is added so that the observed/modeled harvesting dates match? If that is the case, it is then not guaranteed that modeled sowing and maturity days match with the observed ones. An evaluation of this aspect is also required, as I imagine this would affect calibration of the parameters related to leaf pigmentation (and also beta_4 and Z_0). Is there a reason why the authors didn’t force STICS to use observed dates instead of the ones determined via GDD? An alternative could be to retrain STICS so that sowing and maturity provided via GDD match the observed sowing and maturity dates.

L291: why is beta_4 assumed at 50% and not calibrated on observations? Picking a value in the upper range can lead to exaggerating the effects on Esoil. If I remember correctly, the selected sites should provide measured LE fluxes, which could be used to calibrate this parameter.

L375-380: The results section should start with the inferred parameter values, and their 95% CIs. Evaluation on albedo predictions should consider also the performances of the old model: how far do the improvements go? I suggest to modify Figure 2 to include both the predictions from the new model and from the old model (with another color), so that the evaluation can be presented coherently. Furthermore, this section should include the evaluation of Tsurf, H and LE: instead of presenting Figure S9-S10, the authors should conduct an evaluation similarly to the one conducted in Figure 2. This one should also quantify performances of the original model (ORC-D) and of the modified version (ORC-AE).

L386: ORC-AE does not capture alpha_surf evolution at all sites. It is quite clear that the new model leads to a degradation of albedo estimates in two out of five instances (De-Geb2014, FR-Aur-2019). The authors should state this clearly in the paragraph and should better quantify improvements and shortcomings.

L405-436: It seems to me that there is virtually no change in any variable during foliar yellowing. What the phrase as “slightly lower” should be better phrased as “no change”, given also the large variability shown. Residue covering shows some larger effects, but it wouldn’t seem like it’s significant. The authors are requested to better quantify statistically the inferred changes (statistical significance), in order to understand better the extent of the effects of the changes they made in ORCHIDEE-CROP. Furthermore, this section is quite number intensive and hard to follow – I suggest that the authors find alternative phrasings to present their finding. Also, I believe that percentages are seldomly relevant in this context.

Minor comments:

L144-145: I don’t see any reason behind using CRU-JRA at the two ClieNFarm sites rather than ERA-interim – why did the authors decided on a different product in this case instead of using ERA-interim data corresponding to those two locations? The authors should uniform the data across all sites, but if they decide to keep this different meteorological input, they should justify their choice adequately.

L151-167: I suggest aggregating subsections 2.1.3, 2.1.4, 2.1.5, because they are very short. I’d instead merge them in a unique section (called, for instance, “Management practice, bare soil albedo and LAI”).

L220: I thought that the von Karman constant was more around 0.40-0.41. Why using a value this different from what is typically used?

L316: does the value reported indicate mean +- SD? Looks like quite a wide range. Does this come from only 33 values, correct? The authors should not indicate SD, but instead the 95% CI for the mean: I imagine the distribution would not be normal, so I suggest to compute bootstrap CIs. Same goes for the value in L320.

L317: “This delay is close … of 14 days.” Is this to state what the old model is doing? Please better clarify within the paragraph what are the settings for the old model and which are the settings for the new one.

L375-376: the authors claim that R2 of 0.54 is a “good performance”. The adjective is not appropriate and should be reconsidered. Good compared to what? Other models in the literature? Compared to the old model? One could argue that 0.54 is far from being optimal.

L386-387: is FR-Gri2018 in the testing dataset? I thought there were 5 sites and a total of 5 years for testing, but I see that there are 5 years being presented in Figure S4 (and one of them is not FR-Gri2018, but FR-Gri2020). The authors should clarify this aspect. If FR-Gri2018 is in the calibration dataset, Figure 3 can be safely removed, and Figure S4 should take its place instead (with all panels).

Style comments:

L22: The sentence “However, the latter aspect is not yet covered by existing cropping system models” sounds a bit too final. I don’t have enough knowledge on the current state of crop modeling regarding this feature, but I suggest that the authors leave this sentence only if they are 100% sure about it.

L56-57: there’s no need to start a new paragraph here

L70: “Winter wheat … reflectivity”. This sentence doesn’t sound right, as it suggests causality and intentionality where there is none. A sentence that simply describes the species would be enough.

L150-152: Table S2 is mentioned before Table S1. The authors should switch the tables in the SI so that the ordering is correct.

L166-167: What does “but not S2” means? Does it refer to site measurements?

L281-284: Equation formatting needs to be improved

L304: Equation formatting needs to be improved

L311: “between these two phases” instead of “between them”?

L330: not sure if “parametrized” is the right term. Do the authors mean that the new values were included in the new ORCHIDEE-CROP version?

L337: what does “-D” stands for here? What about the other acronyms? Please specify

L341: maybe use ORC-beta_4 instead of just ORC-beta

L370: y overbar is missing

Figures and tables:

Figure S2. In the supplementary there are no limitations. The authors are invited to show more of the years and sites available (or, even better, a summarized evaluation)

Table S1. Wrong date format for tillage

Table 1. This table needs to be simplified greatly. Authors should indicate with just a tick whether there are modifications with respect to the original version or not, and then add in the caption or in footnotes what these changes mean only if they differ with what is described in the methods text.

Hide

ED: Reconsider after major revisions (15 Dec 2025) by Nathaniel Chaney

AR by Ke Yu on behalf of the Authors (04 Feb 2026) Author's response Author's tracked changes Manuscript

EF by Katja Gänger (05 Feb 2026) Supplement

ED: Referee Nomination & Report Request started (06 Mar 2026) by Nathaniel Chaney

RR by Anonymous Referee #3 (27 Mar 2026)

Suggestions for revision or reasons for rejection

The authors effectively addressed my main concerns regarding the improvements achieved with the new model version. In this regard, I find Figure 2 quite compelling. I believe that, at the current state, the manuscript is satisfactory from a scientific content standpoint and has the minimum requirements for publication on GMD.

Nonetheless, there are still many parts that require clarifications, and the study needs many improvements in terms of style, coherence, and notation. In particular, Section 2.1 needs to be greatly improved in clarity, and the authors are required to carefully review all equations.

Please find below a non-exhaustive list of required corrections. The authors are also required to extensively review the overall manuscript by carefully choosing wording and improved logical flow.

Furthermore, I still have some major concerns:
1. A methodological concern remains regarding the “drying climate scenario”, despite the authors’ reply to my previous review. The authors mention that one meteorological year is employed to force the model, but I still fail to see how the 10-year simulations in the dry scenario differ across years. From section 3.4, it seems that results from the dry simulation are the same from one year to the next. If there is barely any difference between the years, then the pairwise comparison between “dry” and “non-dry” years is wrong: in this case, the authors should thoroughly review that section and their overall approach regarding the “drying climate scenario”.
2. It seems to me that the authors claim that residual managing has uncertain implications in terms of climate change mitigation/adaptation (which is reasonable), but it could be useful as an agricultural practice. However, unless I’ve missed something, the authors evaluate differences between ORC-D and ORC-AE, and conclude that in terms of SWC there are no significant differences. I assume that ORC-D here would represent a configuration in which there is no residual managing (bare soil), while ORC-AE maintains the residuals. Based on just statistical significance, it seems that residual managing might be a marginally relevant agricultural practice. Regardless of how the authors intend to address this point, the conclusions should better reflect and convey the hypothetical nature and uncertainties of their evaluation.
3. Where does the 0.5° spatial resolution come from (e.g. L139)? The ORCHIDEE-CROP model is stated to have a 0.5° spatial resolution (L152-153), but it’s unclear how this is relevant, since the model is forced by local measurement sites data. It seems to me that what is being doing is just running the model at sites location: if that’s the case, shouldn’t resampling gridded data at 0.5° be replaced just by interpolation at the measurement site location?
4. L301: So Hrough can be expressed as H_rough=f_(crop,max)×0.5×(1-H_dis ), where 0.5 basically replaces the wheat canopy height in the original formula. So, the formula would be the same as the one above and the only thing that changed is the canopy height considered? The same stands for Z0.

Minor comments:
L117-118: Please add an introductory sentence, explaining that two different sources were used (ICOS and ClieNFarm)
L13 of supplementary: title would be “Quality filtering and preprocessing of surface albedo data at ICOS sites”? To differentiate them from the ClieNFarm sites
L128-129: should this sentence be “Across the selected 10 sites, we chose only the years in which winter wheat was cultivated, and obtained a total of 33 winter wheat years”?
L142: I suggest removing “including precipitation, air temperature, wind speed, humidity, and radiation” – stating “All meteorological variables” would already be enough
L151: If “S2 α_surf” is referring to the Sentinel-2, I suggest using “S2-α_surf” for clarity
L158: “LAI was used primarily to constrain”, isn’t the term ‘primarly’ redundant? Based on what the authors said it was used just to constrain the onset, correct?
L166: why winter wheat and maize instead of just winter wheat?
L178: “𝑆𝑊_𝐼𝑁↓” is used in the formula, but shouldn’t it just be either “SW↓” or “SW_IN”?
L196-200: is this formula correct? At first glance the units do not add up. What are the units for the beta terms?
L210: Please improve formatting of the equation (use bigger parentheses). The same stands for the other equations
L215: I believe that a more standard notation for canopy height could be h_c (or H_c if the authors want to be more consistent with the rest of the notation)
L222: The meaning of “weighted height of vegetation with the maximum fcrop for winter wheat in each grid” is unclear
L225: H_ave appears on both left- and right-hand side of the equation (which I assume it’s an error)
L274: meaning of k3 and k4 should go after eq. (14)
L328: “Data from the remaining five sites with five winter wheat years” – why five sites? Isn’t it just the remaining five years?
L332: “Simulations were performed for the 10 sites”
L345: Fix Table 1 caption
L371: “of the old and new models and observation”. Do the authors mean “the comparative performances of the old and new models against observations”?
L372: “to assess the statistical significance of the results” is unclear. Please state clearly the null hypothesis that is being tested.
L435: “ORC-AE indicates” instead of “ORC-AE suggests”
L505-506: Why would both positive and negative perturbations of k1 lead to enhancements of surface cooling?
L527: I wouldn’t say that the control is “even stronger” – to me it looks like the responses are usually below the 2°C cooling
L531: missing space between “increased” and “Z0”
L548: comma instead of period before “while”
L568: “both and their”?
L621: “while removing the residue cover, temporarily boosting 𝐸𝑠𝑜𝑖𝑙 and LE after the end of residue covering” seems an anacoluthon. Please check wording.
L627-628: Check wording.

Hide

ED: Publish subject to minor revisions (review by editor) (22 Apr 2026) by Nathaniel Chaney

AR by Ke Yu on behalf of the Authors (02 May 2026) Author's response Author's tracked changes Manuscript

ED: Publish subject to minor revisions (review by editor) (07 May 2026) by Nathaniel Chaney

AR by Ke Yu on behalf of the Authors (13 May 2026) Author's response Author's tracked changes Manuscript

ED: Publish as is (03 Jul 2026) by Nathaniel Chaney

AR by Ke Yu on behalf of the Authors (11 Jul 2026) Manuscript

Short summary

Coupling crop and land surface models helps quantify the climate effects of agriculture, but lacks crop-specific management processes. We enhanced a land surface model with time-varying albedo from foliar yellowing and residue cover, improving the simulation of energy and water fluxes. Results show cooler surfaces and slightly wetter soils during residue cover, highlighting how managements improve climate mitigation and adaptation, advancing the development of climate-smart agriculture.