Continuous measurements of greenhouse gases and pollutants were collected using a mobile platform , consisting of analyzers using the Cavity Ring-Down Spectroscopy technique , global satellite positioning unit to record geolocation and vehicle speed, 2-D sonic anemometer to measure wind direction, wind speed, air temperature and relative humidity, and calibration tanks. The following trace gas species were continuously measured from air drawn in at an inlet with a height of 2.87 m: CH4, δ13CCH4, carbon dioxide , carbon monoxide , C2H6. Reported trace gas mole fractions and isotope ratios were corrected using low and high custom gas mixtures that were measured before and after each measurement period. The isotopic values of the gas mixtures were -39.5‰ , -40.7‰ , and -38.5‰ . These gas mixtures contained all the species of interest and were tied to the scale set by the NOAA Global Monitoring Division by measurement against NOAA certified tanks. Isotopic standards were tied to the Vienna Pee Dee Belemnite scale and further calibrated by measuring two standards ranging from -23.9‰ to -68.6‰ with the Picarro 2210-i in the laboratory before the field campaign. Micrometeorological measurements were collected at the reference test site each season, with a 3-D sonic anemometer mounted on a stationary tower near the manure lagoons . Measurements were made at two heights, 2.4 m and 11 m, at a frequency of 20 Hz. For the purposes of our analysis, we only used meteorological data from the 2.4 m tower. On January 15th, 2020, curing cannabis we used a cuboid chamber made of clear PVC to isolate and measure δ13CCH4 from free stall barns and static manure piles from the solid drying area .
The chamber was placed on the free stall barn or manure pile surface and connected to the gas analysis system of the mobile platform with Synflex tubing. For each sample, we collected measurements for ten minutes. We also measured δ13CCH4 from the breath of milking cows, dry cows, heifers, bull calves, and calves in hutches by holding Synflex tubing connected to the mobile platform gas analysis system near the mouths of cows . We measured within 16 cm of milking and dry cows, ~1 m from heifers and bull calves, and ~10 m from calves in hutches.Several corrections to observations were applied for each measurement period. First, observations collected from different instruments were cross-correlated and synchronized to local time . Offsets were recorded between local time and each instrument’s internal clock, which were then used to correct data prior to performing the cross-correlation method. Picarro raw mixing ratio measurements were time synchronized to collocated GPS measurements based on time stamp. Second, a correction was applied based on the lag time between the inlet and instrument reading.Third, trace gas mole fraction and δ 13CCH4 observations were corrected by applying a correction factor from calibrations performed before and after each measurement period.We compared measurements of δ13CCH4 using our mobile laboratory sampling technique using CRDS with analysis of whole-air samples collected at the same time and then analyzed with standard Isotope Ratio Mass Spectrometry . Five whole-air samples of atmospheric CH4 were collected in preconditioned and evacuated 2-L stainless steel canisters with bellow valves, over a period of about one minute .
The canisters were first processed by University of California, Irvine for chemical analysis, and a subsample was then sent to the University of Cincinnati for isotopic analysis with IRMS using a method described in detail by Yarnes . Over the course of the same time intervals, the mobile laboratory continuously measured δ13CCH4 with the CRDS instrument. The differences between δ13C measured by IRMS and CRDS were within the uncertainties of each respective technique . These findings suggest that δ13CCH4 measurements by the mobile laboratory CRDS technique is comparable to the standard IRMS method. We conducted a dilution experiment to analyze the precision of δ13CCH4 sampled with the CRDS instrument at varying CH4 levels similar to what we observed during downwind plume sampling of other dairies in the region. Following a similar method by Miles et al. , a high gas standard with 20.1 ppm CH4 and δ13C-CH4 of -44.35‰ was mixed with zero air using a mass flow controller . The mass flow controllers were used to direct isotopic calibration standard tank into a mixing volume at 20 sccm and mixed with zero CH4 air at 203.3, 181.0, 140.0, 114.00, 20.2 and 13.5 sccm to create target CH4 mole fractions of 1.8, 2.0 , 2.5, 3.0, 10.0 and 12.0 ppm, respectively. To compare with the time interval used to average regional measurements, the final 15 seconds of data for each dilution were averaged to evaluate the precision of the instrument. The standard error of the δ13C-CH4 collected during these tests increased with decreasing CH4 mole fractions . The δ13C end-member from the data collected was within 0.83‰ of the isotopic value of calibration standard tank.Sources of CH4 emissions at the reference test site farm were identified by categorizing atmospheric observations based on proximity to the emission source and wind direction. To evaluate δ 13CCH4 from biogenic sources at the farm scale, observations with CH4 ≤ 30 ppm were selected and averaged by 1-min intervals to minimize uncertainty according to the performance standards of the instrument.
For each source, δ 13CCH4 and the corresponding standard errors were estimated as the yintercept from a weighted linear regression of the inverse of the atmospheric CH4 mole fraction and δ 13CCH4 . Keeling plots were generated for each dairy farm source by applying a weighted linear regression with errors in both the independent and dependent variables based on the York et al. method . To exclude CH4 emissions from fossil-fuel sources, such as from vehicles, which have δ 13CCH4 signatures between -46‰ to -30‰ , we omitted CH4 observations that had corresponding excess C2H6 values > 0.1 ppm and excess CO values > 500 ppb, the 99th percentile from all regional transects . We define excess C2H6 and excess CO as mole fractions above the minimum C2H6 and CO observations for each dairy farm source. At the reference test site, no excess CO measurements above this threshold were detected. For the inverse of CH4, the uncertainty was defined as the mean of the standard errors from the 1-min averaged observations in the weighted linear regression. For δ 13CCH4 observations, we first evaluated the mean of the standard errors from the 1-min averaged observations against the standard error from 1-min averages ofthe standard gas run. Then, we selected the largest standard error of the two as the corresponding uncertainty. In this study, the δ 13CCH4 values reported hereafter are referring to the δ 13CCH4 end-members derived from Keeling plots.Isotopic signatures of CH4 were classified into the following two categories: Dairy Cluster or isolated dairy farms , where there were no major potential sources of CH4 within at least 2 km from the dairy farm. We used 15-s averaged observations to detect CH4 hotspots, weed dryer defined as locations with CH4 levels exceeding 350 ppb above local background. We exclude potential CH4 emissions from fossil fuel sources using the same C2H6 and CO criteria as described above. For each season, we then identified hotspots of CH4 downwind of dairy farms and derived the δ13C end-members with a Keeling plot, using the method described in section 2.5. To ensure the method described in section 2.5 is appropriate for the lower mole fractions observed from downwind sampling of other dairies in the region, we compared the δ13C end-members using the standard error from the CH4 dilution experiment described in section 2.3.4. against the standard error selected using the method described in section 2.5. There was no statistically significant difference between δ13C end-members using Welch’s t-test. Thus, to be consistent with analysis at the farm-scale, the method described in section 2.5 was selected to obtain source δ13C end-members from downwind plume sampling of other dairies. Isotope mixing equations from Fry were used to estimate the fractional contribution of the two CH4 sources, enteric fermentation source areas and manure lagoons,from CH4 hotspots.
Differences in the isotopic signatures from CH4 emissions generated from the free stall barns and corrals may be explained by the types of cattle housed in each area. To further explore this, we conducted isolated breath measurements of different cattle production groups during the winter season and evaluated their diet composition across seasons. Free stall barns only house milking cows and cows within a few days of parturition, while corrals house milk-fed calves in hutches , heifers, bull calves, and dry cows . As shown from the Keeling plots in Figure 2.4, the breath of milking cows and hutch calves were more enriched in δ13CCH4 relative to dry cows and heifers and bull calves . We used feed data collected at our reference test site farm to interpret the variations in δ 13C of CH4 emited from cattle in corrals and free stall barns at the reference test site farm. We found that the types of cattle housed in each area were each fed a distinct type of feed, consisting of C3, C4, or distiller’s dried grains of unknown composition . In all seasons, milking cows were fed a mixture consisting primarily of C3 and C4 feeds , with a small percentage of DDG . Hutch calves were milk-fed and also fed a mixture of C3, C4, and DDG feed, but with a larger percentage of DDG —the diet composition for hutch calves was more variable depending on the season. Bull calves were fed a wide range of C3 , C4 , and DDG feed depending on the month. In contrast, dry cows and heifers were predominately fed a C3 diet with a small percentage of DDG . Given that isotopic measurements of substrates were outside the scope of this study, we assumed that C4 feed had a δ 13C of -12.2 ± 0.3‰ and C3 feed had a δ 13C of -23.6‰ based on reported δ13C of maize and wheat in Chang et al. . For DDG, we assumed an equal mixture of C3 and C4 feed, resulting in a δ13C of -17.9 ± 0.3‰. To estimate the expected δ 13CCH4 for different cattle production groups at the reference test site, we used the linear regression equation derived from the empirical relationship between δ13Cdiet and δ13CCH4 from enteric fermentation of ruminants in Chang et al. . Based on these assumptions, milking cows and hutch calves are projected to emit more enriched δ 13CCH4 values relative to other cattle production groups . Although this pattern generally agrees with our study’s δ 13CCH4 measurements from enteric fermentation source areas, our δ 13CCH4 measurements were often more enriched than expected. The δ 13CCH4 from animal housing is likely impacted by emissions of isotopically enriched CH4 from manure deposited in corrals and free stall barns.The progression of manure from one component of the system to another also influenced the isotopic signature of CH4 at the reference test site. Using a chamber to isolate sources of manure at different stages of the manure management on January 15th, 2020, we observed that a mixture of fresh volatile solids with urine on the floor of free stall barns yielded the most depleted δ13CCH4 . Methane emitted from two separate manure piles at the solid drying area, however, had heavier δ13CCH4 signatures . The more depleted δ13CCH4 observations were from a manure pile that was noticeably drier than the second sample. In comparison, measurements from manure lagoons using the mobile laboratory resulted in δ 13CCH4 of -43.4 ± 0.4‰. Based on our measurement of the oxidation reduction potential , the manure waste stream is anaerobic from cell 1 onward to the holding pond . Prior to that, we expect the waste stream to have varied conditions that include anaerobic and aerobic microsites. Presumably some of the manure on the floors of cattle housing areas is anaerobic, given the continuous presence of water on the floors of free stalls.Stable carbon isotope measurements of CH4 can be a valuable source apportionment technique to distinguish between enteric and manure CH4. At the reference test site farm, we found a clear separation of δ13CCH4 signatures between enteric fermentation source areas and manure lagoons .