Overall reaction:   This manuscript should be published, because it promises advancement in the MATCRO modeling system for a C-4 crop like maize.  However, at present, MATCRO is a quite weak model, combining a good mechanistic leaf photosynthesis system with a very limited plant C and N balance system and even more limiting or non-existent soil N balance.  MATCRO is a quite unbalanced model in terms of components.  The following three major issues need to be addressed to be publishable.

1. The Nfert problems suggest to me that MATCRO is very deficient in having, or totally lacking, a soil organic matter module and lacking in plant N balance. The authors need to come clean on this and state in the text that MATCRO either has or lacks a plant N balance module and soil N module.  Soils were adequately described for water balance processes, but there was no mention of soil organic matter or soil N mineralization.  Basing SLN and Vcmax on Nfert is a very limiting approach, and suggests that soil N supply is ignored.  The authors need to state those limitations and improvements even more strongly.
2. The authors use the term validation. Are Figures 5-10  independent data?  But if so, then what data did you use for calibration?  Was calibration only from literature for individual parameters?  Need to indicate what data was used for calibration as that was never stated.  Otherwise, readers will suspect you used this data for calibration.
3. The statistics of Figures 5, 6, and 7 indicate quite poor performance of MATCRO-Maize and correlation statistics are weak tests. Can we recommend a model that performs that poorly for use by global gridded teams?  Figure 7A and 7B are based on LAI and crop mass over time, but the high correlations there are misleading because of auto-correlation effects of time-series data (give high correlation because it uses time-series values).

Specific comments by line number

l. 83 – Says C-3 here. Should be C-4.

L 101 – why bother with “co-limited” photosynthesis?  That is a C-3 hold-over and probably does not apply to C-4.

l. 120 – Is “0.7𝜇” supposed to be “0.7𝜇mol”? What am I missing?

l. 143 – Explain this better, is system solving iteratively for leaf temperature that satisfies…. “meet the following physical flux equations:” Is that what this does?

Eq 21 looks strange.  “𝐶𝑎−𝑅𝑑”, Rd is a rate in umol/m2/s, but Ca is CO2 concentration.  Does “𝑘𝑝,𝑥𝐶𝑎” make it a rate too.

l. 176-177 Very strange. How can you know “maximum Rubisco carboxylation rate at the canopy level (𝑉𝑐𝑚𝑎𝑥25,𝑥 (𝑙))”?  Strangely worded.  Not really a whole canopy trait at all, because your reference is 𝑉𝑐𝑚𝑎𝑥25(0).   I would think that Vcmax is a characteristic of specific leaf N concentration maybe for upper leaves.  OK, as you describe for 𝑉𝑐𝑚𝑎𝑥25(0) on line 186. 

l. 187-190 – This sentence implies conflict or difference, but in both cases 𝑉𝑐𝑚𝑎𝑥25(0) is based on SLN for all three crops. Re-word to avoid that issue, or delete the whole sentence.

Eq. 27 – I don’t like having two equations for 𝑉𝑐𝑚𝑎𝑥25(0) from two sources.  That does not make sense.

l. 220-235 and Figure 2 – Where are the equations and figures for partitioning to stem? Missing. Not in Table 1 either.  At least mention and say “not shown”, or is stem “by difference”.  Also ear is not the same as grain.  Tell us how you get to grain yield.  Very approximately, grain is 85% of ear at maturity, but grain growth starts later than ear, actually a few days after flowering.  So Kyld is about 0.85???  You use 0.83.  OK

Table 1 –SLW could be somewhat related to SLN.  Please give Tb, 𝑇ℎ, 𝑇𝑜 in Centigrade.

l. 260 – You call this validation. OK, if independent. But then, what data did you use for calibration?  I suspect you used this data for calibration.  Line 288-292 indicates that you calibrated life cycle to AgMIP data.

l. 278-279 – Confusing to go elsewhere for soil data, when you give the soil types of AgMIP study in the Table 2. Re-write.

l. 308-309 – You indicate N fertilization rates. What about N mineralization rates of each soil?

l. 340-341 – I am confused. Here you reduced rubisco “rate” and SLN? On what basis?  How was this justified (was it based on the validation data)?   Apparently, you did calibrate to the data or thought about a possible reason.

Figure 6 for Brazil and others would indicate a problem with temperature parameterization for Vcmax(0), because you have a To that is too low, and even a Th is too low.  You have values typical of a C-3 temperate warm-season crop.

l.355-364 – These statistics and Figures 5, 6, and 7 indicate quite poor performance of MATCRO-Maize. Can we recommend a model that performs that poorly, for use by global gridded teams? Figure 7A and 7B use LAI and crop mass over time which is not warranted because of auto-correlation effects of time-series data (gives high correlation because it uses time-series values).

Figure 8 and 9 really seem to be “blind” evaluation because MATCRO is so much above the observed.  Something is seriously missing here that causes the mis-match.  Figure 8 shows MATCRO doing much better than warranted in drought-prone regions such as West Africa or Mexico or southwestern USA, so is the soil water balance failing or is stomatal conductance effect excessively conserving soil water?  Or is it the “big-leaf” photosynthesis approach, very incomplete handling of N-fert effect on Vcmax, or something else?  Figure 10 could point out issues with the soils for each country and stated N-fert that you used.

Figure 11 (MATCRO usually over-estimates) differs from Figure 7 (where MATCRO under-estimates).  Any reasons for this?

l. 414 and 417 – what do you mean by “changed parameters”. Be more specific, is it what you mentioned on lines 340-341 without justification?

Figure 13 – indicate source of N-fert values used for x-axis

l. 428 – replace “were statistically significant” with “showed statistically significant correlations” I also challenge “relatively well”, as performance was not very good.

l. 433 – “One reason” not “the reason”

l. 450 – Many maize models have LAI growth relatively uncoupled from photosynthesis and C balance. Carbon-driven LAI growth may cause problems.

Go back and confirm that is really how the Brazilian experiment was handled as 𝑁𝑓𝑒𝑟𝑡 = 0.  OR, this indicates that you have problems with getting soil N mineralization simulated.  I did not see a word about SOC of soils.

L, 466 – and soil fertility

Table 4 – I am surprised that the other gridded global models for maize are performing that poorly.  Correlation is a weak test.

l. 535-544 – Nfert problems suggest to me that MATCRO is very deficient in having, or totally lacking in a soil organic matter module and lacking in an semblance of a plant N balance. The authors need to come clean on this and say they lack a plant N balance module and lack a soil N module.

l. 550 – replace “would be” with “are

Comments to the manuscript entitled “Development of the global maize production model MATCRO-Maize version 1.0” by M. Nagata et al.

General comments

The submitted manuscript describes the development of a global gridded crop model for maize, called MATCRO-Maize, that can be coupled with a land surface model. The manuscript is clearly written how the model is constructed and tested. However, given the model deficiencies reported in the manuscript, it seems that the model development is still ongoing. The current version of the model lacks key processes or drivers required to reproduce historical yields. The model calibration appears to be incomplete. These issues could be resolved through further simulation and analysis. For these reasons, I think, the manuscript can be accepted subject to minor revision. I’m looking forward to reading a revised manuscript.

Major concerns

Model validation

1. It is recommended to use multiple gridded yield datasets to validate global gridded crop models (GGCMs) because grid-level yields can vary significantly between datasets, which is a significant source of uncertainty when assessing GGCM performance (Müller et al., 2017, Lin et al., 2021). Currently, annual gridded yield datasets are available for the globe and major producing countries at a 5-arcmin resolution (Su et al., 2022, Cao et al., 2025). In addition to comparing their model simulation with the Global Dataset of Historical Yields (GDHY), authors are encouraged to account for yield dataset uncertainty by comparing their model simulation with a family of recent gridded yield datasets.
2. The validation of crop phenology at the global level is currently lacking. I’m happy to see the model validation result at the site level (Fig. 4). However, the data compared are for only four sites and one year, which is inadequate for concluding the model performance. Gridded crop phenology datasets have recently become available for the globe and some major countries (Luo et al., 2020, Yang et al., 2020, Mori et al., 2023). I strongly encourage the authors to compare their simulation with these datasets.
3. In relation to Comment#2, in the current form, it is unclear how the model parameter values related to crop phenology were determined before the model simulation. The authors state that “We used local daily climate data … and phenological data (planting, flowering, and maturity dates) for model input data at each site. (Line 276-278)”. Did you calibrate the parameter values using the site data and then run the model? If so, this does not constitute model validation because no independent data were used for comparison. I would ask the authors to clarify this point and rerun the model validation if necessary.

Modeling

1. How did you determine Gdd,m (eq.22; the growing degree days at maturity)? Is this a universal value across grid cells worldwide? It is well-documented that Gdd,m varies spatially, with higher values in warmer regions and lower values in cooler regions (Deryng et al., 2011, Mori et al., 2023). I would ask the authors to clarify this.
2. The leaf area index (LAI) simulated by the MATCRO-Maize model appeared to be lower than the site observations (Fig. 5). It is also noticed that the difference in maximum LAI between the sites is smaller in the simulation than in the observations (Fig. 5). It leads to the thought that the maximum value of the specific leaf nitrogen parameterized with annual nitrogen application rate (Nfert) (eq. 29) is rather site-dependent and cannot be applied universally in its current form. This does not mean that publishing this preprint is unjustified. However, readers at least want to know whether underestimation of the seasonal maximum LAI correlates with environmental conditions, such as soil carbon content, soil total nitrogen content, water holding capacity of the soil and so on, in order to seek a possible scaling factor to convert specific leaf nitrogen to LAI. The equation (8) of Hasegawa et al. (2008) for the fraction of canopy cover may help the authors relate specific leaf nitrogen to seasonal maximum LAI (though this equation was developed for rice). If such a correlation analysis provides no insight, then calibrating the scaling factor for each country is another option, as was done by Ai and Hanasaki (2023).
3. The presentation of the relationship between Nfert and yield, as presented in Fig. 12, is a bit misleading and could be improved. As can be seen in Fig. 12 (a), yield increases with an increase in Nfert, but then saturates. The yield response to Nfert, as derived from FAOSTAT, is consistent with literature which attributes recent maize yield growth to delayed leaf senescence (staygreen), morphological change from horizontal to vertical leaf type and increased drought tolerance, and resulting increase in planting density, rather than an increase in N input (Duvick, 2005). These genetics and management improvements have changed maize yield response to N input (Fig. 3 of DeBruin et al. 2017). Therefore, liner regression is inappropriate to describe the Nfert-yield relationship. Consider using a nonlinear regression or locally estimated scatterplot smoothing (LOWESS) instead. More importantly, the presented version of MATCRO-Maize imperfectly represent the Nfert-yield relationship (regardless of whether the data for Egypt is included or omitted). Rather than presenting Fig. 13, I would suggest the authors discuss this limitation of the model.
4. The simulated aboveground biomass was lower than the site observations (Fig. 7). However, the simulated yields at the country level were substantially overestimated. This discrepancy may be due to inaccurate partitioning to harvested organ or to stress factors reducing yield formation. I do understand that there are many factors not considered in the model, such as biotic stresses (pests and diseases, weeds, etc.), as described in Line 554. Nevertheless, recent crop models that are embedded in Earth system models that operate at a global level are encouraged to incorporate some form of parameterization to handle major drivers of historical yield growth even in a simple way (Lombardozzi et al., 2020). Alternatively, please consider calibrating some of the existing parameters to better reproduce historical yields (Ai and Hanasaki, 2023).

Technical corrections

1. 'Production' is generally measured in tones and is calculated by multiplying yield (production volume per unit harvested area and cropping season) by area harvested, in the case of single-season maize (see Box 1 of Wei et al., 2023). However, as MATCRO-Maize does not harvest area, the ”'yield model” is more appropriate than the “production model”.
2. Line 264. I think the correct citation for the GDHY is “Iizumi and Sakai, 2020” rather than “Iizumi, 2019”. Please check what recent literature describes this point (for instance, Data Availability and references of Iizumi et al., 2025).
3. Line 281. Do you mean “AgMERRA” (Ruane et al., 2015), a bias-corrected version of the MERRA reanalysis designed for agricultural applications, rather than the original “MERRA”?
4. Line 173. In agronomic literature, the flowering of maize is generally referred to as 'silking'. The first time it appears, you should mention this, for example, “flowering (known as silking; Dvs,flw)”.

References

• Ai, Z. and Hanasaki, N.: Simulation of crop yield using the global hydrological model H08 (crp.v1), Geosci. Model Dev., 16, 3275–3290, https://doi.org/10.5194/gmd-16-3275-2023, 2023.
• Cao, J., Zhang, Z., Luo, X. et al. Mapping global yields of four major crops at 5-minute resolution from 1982 to 2015 using multi-source data and machine learning. Sci Data 12, 357 (2025). https://doi.org/10.1038/s41597-025-04650-4
• DeBruin, J.L., Schussler, J.R., Mo, H. and Cooper, M. (2017), Grain Yield and Nitrogen Accumulation in Maize Hybrids Released during 1934 to 2013 in the US Midwest. Crop Science, 57: 1431-1446. https://doi.org/10.2135/cropsci2016.08.0704
• Deryng, D., W. J. Sacks, C. C. Barford, and N. Ramankutty (2011), Simulating the effects of climate and agricultural management practices on global crop yield, Global Biogeochem. Cycles, 25, GB2006, doi:10.1029/2009GB003765.
• Duvick, D. N. The contribution of breeding to yield advances in maize (Zea mays L.). Adv. Agron. 86, 83–145 (2005).
• Hasegawa, T., Sawano, S., Goto, S. et al. A model driven by crop water use and nitrogen supply for simulating changes in the regional yield of rain-fed lowland rice in Northeast Thailand. Paddy Water Environ 6, 73–82 (2008). https://doi.org/10.1007/s10333-007-0099-1
• Iizumi, T., Sakai, T. The global dataset of historical yields for major crops 1981–2016. Sci Data 7, 97 (2020). https://doi.org/10.1038/s41597-020-0433-7
• Toshichika Iizumi, Toru Sakai, Yoshimitsu Masaki, Kei Oyoshi, Takahiro Takimoto, Hideo Shiogama, Yukiko Imada, David Makowski, Assessing the capacity of agricultural research and development to increase the stability of global crop yields under climate change, PNAS Nexus, Volume 4, Issue 4, April 2025, pgaf099, https://doi.org/10.1093/pnasnexus/pgaf099
• Müller, C., Elliott, J., Chryssanthacopoulos, J., Arneth, A., Balkovic, J., Ciais, P., Deryng, D., Folberth, C., Glotter, M., Hoek, S., Iizumi, T., Izaurralde, R. C., Jones, C., Khabarov, N., Lawrence, P., Liu, W., Olin, S., Pugh, T. A. M., Ray, D. K., Reddy, A., Rosenzweig, C., Ruane, A. C., Sakurai, G., Schmid, E., Skalsky, R., Song, C. X., Wang, X., de Wit, A., and Yang, H.: Global gridded crop model evaluation: benchmarking, skills, deficiencies and implications, Geosci. Model Dev., 10, 1403–1422, https://doi.org/10.5194/gmd-10-1403-2017, 2017.
• Lin, T.-S., Song, Y., Lawrence, P., Kheshgi, H. S., & Jain, A. K. (2021). Worldwide maize and soybean yield response to environmental and management factors over the 20th and 21st centuries. Journal of Geophysical Research: Biogeosciences, 126, e2021JG006304. https://doi.org/10.1029/2021JG006304
• Lombardozzi, D. L., Lu, Y., Lawrence, P. J., Lawrence, D. M., Swenson, S., & Oleson, K. W., et al. (2020). Simulating agriculture in the Community Land Model Version 5. Journal of Geophysical Research: Biogeosciences, 125, e2019JG005529. https://doi.org/10.1029/2019JG005529
• Luo, Y., Zhang, Z., Chen, Y., Li, Z., and Tao, F.: ChinaCropPhen1km: a high-resolution crop phenological dataset for three staple crops in China during 2000–2015 based on leaf area index (LAI) products, Earth Syst. Sci. Data, 12, 197–214, https://doi.org/10.5194/essd-12-197-2020, 2020.
• Akira MORI, Yasuhiro DOI, Toshichika IIZUMI, GCPE: The global dataset of crop phenological events for agricultural and earth system modeling, Journal of Agricultural Meteorology, 2023, Volume 79, Issue 3, Pages 120-129, Released on J-STAGE July 10, 2023, Advance online publication May 16, 2023, Online ISSN 1881-0136, Print ISSN 0021-8588, https://doi.org/10.2480/agrmet.D-23-00004,
• Alex C. Ruane, Richard Goldberg, James Chryssanthacopoulos, 2015. Climate forcing datasets for agricultural modeling: Merged products for gap-filling and historical climate series estimation. Agricultural and Forest Meteorology, 200, 233-248. https://doi.org/10.1016/j.agrformet.2014.09.016
• Su, H., Willaarts, B., Luna-Gonzalez, D., Krol, M. S., and Hogeboom, R. J.: Gridded 5 arcmin datasets for simultaneously farm-size-specific and crop-specific harvested areas in 56 countries, Earth Syst. Sci. Data, 14, 4397–4418, https://doi.org/10.5194/essd-14-4397-2022, 2022.
• Yang Y, Ren W, Tao B et al., 2020: Characterizing spatiotemporal patterns of crop phenology across North America during 2000-2016 using satellite imagery and agricultural survey data. ISPRS Journal of Photogrammetry and Remote Sensing 170, 156-173. https://doi.org/10.1016/j.isprsjprs.2020.10.005
• Wei, D., Gephart, J.A., Iizumi, T. et al. Key role of planted and harvested area fluctuations in US crop production shocks. Nat Sustain 6, 1177–1185 (2023). https://doi.org/10.1038/s41893-023-01152-2