Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Overall Comments
The manuscript addresses an important problem in BRDF modelling of row crops by incorporating intra-row heterogeneity through three structural control parameters (leaf width, length, and azimuth). The study is relevant and has potential applications in parameter inversion, crop growth monitoring, and yield estimation. The validation against DART and sensitivity analysis are valuable. However, the novelty compared with existing models is not sufficiently highlighted, and the manuscript requires language polishing. Overall, the paper has merit but needs revision before acceptance.

Major Comments

1. The contribution beyond existing models (e.g., 4SAIL-RowCrop, MRTM) should be clarified in the abstract. At present, the distinction from previous work is not sufficiently emphasised.
2. The discussion often remains descriptive (e.g., “good agreement with DART”). The physical reasons for observed discrepancies, particularly the ~5.8% error in the darkspot direction, should be analysed in more depth.
3. Figures 6–9 are difficult to interpret for readers unfamiliar with polar plots. More explanatory captions would help.
4. Limitations: The limitations section should be expanded, especially regarding vertical heterogeneity (e.g., stratified leaf angle, reproductive organs).

Minor Comments

1. The abstract should be written in a more accessible manner and avoid the use of references.
2. Table 1 (Inputs for PROSPECT-PRO model) is unnecessary; the values can be directly provided in the text.
3. Line 40: Replace “surface vegetation” with “terrestrial vegetation.”
4. Line 58: Delete the word “in.”
5. Line 333: “week row effect” should be corrected to “weak row effect.”
6. The conclusions are somewhat repetitive and should be condensed to emphasise the main contributions.

Author Response

Major Comments

Comments 1: The contribution beyond existing models (e.g., 4SAIL-RowCrop, MRTM) should be clarified in the abstract. At present, the distinction from previous work is not sufficiently emphasized.

Response 1: Thank you very much for this insightful comment. We have carefully revised the abstract to explicitly clarify our novel contributions beyond existing models such as 4SAIL-RowCrop and MRTM. As noted in Line 37, the updated abstract now emphasizes two main advancements: (1) the introduction of a framework that enables pronounced spatial heterogeneity and high tunability of LAVD, and (2) the capability to accurately simulate row crops across their entire growth cycle, thereby bridging discrete row structures with continuous developmental stages.

 

Comments 2: The discussion often remains descriptive (e.g., “good agreement with DART”). The physical reasons for observed discrepancies, particularly the ~5.8% error in the darkspot direction, should be analyzed in more depth.

Response 2: We sincerely appreciate this valuable comment emphasizing the need for a deeper physical interpretation beyond descriptive comparisons. We fully agree that understanding the mechanisms behind the discrepancies is essential for scientific rigor. In response, we have replaced the previously reported root mean square error of 5.8% with the maximum error of 22.6% to better highlight the extent of deviation. Furthermore, we have added a dedicated analysis section (Page 18, Line 486) and an accompanying figure (Page 19, Figure 15) to examine the correlation between LAVD spatial heterogeneity and BRDF characteristics. This new analysis indicates that the discrepancies—particularly the larger errors in the darkspot direction—are primarily attributed to variations in soil and leaf visibility, which are in turn governed by the spatial heterogeneity of LAVD that our updated model explicitly represents.

 

Comments 3: Figures 6–9 are difficult to interpret for readers unfamiliar with polar plots. More explanatory captions would help.

Response 3: We sincerely appreciate the reviewer’s constructive suggestion to enhance the readability of Figures 6–9 for readers who may be less familiar with polar plots. In response, we have added a paragraph at Line 351 to introduce the concept of the polar map and improve interpretability. The new text reads:
“Considering the directionality of BRDF, it is presented on a polar map, where the coordinate system represents zenith and azimuth angles. The angular coordinate corresponds to the azimuth angle, starting from north and increasing clockwise, while the radial distance from the center represents the zenith angle (0° at the center and 90° at the outer circle). This projection allows for an intuitive visualization of the spherical data on a two-dimensional plane, clearly illustrating the directional characteristics.”

 

Comments 4: Limitations: The limitations section should be expanded, especially regarding vertical heterogeneity (e.g., stratified leaf angle, reproductive organs).

Response 4: We are grateful for this valuable suggestion to strengthen our limitations section. The limitations have been extended and classified into two categories: vertical heterogeneity and dynamic evolutions. They are described in detail in the text (Section 4.3, starting at Line 517). 

 

Minor Comments

Comments 1: The abstract should be written in a more accessible manner and avoid the use of references.

Response 1: We agree that the original abstract was not academic and accessible. The references are all deleted. The abstract is revised to better focus on the research problem, our methodology, key findings, and the main contributions of our work.

 

Comments 2: Table 1 (Inputs for PROSPECT-PRO model) is unnecessary; the values can be directly provided in the text.

Response 2: We agree that integrating the parameters into the text improves the manuscript's conciseness. Accordingly, we have removed Table 1 and directly provided the input values for the PROSPECT-PRO model in the text at Line 282.

 

Comments 3: Line 40: Replace “surface vegetation” with “terrestrial vegetation.”

Response 3: Thank you for this precise terminology suggestion. We agree that "terrestrial vegetation" is more accurate, and we have replaced "surface vegetation" accordingly at the beginning of introduction:” Row crops, a significant category of terrestrial vegetation, are extensively distributed across the world.”

 

Comments 4: Line 58: Delete the word “in.”

Response 4: We thank the reviewer for the careful correction. We have deleted the word “in” and add a temporal adverbial (Line 69). The sentence is also optimized and now reads:” As early as 1972, Jackson et al. [10] demonstrated through experiments and simulations that neglecting row architecture induces significant reflectance bias for row crops canopy.”

 

Comments 5: Line 333: “week row effect” should be corrected to “weak row effect.”

Response 5: We sincerely apologize for this careless spelling error and feel grateful to the reviewer for catching this. The typo "week row effect" has been corrected to "weak row effect" at Line 368 now. We have also thoroughly proofread the manuscript to correct other minor grammatical and spelling errors.

 

Comments 6: The conclusions are somewhat repetitive and should be condensed to emphasis the main contributions.

Response 6: Thanks for your professional suggestion. We have refined the part of conclusions and highlighted the main contributions of the new model. The new text now focuses on the model's novel parameterization for capturing heterogeneity, its accuracy throughout the row crops life cycle, and the critical impact of LAVD heterogeneity on BRDF simulations.

Reviewer 2 Report

Comments and Suggestions for Authors

Major Points

1. How can the three control parameters of the model be determined in practice? There should be some discussion.

2. DART is used as the reference in the evaluation of the new model, which is aimed at “accurate BRDF simulations“. Question is, why not just use DART for simulations? What are the advantages of the proposed model over DART?

3. The error value “5.8%” only occurs in abstract and conclusions. This is not acceptable. There should not be numbers that do not occur in the main body but only occur in abstract/conclusions.

4. Besides the existing polar plots, consider adding Yan Model’s values in Table 3 (or a separate table).

Minor Points

1. L88: citation form of Liou et al. and Wang et al. is different from what is used in the rest of the manuscript.

2. L100: the statement is too absolute. Either explain in detail or use a softer tone instead.

3. L154: incorrect citation for Matti et al. [5] is Mõttus et al.

Author Response

Major Points

Comments 1: How can the three control parameters of the model be determined in practice? There should be some discussion.

Response 1: We sincerely thank the reviewer for this important and insightful comment. In response, we have added a new subsection (Section 4.3, starting at Line 518) to discuss the practical determination of the three control parameters. Regarding the leaf width control parameter β, Mõttus et al. identified it as a species-specific parameter that can be estimated through direct measurements. Extending this reasoning, we suggest that all three control parameters are likely species-specific, as different species exhibit distinct leaf shapes, lengths, and azimuth distributions. Therefore, these parameters can be determined empirically through targeted field or laboratory measurements.

 

Comments 2: DART is used as the reference in the evaluation of the new model, which is aimed at “accurate BRDF simulations“. Question is, why not just use DART for simulations? What are the advantages of the proposed model over DART?

Response 2: We sincerely thank the reviewer for this critical and insightful question, which provides an excellent opportunity to clarify the necessity and advantages of our proposed model. DART is indeed a well-established, high-accuracy 3D radiative transfer model, and for this reason, we used it as a rigorous reference for validation. However, despite its accuracy, DART’s detailed geometric construction and high computational demand limit its practicality for large-scale or inversion-oriented applications. In comparison, our proposed model offers the following advantages:

1. Scene Parameterization
   DART requires the prior construction of complex, scene-specific 3D geometries. In contrast, our model quantitatively represents a row-crop scene using only three control parameters for heterogeneity (e, β, and ψ), along with fundamental canopy structural parameters (LAI, H, L, etc.). This formulation explicitly defines and controls the spatial heterogeneity of LAVD without depending on detailed 3D reconstruction, thereby transforming heterogeneity from an implicit geometric attribute into a set of portable and quantifiable parameters.
2. Parameter Inversion Capability
   The explicit and quantitative parameterization makes our model inherently suitable for inversion from remote sensing observations. While DART excels in forward radiative transfer simulations, its structural complexity and computational cost render parameter inversion highly challenging in practice.
3. Computational Efficiency
   By abstracting the geometric and radiative complexities of the scene into a few interpretable parameters, our model achieves substantial computational efficiency, enabling large-scale or multi-temporal simulations that would be computationally prohibitive with DART.

 

Comments 3: The error value “5.8%” only occurs in abstract and conclusions. This is not acceptable. There should not be numbers that do not occur in the main body but only occur in abstract/conclusions.

Response 3: We sincerely thank the reviewer for pointing out this important issue. We apologize for the earlier oversight of presenting the “5.8%” error value only in the abstract and conclusions without corresponding discussion in the main text. In the revised manuscript, we have replaced the previously reported root mean square error of 5.8% with the maximum error of 22.6% to better highlight the discrepancy, and this value is now explicitly mentioned in Line 473. Furthermore, we have added a new subsection (Line 487: Correlation between LAVD spatial heterogeneity and BRDF) to provide a more in-depth analysis of the sources of this error.

 

Comments 4: Besides the existing polar plots, consider adding Yan Model’s values in Table 3 (or a separate table).

Response 4: Thank you for the suggestion. We add the equivalent half-width (L_e) of row structure used for Yan model in Table 1 (Line 323). L_e is the average length of leaves in the scene

 

Minor Points

Comments 1: L88: citation form of Liou et al. and Wang et al. is different from what is used in the rest of the manuscript.

Response 1: Thank you for pointing this out. We have deleted this and add new citations (Line 98).

 

Comments 2: L100: the statement is too absolute. Either explain in detail or use a softer tone instead.

Response 2: Thank you for the valuable suggestion. We agree that the original statement was overly absolute and somewhat sketch. To address this, we have rewritten the paragraph (now Line 111~120) to adopt a softer tone and explain the limitations of existing models in detail.

 

Comments 3: L154: incorrect citation for Matti et al. [5] is Mõttus et al.

Response 3: Thanks for identifying this citation error. We have corrected it to "Mõttus et al." at line 178 and checked the formatting of all other references for consistency and accuracy. (now Line 175).