Introduction

Ammonia (NH3) is a significant air pollutant, especially in combination with acid gas production from fossil fuel combustion, because the resulting acid–base reaction potentially leads to an air quality problem in the form of haze and respirable particulate matter (PM). The link between PM and increased mortality is well established (Pope et al. 2002; Cohen et al. 2005). Ammonia emissions’ estimates from swine manure treatment lagoons, as a percent of feed nitrogen (N) input, have been reported to vary from 36 to 71 % (Doorn et al. 2002a; Hatfield et al. 1993; USEPA. 2004). From a systems’ analysis approach using the USEPA National Emissions Inventory (USEPA, 2004), the addition of all NH3 emissions’ components, such as from housing (22 %), lagoons (43 %), field application of manure (23 %), N leaving as animal protein (30 %, from host data), suggest that more than 100 % of the N entering the farm system is leaving the farm as NH3 volatilization plus animal product. Recent studies in North Carolina (NC) (Harper et al. 2004b), the Georgia Coastal Plains (GA) (Harper and Sharpe 1998; Harper et al. 2000), and the Central Great Basin (CGB) (Harper et al. 2010; Weaver et al. 2012) regions have shown that swine lagoons emit significantly less NH3 than previously and currently thought. Much of the N estimated as NH3 gas emissions has been found to be converted to dinitrogen gas (N2) (Harper et al. 2000, 2004b; Weaver et al. 2012), representing an even larger discrepancy for the N balance of farm systems suggested by the USEPA. This aspect of dinitrogen emissions, not considered in most of the estimates of NH3 emissions from animal feeding operations (AFOs), highlights the fact that the N cycle in lagoons is not fully understood. Benign N2 emission from lagoons is a pathway of N emissions is that is significant and must be considered in the total N balance of AFOs. When the National Emissions Inventory (USEPA 2004) NH3 emissions values are combined with published (measured) N2 emissions (Harper et al. 2000, 2004a, b; Weaver et al. 2012), in many cases more N as NH3 plus N2 is emitted than is excreted by the animals, suggesting the need to reevaluate emissions’ estimates.

Many of the current NH3 emissions’ estimates are based upon chamber measurements. A number of studies using dynamic chamber measurements (Aneja et al. 2000; Blunden and Aneja 2008) have led to higher emission estimates than found by micrometeorological measurements (Harper and Sharpe 1998; Harper et al. 2000, 2004b, 2010). Doorn et al. (2002b) pointed out that studies with dynamic chambers led to emission factors 2.3 times higher than studies with micrometeorological techniques, while others (Shah et al. 2006; Rochette et al. 1992; Harper 2005; Harper et al. 2010; 2011 ) stated that chamber techniques are not even suitable for developing emission factors as they create conditions at the water surface that overestimate NH3 emissions. Based on all of the evidence (Harper et al. 2000, 2004b; Weaver et al. 2012) and discussions regarding the physical chemistry of highly anaerobic systems (van Clemput 1972, 1997), it seems very plausible that NH3 emissions from lagoons are lower than indicated by current emission factors and a significant fraction of N is emitted as N2.

There are complex interactions between carbon (C) and N compounds during manure processing by microbial and chemical processes. While little emissions’ research for methane (CH4) and carbon dioxide (CO2) has been accomplished (Sharpe et al. 2001; DeSutter and Ham 2005) in AFOs, Harper et al. (2000; Table 1; 2010) found interesting correlations between emissions of NH3, CH4, nitrous oxide (N2O), and CO2 from manure-processing lagoons. These and other studies (Harper et al. 2010) show that manure management aimed at reducing the emissions of one gas could have the undesired consequence of increasing emissions of other gases. In these studies, manure lagoons with a high rate of methanogenesis also converted significant amounts of ammonium (NH4 +) to benign N2 gas with little or no N2O produced [in the lagoons with the highest rate of methanogenesis, atmospheric N2O was actually absorbed by the lagoon (Harper et al. 2000)]; however, when methanogenesis decreased, smaller emissions of N2 occurred and higher rates of N2O were produced. Harper et al. (2010) also showed that removing organic material from swine production farms for biogas production reduced CH4 emissions by 47 % (the reduction resulted in a 44 % decrease in radiative forcing gases) from the biogas farms while increasing NH3 emissions from the biogas farms by 46 %, a substantial increase in air-quality emissions. Weaver et al. (2012) also showed similar results. The above studies suggest there is a relationship between the amount of methanogenesis and conversion of NH4 + to N2 gas in manure-processing lagoons. Thus, the main purpose of this study was to measure biological gas emissions from six manure-processing lagoons within a three-county area of eastern NC, a farm in GA, and in a large swine operation in the CGB (15 farms), and to evaluate the relationship between methanogenesis (CH4 production) and conversion of organic and inorganic N to N2 gas (N2 production).

Table 1 Summary of lagoon ebullition studies for methane, carbon dioxide, and dinitrogen gases used in this report
Full size table

Materials and methods

In the 22 swine lagoons studied from all regions, undecomposed organic material (manure) from animal production houses is pumped to lagoons where the organic materials settle to the bottom forming a layer of semi-solid organic material which is anaerobically decomposed producing gas. Gas bubbles emitted from the sludge layer in each of the lagoons, were trapped in six collectors (Fig. 1) randomly located within each of six areas of the lagoon. These gas collectors do not interfere with the emission process, as with NH3 chambers (Harper, 2005). On a short-term basis (< ~2 weeks), ebullition gases, the result of biological and thermodynamic processes, are emitted from the lagoon bottom and are not affected by climatic events at the lagoon surface; however, NH3 emissions are highly influenced by the physical processes of water surface turbulence and temperature. The collectors were made of 20-L, open-bottom carboys (0.275 m diameter) with flotation collars at the top of the carboys (Fig. 1) and tethered to the lagoon bottom to collect the mass-flow gases (bubbles) before they reached the water–air interface. All air was removed from the collectors at placement. Water in the collectors was displaced by the ebullition gases over time, visually measured on a graduated scale on the collector periodically to determine gas mass-flux. Gases were transferred from the collectors using sample lines flushed with the gases from the collectors and then subsequently attached to evacuated six-L SUMMA canisters. The SUMMA canister samples were then transported to a laboratory where gas samples were analyzed by gas chromatography (Harper et al. 2004b; Weaver et al. Weaver et al. 2012). No N2O was found in the collectors via GC. In other studies no N2O emissions were found from anaerobic lagoons using atmospheric transport techniques and tunable diode laser spectroscopy (Harper et al. 2000, 2004a, b). Samples of helium (He) injected into the collectors showed a sampling procedure error of about 1 % due to atmospheric N2 contamination (see Harper et al. 2004b). Further, modeling studies showed the theoretical maximum contamination from the atmosphere would be <5–10 % (De Visscher and Harper 2005, unpublished data). Gas fluxes were determined by measuring the amount of gases collected divided by the time between measuring intervals (collection volumes were measured as ebullition necessitated, normally from two to three times per week in summer and weekly or bimonthly in winter) and then multiplying the emissions by the measured concentrations of each gas. This sampling protocol has been used extensively and further description of the measurement technique may be found in Harper et al. (2000, 2004b) and Weaver et al. (2012).

Fig. 1
figure 1

Gas collectors constructed from open-bottom 20-L carboys. Graduations for evaluation of gas volumes located on the side of the carboy

Full size image

A summary of all farms in this study is included in Table 1. Fifteen farms of four different types in the CGB were sampled during 2002–2006: two sets (2002–2003 and 2004–2006) of three each F farms; another set of three F farms with organic matter removed for biogas production (2004–2006); one set of three each of nursery (N) farms (2002–2003) and sow (S) farms (2002–2003). Data from 2004 to 2006 are from an earlier published study (after Weaver et al. 2012). Six farms were sampled during 1999–2001 for N and C emissions in NC including three farrow-to-wean (FW), two finisher (F), and one farrow-to-finish (FF) farms. Data from an F farm in GA during the period of 1994–1998 were included (after Harper et al. 2000). Farm animal numbers ranged from 1,400 to 12,000.

Farms were selected in three geographical areas: fifteen in the CGB, six in three NC counties, and one in GA, to evaluate the effect of management on biogas emission rates (subject to host availability). The farm types included F, FW, and FF farms with input feed protein ranging from 13 to 17 % (feed N from 2.1 to 2.7 %). Three sow farms in the CGB were selected for comparison to production farms. Feed input, feed analysis, animal numbers and weights, number of animals sold, and other management information were supplied by the host owners/managers where available. Lagoon temperature was measured 2.5 cm below the water surface and within the sludge layer with micro temperature-loggers [Onset Computer Corp, Bourne, ME (Note: commercial names are included for the benefit of the reader and do not imply endorsement by the authors or their host institutions)]. The lagoons typically never formed crusts on the surface and were well mixed as demonstrated by near uniform temperatures from the top to the top of the sludge layer in lagoons of the CGB study. Since it is not appropriate to calculate NH3 emissions from chamber systems (Harper 2005), lagoon NH3 emissions were calculated from pH, NH4 + measurements of effluent samples (collected in bottles at the surface of each lagoon), surface lagoon temperatures, and wind speeds measured at 1.5 m height (from a metereological station on site), and a lagoon NH3 emissions model by De Visscher et al. (2002). Housing NH3 emissions were estimated from a model developed by Harper et al. (2004a) for North Carolina swine farms.

Effluent and sludge layer samples collected were frozen immediately and shipped to a laboratory for analysis of NH4 +, nitrate (NO3 −), nitrite (NO2 −), and pH (for a description of analysis procedures see Harper et al. 2000, 2006). All lagoons were sampled similarly on a monthly basis.

The precision of biogas emission measurement was evaluated using the absolute value of the coefficient of variation, or relative standard deviation (RSD), obtained by dividing the standard deviation by the mean. To evaluate precision of the individual carboy measurements, the daily carboy emissions of the six lagoon carboys were used to calculate the daily average for the lagoon along with its RSD. The daily RSDs of the lagoons were averaged to calculate the average daily RSD and standard deviation of the daily RSD. To evaluate the precision of the farm lagoon emissions, a similar procedure was followed. Individual farm lagoon emissions are the average of the six carboy measurements. As lagoon emissions from three identical farms were measured for each farm type, the average individual farm emissions per farm type could be determined as well as the standard deviation of the individual farm emissions to calculate daily, monthly and yearly RSDs for each farm type. The RSDs for each farm type were averaged and a standard deviation was subsequently determined for daily, monthly and yearly RSD of individual farm emissions.

Results and discussion

Precision of biogas emission measurements

Average annual gas emissions (total component and percent of total) increased as the amount of farm input-feed per size-of-lagoon increased (i.e. increased manure C with respect to lagoon processing size). Biogas production varied substantially among the six collectors on each lagoon site. The RSD between collectors on a single lagoon, on a daily basis, was 48 ± 13 %. While there was considerable variability between individual collector’s measurements, the variability of biogas emissions measurements was much less between lagoons when the six collectors were averaged. For example, the average RSDs of lagoon daily biogas emissions (average of six collectors) from lagoons of identical farms in the CGB were 23 ± 2 %. The variability of measurements between identical farms decreased even further when compared on a monthly (average RSD = 14 ± 6 %) or yearly basis (8.8 ± 6.0 %). We interpret this to mean that the 6 collectors are adequate to determine representative emission measurements on a yearly basis. Individual gas emissions showed regression relationships vs. feed input (R2) greater than 0.67 for total component emissions (Fig. 2a) and greater than 0.86 for percent of all component gas emissions versus feed input (Fig. 2b).

Fig. 2
figure 2

Average annual lagoon methane, dinitrogen, and carbon dioxide emissions as emissions per unit area of lagoon surface (a) and percent of total gas emissions (b) with respect to feed input per lagoon size in North Carolina

Full size image

Climate/temperature effects

When comparing biogas production between farms, temperature effects in the lagoon sludge must be considered. Farms from the CGB were included to test the robustness of the trends in biogas production and their relationship to feed input and temperature. Each system monitored in the CGB was comprised of three identical farms allowing for quantification of the variance in the data. Average monthly CH4 production was directly related to sludge (where most of the processing occurs) temperature (Fig. 3a). On an annual basis, measured sludge temperatures were found to be within one degree (0.8 °C higher) of the average annual air temperature at 1.5 m height (Harper and Weaver, unpublished data), and we suggest that air temperature can be used as a surrogate temperature for the sludge. When CH4 production was plotted versus average monthly air temperature (Fig. 3b), the gas production dependence upon monthly air temperature was almost as good as sludge temperature (R2 values similar). Additionally, the dry climate of the CGB causes much higher evaporation rates and results in different management of swine lagoons.

Fig. 3
figure 3

a Average monthly methane production (of three farms) in relation to the sludge temperature at the bottom of the lagoons (where most of the decomposition occurs) over 2 years in the Central Great Basin (CGB). b Average monthly methane production (of three farms) in relation to the air temperature over 2 years in the CGB

Full size image

Feed input effects

Data from the NC farms were used to test for the effects of feed input on biogas production. The NC data demonstrated that total biogas emissions (kg gas ha−1 d−1) increased linearly (Fig. 2a) with daily feed input per lagoon size (kg feed d−1 ha−1) (R2 = 0.67). Component gas emissions all increased linearly with feed input but CH4 had the largest increase with feed input per lagoon size (R2 = 0.78). Carbon dioxide and N2 gas emissions also increased linearly but at smaller rates than CH4 (with correlations of R2 = 0.76 and 0.32, respectively). Lower correlations for N2 gas can be partially explained by a change in composition of biogas with feed input (Fig. 2b) where CH4 and CO2 emissions, as the percentage of total gas production, increased and N2 emissions decreased with respect to increased daily feed input rates (kg feed d−1ha−1). The N2 gas produced from the conversion of NH4 + to N2, was not positively correlated with feed input (as was CH4 and CO2) since N2 is produced via a different mechanism than methanogenesis (Weaver et al. 2012).

Feed input values from the CGB could not be used to predict emissions (using the linear relationships determined in Fig. 2a) in other areas due to large differences in lagoon temperatures and to very different animal and manure management. In the CGB lagoons, no effluent was discharged from the lagoon system to maintain water levels (evaporation was sufficient); thus, organic matter was diminished only by anaerobic decomposition and all lagoon N was removed either via NH3 volatilization and/or conversion of NH4 + to N2 gas (Harper et al. 2000, 2004b; Weaver et al. 2012). Harper et al. (2000) found no N2O emissions from swine anaerobic processing lagoons (indeed, there was absorption of N2O from the atmosphere by the lagoon). Additionally, because the feed input (kg feed ha−1d−1) was similar between lagoons in CGB (Table 1), the relationship between feed input and emissions could not be tested in the CGB.

The relationship between NH4 + concentration and gas emissions was evaluated in the NC lagoons. Similar to lagoon biogas emissions in a GA study (Harper et al. 2000), as NH4 + concentration increased across the six NC lagoons, total and individual gas emissions increased. However, the increased emissions effect was due to an increase in manure availability resulting in more biological decomposition from more feed input (Fig. 2a). Additionally, NH4 + concentrations also increased with more biological decomposition. Consequently, higher NH4 + concentrations and gas emissions are both correlated to feed input and not necessarily to each other.

Mechanisms for N2 production

When Harper et al. (2000) could not balance the feed N input and all forms of N output (including meat, lagoon NH3 volatilization, field application NH3 losses, field denitrification losses of N2 and N2O emissions, lagoon N2O emissions, etc.), they suggested the possibility that some of the NH4 + may have been converted to N2 during manure-processing and that different reactions were involved, depending on the N form and concentration. With higher NH4 + concentration and biological activity (i.e. CH4 production) their studies suggested that the N2 production may have occurred via ‘chemical denitrification’ (Van Cleemput 1997). Thermodynamics and the Gibbs free energy of reaction for chemical denitrification (Van Cleemput 1972) suggest that spontaneous conversion of NH4 + to N2 may occur in animal manure lagoons (Harper et al. 2004b, Table 7). It is possible that there is some biological denitrification in the lagoons, but we think it is small since we measured little NO3 − (<0.1 mg NO3 −-N L−1). Furthermore, dissolved oxygen (O2) concentrations (mean of about 0.1 % dissolved O2 across all the primary lagoons) can barely support autotrophic nitrification even under otherwise optimal conditions. We did not find NO2 −, an intermediate step in biological nitrification/denitrification, in any of the primary lagoons. Zhang (2003) in studies of an anaerobic sludge reactor also found almost all nitrite removed (97–100 %) with gas contents of 89, 8, and 3 % of N2, CH4, and CO2, respectively. These and other anaerobic laboratory studies (Harper et al. 2001, unpublished data) showed similar conversion of solution NH4 + to N2 gas. Studies of swine lagoons by Hunt et al. (2010) found similar conclusions to Harper et al. (2000, 2004) finding little N2O (produced from incomplete denitrification) being part of the system N balance. They also found there was a lack of sufficient denitrification enzyme activity (DEA) within the wastewater to support large N2 losses via classical nitrification and denitrification.

There are other possible microbial processes to explain the N2 production (Thamdrump 2012). Like classical denitrification and the anaerobic ammonia oxidation bacterial process (ANAMMOX), the full extent of conversion of NH4 + to N2 remains unclear (Ettwig et al. 2009). Kartal et al. (2011) has recently presented strong evidence to explain the ANAMMOX mechanism for conversion of NH4 + to N2 production; meanwhile, in this paper we demonstrate that the simple conversion of NH4 + to N2 is thermodynamically favorable later in the manuscript.

Harper et al. (2000) showed that as lagoon NH4 + increased, NO3 − and dissolved O2 decreased, while N2 and CH4 emissions increased. Other studies have shown that when organic C is removed for biogas production, methanogenesis is reduced and the lagoon NH4 + content is increased (Amon et al. 2005) and measured whole-farm NH3 emissions are increased (Harper et al. 2010). The above studies had treatments which reduce methanogenesis or lagoon NH4 + concentration but the studies in this research compare emissions from normal animal management and manure processing systems. The three lagoons in which organic matter was removed for biogas production were included to provide an additional comparison for the effect of reducing decomposition and methanogenesis.

Thermodynamic relationships

The net effect of all these studies suggests that as methanogenesis is decreased, conversion of NH4 + to N2 is decreased. We think the causal relationship between methanogenesis and NH4 + to N2 conversion is thermodynamically favored, while competing with other NH4 + oxidizing reactions. The following reactions are considered:

$$ {\text{NH}}_{ 4}^{ + } \left( {\text{aq}} \right) \, + \, 0. 7 5 {\text{ O}}_{ 2} \left( {\text{g}} \right) \, \to \, 0. 5 {\text{ N}}_{ 2} \left( {\text{g}} \right) \, + {\text{ H}}^{ + } \left( {\text{aq}} \right) \, + { 1}. 5 {\text{ H}}_{ 2} {\text{O}}\,\left( {\text{l}} \right) $$
(1)
$$ {\text{NH}}_{ 4}^{ + } \left( {\text{aq}} \right) \, + {\text{ 2 O}}_{ 2} \left( {\text{g}} \right) \, \to {\text{ NO}}_{ 3}^{-} \left( {\text{aq}} \right) \, + {\text{ 2 H}}^{ + } \left( {\text{aq}} \right) \, + {\text{ H}}_{ 2} {\text{O}}\,\left( {\text{l}} \right) $$
(2)
$$ {\text{NH}}_{ 4}^{ + } \left( {\text{aq}} \right) \, + { 1}. 5 {\text{ O}}_{ 2} \left( {\text{g}} \right) \, \to {\text{ NO}}_{ 2}^{-} \left( {\text{aq}} \right) \, + {\text{ 2H}}^{ + } \left( {\text{aq}} \right) \, + {\text{ H}}_{ 2} {\text{O}}\,\left( {\text{l}} \right) $$
(3)

Reaction (1) could represent either a chemical denitrification step, or a microbial process. Without more direct evidence no distinction can be made between a chemical and a microbial process. Hence, we simply refer to reaction (1) as a “conversion” without specifying its nature. Reaction (2), or nitrification, is discussed below. Reaction (3) is significant as no appreciable concentrations of NO2 − were determined in any of the lagoons. The nitrite ion is key for the anaerobic oxidation of NH3 (ANAMMOX) as NO2 − must be present [Eq. (4)]:

$$ {\text{NH}}_{ 4}^{ + } \left( {\text{aq}} \right) \, + {\text{ NO}}_{ 2}^{-} \left( {\text{aq}} \right) \, \to {\text{ N}}_{ 2} \left( {\text{g}} \right) \, + {\text{ 2 H}}_{ 2} {\text{O}}\,\left( {\text{l}} \right) $$
(4)

The Gibbs free reaction energy ΔrG of the three reactions was calculated under the following conditions: NH4 + concentration 1500 mg L−1, NO3 –-N concentration 0.1 mg l−1, NO2 −-N concentration 0.1 mg L−1, pH 8, N2 partial pressure 81 kPa, O2 partial pressure 0.1–10−15 bar. The calculation is similar to that of Harper et al. (2004b) except that the speciation between NH3 and NH4 + was not considered (NH4 + is the dominant species and its concentration does not influence the relative ΔrG between the three reactions).

This relationship is illustrated in Fig. 4. A negative value of ΔrG indicates that the reaction is thermodynamically favorable. It is clear that the formation of NO3 – from NH4 + is thermodynamically more favorable than N2 or NO2 – formation at O2 partial pressures above 10−8 bar when other concentrations remain the same. At lower O2 partial pressures N2 formation is thermodynamically more favorable than the formation of NO3 – and NO2 – from NH4 +. This might explain why N2 production and CH4 production are correlated. Methanogenesis is only possible at extremely low O2 concentrations, and under these conditions N2 production is thermodynamically more favorable than NO3 – production. This should not be interpreted as conclusive evidence, as both reactions are thermodynamically favorable in all conditions considered; and, other factors like kinetics play a role as well. The presence of an electron donor (organic material) removes oxygen to the point where NO3 − production becomes thermodynamically less favorable than N2 production. Kinetically, nitrification has an estimated saturation constant of 0.5 mg L−1 according to the standard activated sludge model, ASM3 (Gujer et al. 1999). The ASM3 model predicts that nitrifiers cannot maintain their activity at oxygen concentrations below 0.026 mg L−1 (6.3 × 10−4 bar) under otherwise optimal conditions (i.e., in the absence of any other limiting factor).

Fig. 4
figure 4

Gibbs free energy of reaction for N2 production and NO3 − production from NH4 + versus O2 partial pressure (NH4 + concentration 1,500 mg L−1, NO3 − concentration 0.1 mg L−1, pH 8, N2 partial pressure 81 kPa)

Full size image

The sensitivity of the thermodynamics of reactions (1) (2), and (3) to the variables that were kept constant in the above analysis was investigated. The sensitivity of ΔrG to any of the reactants or products was determined to be less than 11.42 kJ mol−1 for any 100-fold change in concentration (or 2 pH units). It is concluded that the thermodynamics of NH4 + oxidation is only slightly sensitive to pH and concentrations of NO3 −, NO2 −, N2, and NH4 +, so a possible uncertainty of any of these variables will not invalidate the analysis.

Comparison of system N emissions

The relative N emissions (ratio of N emitted to feed N input) from the farms in a geographical area (in NC) are shown in Fig. 5 (volatile NH3-N from housing and lagoons, NH4 +-N conversion to N2, protein-N, and unknown-N). Measured N2 emissions were not consistent within farm types or across all farm locations. The smallest N2 flux occurred in a farm (FW #3) which also had the highest estimated housing NH3 emissions. Inversely, the largest N2 flux was in a farm (FF) with the smallest housing NH3 emission losses [housing NH3 emissions were only slightly linearly correlated with N2 emissions across all farms, R = 0.63 (R2 = 0.40, n = 6)]. Although not conclusive, the inverse relationship suggests that increased N loss as NH3 will reduce N2 emissions.

Fig. 5
figure 5

Comparison of N emissions from farms due to housing NH3 losses, lagoon NH3 and N2 losses, and protein N removal from the farms (F finisher, FF farrow to finisher, and FW farrow to wean)

Full size image

Statistical models for gas emissions

The correlation of methanogenesis and N2 emissions (R2 = 0.78) was quite good across the lagoons studied in NC leading us to consider if the relationship (y = 0.23x) would be comparable across wide geographical regions of the U.S., as well as with management practices. Methane and N2 emissions were combined (Fig. 6) with the studies in the CGB and from a previous study in GA (Harper et al. 2000). The relationship between CH4 and N2 emissions were surprisingly similar changing the overall correlation only slightly, R2 = 0.71, and a linear relationship of y = 0.20x, suggesting a near-universal relationship between methanogenesis and conversion of NH4 to N2 in highly anaerobic conditions (comparing NC results to all results). The correlation of fluxes was significant at the 2 % level (t = 3.76). A linear relationship can be inferred from the data:

$$ F_{{{\text{N}}_{2} }} = B F_{{{\text{CH}}_{4} }} $$
(5)

with F i the flux of compounds i in kg ha−1 d−1 and B an empirical coefficient. Based on simple linear regression, the value of B = 0.20 is found because of the similar compositions of gas from individual systems.

Fig. 6
figure 6

Average annual dinitrogen production due to anaerobic decomposition in relation to methane production at six farms in North Carolina, 15 in the Central Great Basin (CGB, all data points from the CGB are the average of three identical farms), and one in Georgia (from Harper et al. 2000). The two data points from CGB –F(05) are from Weaver et al. 2012. The Sow farm in the CGB was not used in the relationship since the animal and waste management systems were very different (see X data point)

Full size image

The S farms were not included in the relationship (see X data point) since the animal size and management, feed input, and manure and urine management were very different.

Gas emissions will vary with respect to farm management (feed input, animal weight, etc.) and climatic conditions. As such, it is difficult to directly compare emissions from different locations. Farm management factors most correlated (and data most likely available) are feed input and size of animal. The climatic factor which most affects the biological decomposition of sludge is the temperature of the biological material in the lagoon anaerobic layer (i.e., sludge temperature, see Fig. 3a). Measurements were used from all the farm systems in the CGB to correct for temperature effects by correlating monthly air temperature with monthly gas emissions as discussed previously. The dependence upon feed input per surface area was estimated from NC data where there was no significant temperature difference between farms. Annual CH4, N2, and CO2 emissions (kg gas component ha−1 d−1) were estimated from lagoons by the following relationships:

$$ {\text{CH}}_{ 4} = \, \left( {0.0 2 3 \times{\text{ FIS }} - { 25}} \right) \, \times \, \left( {0.0 3 9 { } \times {\text{ T}}_{\text{a}} + \, 0. 2 6} \right) $$
(6)
$$ {\text{N}}_{ 2} = \, \left( {0.00 3 9 \times{\text{ FIS }} + { 1}. 3} \right) \, \times \, \left( {0.0 3 3 { } \times {\text{ T}}_{\text{a}} + \, 0. 4 1} \right) $$
(7)
$$ {\text{CO}}_{ 2} = \, \left( {0.00 2 7 \times{\text{ FIS }} - { 7}. 4} \right) \times (0.0 40 \times {\text{T}}_{\text{a}} + \, 0. 2 4) $$
(8)

where FIS is the annual average daily feed input per lagoon surface area (kg feed d−1 ha−1) and Ta is the average annual air temperature (°C) at the site. Temperature corrections were standardized to the average annual air temperature in the NC studies (18.85 °C). When these relationships were used to estimate CGB gas emissions, estimated CH4 emissions were 74 ± 24 % high, CO2 emissions were 58 ± 13 % low, and N2 emissions were 49 ± 42 % high compared to measured emissions.

Using information on the variables measured, we analyzed the data to determine the variables most related (and possibly causal), not already mentioned, to the conversion of NH4 + to N2 for the studies in NC. The amount of feed per average animal weight (and C input) had the highest correlation with N2 emissions (R2 = 0.87). This is not surprising as feed per animal correlates highly with C and N lagoon input (and consequent increased methanogenesis), along with NH4 conversion to N2, across studies over three states (Fig. 6).

Conclusions

In summary, gas emissions were measured in six anaerobic, manure-processing swine lagoons across NC, 15 in the CGB, and one in GA. Conversion of NH4 + to N2 was observed in all lagoons and a correlation was found between methanogenesis (CH4 emissions) and conversion of ammoniacal N to benign N2 gas. Anaerobic digestion not only decomposes organic C to CH4, but also organic N to NH4 conceptually leading to an increase in NH4 concentration and, as a consequence, a potential increase in NH3 emissions. However, we find in these studies that a reduction of C causes an increase in NH3 emissions, rather than a decrease, since NH4 is not converted to N2. Dinitrogen emissions were seen to linearly increase with methanogenesis (CH4 production), further explaining why removal of organic material from lagoons for biogas production would increase NH3 emissions from lagoons, a phenomenon which has been seen in other studies (Harper et al. 2010; Weaver et al. 2012). A causal effect for the relationship between methanogenesis and the potential conversion of NH4 + to N2 is explained based on thermodynamics. Dinitrogen emissions can be estimated across all regions utilizing CH4 emissions (if available). The highest correlation between normally-obtained management variables and N2 emissions was input-feed per animal-weight which provides the organic C for methanogenesis. Simple statistical regression models including average annual feed input and annual average air temperature were developed which explained most of the N2 emissions variability and had an acceptable error when tested against other lagoons. These studies provide the capability to estimate farm lagoon CH4, CO2 and N2 emissions from normally-available farm input and local climate data. Further investigations into the mechanisms of NH4 + to N2 conversion and into the variability of CH4 emissions are needed.