Summertime Airborne Measurements of Ammonia Emissions From Cattle Feedlots and Dairies in Northeastern Colorado

Phase One of the Transportation and Transformation of Ammonia (TRANS2Am) field campaign took place in northeastern Colorado during the summer of 2021. One of the goals of TRANS2Am was to measure ammonia (NH3) emissions from cattle feedlots and dairies. Most of these animal husbandry facilities are co‐located within oil and gas development, an important source of methane (CH4) and ethane (C2H6) in the region. Phase One of TRANS2Am included 12 near‐source research flights. We present estimates of NH3 emissions ratios with respect to CH4 (NH3 EmR), with and without correction of CH4 from oil and gas, for 29 feedlots and dairies in the region. The data shows larger emissions ratios than previously reported in the literature with a large range of values (i.e., 0.1–2.6 ppbv ppbv−1). Facilities housing cattle and dairy had a mean (std) of 1.20 (0.63) and 0.29 (0.08) ppbv ppbv−1, respectively. We also found that only 15% of the total ammonia (NHx) is in the particle phase (i.e., NH4+ ${\text{NH}}_{4}^{+}$ ) near major sources during the warm summertime months. We examined the evolution of NH3 in one plume that was sampled at different distances and altitudes up to 25 km downwind and estimated the NH3 lifetime against deposition and partitioning to the particle phase to be 87–120 min. Finally, we calculated estimates of NH3 emission rates from four optimally sampled facilities. These ranged from 4 to 29 g NH3 · h−1 · hd−1.

• We report ammonia emission rates and emission ratios with respect to methane for beef cattle feedlots and dairies in Colorado • The NH 3 EmR estimates are generally larger than prior estimates.NH 3 EmR estimates for feedlots are four times larger than those for dairies • NH 3 inferred emission rates for cattle feedlots range from 4 to 29 g NH 3 • h −1 • hd −1 for sampling temperatures between 19 and 27°C Supporting Information: Supporting Information may be found in the online version of this article.

10.1029/2023JD039043
2 of 20 Most of the NH 3 emissions in the United States are attributed to fertilizer application (21%) and livestock (59%) (Environmental Protection Agency (EPA), The National Emission Inventory (NEI), 2017).The United States is the largest producer of beef in the world and 80% of the country's beef cattle and dairy cows are concentrated in the Great Plains region and in parts of the Corn Belt, Southwest, and Pacific Northwest (USDA, 2012).NH 3 from livestock is emitted to the atmosphere through biological and chemical bacterial decomposition of excreted N. Livestock eat protein-and N-rich feed to yield desirable N-rich products (i.e., meat, milk, and eggs).However, most of the N in the feed (70%-95%) is eliminated through excretion rather than converted to the N-rich products, resulting in large emissions of NH 3 and other N compounds (i.e., nitrous oxide [N 2 O]) (Huntington & Archibeque, 2000).Other gasses emitted from animal feeding operations include three greenhouse gasses: methane (CH 4 ), N 2 O, and carbon dioxide (CO 2 ).CH 4 and N 2 O have 28 and 265 times the 100-year Global Potential Warming of CO 2 , respectively (IPCC, 2021).Livestock emits CH 4 through enteric fermentation.Emitted NH 3 and CH 4 from livestock vary greatly depending on diurnal and seasonal cycles and the number of animals in each facility.Eventually, NH 3 emitted into the atmosphere will deposit to the water or soil surfaces as gas or particles or through precipitation events.Although, as mentioned above, NH 3 deposition could negatively affect ecosystems such as water eutrophication and soil acidification, in N-poor ecosystems such as N-starved farmlands adjacent to CAFOs, NH 3 deposition could have beneficial fertilization effects (Preece et al., 2017).Manure treatment and storage methods greatly influence the emissions of NH 3 and CH 4 (Eilerman et al., 2016).Proposed strategies to reduce the emission of NH 3 to the environment include reducing high protein feed, frequent removal of manure and separation of it from urine, filtration of emissions from confinement facilities (i.e., scrubbers/ filters), controlling conditions to keep low temperatures and low pH in the manure, and injection or incorporation of NH 3 into the soil soon after fertilizer application (Ndegwa et al., 2008).
Historical challenges in measuring NH 3 (i.e., Miller et al., 2014;Roscioli et al., 2016) have limited progress on emissions estimates for this pollutant.Recent advances in measuring NH 3 , including in-situ measurements (e.g., Ellis et al., 2010;Pollack et al., 2019;Roscioli et al., 2016) and satellite retrievals (e.g., Cady-Pereira et al., 2023;Van Damme et al., 2018), have increased our awareness of the importance of measuring NH 3 from large point sources.The availability of new observations has helped models to represent NH 3 emissions in the United States (Zhu et al., 2015) and worldwide (Clarisse et al., 2009) better and shine light on the potential underestimation of emission inventories on NH 3 emissions (e.g., Heald et al., 2012;Nowak et al., 2012).Considerable deposition of N in and around sensitive ecosystems has been identified as one of the leading problems of NH 3 emission in the United States (Benedict et al., 2013a(Benedict et al., , 2013b) ) and globally (Liu et al., 2022).Due to the large number of NH 3 emission sources and the uncertainty in the magnitude of their NH 3 emission, more detailed measurements are needed.
Here, we report on summertime airborne observations of NH 3 and CH 4 collected over northeastern Colorado during Phase One of the Transportation and Transformation of Ammonia (TRANS 2 Am) field intensive.We use these data to produce (a) a summary of summertime NH 3 emission ratios with respect to CH 4 representing 29 beef cattle and dairy facilities and their dependence on temperature and time of day, (b) estimates of the differences between emission ratios associated with beef cattle and dairy cows facilities, and (c) NH 3 emission rate estimates as a function of maximum animal capacity for select comprehensively sampled facilities.

Study Region
Colorado has a large number of livestock operations (Figure 1a), and the majority and largest facilities (in terms of maximum reported animal capacity) in this region house beef cattle and dairy cows (blue and pink dots, respectively, in Figure 1).Over 1 million animals are clustered in counties in the northeast part of the state (i.e., Larimer, Weld, Morgan, Washington, Yuma, Logan, and Phillips).While many of these operations are large sources of trace gases, separating and quantifying the emissions from individual facilities is difficult because many facilities are located in close proximity to the dense oil and gas development throughout much of the area (i.e., Denver-Julesburg basin-black dots in Figures 1a and 1b), as well as large urban centers (i.e., vehicle traffic and industrial sources along the Colorado Front Range corridor).Rocky Mountain National Park (RMNP) and other sensitive high-alpine areas are located directly west of the polluted Colorado Front Range.N deposition in this area is dominated by reduced N during upslope events (easterly winds) that carry emissions from the eastern plains to the mountain ecosystems (Benedict et al., 2013a(Benedict et al., , 2013b;;Li et al., 2016;Pan et al., 2021).Agricultural Several recent measurement campaigns have aimed at characterizing emissions from animal husbandry in the region.For example, Eilerman et al. (2016) reported on a year-long ground-based survey of four facilities housing beef, dairy, and sheep, and this report summarizes their diurnal and seasonal variations.The study highlights the strong relationship between NH 3 emissions and time of the day.Kille et al. (2019) apportioned CH 4 emissions in the region to either oil and gas or agriculture using ethane (C 2 H 6 ) and NH 3 as tracers for each of these sources, respectively.This resulted in NH 3 emissions ratios with respect to CH 4 for the region.The Ammonia Phase Partitioning and Transport (APART) field campaign was the proof of concept field study leading to TRANS 2 Am.They characterized plumes downwind of five beef cattle facilities during November 2019, showing that NH 3 near-source emissions can be tracked using airborne platforms.They found that large NH 3 emissions ratios can be observed during cooler temperature conditions in the region (McCabe et al., 2023;Pollack, McCabe, et al., 2022).Finally, Golston et al. (2020)

10.1029/2023JD039043
4 of 20 and found a large underestimation in emissions inventories (i.e., NEI and EDGAR) for NH 3 and CH 4 as well as significant site-to-site variability for NH 3 and CH 4 emissions.The NH 3 emission ratios with respect to CH 4 from all the studies listed above range between 0.17 and 2.7 ppbv ppbv −1 .Even fewer studies (Golston et al., 2020;Kille et al., 2017;McCabe et al., 2023) report emission rates of NH 3 (rather than emissions ratios that are normal- ized by CH 4 ) in the region.The few studies reporting emission rates of NH 3 highlight the need for more systematic measurements to estimate this magnitude in the region (Golston et al., 2020;Kille et al., 2017;McCabe et al., 2023).Because of the large number of facilities in the region and the large variability in their near-source emissions and evolution, large uncertainties remain on what NH 3 emissions from livestock are in Colorado.

Campaign Overview
The TRANS 2 Am field campaign occurred over two phases: (a) 27 July 2021 to 23 August 2021 and (b) 16 August 2022 to 2 September 2022.The field campaign was halted abruptly in August 2021 when the plane was damaged by a collision with a bird and then resumed in 2022.Here, we focus on data collected in 2021.During both phases of TRANS 2 Am, the University of Wyoming King Air (UWKA) was based at Laramie Regional Airport (KLAR) in Laramie, WY, and was deployed to the northern Colorado Front Range. Figure 1a shows the study region, and Figure 1b shows the facilities for which we provide NH 3 emission estimates.The flight patterns associated with TRANS 2 Am were designed to meet two sets of objectives.The first set of objectives focuses on near-source emissions and evolution, and the second set of objectives focuses on the regional transport of reduced N into the nearby Rocky Mountains.This manuscript focuses on the first set of objectives.Figure S1  ).The following sections provide details on the instrumentation and flight patterns deployed during TRANS 2 Am.

Gas-Phase NH 3
NH 3 was measured using a Colorado State University (CSU) owned and operated commercial (Aerodyne Research, Inc.), single-channel, quantum-cascade tunable infrared laser direct absorption spectrometer (QC-TILDAS) operating at 967 cm −1 with and effective path length of 76 m (Ellis et al., 2010;McManus, 2010;McManus et al., 1995;Zahniser et al., 1995).The NH 3 instrument was utilized aboard the UWKA during APART, and details of the instrumentation are available in Pollack, McCabe, et al. (2022).Briefly, the spectrometer uses a direct absorption technique combined with a high sample flow rate (>10 SLPM) to achieve a fast (up to 10 Hz) collection of absolute NH 3 mixing ratios.The NH 3 QC-TILDAS is operated with a heated inertial inlet to provide filter-less separation of particles >300 nm from the sample stream (Ellis et al., 2010).Prior studies show active continuous passivation of the instrument flow path with a strong perfluorinated base improves the time response of the NH 3 QC-TILDAS on mobile platforms (Roscioli et al., 2016).However, we found that a response time of 1-3 s associated with a 90% recovery in NH 3 signal could be maintained during TRANS 2 Am flights without passive addition by regularly cleaning the instrument sampling surfaces between flights (Pollack et al., 2019).The NH 3 TILDAS is mounted on a vibrationally isolated apparatus and a constant high-frequency vibration is applied to the laser objective to wash out etalon fringe effects due to motion in flight, and thus there is minimal impact of motion sensitivity on instrument precision (Pollack et al., 2019).An injection-style aircraft inlet allows calibration and passivation gases to be introduced into the sample stream within a few centimeters of the inlet tip.The QC-TILDAS was calibrated on the ground between flights via standard addition to the sample stream with a known concentration of NH 3 generated from a temperature-regulated permeation tube.The instrument was regularly zeroed in flight by overflowing the inlet tip with a bottled source of NH 3 -free, ultrapure (or "zero") air.The emission rate of the permeation device was calibrated before and after the flight intensive by the NOAA ultraviolet (UV) optical absorption system (Neuman et al., 2003).As reported by Pollack et al. (2019), adding individual uncertainties in quadrature resulted in a combined uncertainty of ±12% of the measured mixing ratio.During TRANS 2 Am, the 1-Hz NH 3 measured mixing ratio had a 1-Hz precision in flight of 60 pptv corresponding to a 3-sigma detection limit of 180 pptv detection limit.Thus, the overall uncertainty of the instrument is reported as ±12% of the measured mixing ratio plus the 180 pptv detection limit.

Gas-Phase HNO 3
Similar to NH 3 , HNO 3 was measured using a commercial (Aerodyne Research, Inc.), single-channel, QC-TILDAS but operating at 1,723 cm −1 with an effective path length of 76 m.The HNO 3 instrument is owned by Aerodyne and was operated by CSU during the TRANS 2 Am field campaign.To make space for the complete payload and to maintain the >10 SLPM sample flow rate for up to 10 Hz collection, the NH 3 and HNO 3 instruments shared a common aircraft inlet, inertial inlet, and pumping system.Like the NH 3 instrument, the HNO 3 instrument was calibrated on the ground between flights via standard addition to the sample stream with a known concentration of HNO 3 generated from a temperature-regulated permeation device (Kin-Tech; verified by the NOAA UV optical absorption system; Neuman et al., 2003).NH 3 and HNO 3 calibrations were performed individually with copious flushing of the sampling surfaces of the common inlet before the application of the other calibrant.The HNO 3 instrument was regularly "zeroed" with a bottled supply of ultrapure air in flight.Like NH 3 , the HNO 3 instrument time response can be improved using active continuous passivation of the sampling surfaces using a strong acid (Roscioli et al., 2016).However, passive addition is not possible when using a combined sample flow path with NH 3 .The typical time response of the non-passivated HNO 3 instrument is ∼70 s for a 90% recovery in signal (Roscioli et al., 2016).During TRANS 2 Am, the time resolution of the HNO 3 instrument was degraded to ∼500 s for 90% signal recovery owing to the use of a common inlet and the HNO 3 QC-TILDAS being positioned downstream of the NH 3 QC-TILDAS in the flow path.Given this long-time response in mind, future comparisons between HNO 3 and other species will require convolution of the fast measurements to the slower HNO 3 data.All the same, HNO 3 data were collected at 10 Hz and averaged to 1 Hz during Phase One of TRANS 2 Am.The 1-Hz precision was 185 pptv, corresponding to a three-sigma detection limit of 555 pptv at 1 Hz.The HNO 3 spectrometer was also mounted on vibration isolators and a constant high-frequency vibration was applied to the laser objective, and thus motion sensitivity in flight had a minimal impact on precision.The uncertainty related to the 1 Hz samples is ±20% of the measured mixing ratio plus the 555 pptv detection limit.

C 2 H 6
C 2 H 6 measurements were collected at 1 Hz using a University of Wyoming-owned and operated commercial spectrometer (Aerodyne Research, Inc., Ethane Mini Trace Gas Monitor) employing a similar tunable infrared laser direct absorption spectroscopy (QC-TILDAS) technique as the NH 3 and HNO 3 instruments (Zahniser et al., 1995).The C 2 H 6 instrument uses a 2,990 cm −1 distributed feedback tunable diode laser, a multipass cell with a path length of 76 m (McManus et al., 1995) and an infrared detector.The C 2 H 6 QC-TILDAS is described in detail in Yacovitch et al. (2014).The instrument was zeroed periodically in flight with UZA and calibrated on the ground between flights using a high-accuracy (2.09 ± 0.01 ppb) standard purchased from NOAA ESRL.The 1-Hz precision in flight was 90 ppt resulting in a three-sigma detection limit of 270 ppt.

CH 4 , CO, CO 2 , and H 2 O
CH 4 , CO, CO 2 , and H 2 O were measured simultaneously using a University of Wyoming owned and operated Picarro G2401-m flight-ready analyzer.The instrument samples each species in rotation at ∼0.3 Hz.This closed-path instrument employs infrared cavity ring-down spectroscopy.Ambient air is pumped at a flow rate of 600 mL min −1 into an optical cavity that is maintained at 45°C and 140 Torr (Crosson, 2008).Ultra-high reflectivity mirrors allow for multiple passes in the cavity, creating an effective path length of >10 km and leading to high measurement sensitivity (Crosson, 2008).Precision was 30 ppb for CO, 200 ppb for CO 2 , and 2 ppb for CH 4 with low drift.The stated low drift for a 24-hr period is 1.5 ppb for CH 4 .Most of the flights of TRANS 2 Am were 4 hr long, which results in <2 ppb of drift (bellow the noise of the instrument).The instrument was zeroed using a bottled supply of UZA and periodically calibrated on the ground between flights with a high-precision NOAA ESRL standard.

Aerosol Composition
Cations, anions, organic acids, and carbohydrates were measured using a Particle-into-Liquid Sampler (PILS) coupled with a fraction collector.This system allows for the collection of liquid samples for offline analysis by ion chromatography.The PILS collects ambient particles into purified water.After particles are grown inside the body of the PILS by mixing the cool airflow with hot steam, the particles are collected by an impactor, and then washed off by a continuous flow of liquid passed over the impactor, providing a liquid sample for analysis (Orsini et al., 2003).The PILS sampled from the NCAR-University of Wyoming Aerosol Inlet mounted to the roof of the King Air (Snider et al., 2018).The PILS size-cut was provided by a non-rotating MOUDI impactor stage with a 50% transmission efficiency at 1 atm ambient pressure of 1 μm (PM1) (Marple et al., 1991).The flow rate for the PILS was 15 LPM pulled off of the main aerosol inlet line.Sodium carbonate and phosphorous acid-coated denuders were placed upstream of the PILS to remove gaseous interferences.A valve upstream of the PILS was manually closed for 10 min, forcing the airflow through a HEPA filter to obtain a measurement of the background in near real-time.
The liquid sample from the PILS was sent to a Brechtel Fraction Collector to collect samples for offline analysis (Sorooshian et al., 2006).The PILS liquid flowrates were set and the fraction collector operated similarly to the 10.1029/2023JD039043 7 of 20 approach used during WE-CAN (Western Wildfire Experiment for Cloud Chemistry, Aerosol Absorption, and Nitrogen) to obtain ∼1.2 mL of liquid sample every 2 min (Sullivan et al., 2022).Pre-loaded carousels were manually switched during flight.After each flight, the vials were unloaded, recapped with solid caps, and transported to CSU in coolers with ice packs to be stored in a 2°C cold room until analyzed.
Each fraction collector vial was brought to room temperature and analyzed for cations, anions, organic acids, and carbohydrates.Only  NH + 4 data is used in the analysis presented here.A Dionex ICS-3000 ion chromatograph was used to measure  NH + 4 .A Dionex IonPac CS12A analytical column (3 × 150 mm 2 ) employing an eluent of 20 mM methanesulfonic acid at a flow rate of 0.5 mL/min was used.The injection volume was 190 μL with a complete run time of 17 min.Concentrations were blank-corrected using the average of all background samples collected during a specific flight.The limit of detection for  NH + 4 was 0.001 μg/m 3 .

Near-Source Sampling Approach
The approach to sample and follow plumes from specific large animal husbandry sources encompassed four steps and an example of this approach is provided in Figure 1c.(a) The UWKA characterized the planetary boundary layer (PBL) after take-off by climbing to the top of the PBL and during descent while approaching the region with the target facility.(b) Once the pilot visually identified the target facility, the UWKA circled it at ∼300 m (∼1,000 ft) agl to identify any obstacles and determine the plume outflow direction.When no obstacles were identified, the UWKA proceeded to perform an additional circle of the facility at ∼150 m (∼500 ft) agl.During these maneuvers, the aircraft remained ∼1 km from the edge of each facility to limit the noise exposure for the animals.(c) Once the plume outflow location was determined, the UWKA completed a set of stacked boxes downwind at different vertical levels.Vertical altitudes were determined to optimize time and sampling throughout the PBL.The vertical distance between flight legs was ∼150 m (∼500 ft).The closest and furthermost legs of the boxes were located ∼5 and 10 km downwind of each facility, respectively.These distances shifted slightly for safety considerations as needed (i.e., air traffic control and obstacles).(d) When plumes were clearly detected 10 km downwind, and time allowed, another set of stack boxes was completed further downwind.Note that this sampling approach was designed to optimize samples of vertical "curtains" used to calculate emission rates rather than following a particular parcel of air downwind from the emission source.The data presented here cannot be considered pseudo-Lagrangian sampling.Figure 1d shows an example transect of NH 3 and CH 4 produced from the sampling approach outlined above.Note the large co-located enhancements in both species (i.e., NH 3 and CH 4 ).

Analysis Approach and Calculations
In this study, we use three methods to estimate NH 3 emission ratios relative to CH 4 (NH 3 EmR) for targeted facilities.(a) We determine the average NH 3 normalized excess mixing ratio (NEMR) per horizontal plume transect (Section 2.5.2).(b) We estimate NH 3 EmR from an Ordinary Least Squares (OLS) slope calculated using the observed NH 3 and CH 4 mixing ratios within plume conditions (Section 2.5.3).(c) We determine the NH 3 EmR using the OLS slope that is refined by removing CH 4 mixing ratios associated with emissions from oil and gas operations (Section 2.5.4).All three methods utilize observations identified during transects 10 km or closer to the targeted facility.

Transect Identification
To estimate NH 3 EmR, we identified all plume transects within 10 km downwind of each facility (a total of 232 transects: 156 for cattle facilities and 76 for dairy facilities).The transects are characterized by co-located enhancements with respect to background air for NH 3 and CH 4 , as shown in Figure 1d.CH 4 can be considered a conserved tracer because its lifetime against oxidation by the hydroxyl radical (OH), which is its main sink process, is sufficiently long (∼8.3 years) such that it does not undergo any significant loss process in the timescales relevant to this study (i.e., minutes to hours) (Holmes, 2018).Each of our transects includes at least 10 data points of background air or out-of-plume observations on either side of a plume, which is necessary to calculate the enhancement of NH 3 and CH 4 used to calculate the NH 3 NEMR (see Section 2.5.2).Note that in some transects, background mixing ratios differed on each side of a plume.
We calculated the physical age of each intercepted transect downwind of the facility using the distance of the transect downwind from the facility divided by the average wind speed for that transect.We note that most of our intercepted transects fall within the first hour of physical age (see Figure S2 in Supporting Information S1).We also estimated an emission time for each transect by subtracting the physical age for each transect from the average time at which that transect was sampled.A summary of these calculations is presented in Figure S3 in Supporting Information S1.The data collected during TRANS 2 Am largely represent plumes that have been emitted between mid-morning to early afternoon.
We include transects up to 10 km downwind of a given facility for the NH 3 EmRs calculations described in Sections 2.5.2-2.5.5.Table S1 and Figure S4 in Supporting Information S1 summarize how calculated values vary when they are based on different subsets of the data.Briefly, overall estimates of NH 3 EmRs are slightly higher (2%-7%) when they are based only on transects collected <4 km downwind versus including more data collected further downwind.Next, we describe the three methods used to calculate the NH 3 EmRs, including a detailed explanation of the regression model to isolate oil and gas CH 4 emissions from agricultural CH 4 emissions (Section 2.5.4). Figure 2 shows a schematic of the methods used for the NH 3 EmR calculations.

NH 3 Normalized Excess Mixing Ratios
For each identified transect described above, we calculate the NH 3 NEMR using Equation 1. (1) In Equation 1, transect-specific in-plume values are defined as the average of 1 Hz observations where NH 3 is > the 25th percentile for that transect.Background values are defined as the average of the observations where NH 3 is ≤ the 25th percentile.Using Equation 1 results in one NEMR value for each transect.A sensitivity analysis using different backgrounds (see Figure S5 in Supporting Information S1) shows that choosing a lower percentile for background values (i.e., fifth percentile) might result in lower plume average values, underestimating the NH 3 EmR by ∼5%.Calculating transect-specific NH 3 NEMRs allows us to look at each plume interception independently as a function of time of day, distance from a facility, and vertical location.Note that some of the 9 of 20 analysis in this manuscript uses NH 3 NEMR for individual transects rather than the average NH 3 NEMR within 10 km from the facility (i.e., NH 3 EmR).

OLS Regression
We use the slope calculated from an OSL linear regression of NH 3 versus CH 4 for the observations from all the transects within 10 km as one estimate of the NH 3 EmR.This method uses only in-plume observations.A sensitivity analysis of this method using both background and in-plume observation versus only in-plume observations shows very similar results (Figure S6 in Supporting Information S1).

Linear Regression Analysis With Multiple Predictors to Eliminate Influence of CH 4 From Oil and Gas Sources
Given the close proximity of oil and gas operations to agricultural facilities, we used co-measured C 2 H 6 to account for the influence of this potential additional CH 4 source.This is likely a more important issue for correctly interpreting aircraft observations than those collected by vehicles with close access to the perimeter of target facilities; however, Kille et al. (2017) were able to consistently quantify a positive ethane flux out of one of the three dairy sites they sampled using the CU Solar Occultation Flux instrument onboard a mobile vehicle.
Following Kille et al. (2019), we use the linear model module from the sklearn python package version 1.0.2 to perform a linear regression analysis to the time series of ΔCH 4 (Equation 2) for each facility or group of nearby facilities sampled during each Research Flight (RF) (see Table S2 in Supporting Information S1 for details).This method helps us account for or eliminate the percentage of CH 4 in each observation that is attributed to oil and gas operations in the time series selected for the analysis (Equations 3-5).We use ΔC 2 H 6 as the predictor for oil and gas emissions, ΔNH 3 as the predictor for agricultural (i.e., livestock) emissions, and ΔCH 4 as the predictand.This method assumes that oil and gas and livestock operations are the only sources of C 2 H 6 and NH 3 in the region, respectively.
The regressions coefficients β 1 and β 2 (ppbv ppbv −1 ) represent the ΔCH 4 /ΔC 2 H 6 and ΔCH 4 /ΔNH 3 ratios, respectively.The coefficient β 0 represents the excess CH 4 above the background that cannot be attributed to any of the two sources.All ΔX (CH 4 , C 2 H 6 , or NH 3 ) have been calculated using a background, defined as the average of all the data below the 10th percentile of the time series selected for the analysis.We note that using ΔX instead of absolute mixing ratios for each trace gas does not change the values for β 1 and β 2 , it only changes the β 0 coefficient, which is related to the regional background selected.Table S2 in Supporting Information S1 shows details for the linear regression analysis, including the regression score, the regression coefficients, and the facilities name and RF that are included in the analysis.We only used model outputs with regression scores (R 2 ) above 0.4.Once β 0 , β 1 , and β 2 have been defined from the linear regression model, we use them to calculate the percentage contribution from each emission source (β 0 = other/unexplained CH 4 above regional background, β 1 = CH 4 from oil and gas, and β 2 = CH 4 for agricultural emissions) for each 1-Hz CH 4 observation using Equations 3-5.An example of the percentage contribution from each emission source to CH 4 is shown in Figure S7 in Supporting Information S1 for the same transect shown in Figure 1d. (3) (5)

Emission Rate Calculations
We calculated NH 3 emission rates in grams of NH 3 per hour per head of cattle (g NH 3 • h −1 • hd −1 ) for four facilities (F01, F04, F19, and F27/28) sampled under ideal wind conditions (i.e., winds >4 ms −1 ) with a consistent direction, sampling boxes located perpendicular to the wind direction, and minor influence from other emission sources (i.e., other feedlots or dairies)) following the methods by Hacker et al. (2016).The reference frame of the plumes was rotated using the prevalent wind direction to minimize the crosswind component and maximize the perpendicular wind component.We used different downwind sections (i.e., curtains) of each plume to get multi ple independent emission rate estimates per facility (i.e., Figure 1c shows three transects [i.e., curtains] collected downwind of F01 and Figure 6a shows seven transects collected downwind of F04).The 1-Hz data, including calculated instant fluxes (μg NH 3 • m −2 • s −1 -instantaneous P, T conditions), were averaged to a 200 × 100 m (horizontal × vertical) grid.We assume that the layer near the surface is the same as the data collected at the lowest sampled layer (e.g., ∼150-300 m AGL) corrected by the average topography in each grid cell.We assume that relevant concentration values at the top of the planetary boundary layer height (PBLh) were 10% of the highest available sampled altitude.The PBLh was calculated using potential temperature, water, and wind vertical profiles during descent while approaching the target facilities, following similar methods detailed in Cazorla and Juncosa (2018).We apply a simple linear interpolation (smooth factor = 1) to complete grids without observations.Finally, we integrated the instant fluxes across the total curtain area to obtain a total emission rate per facility at different distances.The maximum animal capacity per facility is based on a livestock registration and permitting database maintained by the Colorado Department of Public Health and Environment (CDPHE, 2017).

Emission Ratios for Cattle Feedlots and Dairies
Figure 3 shows NH 3 EmR for 29 facilities housing (a) beef cattle or (b) dairy cows sampled during Phase One of TRANS 2 Am. Figure 3 shows that there is a large variability in NH 3 EmR between different facilities.Specifically, beef cattle facilities have NH 3 EmR ranging from 0.1 to 2.6 (average 1.2) ppbv ppbv −1 and dairies have lower NH 3 EmR ranging from 0.2 to 0.5 (average 0.3 ppbv ppbv −1 ).There is a significant difference between the average NH 3 EmR associated with beef cattle versus dairy facilities.During this study, the NH 3 EmR associated with dairies was on average, four times less than the NH 3 EmR associated with beef cattle (0.3 vs. 1.2 ppbv ppbv −1 ).This difference has been observed previously (Eilerman et al., 2016).Factors that contribute to this pattern include differences in production (milk vs. beef), feeding products, and differences in CH 4 emissions between beef cattle and dairy (Golston et al., 2020).In this data set, dairy NH 3 EmRs are less variable, but this could also be explained by the fewer observations.Note that the different methods used to calculate NH 3 EmR provide similar results with few exceptions.2016) and Golston et al. (2020) for both cattle and dairy facilities.This is only partially explained by the time of day of our sampling.The vast majority of the sampling occurred between 10 a.m. and 3 p.m. LT when the diurnal profile of NH 3 EmR typically peaks (see Eilerman et al. (2016), Golston et al. (2020) and Section 3.2).However, a close comparison to the work of Golston et al. (2020) indicates that we observed higher NH 3 EmR during the afternoon peak in NH 3 emissions.Our average values are also higher than those reported by Eilerman et al. (2016) for summer only and for cattle specifically.(b) Before the sampling by Golston et al. (2020), estimates of NH 3 EmR in this region were limited to a handful of facilities.Our data set represents a dramatic increase in the number of facilities that can be used to estimate NH 3 EmR.While each prior study cited in Table 2 indicates a substantial step forward in terms of its technical and/or methodological approach, more observations are likely still needed to account for the true variability in NH 3 EmR, particularly outside of the warm summer months.

Temperature Dependencies
In general, Figure 3 shows that the highest NH 3 EmR observed during Phase One of TRANS 2 Am were associated with the highest temperatures.Figure 4 explores the relationship between NH 3 EmR and temperature further., 2017).Bars with the smallest width represent those facilities for which there is no available information about animal capacity (N = 3; see Table 1 for more details).The different symbols represent different methods for estimating the NH 3 EmR.Circles indicate values calculated using the NH 3 NEMR method described in Section 2.5.2.Squares represent values calculated using in-plume observations and the OLS regression method as described in Section 2.5.2 (NH 3 vs.CH 4(ag + others) OLS); triangles indicate values calculated using the OLS regression method with the multiple-predictors linear regression method for isolating CH 4 emissions associated with agricultural sources (Sections 2.5.3 and 2.  NH 3 EmR for beef cattle generally increases with increasing temperatures, consistent with prior work documenting an exponential relationship between NH 3 EmR and temperature for livestock facilities (Eilerman et al., 2016;Golston et al., 2020).In general, we observe a weak overall relationship between NH 3 EmR estimates and temperature for facilities housing beef cattle.The relationship between NH 3 EmR and temperature is especially hard to assess for the few NH 3 EmR estimates for dairies (panel d).Note that TRANS 2 Am collected data in a small range of temperatures compared to those observed year-round in Colorado (Figure 4 panels a and c).Most of the data were collected during hot and dry conditions, and few observations were collected during hot and humid conditions, usually after precipitation events.Broadly consistent with our findings, the data presented by Golston et al. (2020) that were collected at temperatures >∼25°C also shows a large spread in NH 3 EmR ranging from near 0 up to almost 2 ppbv ppbv −1 .In both data sets, variability in NH 3 EmR appears larger in this uppermost temperature range.
The observed relationship between NH 3 EmR and temperature for beef cattle facilities is shown with the dashed line in panel (b).The temperature dependence of NH 3 emissions was derived using the principle that volatilization of NH 3 increases with higher temperature (Eilerman et al., 2016;Sander, 1999;Sutton et al., 1994).Briefly, the NH 3 compensation point for volatilization varies as a function of temperature and pH of the solution, both unknown for this data set.Here, we have used atmospheric temperatures as a proxy for soil temperature and coefficients that remain mostly constant with temperature and pH, allowing for a semi-empirical fit of the observations to the model (see Equations S1 and S2 in Supporting Information S1 for more details).Panel (b) shows that one of the highest estimates NH 3 EmR corresponds to the highest temperatures and lowest relative humidity in the range of observations (i.e., 2.2 ppbv ppbv −1 , 31°C, 21%).However, we observed a larger NH 3 EmR (i.e., ∼2.6 ppbv ppbv −1 ) from an opportunistic sample at a lower temperature and higher relative humidity (i.e., 25°C, 27%).The few samples associated with lower temperatures (purple bars in Figure 3), higher relative humidities, and lower NH 3 EmR (i.e., facilities sampled during RF02 and RF08) were collected after regional precipitation events.Overall, most of the reduced nitrogen (NH x = NH 3 +  NH + 4 ) in the near-source sampling is found in the gas phase as NH 3 (see Figure S10 in Supporting Information S1). Figure S10 in Supporting Information S1 shows

Table 2 Comparison of Molar Emission Ratios and Emission Rates in the Colorado Front Range From Previous Publications
that the partitioning of NH 3 to the particle phase as  NH + 4 is, on average, less than 15% of the total NH x for the data presented in this study.The few exceptions (RF02, RF03, and RF08) are those sampled after precipitation events, which, in general, have lower NH 3 mixing ratios (see Figure S11 in Supporting Information S1).To further explore the relationship between NH 3 emissions and temperature and based on previous observations that found a strong correlation between NH 3 emissions and time of the day (Eilerman et al., 2016), Figure 5 shows transect-specific NH 3 NERM (see Section 2.5.2) as a function of sampling time.
Figure 5 shows a diurnal pattern of NH 3 NEMR with higher values in the mid-afternoon and lower values in the morning and evening periods.The highest NEMRs were observed between 12 and 4 p.m.More quantitative information can be drawn from Figure 5 by examining some of the specific flights where that same facility was repeatedly sampled at different times and temperatures.F13, F14, F15, and F16 were sampled during both RF06 (pink dots in Figure 5) and RF14 (dark purple dots in Figure 5).RF06 was a mid-morning flight (10 a.m. to 1 p.m.) with average plume interception temperatures and relative humidities of ∼23°C and 34%.RF14 was an afternoon flight (1-4 p.m.) with average plume interception temperatures and relative humidities of ∼30°C and 21%.The NH 3 EmR estimates for these four facilities are 0.14-1.06ppbv ppbv −1 higher for RF14 (hotter and drier) than they are for RF06 (colder and more humid).F04, one of the biggest facilities sampled during Phase One of TRANS 2 Am (see Figure 6 for details), was sampled during RF03 (light green dots in Figure 5) and RF13 (light purple dots in Figure 5).RF13 was a mid-morning flight (10 a.m. to 2 p.m.) with higher average temperature and relative humidity than RF03 (27°C vs. 25°C and 31% vs. 27%), a late afternoon flight (4-6 p.m.).The NH 3 EmR were higher during RF13 (hotter and more humid) than RF03 by 0.27 ppbv ppbv −1 (Figure 7).We also observed one case where higher temperatures did not produce a higher NEMR for a given facility.F01 was sampled twice (RF01 [light blue dots in Figure 5] and RF09 [light orange dots in Figure 5]) at roughly the same time of the day; the average temperature was slightly higher during RF09 (24.5°C) than during RF01 (23.2°C).However, the relative humidity was considerably higher during RF09 (45%) than during RF01 (30%).Despite a similar time of day and temperature range, NH 3 EmR estimates are lower for RF09 than they are for RF01 (1.00 vs. 1.972 ppbv ppbv −1 ), suggesting that drier conditions favor NH 3 volatilization and emissions.The NH 3 emissions estimated for all these facilities and their differences in the context of different temperatures, relative humidities, and sampling times reflect the variability of NH 3 emissions for the same facility and their dependency on temperature, relative humidity, and time of day.

Plume Evolution Case Study
Loss of gas-phase NH 3 in plumes advected from feedlots can occur from dry deposition or chemical transformation of NH 3 to particle  NH + 4 .Under the assumption of constant emissions relative to CH 4 and pseudo-Lagrangian sampling, we can use the ratio of NH 3 to CH 4 downwind of the CAFOs to constrain the loss of gas-phase NH 3 , since both NH 3 and CH 4 are diluted similarly downwind of a source, but CH 4 does not undergo significant chemical losses in the temporal scales of this study (i.e., hours).Changes in the ratio of NH 3 to CH 4 downwind of the CAFOs can be used to calculate the total loss of NH 3 due to deposition or partitioning to the gas-phase (Lassman et al., 2020).
Across the data set presented in this work, we identified only one instance with substantial decay of NH 3 beyond 10 km from the facility: F04 during RF13.Note that the plume downwind of this facility did not decay similarly on the other sampling day (i.e., RF03).F04 was sampled on 23 August 2021, between 12:30 and 2 p.m. LT.F04 is one of the largest facilities housing beef cattle sampled during TRANS 2 Am, with a reported animal maximum capacity of 100,000 animals.The plume intercepted from F04 during the RF13 shows the largest NH 3 mixing ratios observed throughout Phase One of TRANS 2 Am with values up to 440 ppbv of NH 3 .The plume from F04 was intercepted up to 25 km downwind (Figure 6).The sampling included circles around the facility (as close as 2 km) and three distinct vertically stacked boxes at 4-6 km downwind, 11-14 km downwind, and 17-23 km downwind.The first two stacked boxes were executed at ∼175, 325, and 450 m agl.The last one was executed only at 175 and 325 m agl.The PBL on this day contained two inversions.The lowest inversion was identified at ∼500 m agl, and a second one was located at ∼1,300 m agl (see Figure S12 in Supporting Information S1 for vertical profiles).Above the latter

NH 3 Emission Rates
Figure 7 shows inferred NH 3 emission rates for F01 during RF01, sampled from three distances downwind.The inferred emission rate magnitudes vary from 10 to 26 g NH 3 • h −1 • hd −1 depending on where the calculation is performed.NH 3 inferred emission rates for F01, F04, F19, and F27/F28 range from 4 to 29 g NH 3 • h −1 • hd −1 .The average (std) is 14.4 (6.62) g NH 3 • h −1 • hd −1 .All these facilities house beef cattle.Similar to the example shown in Figure 7 (F01), the estimated NH 3 emission rates for F04, F19, and F27/28 depending on where the calculation is performed downwind from each facility.In general, there is no consistent relationship between the NH 3 emission rate and distance downwind (see Figure S13 in Supporting Information S1).For more details, refer to Table 1 and Figures S14-S16 in Supporting Information S1.The emission rates for F27 and F28 are combined estimates since these facilities are within ∼8 km of each other, and their plumes merged during sampling.Note that the NH 3 emission rates estimates for F04 during RF13 are presented for curtains >10 km from F04, where substantial NH 3 loss was observed (see Figure 6).If the curtains >10 km are not included, the average NH 3 emission rate for F04 during RF13 increases from 11.8 to 17 g NH 3 • h −1 • hd −1 which is closer to the estimates of the other two larger facilities (F01 and F27/28).Unlike NH 3 emissions ratios, inferred NH 3 emission rates do not require correction for CH 4 emissions.
The average of our estimates is higher than those reported in previous studies (i.e., 2.64-12 g NH 3 • hd −1 • h −1 Golston et al., 2020;Kille et al., 2017;Shonkwiler & Ham, 2018;Sun et al., 2015) with one exception that reports NH 3 emission rates 14 (±2) g NH 3 • hd −1 • h −1 for one facility (F04 in this study) (McCabe et al., 2023).Differences between our findings and previous results could be explained by (a) a single sample per facility, (b) the different methodology sample NH 3 (i.e., airborne observations vs. ground observations), or they may more fully reflect the variability in emissions in space and time.Our estimates are also larger than those reported from other regions.For example, emission rates measured for facilities in Alberta, Canada, report values of 5.83 and 8.5 g NH 3 • h −1 • hd −1 , which are in the lower range of what we observed in northeastern Colorado (McGinn et al., 2007;Staebler et al., 2009).Shonkwiler and Ham (2018) summarize other studies of NH 3 emission rates in other places in the United States and worldwide during different seasons.NH 3 emissions rates estimates of 2.6 and 5.75-8 g • h −1 • hd −1 have been observed downwind of small feedlots (<1,200 beef cattle) in China (Yang et al., 2016) and Australia (Denmead et al., 2008), respectively.

Conclusions
Here, we report on summertime airborne observations of NH 3 and CH 4 collected over northeastern Colorado during the first phase of the TRANS 2 Am field intensive.We intercepted plumes downwind of 29 dairies and cattle feedlots, and we used these observations to infer NH 3 emissions ratios (EmR) with respect to CH 4 and NH 3 horizontal emission rates for a subset of facilities.We show the following: • NH 3 EmR during August 2021 ranged from 0.1 to 2.6 ppbv ppbv −1 , which are larger than those presented in previous literature.The NH 3 EmR associated with beef cattle feedlots are, on average, four times larger than NH 3 EmR associated with dairies (1.2 vs. 0.3 ppbv ppbv −1 , respectively).• The study region presents a complex combination of emissions sources, especially for CH 4 , which has substantial emissions from both oil and gas operations and agriculture intermixed.We present four estimates of NH 3 EmR with and without correction for CH 4 emissions from oil and gas operations.Estimates that account for intermixed oil and gas CH 4 emissions (NH 3 NEMR and OLS regression with agriculturally specific CH 4 ) are higher than NH 3 EmR based on methods that do not account for intermixed oil and gas CH 4 emissions (OLS regression with uncorrected CH 4 ).Accounting for CH 4 emissions from oil and gas operations is likely more important for aircraft observations than ground-based observations that sample immediately adjacent to facilities.However, prior work that has not accounted for other emissions sources may underestimate NH 3 EmR in the region.• In the region immediately downwind of the diaries and cattle feedlots we sampled, particle phase  NH + 4 accounted for <15% of the absolute NH x on average.• Prior work has reported correlations between NH 3 EmR and temperature.However, data from Phase One of TRANS 2 Am represent a relatively small temperature range.Even though we document a general trend of increasing NH 3 EmR with temperature, we also observed high NH 3 EmR at lower temperatures with high relative humidities.• NH 3 NEMRs have a relationship with time of day, with higher NH 3 NEMRs between 12 and 4 p.m. LT.For five of the six facilities sampled on different days, NH 3 NEMRs are higher for samples collected under hotter and drier conditions as well as later in the day.• We document NH 3 decay relative to CH 4 within the plume associated with the largest facility (F04) sampled during Phase One of TRANS 2 Am.This plume contained the highest NH 3 mixing ratios observed during the campaign.The plume encountered downwind of F04 during RF13 had an e-folding time of 108, 87, and 119.3 min for samples collected downwind at 175, 325, and 450 m agl.While the decay was clear on RF13, the plume downwind of F04 did not show similar decay during its sampling on RF03.• Our inferred horizontal emissions rate estimates for NH 3 for four beef cattle facilities range from 4 to 29 g NH 3 • h −1 • hd −1 with an average (std) of 14.4 (6.62) g NH 3 • h −1 • hd −1 .We observe larger NH 3 emission rates (g NH 3 • h −1 • hd −1 ) compared to other studies, with one exception.We do not find a relationship between maximum reported animal capacity and inferred horizontal NH 3 emissions rates.
Phase One of TRANS 2 Am substantially increases the number of in situ measurements of NH 3 emissions and their relationship with CH 4 from facilities housing beef cattle and dairy cows in northeastern Colorado.These data represent the warmest part of the seasonal cycle.Future work should focus on colder, more humid conditions characteristic of other times of the year.These data show the variability associated with NH 3 emissions from CAFOs during hot and dry conditions.Forthcoming papers will discuss similar data collected during the second phase of the campaign conducted in August 2022 and data from research flights focused on the evolution of large regional plumes as they move from the polluted Front Range up into the Rocky Mountains.
west cross over urban centers where other urban pollutants (i.e., nitric acid [HNO 3 ]) are available for forming fine particulate matter (PM 2.5 ) through ammonium nitrate (NH 4 NO 3 ) formation.
Figure 1.(a) Map of northeastern Colorado showing the large number of livestock facilities in the region.The different colors signify the type of animal housed at each facility, and the size of the marker is proportional to the maximum animal capacity.Black dots indicate the locations of active oil and gas wells as of 2015.(b) Same map as (a) but only including facilities sampled systematically during Phase One of TRANS 2 Am.(c) Flight track of the UWKA colored by NH 3 (ppbv) during the sampling of Facility 1 (F01; refer to Table1for more information) on 2 August 2021.This example flight track is representative of the general sampling strategy used during TRANS 2 Am.Letters (i, ii, iii) refer to different vertical transects used for emission rate calculations (refer to Figure6and Section 2.5.5 for more information).(d) Example of a transect with colocated enhancements of NH 3 (ppbv; purple) and CH 4 (ppmv; green) versus horizontal distance from the facility.Note that we include non-plume background values in each transect (i.e., tails on each side of each transect).

Figure 2 .
Figure 2. Schematic of the methodology used to calculate the NH 3 emissions ratios with respect to CH 4 (NH 3 EmR) showing the three methods presented in Figure 3.

Figure 3 .
Figure 3. Average NH 3 emission ratios with respect to CH 4 (NH 3 EmR) for (a) beef cattle and (b) dairy facilities.NH 3 EmRs are calculated using all transects <10 km from the targeted facility.The number of transects used is shown in gray to the left of the colored bars.Bars are colored by the average ambient air temperature (°C) measured from the aircraft during the time of sampling, and bar width represents the maximum animal capacity reported(CDPHE, 2017).Bars with the smallest width represent those facilities for which there is no available information about animal capacity (N = 3; see Table1for more details).The different symbols represent different methods for estimating the NH 3 EmR.Circles indicate values calculated using the NH 3 NEMR method described in Section 2.5.2.Squares represent values calculated using in-plume observations and the OLS regression method as described in Section 2.5.2 (NH 3 vs.CH 4(ag + others) OLS); triangles indicate values calculated using the OLS regression method with the multiple-predictors linear regression method for isolating CH 4 emissions associated with agricultural sources (Sections 2.5.3 and 2.5.4) (NH 3 vs.CH 4(ag) OLS).Error bars show the standard deviation between the three methodologies.Filled and open symbols show estimates for facilities sampled systematically and opportunistically, respectively.Note that those facilities sampled opportunistically (open symbols) do not follow the sampling strategy (i.e., spiral + boxes downwind) outlined in Section 2.4.1 and usually include only a few transects (1-4).The dotted and dashed lines in Figures2a and 2brepresent the average NH 3 EmR for beef and dairy cattle, respectively, during Phase One of TRANS 2 Am.(c) Average NH 3 EmR estimates for all facilities (black), separated by beef (blue) and dairy (pink).
Figure 3. Average NH 3 emission ratios with respect to CH 4 (NH 3 EmR) for (a) beef cattle and (b) dairy facilities.NH 3 EmRs are calculated using all transects <10 km from the targeted facility.The number of transects used is shown in gray to the left of the colored bars.Bars are colored by the average ambient air temperature (°C) measured from the aircraft during the time of sampling, and bar width represents the maximum animal capacity reported(CDPHE, 2017).Bars with the smallest width represent those facilities for which there is no available information about animal capacity (N = 3; see Table1for more details).The different symbols represent different methods for estimating the NH 3 EmR.Circles indicate values calculated using the NH 3 NEMR method described in Section 2.5.2.Squares represent values calculated using in-plume observations and the OLS regression method as described in Section 2.5.2 (NH 3 vs.CH 4(ag + others) OLS); triangles indicate values calculated using the OLS regression method with the multiple-predictors linear regression method for isolating CH 4 emissions associated with agricultural sources (Sections 2.5.3 and 2.5.4) (NH 3 vs.CH 4(ag) OLS).Error bars show the standard deviation between the three methodologies.Filled and open symbols show estimates for facilities sampled systematically and opportunistically, respectively.Note that those facilities sampled opportunistically (open symbols) do not follow the sampling strategy (i.e., spiral + boxes downwind) outlined in Section 2.4.1 and usually include only a few transects (1-4).The dotted and dashed lines in Figures2a and 2brepresent the average NH 3 EmR for beef and dairy cattle, respectively, during Phase One of TRANS 2 Am.(c) Average NH 3 EmR estimates for all facilities (black), separated by beef (blue) and dairy (pink).

Figure 4
Figure 4 shows the NH 3 EmR estimates as a function of temperature (°C) and relative humidity (%) for beef cattle (top panel) and dairy cows (lower panels).Panels (a) (cattle) and (c) (dairy) show the range of temperature and relative humidity of the data presented here, colored by NH 3 EmR and sized by facility maximum capacity.Panels (b) (cattle) and (d) (dairy) show NH 3 EmR versus temperature colored by relative humidity.
Bold values denote highlight the values of this study.

Figure 4 .
Figure 4. NH 3 EmRs (ppbv ppbv −1 ) estimates for individual facilities as a function of temperature and relative humidity for beef cattle (top panels) and dairy cows (lower panels).

Figure 5 .
Figure 5. Transect-specific NH 3 NEMR as a function of local sampling time.The points represented by the boxplots are colored by research flight (RF).Points for RF01 (light blue) and RF02 (dark blue) were made larger intentionally to prevent masking from other data points.The lower and upper ends of the boxes span from quartile 1 (Q1) to quartile 3 (Q3).The whiskers correspond to each box edge (Q1 or Q3) ±1.5 the interquartile range (IQR: Q3-Q1).If no outliers are present, the whiskers represent minimum and maximum values.The dashed rhombuses show the sample mean (middle line) and the standard deviation (corners).N indicates the number of transects in each box plot.Note that all transects, including those >10 km and all animal types, have been included.

Figure 6 .
Figure 6.(a) Flight track of the UWKA colored by NH 3 (ppbv) during sampling of F04 on 23 August 2021.As in Figure1, the colored and sized dots represent agricultural facilities housing different animals, and the black dots signify oil and gas operations as of 2015.Letters (i-vii) refer to different vertical transects used for emission rate calculations (refer to FigureS16in Supporting Information S1 and Section 2.5.5 for more information).(b) Transect-specific NH 3 NEMR as a function of altitude above ground level (m agl) and distance downwind from the center of the facility (km).(c) Normalized NH 3 NEMR with respect to the maximum value in each altitude bin (∼175, 325, and 450 m agl) as a function of the traveled time since emission (in minutes).Traveled time is calculated as the distance of each transect from the targeted facility divided by the average wind speed for that transect.Lines signify the linear fit for each altitude bin.E-folding time (time at which normalized NH 3 NEMR drops below 1/e) was determined by the linear fit.

Figure 7 .
Figure 7. Example NH 3 emissions rates determined for F01 sampled during RF01 via transects at downwind distances of (a) 0-3.1 km, (b) 3.7-5.7 km, and (c) 17.6 and 19 km.The color bar shows inferred instantaneous horizontal fluxes in each grid cell in μg NH 3 m −2 s −1 .Grid cells are 200 (horizontal) × 100 (altitude) m.The number in the lower right corner shows the total inferred emission rates in g NH 3 • h −1 • hd −1 .Panels show simple linear interpolation using the python package scipy.interpolate, with a smooth factor of 1.

Table 1
Summary of Facility Information, Meteorological Conditions, NH 3 EmR, and Emission Rates for Beef Cattle [C] and Dairy [D] Facilities Sampled During Phase One of TRANS 2 Am In general, those methods used to correct CH 4 emissions from other sources (i.e., NH 3 NEMR and OLS regression corrected for CH 4 ) produce similar values, but those values are higher than values produced using methods that do not correct for CH 4 emissions from other sources (OLS regression without correcting for CH 4 ) (see FiguresS4 and S6in Supporting Information S1).Differences in NH 3 EmR from facilities housing beef cattle versus dairy cows could result from higher emissions of NH 3 from facilities housing beef cattle and/or lower CH 4 emissions from facilities housing dairy cows.A t-test of the distribution of NH 3 and CH 4 for facilities housing beef cattle and dairy cows shows that in this data set, both are true (see FiguresS8 and S9in Supporting Information S1).Facilities housing beef cattle have higher mean NH 3 mixing ratios (25.4 vs. 12.3 ppbv) and lower CH 4 mixing ratios (1.978 vs. 1.984 ppmv) compared to dairies.Table 2 compares our observed NH 3 EmR (and fluxes see Section 3.4) observed in our study to prior observations in the Colorado Front Range.There are several key points from this comparison.(a) Our calculated NH 3 EmR are higher than both Eilerman et al. ( Eilerman et al. (2016).(2022), McCabe et al. (2023)report NH 3 EmR estimates during November 2019 under colder conditions (median of 15°C) for five beef cattle facilities in the same study region.Their estimates range from 0.8 to 2.7 ppbv ppbv −1 .Similar to whatEilerman et al. (2016)observed, this suggests that large NH 3 EmR exist under colder conditions.