the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 05 Feb 2024
Applicability of the inverse dispersion method to measure emissions from animal housings
Abstract. Emissions from agricultural sources substantially contribute to global warming. The inverse dispersion method has been successfully used for emission measurement from various agricultural sources. The method has also been validated in multiple studies with artificial gas releases mostly on open fields. Release experiments from buildings have been very rare and were partly affected by additional nearby sources of the same gas. What is also lacking are specific release studies for naturally ventilated animal housings. In this study, a known and predefined amount of methane was released from an artificial source inside a barn that mimics a naturally ventilated dairy housing. For concentration measurements, open-path devices (OP) with a path length of 110 m were placed in downwind direction of the barn at a distance of 50 m, 100 m, 150 m, and 200 m and additionally, a 3D ultrasonic anemometer (UA) was placed in the middle of the OP paths at 50 m, 100 m and 150 m. Upwind of the barn, an additional OP and an UA were installed. The median recovery rates of the experiment depending on the used OP and UA combination ranged between 0.56–0.71. It is concluded that for the present study case, the effect of the building and a tree in the main wind axis led to a systematic underestimation of the inverse dispersion method derived emission rate probably due to deviations of the wind field and turbulent dispersion from the ideal assumptions.
- Preprint
(2264 KB) - Metadata XML
-
Supplement
(3866 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on amt-2023-258', Anonymous Referee #2, 07 Feb 2024
Review of “Applicability of the inverse dispersion method to measure emissions from animal housings” by Bühler et al., submitted to Atmospheric Measurement Techniques.
General Comments
The simplicity of an idealized IDM technique for calculating gas emissions (relative to other approaches) makes the technique highly valuable, particularly if it can be accurately applied in situations that do not meet the “idealized” standard of horizontally homogeneous wind flow. This study looks at one such situation: emissions from animal barns. It examines the accuracy of an IDM calculation of barn emissions (Q) versus the distance from the barn where concentration (C) is measured. The idea is that while the C-Q relationship is affected by the wind complexity near the barn that results in IDM errors, as the distance of the C measurement increases the IDM error is reduced. At some threshold distance D from the barn, the idealized IDM calculation becomes suitably accurate. Exploration of the threshold D is useful given the wide range of emission sources accompanied by some type of wind complexity. The subject matter is suitable for AMT and will be a valuable contribution to efforts to understand the limitations of IDM (and how IDM can be used to provide accurate results).
This study follows from a handful of others having similar objectives. Earlier work has suggested that D depends on the height of obstacles (h) around the source (e.g., barn height or fence height). In this manuscript the authors conduct a tracer release study and found that for the range of downwind locations examined (D/h < 21 or < 12) that Q was biased to underestimate the true gas release rate. I have some general comments about the presentation and interpretation of the results. These are:
1. Previous studies suggested the threshold distance D (for idealized-IDM accuracy) relates to the dominant height h of obstacles around the source. These obstacles could be barns, trees, or fences. The idea being that the wind flow disturbance extends over a downwind distance scaling with h. In this experiment, the largest obstacle seems to be the large tree beside the barn. If the authors want to compare their results with earlier studies, they should focus on D scaled with the tree height (not the barn height). The authors are aware of this, but it should be stated more clearly.
2. The basis for arguing for a threshold distance D (for idealized-IDM accuracy) is that the wind disturbance caused by an obstacle has limited spatial extent, and far downwind of the disturbed zone the resulting gas plume will become indistinguishable from the plume exposed to undisturbed flow, so at some distance dispersion can be accurately calculated based on the undisturbed wind. In this thinking the proper wind measurement location is upwind of the disturbance (or very far downwind). Measurement locations in the disturbed zone have no purpose in this paradigm. In this study the authors measure wind in the disturbed zone and treat these as another valid option for use in IDM. I would like to see the authors recognize these locations do not fit the hypothesis espoused in earlier studies. I am not requesting they be eliminated, but they should not be taken as “equal” to the upwind measurements (unless the authors want to describe a different IDM paradigm).
3. The authors published an earlier paper critical of the accuracy of the Boreal GF3 laser for measuring gas concentration. It would be good to see comments reconciling this earlier critique with the use of GF3’s for this study. I am concerned about the inaccuracy of the GF3’s for this study. I am not convinced by authors broad statement that laser calibration can be excluded as a cause of Q error.
4. The section on “Plume modeling and wind field rotation” has limited usefulness and I suggest it be eliminated. The wind field downwind of the barn is clearly complicated and the barn plume structure will be impacted by that complexity. This is not represented in the IDM calculations of the plume maps. Using different anemometer locations in the different maps calculations (and then assuming a horizontally uniform field) does not address the problem – all the calculations are wrong in detail. There is little to be learned with these plume maps and they are deceptive. This material is not critical to the main objective.
5. The main results of this study (in terms of elucidating a threshold D for IDM) are broadly consistent with earlier work, and I would like the authors to more clearly state this (i.e., in the conclusions). Earlier studies suggest that D/h ranges from 5 to 30 over a range of circumstances. This study suggests D/h > 12. This broadly fits within earlier work. There is no reason to expect a universal D/h value (“Moving 10h (or any specific distance) downwind of an obstacle is unlikely to be a universal threshold for ignoring wind disturbances …”; Flesch et al. 2005, Deducing ground-to-air emissions from observed trace gas concentration, J Applied Meteorol). It would be good to more clearly place the study results into these earlier lines of thinking.
Specific Comments
6. Ln 29: “The IDM is a micrometeorological method that combines concentration measurements up- and downwind of the spatially defined source with an atmospheric dispersion model”. IDM is more flexible than described here. The input concentration measurements do not have to be an upwind-downwind pair. Bai et al. (2023, Measurement of long-term CH4 emissions and emission factors from beef feedlots in Australia, Atmosphere) is an example where vertically separated concentration measurements above the source were used to calculate emissions.
7. Ln 31: “The IDM with a bLS model has been verified in multiple release experiments on open fields that reflect ideal conditions in terms of Monin-Obukhov-Similarity theory …”. It would be good to remind the audience of situations when MO similarity theory should be theoretically accurate, e.g., a horizontally homogeneous surface layer, where the source-to-sensor distance would be less than order 1 km.
8. Ln 103: “ … factory calibration were applied …”. The authors previously published a very interesting paper on problems with the Boreal GF3 lasers used in this study: e.g., “Application with paired devices needs an intercalibration of the devices. However, it remains unclear to what extent a side-by-side intercalibration can be transferred to the actual measurement setup, since relocation of the devices might cause systematic changes, as indicated by the different regression coefficients for different intercomparison campaigns.” I would like the authors to comment on the capability of the GF3 to accurately measure CH4 concentrations in this study, in the context of their earlier criticisms.
9. Section 3.3. “Plume modeling and wind field rotation”. As mentioned earlier, I think this section has limited usefulness. There is little to be learned from the plume maps, and I believe they are deceptive. I would delete. I suggest a simpler example to illustrate the wind complexity (what about a simple wind vector plot of the wind measurement locations for a small number of periods?).
10. Ln 219: The authors discuss how the wind varies downwind of the barn, and reference earlier recommendations regarding how far an IDM concentration measurement should be made downwind of a barn. These earlier recommendations were made based on the “barn or other dominate obstacle height”. It is perhaps unfair to apply these earlier recommendations based on barn height – earlier authors would argue the larger tree is the dominate obstacle height.
11. Ln 241: “A bias in the results due to biases in the intercalibration of the OP or in the amount of released gas could be excluded. When the barn was excessively vented after the CH4 release, no increase in the downwind CH4 concentration could be observed, indicating that no CH4 was kept back inside the barn.” This conclusion is too strong. A bias in the absolute concentration measurement (of all lasers) would certainly lead to biased estimates of the emission rate. Errors in gas release rate would similarly bias the results.
Technical Corrections12. Ln 82: “… and a 20 m long with …” should this be a “20 m long tubing with …”?
13. Ln 107: “The concentrations between the five OP were inter-calibrated …” How were the lasers inter-calibrated? Did the regression fit only a multiplier (slope) or a slope & offset?
14. Ln 124: For the anemometer placed upwind of the barn … what was the distance from the anemometer to the closest upwind obstacle (trees? building? Equipment?), both in absolute distance and distance scaled with the obstacle height?
15. Ln 137: “… the concentration upwind of the source is equally measured.” What does “equally” mean here? Maybe it should be “also” measured?
16. Ln 222: “… if the 15 m high tree is considered as the relevant flow disturbance …” Because the tree is located southwest of the barn, is the “fetch” the distance from the tree to the downwind laser, or the distance from the barn? Clarify.
Citation: https://doi.org/10.5194/amt-2023-258-RC1 -
RC2: 'Comment on amt-2023-258', Anonymous Referee #1, 20 Feb 2024
General comments:
This article on the applicability of the inverse dispersion method for measuring emissions from animal housing is exemplary in its writing and argumentation. The clarity and coherence of the content make it a valuable contribution to the field. However, the review identifies two notable drawbacks – (1) the challenge of frequently suboptimal atmospheric conditions and the limited range of conditions considered, (2) the use of the threshold distance/fetch which is usually based on the dominant obstacle height (rather than the source height). Concerning the latter one, I would argue that the tree (15 m) is the dominant wind obstacle rather than the barn itself. The discussion (chapter 4) would be further enriched by incorporating more literature, providing a more comprehensive context for readers. Despite these considerations, the article stands out as a well-crafted and insightful exploration of emission measurement methods.
Specific comments:
line 25: The introduction could benefit from more recent literature (e.g. instead of Stocker et al., 2013 and Gerber, 2013)
line 100: please provide further information on the device specific relationships as it is important for the accuracy of the concentration measurements
line 110: please define the input parameters for the model
lines 120/145/215/240: An important scale to determine the distance between source and the downwind measurement location is the height of the largest wind obstacle. When comparing to other studies, it is essential to include the fetch based on the tree height rather than the barn height.
Table 3: I find the term “All UA” and “All OP” misleading. Isn’t it a mean/median value of the considered options?
Line 205: I appreciate the approach of a sensitivity analysis. But then other parameters should be considered as well (and not only the rotation of the wind direction).
Line 225: Despite the influence of the barn and tree, the recovery rates did not substantially differ.
Line 245: It may be worth to conduct a sensitivity analysis for several parameters instead of using only one – the rotation of the wind direction
Line 275: Delete the second “with” in the sentence “Other IDM studies have shown ….”
Citation: https://doi.org/10.5194/amt-2023-258-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 223 | 34 | 11 | 268 | 16 | 12 | 11 |
- HTML: 223
- PDF: 34
- XML: 11
- Total: 268
- Supplement: 16
- BibTeX: 12
- EndNote: 11
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1