Study of the Distribution Characteristics of the Airflow Field in Tree Canopies Based on the CFD Model
Round 1
Reviewer 1 Report
Dear Authors,
The manuscript focus is on modeling trees in CFD to account for the airflow through them for spray distribution. This is an important topic and I believe you have a good draft with explaining the problem and proposing a solution using porosity model.
There are some points which should be discussed in detail when it comes to numerical simulation so that the readers understand the work and the approach.
1) Did you do any mesh dependency study? What happens when you increase the mesh density for your domain? What happens if you increase your domain size?
2) Why was the pressure at all boundaries taken to be zero and not atmospheric? Could you run it with atmospheric pressure?
3) What about cross-wind conditions? How have you addressed it?
4) The relative error between numerical simulation and experiment seems too high (20%). An acceptable value is less than 10%. I do like the regression analysis to show the relevance. What can you do to bring this error down to say 15%?
5) What about the trees in the other rows? This study looks at isolated row only.
Author Response
Dears reviewers,
Thank you for your comments on my article during your busy schedule. The replies to each question are as follows:
(1)A grid independence test was added in Section 2.5 to increase the reliability of the data. In addition, the selection of calculating domain directly affects the research space. Increasing the computing domain will undoubtedly increase the amount of computation. Through the preliminary experiment, the length of 14m, width of 16m and height of 12m are determined as the computing domain.
(2)In this study, the arc surface of the sprayer was set as the velocity inlet, while there was no wall surface (except the ground) in the real operating environment, so the other boundaries were set as the pressure outlet. When the atmospheric pressure at the outlet was set to 0, it indicated that in this calculation domain, the wind dissipated under the action of canopy and air flow field. Therefore, the atmospheric pressure at any position and at any time can be calculated through the simulation model.
(3)In this study, the longitudinal wind speed of the outlet of the fan is more than 10 m/s, which was dominant in the canopy. Of course, there are cross-wind in the atmosphere, but due to their transient instability (direction, magnitude), the transversal winds were not considered in this model.
(4)In the validation test, multiple sets of data were added under the same operating conditions. The velocity in the canopy and the simulated value were evaluated again and the final calculated error was 13.68%.
(5)The 3WQF-1000 orchard air-assisted sprayer works for the tree on both sides of the sprayer at the same time, and the wind field is uniform and symmetrical. Therefore, the computing domain of the geometric model is defined as a one-sided tree environment in numerical simulation to avoid excessive computing costs.
Reviewer 2 Report
Dear Authors and Editors
I am not a specialist in agronomy, but I do work with CFD. The submitted manuscript is very interesting for me and it seems to be useful too. The research approach seems to be reasonable, however, I cannot recommend the paper for publication in the present form.
My first major concern is poor language, which makes understanding of the paper difficult. I would advise the authors to find a person, which will improve their language and also a person that is closer to CFD community that will smooth out some unusual expressions.The main example: airflow from the sprayer is not"wind". So I was quite confused before I have found the meaning of the "wind" as used in the manuscript.
My second concern: the CFD analysis is not adequately described, although the basic approach is sound. Please describe the mathematical model first (no need to write Navier-Stokes equations), but you start with turbulence model eq.s. (6), (7) first and AFTER that you specify the porosity model and the sources for k and eps transport equation - Eqs. (1), (2), (3), (5). I think than the sources (1)-(5) should appear in k-eps equations (6) and (7) inside the porosity volume.
I miss the mesh refinement sensitivity.
Further comments are under yellow painted text in the attached pdf file.
Parts of the yellow text without any comment point to the poor language. Only some of the poor sentences are marked.
Comments for author File: 
Comments.pdf
Author Response
Dear reviewers:
Thank you for your comments on my article during your busy schedule. The replies to each question are as follows:
(1)To response to the poor language proposed by reviewers, professionals have been invited to polish the language.
(2)This part has been adjusted in order in the paper. First the turbulence model, then the porous medium model.
(3) To increase the mesh refinement sensitivity, mesh independence test and mesh quality evaluation are added in section 2.4 and 2.5.
Reviewer 3 Report
The paper presents the investigation based on computational fluid dynamic approach of airflow of air-assisted sprayers in order to provide plant protection in China. The authors used computational fluid dynamics (CFD) method, k-ε turbulence model and SIMPLE algorithm to solve equations. 3D simulated model of the spatial distribution of the airflow field in and around the canopy of trees was established based on porous model in this paper. The paper is well written, interesting, well organized and results are presentative.
In my opinion, the manuscript deserves to be published in your journal. I have only MINOR revision suggestions.
Comments:
1.. Some applications in wind engineering concerning leafs and trees have already been published . Look at papers: Guan, D., et al., A Wind-Tunnel Study of Windbreak Drag, Agricultural and Forest Meteorology, 118 (2003), 1-2, pp. 75-84
Manickathan, L., et al., Parametric Study of the Influence of Environmental Factors and Three Properties on the Transpirative Cooling Effect of Trees, Agricultural and Forest Meteorology, 248 (2018), Jan., pp. 259-274
2. You can use other turbulence models, like RNG k-ε model, and Reynolds stress model (RSM) turbulence model. and compare the results. In that case statistical analysis can be usefull.
3. You can add some information concerning grid, grid refinement tests and CFD model parameters.
4. Figure 12 has letters in Chinese. Please, translate it into English.
Author Response
Dear reviewers:
Thank you for your comments on my article during your busy schedule. The replies to each question are as follows:
（1）These two articles provide reliable theoretical support for the study of drag coefficient in canopy and have been cited in relevant parts of the paper.
（2）It is important to study whether there are differences between different turbulence models on the velocity distribution of wind in the canopy for the stability of CFD model calculation. According to the relevant literature, the standard k-ε model can provide high prediction accuracy. Therefore, this paper adopts the standard k-ε model to simulate the airflow.
（3）In the last paragraph of section 2.4, the skewness was used to examine the tetrahedral mesh quality, which can be used for numerical simulation. Meanwhile, section2.5 of mesh independence test was added.
（4）Chinese has been modified to English.
Round 2
Reviewer 1 Report
Dear Authors,
The changes look good and manuscript is ready to be accepted for publication.
Author Response
Thank you for your comments on this paper .
Reviewer 2 Report
Authors have taken into account many of my remarks and ignored some others.
I still believe that "wind speed" should be replaced with "air velocity" because the authors did not make a study when external wind is blowing. So the expression is confusing. Figure 10 is still 3D. 2D image would be much clearer.
Author Response
Thank you for your comments on this paper.According to your suggestion, wind speed has been replaced with air speed. Figure 10 was changed to 2D.
