This study was conducted at a free-stall dairy operation in the Texas Panhandle (Figure 1). This dairy had three naturally ventilated free-stall barns with open sides and ends and housed a total of 3200 milking cows. The open-lots at this free-stall dairy were allocated for heifers and dry cows and were provided with feeding and watering systems. The open-lots contained about 3750 heifers and the surface was mostly dry with portions containing scattered feed, feces, and urine. No air sample from an open-lot pen with dry cows was collected and dry cows were not used to calculate emissions rates. Manure accumulated in the open-lot was removed by tractor mounted scraper twice a week and cleaned the open-lots twice a year. The scraped manure was stockpiled and either applied in the crop fields or pastures or composted on-site. The runoff from the open-lot during a rain event was channeled to a runoff pond that was separate from the treatment lagoon.

Manure that accumulated in the free-stall barns was flushed every two hours from storage tanks that were recycled wastewater from the secondary lagoon. The flushed liquid manure then channeled to the gravity settling basin where most of the solids were separated. The separated liquid goes to an inclined screen mechanical solid separator for further separating liquid and solids. Screened solids were composted in windrows and used as bedding for loafing pens. The liquid manure from the mechanical separator was channeled to the primary anaerobic lagoon. Then, effluent from the primary lagoon was conveyed to secondary lagoon through a pipe outlet. Air sampling using flux chamber was performed for five consecutive days during daylight hours (9:00 am to 7:00 pm) and concentration values reported for that period. These are the snapshot data of 5 consecutive days in summer, representing typical summer climate of the region. However, concentration values beyond this period might be different. Four measurements were taken 1 h after flushing and another 4 measurements were taken just before the next flush. A total of 74 samples were collected from the barn (manure lane and bedding area), loafing pens, open-lot, the settling basin, lagoons, and compost piles within the dairy operation (Table 2). Measurements were taken randomly at 8–14 locations in each GLAS. Seventy four chromatograms (visible records in a graph) showing the result of separation of the GHGs of an air mixture were collected during a week-long sampling campaign in summer from this dairy

At the feedyard, air samples were collected for five consecutive days during daylight hours (9:00 am to 7:00 pm) in August 2010. The feedyard fed approximately 40,000 head of cattle in the growth and finishing stages of beef production. Four corner pens of the feedlot were selected as sampling locations (Figure 2). Before sampling was initiated, an imaginary 6 × 6 grid was made for each pen/GLAS and each cell was assigned a grid location. The numbers selected corresponded to the cell where the flux chamber was to be placed. Air samples were collected randomly at 8–20 locations in each GLAS corresponding to the total number of grid cells (e.g., 1 to 36 for a 6 × 6 grid). The size of the entire pen surface, compost pile, and runoff pond were 910,000; 49,000; and 59,000 m², respectively (calculated from the Google map). The feedlot pen surface was mostly dry with portions containing scattered feed, feces, and urine. The manure solids on feedyard pens are scraped after each growing cycle and the manure is hauled to the compost area. The surface runoff of the feedlot pen surface was channeled to the runoff pond. Later, wastewater from the runoff pond piped or pumped to the runoff holding pond. A total of 87 air samples were collected from four corner pens of a large feedlot, runoff holding pond, and compost pile.

Detailed ration information was not available due to proprietary concerns. According to data supplied by the dairy, the diets for milking cows were mainly comprised of forage (wheat hay, wheat silage, alfalfa hay, alfalfa hay chop, and corn silage), grain (broken corn, corn rolled, corn pericarp, and cotton seed), protein (cotton seed meal, beet pulp, canola, and dried distillers grain), vitamin (A & D), and water. The ration constituted about 43.2, 32.9, 2.3, 18.0, and 3.6% of these constituents on an as fed basis, respectively. Similarly, heifers in open-lots were fed 93.6% forage, 2.3% grain, 3.7% protein, and 0.44% mineral on an as fed basis. On the other hand, forage (corn silage), grain (steam flaked corn), protein (wet distillers corn and condensed corn distillers), fat, beef finisher, and trace elements (rumensin, tylan, vitamin A & D) were 19, 57.2, 20.1, 0.8, 2.1, 0.8% on an as fed basis, respectively, for the feedyard. The dairy cattle ration contained 24.2% (on an as fed basis) more forage and 24.3% less grain than the feedlot ration. The protein contents were similar for both milking cows and feedyard rations. Dairy and feedyard animals were fed a mixed ration twice daily at around 7:00 and 14:00 h. In Texas, average weight of cow/heifers, dairy-milk cows, and finishing beef were 454, 590, and 546 kg, respectively [25].

The median CH₄ concentrations measured from the primary lagoon, settling basin, and secondary lagoon (3739; 1457 and 458 ppm_v, respectively) were significantly different (p < 0.05) and higher than those from other GLAS (Table 2). Cumulative, these three sources constituted about 23% of the total GLAS area (Figure 5). High temperatures during the summer were the main factor leading to higher CH₄ emissions from the settling basin and lagoons. Previous research reported that methanogenisis and subsequent methane production in the anaerobic settling basin and lagoons is strongly dependent on temperature that accelerated anaerobic decomposition of organic matter [32]. Weiske et al. [33] also reported that increased microbial activity due to higher temperature during the summer amplified the CH₄ production in a slurry-based manure management system. The highest median concentration was measured from the primary lagoon followed by the settling basin, secondary lagoon, open-lot, and loafing pen. Cumulatively, those five sources constituted about 88% of the total GLAS area (Figure 5). Concentrations measured at the manure lane, bedding area, and compost piles were significantly different but lower than those measured from other GLAS. Although not statistically different, median CH₄ concentrations measured from the loafing pen and open-lot (11 and 13 ppm_v, respectively) trended higher than those in manure lane, bedding area, and compost piles. This was likely due to fermentation induced by the anaerobic environment of the loafing pen and open-lot pen, and anaerobic soil layers produced immediately under the manure's top surface [5]. The average CH₄ concentration was 0.36 ppm_v from compost piles, which was lower than other sources, likely because the manure was already partially degraded.

The highest median CO₂ concentration was measured from the primary lagoon followed by the settling basin, open-lot, and secondary lagoon. Cumulatively, these four sources constituted about 68% of the total GLAS area (Figure 6). Higher CO₂ concentrations from those four GLAS were emitted due to anaerobic degradation of manure with high temperature in the summer. The median CO₂ concentration measured from the primary lagoon was significantly higher than those from the other GLASs (p < 0.05). Carbon dioxide concentrations between the primary lagoon and settling basin (2758 and 2641 ppm_v, respectively), and settling basin and open-lot (2641 and 1269, respectively) were statistically similar (p > 0.05); and significantly higher than those from the manure lane, bedding area, loafing pen, and compost piles. There were no significant differences in measured CO₂ concentrations between the open-lot and loafing pen (p > 0.05). Similarly, the median CO₂ concentration among manure lane, bedding area, and compost piles were significantly different and lower than those measured from other GLASs (p < 0.05).

The highest average N₂O concentration was measured from the bedding area, open-lot, and loafing pen. They were statistically similar (p > 0.05). Nitrous oxide concentrations measured from those three GLASs were significantly higher and different than those in the manure lane, settling basin, primary and secondary lagoons, and compost piles. There were no significant differences in the median N₂O concentrations between the primary and secondary lagoons; settling basin and compost piles; and manure lane and settling basin (p > 0.05; Table 2). Average atmospheric concentrations of CH₄, CO₂, and N₂O are approximately ∼1.7, ∼379, and ∼0.3 ppmv [2]. Nitrous oxide concentrations measured from the manure lane, settling basin, primary and secondary lagoons, and compost pile (0.1, 0.16, 0.06 and 0.04, and 0.2 ppm_v, respectively) were much lower than those in the ambient (background) atmospheric concentration (∼0.3 ppm_v). Low N₂O concentrations were due to the diluting effect of sweep air (zero grade air used to generate flux) inside the flux chamber since those GLASs contributed very little to emit N₂O. The freshly deposited manure in the manure lanes was the anaerobic product of enteric fermentation inside the rumen, which contains low N₂O. Similarly, a true anaerobic condition in the slurry (settling basin) and liquid manure in the lagoon produced low N₂O emissions. This was because N₂O is formed during aerobic nitrification and anaerobic denitrification. These results showed a good agreement with previous work by Osada et al. [34]. They observed that slurry manure emits a small amount of N₂O due to the poor aerobic conditions. Higher N₂O concentrations in the open-lot, bedding area, and loafing pen compared to other GLAS were likely due to occasional mixing of urine with manure and soil (open-lot pen), compost or separated solids (loafing pen), and sand (bedding materials) under aerobic nitrification process.

The aggregate CH₄ ER for this dairy was 836 g hd⁻¹ d⁻¹ in the summer (Table 3). The high aggregate CH₄ ER was mainly due to a higher methane concentration measured from the anaerobic lagoon at the dairy operation (Table 2). Methane production in an anaerobic lagoon mainly depends on the amount of degradable organic material in the wastewater and temperature. Previous researches also indicated that methanogenisis and subsequent methane production in the anaerobic manure storage and treatment structures such as lagoons, settling basins, and slurries were strongly dependent on high temperature [31–33]. The anaerobic primary lagoon contributed approximately 80% of the aggregate CH₄ ER. Based on the USEPA estimate [1], the aggregate CH₄ ER for this dairy was 238 g hd⁻¹ d⁻¹. The direct field CH₄ ER measurement was about 3.5 times higher than that of EPA estimate. Previous research reported an underestimation of CH₄ emission from anaerobic lagoon by 65% using the EPA method [36]. The median CH₄ ER estimated from the primary lagoon was significantly higher than those from the other GLAS (p < 0.05). This source alone contributed about 80% of the overall CH₄ emission and comprised about 8% of the total GLAS (Figure 5). The highest average CH₄ ER was estimated from the primary lagoon, followed by the secondary lagoon, settling basin, and open-lot pen. Methane production occurred during anaerobic manure decomposition with favorable high temperatures in the summer that increased manure bacterial activity and produced high CH₄ concentrations compared to other GLAS. Higher ERs from the settling basin, primary lagoon, and secondary lagoon were likely due to high CH₄ concentrations despite the relatively small footprint areas. Those three GLAS contributed about 98% of the overall CH₄ emission and comprised about 24% of the total emitting area (total GLAS). Therefore, abatement and management practices that address emissions from these sources will likely be most effective for reducing facility-wide emissions. There were no significant difference in CH₄ ERs between the settling basin and open-lot pen (p > 0.05). In contrast, there were significant differences in CH₄ ERs among the manure lane, bedding area, loafing pen, and compost piles (p > 0.05).

The overall calculated CH₄ ER for this dairy was 5573 g hd⁻¹ d⁻¹ in the summer (Table 3). The average CO₂ ER estimated from the open-lot pen was significantly higher than those from the other GLAS (p < 0.05). This source alone contributed about 51% of the overall CO₂ emission and comprised about 45% of the total GLAS (Figure 6). The highest average CO₂ ER was estimated from the open-lot, followed by primary lagoon, secondary lagoon, and loafing pen. These four GLAS altogether contributed about 94% of the overall CO₂ emission and comprised about 81% of the total emitting area (total GLAS). Therefore, abatement and management practices to reduce CO₂ emissions can be directed to those four GLASs. It is likely that the anaerobic decomposition of the manure in the primary lagoon during summer resulted higher CO₂ ER and this source (GLAS) alone contributed about 24% of the overall emission and comprised of 8% of the total emitting area (Figure 6). Higher CO₂ in the open-lot and the loafing pen may be due to the incomplete anaerobic decomposition of the mixture of manure, urine, and soils (open-lot pen); and mixture of manure, urine and separated solids (loafing pen). There were no significant difference in CO₂ ERs among the loafing pen, and primary and secondary lagoons. The CO₂ ERs from those three sources were significantly higher than those estimated from the manure lane, bedding area, settling basin, and compost piles (p < 0.05).

The overall estimated N₂O ER for this dairy was 3.4 g hd⁻¹ d⁻¹ in the summer (Table 3). Based on the USEPA estimate [1], the aggregate N₂O ER for this dairy was 0.7 g hd⁻¹ d⁻¹. In-field measurements of CH₄ emissions were about 5 times higher than EPA estimates. Similar to CH₄ emissions, N₂O emission estimates may have been underestimated by the USEPA method. The average N₂O ER estimated from the open-lot was significantly higher than those from the other GLASs (p < 0.05). Higher N₂O from the open-lot pen compared to other GLAS were likely due to the mixing of urine with manure and soil that facilitated aerobic nitrification process. This source (open-lot) alone contributed about 72% of the overall N₂O emission and comprised about 45% of the total GLAS (Figure 7). The highest average N₂O ER was estimated from the open-lot, followed by the loafing pen, and bedding area (p < 0.05). These three GLAS altogether contributed about 96% of the overall N₂O emission and comprised about 61% of the total emitting area (Figure 7). Therefore, abatement and management practices that address emissions from these three sources will likely be most effective to reduce N₂O emissions for this dairy. There were no significant difference in N₂O ERs among the manure lane, and primary and secondary lagoons.

At the feedyard, four corners of a large feedlot were sampled for GHGs. The minimum and maximum median volumetric concentrations of GHGs measured at various pen surfaces during the summer were 2.8 to 5.3, 668 to 1054, and 0.31 to 0.41 ppm_v for CH₄, CO₂, and N₂O, respectively (Table 4a). The highest average CH₄ concentration was measured from the SE corner pen, which had relatively more manure on the surface compared to other pens (p < 0.05). No significant differences in median CH₄ concentrations among the NW, SW, and NE corner pens. No significant differences in CO₂ concentration were detected between NE and SE corner pens and they were significantly higher than those measured from NW and SW corner pens (p < 0.05). Carbon dioxide concentrations between NW and SW corner pens were statistically similar (p > 0.05). Similarly, no significant differences in N₂O concentrations between NW, SW, and NE corner pens, but they were significantly higher than that from the SE corner pen. The measured gas concentrations within and among the pens were found to vary widely as indicated by quartiles (Table 4a), likely due to spatially variable manure loading rates, fluctuations in temperature, and variations in populations of microorganisms that were responsible for gaseous emissions at various pen surfaces. The pooled data (70 observations) from four pens were used to represent the concentration and emission rate from the entire feedlot pen surface.

The highest CH₄ concentration was measured at the runoff holding pond, followed by the pen surface (feedlot) and compost piles. Median CH₄ concentrations in the runoff holding pond were significantly different and higher than those in the pen and compost pile areas. The average CH₄ concentration measured from the runoff holding pond was about 22 and 57 times higher than those from the pen surface and compost pile areas (Table 4b). The average methane (CH₄) concentration measured over the pen surface was about 2.6 times higher than that measured at the compost pile. Higher CH₄ concentrations in the anaerobic runoff holding pond were likely due to high temperature during summer time. Previous research reported that methanogenisis and subsequent methane production in anaerobic settling basins and lagoons is strongly dependent on temperature [31]. CH₄ concentrations measured from the compost pile were the lowest among the sources (Table 4b). The decomposed organic manure in the compost piles produced low CH₄ ER.

Median CO₂ concentrations were also highly variable within each GLAS as indicated by the large quartiles (Table 4b). Unlike CH₄ concentrations, median CO₂ concentrations measured from the pen surface were significantly different and higher than those measured from the runoff holding pond and compost pile area. The median CO₂ concentration measured from the pen surface was about 2 and 1.5 times, respectively, higher than those from the runoff pond and compost pile areas (Table 4b). Median CO₂ concentrations were higher at the compost pile area compared to the runoff holding pond, although, there was no significance difference in CO₂ concentrations between those two GLAS. Higher CO₂ concentrations above the compost pile were likely due to high temperatures during summer time.

Similar to CH₄ and CO₂ concentrations, N₂O concentrations were also highly variable within each GLAS as indicated by the quartiles (Table 4b). The highest median N₂O concentration was measured from compost piles, followed by the pen surface and runoff pond. Median N₂O concentrations in those three sources were significantly different (p < 0.05). Median N₂O concentrations measured at the compost pile were about 2.8 and 14 times higher than those in the pen surface and runoff pond, respectively (Table 4b). The low median N₂O concentration above the runoff holding pond was due to a true anaerobic condition since N₂O is formed during aerobic nitrification and anaerobic denitrification. Previous work also reported that slurry manure emits a small amount of N₂O due to the poor aerobic conditions [34].

The aggregate ERs for CH₄, CO₂, and N₂O were 836; 5573; and 3.4 g hd⁻¹ d⁻¹, respectively, for the dairy operation during the summer excluding enteric CH₄ emissions. Similarly, the aggregate summer ERs for CH₄, CO₂, and N₂O were 3.8; 1399; and 0.68 g hd⁻¹ d⁻¹, respectively, estimated from the feedyard. The aggregate, CH₄, CO₂, and N₂O ERs from the dairy facility during the summer were about 220, 4, and 5 times higher, respectively, than those from the feedyard operation. Greenhouse gas ERs are also expressed in CO₂ equivalent (CO₂e) as kg of CO₂e hd⁻¹ d⁻¹ taking into account all the compounds (CH₄, CO₂, and N₂O). To calculate CO₂e, conversion factors of 25 and 310 were used for CH₄ and N₂O, respectively. Aggregate GHG ERs estimated from the dairy and feedyard operations were about 27.5 and 1.7 kg hd⁻¹ d⁻¹ CO₂e, respectively, when all three compounds (CO₂, CH₄, and N₂O) were used in estimating CO₂e. The aggregate summer GHG CO₂e in terms of kg hd⁻¹ d⁻¹ estimated from the dairy was about 16 times higher than that from the feedyard. It was estimated that about 76% of the aggregate GHG CO₂e was obtained from CH₄ and 4% from N₂O at this dairy operation. In contrast, the contributions of CH₄ and N₂O to the aggregate GHG CO₂e were 6 and 12%, respectively, from the feedyard operation. This was because the low CH₄ emissions from the runoff holding pond attributed to lower aggregate CO₂e from the feedyard operation.

Based on the GHG ERs for this dairy with 6,950 animals (milking cows and heifers), an estimated total GHG emission during the summer was approximately 191 tons of GHG CO₂e per day. For the feedyard with 40,000 cattle heads, an estimated summer GHG emission was about 68 tons of CO₂e per day. The summer GHG emission estimated from the dairy operation was approximately 3 times higher than that from the feedyard operation. Greenhouse gas emissions depend on animal type and growth stage due to different diets, feed conversion mechanisms (feed digestibility), and the manure management [3]. Higher aggregate GHG CO₂e emissions from the dairy compared to the feedyard were driven by differences in CH₄ emissions. This was due to different manure handling and management processes such as the use of an anaerobic settling basin and lagoons. These sources inherently emit more CH₄ than a feedyard operation. Phetteplace et al. [22] also reported higher aggregate GHG emissions (in CO₂e) from dairy operations relative to feedyards.

The aggregate GHG ER from this study was also compared with that estimated based on the USEPA proposed rule. According to the proposed rule, facilities annually emitting 25,000 metric tons of CO₂e or more report their GHG emissions to the EPA [1]. Emission rates were estimated based on the annual GHG CO₂e emission limit (25,000 metric tons of CO₂e) imposed by the USEPA and the corresponding threshold population for a particular manure management operation [38]. Per USEPA recommendation henceforth USEPA ER, for a particular type of operation (e.g., dairy or feedyard) with manure management system types (e.g., for dairy: cows with uncovered anaerobic lagoon with solid separation, heifers and calves are on dry lot and for feedyard: steers and heifers on feedlot, and manure composted), the average annual ambient temperature °C (region), and animal population threshold with respect to the emission cap (25,000 metric tons of CO₂e) were used to estimate the USEPA ER [38]. In this research, a region with an annual average ambient temperature of about 14.5°C [39] and animal population threshold (5,500 and 59,100 for the type of dairy and feedyard, respectively) level below which facilities were not required to report emissions from manure management was chosen [38].

On a “per head” basis, the aggregate USEPA GHG ER (based on annual average) was about 12.45 and 1.16 kg hd⁻¹ d⁻¹ CO₂e for dairy and feedyard operations, respectively [37]. The annual USEPA GHG ER in kg hd⁻¹ d⁻¹ CO₂e for dairy was about 11 times higher than that of the feedyard. Estimated aggregate summer ERs from this study at dairy and feedyard operations were 27.5 and 1.7 kg hd⁻¹ d⁻¹ of CO₂e, respectively. The summer GHG ER at the feedyard based on direct measurement was about 1.5 times higher than the annual USEPA ER based on annual average (1.7 and 1.16 kg hd⁻¹ d⁻¹ CO₂e, respectively). Similarly, aggregate summer GHG CO₂e at the dairy obtained from the direct measurement was approximately 2.2 times higher than that of the USEPA ER. A recent study reported CH₄ ER from anaerobic lagoons estimated using the USEPA approach may be underestimated by 65% [36]. Higher CH₄ concentrations (Table 2) from anaerobic liquid manure treatment sources (settling basin, primary and secondary lagoons) contributed to high aggregate CO₂e from the dairy operation.

The proposed USEPA emissions threshold table does not include CO₂ [1]. Therefore, GHG ERs in terms kg of CO₂e hd⁻¹ d⁻¹ was also estimated in our study using CH₄ and N₂O only. Greenhouse gas ERs estimated from the dairy and feedyard operations were about 22 and 0.31 kg hd⁻¹ d⁻¹ CO₂e when CO₂ was not included in estimating CO₂e. According to the new calculation, the aggregate dairy GHG ER based on direct measurements was about 1.8 times higher than that of the annual USEPA ER. In contrast, the USEPA ER for feedyard GHG (1.16 kg CO₂e hd⁻¹ d⁻¹ was about 3.7 times higher than that obtained with the direct measurement (0.31 kg CO₂e hd⁻¹ d⁻¹). It is well recognized that GHG emissions vary with weather conditions especially with temperature change. Higher temperature in the summer accelerates GHG emissions from same anaerobic manure management operations than those in the winter with lower temperature. Previous research reported that the overall CH₄ and N₂O ERs during the summer were approximately 5.3 and 2.2 times higher than those in the winter for a free-stall dairy in central Texas [40].