| 28 Jul 2026
Novel mobile measurements of ammonia and methane: distinguishing agricultural and traffic sources in the Netherlands
Abstract. A mobile atmospheric measurement platform combining an open-path ammonia (NH₃) analyzer (HT‑8700E) and a closed-path tunable infrared laser direct absorption spectroscopy (TILDAS) multi-gas system was deployed to enable simultaneous, high-frequency observations of NH₃, methane (CH₄), and carbon monoxide (CO) under real-world conditions. The open-path configuration minimizes inlet-related artefacts and enables fast-response NH₃ measurements. A comprehensive measurement framework is presented, including instrument integration, field calibration and validation using co-located reference measurements, data synchronization, plume identification, background removal, and plume-integrated enhancement analysis.
The performance of the platform was evaluated during a 160 km mobile campaign in the Veluwe region of the Netherlands, a Natura 2000 nature reserve surrounded by intensive livestock farming. During the campaign, plumes from 49 individual sources and approximately 630 transient traffic-related events were detected, resolving short-lived concentration enhancements across agricultural, industrial, and traffic-dominated environments. Plume-integrated enhancement ratios (ΣNH₃:ΣCH₄) showed systematic differences among source types, with low ratios for livestock farms (∼0.04 for cattle and goats), intermediate values for industrial sources, and substantially higher ratios for manured fields (>2) and traffic plumes (>4). In addition, NH₃ emission factors derived for traffic increased with driving speed from approximately 0.12 g kg⁻¹ fuel at low speeds to ~0.28 g kg⁻¹ under highway conditions.
These results demonstrate that simultaneous mobile measurements of NH₃, CH₄, and CO provide an effective approach for detecting and differentiating emission sources under real-world conditions. The combined observations reveal distinct source signatures and potential gaps in existing emission inventories. Although uncertainties remain due to atmospheric variability, calibration, limited temporal sampling, and overlap in source-specific ratios, the methodology shows strong potential for high-resolution characterization of heterogeneous NH₃ emissions through repeated mobile monitoring campaigns.
Review of “Novel mobile measurements of ammonia and methane: distinguishing agricultural and traffic sources in the Netherlands” by J. Zhang et al.
Summary:
This manuscript presents simultaneous measurements of NH3, CH4, CO, N2O, and other trace species using a mobile laboratory to distinguish between agricultural and non-agricultural emission sources in the Netherlands. Although the dataset is constrained by limited sampling (4–5 hours on a single day), this study provides a valuable contribution to the literature, where mobile-platform, multi-species strategies for source separation are currently underrepresented. While this work establishes a useful foundation for future studies, the manuscript requires additional detail in specific sections and refinement to improve clarity. I recommend publication in AMT following the suggested revisions below.
General Comments:
While the specific deployment of this instrument suite aboard this mobile laboratory in this region may be unique, characterizing the entire approach as “novel” in the title seems misleading. Given a history of similar mobile sampling techniques reported in the literature, I suggest adjusting the title to more accurately reflect the study’s objectives. See additional comments below.
The manuscript lacks a dedicated section on measurement uncertainty. Reported results should incorporate propagated uncertainties associated with measurement systems, analysis methods, and statistical variability.
The manuscript neglects to address the impact of NH3 gas-particle transformation between emission and sampling. This is an important consideration for atmospheric NH3 sampling given that atmospheric lifetime and transformation timescales can be quite short under certain conditions (like those during the sampling period). The implications of this process should to be discussed accordingly in the context of this work.
The main paper lacks sufficient cross-referencing to the SI.
Specific Comments:
Abstract, L23 ish: It should be acknowledged that the sample size for traffic plumes is significantly smaller than for other sources.
P2, L36: Expand the introduction to include NH3’s role in fine particle formation and its broader implications for regional air quality, rather than focusing solely on its contribution to reactive nitrogen deposition.
P2, L70: Clarify that Sun et al. (2014) focused on US environments, while Sun et al. (2017) provided a US-China comparison. Consider adding Miller et al., 2015, https://agupubs.onlinelibrary.wiley.com/doi/10.1002/2015JD023241.
P3, L78: The manuscript overstates the novelty of the mobile sampling framework. While the specific deployment of this instrument suite in the Netherlands may be unique, the broader approach of using open- and closed-path in-situ spectrometers on a mobile platform to distinguish between agricultural and non-agricultural sources is not new and there are publications in the literature reflecting this. Some examples are already identified and referenced in this work (e.g., Golston and Sun). Since there are a finite number of these publications, I recommend acknowledging and expanding the literature review to include:
Miller et al., 2015, https://agupubs.onlinelibrary.wiley.com/doi/10.1002/2015JD023241.
Eilerman et al., 2016, https://doi.org/10.1021/acs.est.6b02851.
Juncosa Calahorrano et al., 2024, https://doi.org/10.1021/acs.est.3c10902.
Tao et al., 2015, https://link.springer.com/article/10.1007/s00340-015-6069-1.
Thornhill et al., 2010, https://acp.copernicus.org/articles/10/3629/2010/.
Section 2.1: Some additional information about the study area would be helpful for readers unfamiliar with the types of sources in this region. For example, it would help to clarify the nature of local livestock operations (e.g., dairy vs. beef; confined/concentrated operations vs. open pasture). Furthermore, identify other potential regional emission sources, such as wetlands, which could significantly influence NH3, N2O, and CH4 signals.
Section 2.2.1: This section could benefit from additional details around the following concepts:
P5, L122 says that Met data was acquired from a nearby airport. Is there a reason the mobile lab did not have a met sensor onboard during drives?
Section 2.3.1: It would help to have more details in the SI about how the open-path NH3 system was calibrated. This is also a good place to explicitly discuss accuracy and precision (or total uncertainty) of the measurements used in this analysis compared to the other instrument mentioned on L158. Along the same lines, what are the uncertainties of the other supporting measurements from the TILDAS?
P6, L161: For measurements with OSS<10%, what does this translate to in terms of a detection limit in ppb or ug/m3?
P6, L171: What are the other potential sources you found? Can you list them here or later in the results and discussion?
P6, L178: The findings of Lassman et al., 2020, https://doi.org/10.1016/j.agrformet.2020.107989, are an important consideration for this work and would be worth discussing in the framework of this analysis for context.
P7, L208: Can you justify using this approach over the other methods, particularly regression analysis? Can you perform a sensitivity test to show how the results and uncertainties are impacted if one method is used over another? Could the differences in methodology of analysis be used to provide some estimate of the overall uncertainties on the results?
P8, L245: Can the database be further disaggregated to distinguish between dairy and beef cattle, which can have very different NH3:CH4 ratios?
P9, L263: How do processes like respiration and uptake influence the use of CO2 in this context?
P9, L266: Is this a reasonable assumption given that enhanced NH3 emissions have been associated with diesel vehicles outfitted with selective catalytic reduction (SCR) systems? What did the on-road and off-road fleet look like in this region (e.g., are vehicles predominantly gasoline, diesel, EV, hybrid) during the time period of this study? Have SCR systems been implemented in this area? If yes, when did this happen in relationship to the timeframe of the Burgard 2006 study? See also, Farren et al., 2020, https://doi.org/10.1021/acs.est.0c05839.
P10, L292: Are wetlands a potential source in this area?
P10, L299: Simultaneous enhancements in N2O corroborate this statement. See Eilerman et al., 2016, https://doi.org/10.1021/acs.est.6b02851.
Figure 5: Can you further refine dairy vs beef cattle as the ratios can be quite different for these types of facilities? For example see Juncosa Calahorrano et al., 2023, https://doi.org/10.1029/2023JD039043.
P14, L351: Could a simple Multivariate Linear Regression model (or a Positive Matrix Factorization model if there are enough observations) be applied to these measurements for a source apportionment analysis (maybe something like CO = traffic, N2O = manured fields, NH3 = cattle livestock, CH4 = enteric fermentation, ethane = petrochemical)? MLR examples, include Kille et al., 2019, https://doi.org/10.1029/2019GL082132; McCabe et al., 2023, https://acp.copernicus.org/articles/23/7479/2023/; Pollack et al., 2022, https://doi.org/10.1021/acs.est.1c07382; and Lill et al., 2025, https://doi.org/10.1029/2025JD044019.
Section 3.4 (and Figure 7): It would help to report and compare absolute values of NH3 and CH4. One way to exemplify this could be to add a box plot (or violin plot) of the observed mixing ratios for all species. This could be a multipaneled figure broken down by source like Figure 7.
Table 1: Are traffic NH3 to CH4 ratios biased by low CH4 emissions from vehicles? How would a similar analysis of NH3 to CO ratios change your conclusions about agricultural vs. non-agricultural sources in this region?
Section 3.5: See comment above about SCR. Also, did you ever just sit and sample in an exclusively urban area to characterize predominantly traffic-related ratios?
P21, L459: I disagree with this phrase: “an advantage not achievable with closed-path instruments“. Prior literature shows that closed-path NH3 TILDAS instruments are capable of achieving adequate response times for this type of sampling and analysis (See Pollack et al., 2019, https://doi.org/10.5194/amt-12-3717-2019; Roscioli et al., 2015, https://doi.org/10.1021/acs.jpca.5b04395; and others coming soon using a mobile lab). A more objective discussion of the advantages and disadvantages of open and closed-path instruments is recommended. The differences in data coverage reported in this analysis on P10, L304 highlights some of those differences.
P21, L477: Should be extent not extend.
Section 4.4: You could similarly compare NH3 to carbon ratios with NH3:CO ratios reported for North American cities by Lill et al., 2025, https://doi.org/10.1029/2025JD044019.
Figure S1: The color scales are not particularly useful for distinguishing enhancements. One way to highlight datapoints, especially if one or two high points are skewing the scale, is to reduce the full-scale range and report an off-scale maximum.
Figure S3: Add a 1:1 line to guide the eye.