Computational Fluid Dynamics Modeling of Ammonia Concentration in a Commercial Broiler Building
Round 1
Reviewer 1 Report
Comments to the manuscript 2370537 “Computational fluid dynamics modeling of ammonia concentration in a commercial broiler building”, this study developed a model to predict the ammonia (NH3) emissions generated in the poultry production systems. The topic is interesting but I have few comments and observations.
To begin with, the title looks like authors developed the ammonia estimation model, but They did not. I suggest to rewrite the title.
Introduction section
Line 42-44. Do you know what are the approximate values of GHG (CO2, CH4 and N2O) emitted for a commercial broiler building or by hen in a specific time? Also you should consider how litter is management
Materials and methods
What is experimental unit? The facility, the NH3 sensor?
Once the fans are working in one way direction and you located four NH3 sensors across the facility (broiler building), the NH3 measure is the same in the four sensors?
What was the reason to not use a statistical tool?
Table 1. I wonder if diet is important to predict NH3, in this case crude protein content. How the model works to predict the ammonia emissions? Are there mathematical equations?
In total how many observations do you have a day, why the number of days were not increased?
Line 182 – 185. This paragraph looks more adequate for the Discussion section
Line 230. You said that mesh 3 and 4 were close enough, based on what??? To me the four meshes are close enough.
In figure 5. I wonder, why W1 or W2 are not displayed?
Results
In figure 4. Could you used the Mean Square Prediction Error (MSPE) in your data?
Conclusions
Line 400-408 these paragraphs are not a conclusion. Same in line 409.
Author Response
Reviewer #1:
Comment: To begin with, the title looks like authors developed the ammonia estimation model, but They did not. I suggest to rewrite the title.
Reply: In the authors opinion, the title effectively corresponds to the work that was carried out: “CFD modeling”. The title does not include the word “development”. In that sense, we think that the title is appropriate.
Introduction section
Comment: Line 42-44. Do you know what are the approximate values of GHG (CO2, CH4 and N2O) emitted for a commercial broiler building or by hen in a specific time? Also you should consider how litter is management
Reply: The broilers are housed in buildings with a solid floor and the litter material (pine shavings or rice hulls) is removed at the end of each fattening cycle (30 to 42 days) (reference [3] in the paper). The average annual gas emission rates from broiler houses using new litter material in each production cycle varied greatly among European countries, with values from 0.06 to 0.45 g day−1 broiler−1 for NH3, 0 to 46 mg day−1 broiler−1 for N2O, 55.2 to 98.4 g day−1 broiler−1 for CO2 and 0 to 50 mg day−1 broiler−1 for CH4 (Pereira et al., 2023).
This last paragraph was included in the manuscript in the Introduction Section. The new reference was also included.
Materials and methods
Comment: What is experimental unit? The facility, the NH3 sensor?
Reply: The experimental unit is described in the manuscript: it is the commercial broiler building used is this study. The measurement equipment of ammonia concentration is described in section: 2.2. Field measurements
Comment: Once the fans are working in one way direction and you located four NH3 sensors across the facility (broiler building), the NH3 measure is the same in the four sensors?
Reply: The concentration of NH3 was measured in each one of the NH3 sampling point. In this manuscript, the measured average NH3 concentration in four monitoring points is compared with numerical results, also with NH3 average in four points to the correspondent location. However, the NH3 concentration is different in each monitoring point.
Comment: What was the reason to not use a statistical tool?
Reply: As referred in the manuscript, the simulations were performed for a particular time instant of the broilers’ growing cycle, as shown in Table 2. In that sense, a statistical analysis is not appropriate.
Comment: Table 1. I wonder if diet is important to predict NH3, in this case crude protein content. How the model works to predict the ammonia emissions? Are there mathematical equations?
Reply: The crude protein content in the diet influences the NH3 emission and is considered a mitigation measure to decrease such losses. In this study, we used a fixed value of crude protein content (19.2%) to simulate the standards diets provided by the integrator under real conditions.
In the CFD model, a source of ammonia (NH3 Litter emission [kg/(m2.s)] in Table 2) is imposed on the surface representing the chicken litter, equal to the average value that was experimentally measured throughout the volume of the poultry house, for the particular instant to which the simulation corresponds. This was the average value experimentally measured at the four sampling points.
Comment: In total how many observations do you have a day, why the number of days were not increased?
Reply: There is one observation per hour at each of the NH3 sampling points. Contrary to the experimental study, in this case, numerical simulations were only performed for a fixed instant (hour) of each season that was found to be representative. Our goal was to compare the numerical simulation results with the experimental measurements for specific times, and thus evaluate the predictive ability of the numerical model.
Comment: Line 182 – 185. This paragraph looks more adequate for the Discussion section
Reply: The text included in this paragraph (Line 182 – 185) describes to the turbulence model used in the present modeling, that is the propose of the section 3.1. We maintain the paragraph.
Comment: Line 230. You said that mesh 3 and 4 were close enough, based on what??? To me the four meshes are close enough.
Reply: The average NH3 volume fraction values results showed a convergence behavior with the number of nodes, as seen in figure 3. Based on these results, mesh 3 was adopted. Using a more refined calculation mesh would greatly increase the computational effort without significant benefit to the calculation results.
Comment: In figure 5. I wonder, why W1 or W2 are not displayed?
Reply: Figure 5 aims to show that the flow pattern is very dependent on the simulation conditions (initial and boundary conditions). In this case, we show the difference of the flow pattern of the summer (S1) and mid-season (MS2) situation. W1 and W2 conditions were not included because they do not bring relevant additional information.
Results
Comment: In figure 4. Could you used the Mean Square Prediction Error (MSPE) in your data?
Reply: As mentioned, for each condition there is only one simulation result, so no statistical treatment can be done.
Conclusions
Comment: Line 400-408 these paragraphs are not a conclusion. Same in line 409.
Reply: These paragraphs introduce general information to the reader, prior to synthesize the results. We believe it makes sense to include these paragraphs at the beginning of the conclusions section.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
In general
This manuscript is written clearly, transparently and logically. However, not all procedures were performed to compare experimental and model results. To clarify the comparison of experimental and model data, it is necessary to give the normalized amount of ammonia (NH3) in kilograms produced in winter, summer and off-season in a commercial broiler building. The rationing procedure should include: per broiler, per day, per month in each season and the total number of kilograms per year. The results obtained should be presented in the form of a table.
In specific
Lines 324-327. It was written “…Although the vertical scale of the graph is limited to 30 mg/m3 for better depicting the evolution of NH3 concentration in the first half of the tunnel, NH3 concentration reaches values larger than 40 mg/m3 for configurations S1, S2, and W1”.
Figures 8 and 9 should illustrate all data without any restrictions.
Author Response
Reviewer #2:
Comments and Suggestions for Authors
In general
This manuscript is written clearly, transparently and logically. However, not all procedures were performed to compare experimental and model results. To clarify the comparison of experimental and model data, it is necessary to give the normalized amount of ammonia (NH3) in kilograms produced in winter, summer and off-season in a commercial broiler building. The rationing procedure should include: per broiler, per day, per month in each season and the total number of kilograms per year. The results obtained should be presented in the form of a table.
Reply: The simulations were performed for a particular time instant of the broilers’ growing cycle, as shown in Table 2. The extension proposed by the reviewer would be a completely different approach that we might consider in a future study, depending on the availability of experimental data.
In specific
Comment: Lines 324-327. It was written “…Although the vertical scale of the graph is limited to 30 mg/m3 for better depicting the evolution of NH3 concentration in the first half of the tunnel, NH3 concentration reaches values larger than 40 mg/m3 for configurations S1, S2, and W1”. Figures 8 and 9 should illustrate all data without any restrictions.
Reply: The maximum vertical scale limits were limited to 30 [mg/m3] for a better resolution of the graph.
The ordinate scales of the graphs in figures 8 and 9 have been changed and the full results are now presented.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Accepted
