Greenhouse Gas Emissions Non-Cattle Confinement Buildings: Monitoring, Emission Factors and Mitigation

Worldwide environmental issues are dominated by climate change, especially by the increase in greenhouse gas (GHG) emissions (UNDP, 2007). The rise in of GHG concentrations in the atmosphere has become a major environmental concern as revealed in the Kyoto Protocol (AAFC, 2000). Besides contributing to global warming by absorbing infrared radiation, carbon dioxide (CO2), methane (CH4) and nitrous oxide (N2O) have been declared the most harmful gases for ecosystems, apart from ammonia (NH3) (Pain, 1998; Copeland, 2009).

The majority of the pig and poultry operations are mechanically ventilated whereas confined cattle facilities are primarily naturally ventilated. The methodology to estimate gas emissions depends on wind speed, direction, building opening orientation and outsideinside temperature differential.
The inlet and outlet gas concentrations play an important role. Monitoring methodologies differ among countries. The air exchange rate of an animal housing facility must be measured accurately. This would apply to both mechanically ventilated (MV) buildings and naturally ventilated (NV) facilities or a combination of the two ventilation systems. Buildings with combined systems are commonly referred to as hybrid ventilation systems (HV). Many methods have been developed to measure ventilation rates from animal housing facilities.
The following sections will address these topics with a complete description of an experimental setup used by the authors for sampling and analysis of GHG emitted by a number of non-cattle confinement buildings. Also best methods for measuring the air flow rate in both MV and NV barns are suggested.

Wind direction
C in C out T out , RH, P atm N animals , T in

Atmospheric concentration gas analysis
Since the ambient air is used for ventilation of animal buildings, the determination of gas emissions from agricultural activities initially requires a measurement of the ambient air concentration. The contribution of farming systems under study may affect the GHG ambient concentrations near the facility. The measurement of atmospheric concentration need to use equipment having great sensitivity and selectivity like those utilizing optical properties of gas such as Fourier transform infrared spectroscopy (FTIR), photoacoustic spectroscopy (PAS) and non dispersive infrared analyser (NDIR) or separation techniques like chromatography with selective detectors (Neftel et al., 2006).
Since agricultural and forest soils are involved in gas exchange with the atmosphere and many agricultural activities like animal husbandry are significant sources of CH 4 and N 2 O, several studies were conducted to quantify emissions from soils and livestock buildings. In those projects, a number of researchers primarily interested in the characterization of emissions from livestock buildings often work with PAS in the infrared (Blanes-Vidal et al., 2008;Cabaraux et al., 2008;Philippe et al., 2007). While for other authors interested in analytical development or atmospheric flux measurements from soils, separation by chromatography seems to be the preferred means of detection and quantification of trace gases in ambient air (Loftfield et al., 1997;Sitaula et al., 1992;Weiss, 1981;Blackmer & Bremner, 1977). The air is mainly composed of N 2 , oxygen (O 2 ) and argon (Ar) with several others gases in trace concentrations like CO 2 , CH 4 and N 2 O. These components can be separated by chromatography and detected by different detectors more or less specific to the target gas. The technique is simple, proven and allows the simultaneous quantification of CO 2 , CH 4 and N 2 O in the gaseous effluents discharged to the atmosphere. Compared to other techniques having the required sensitivity, like most modern spectroscopic techniques, the chromatography is known to produce reliable results and can be envisaged as a moderate to low cost technique with easy apparatus implementation.
These three GHG are easily separated at low temperature on a column filled with porous polymers Porapak Q or Chromosorb 102 (Cowper & DeRose, 1983). However, the analysis strategy depends on the detectors used and additional gases to be separated and quantified in the sample. Methane can be precisely measured by a flame ionization detector (FID) on a wide range of concentrations ranging from parts per million to volume percent (Cowper & DeRose, 1983) which is suitable for measuring emissions from a livestock building where CH 4 concentrations will range from atmospheric pressure of 1,7 ppmv (Brasseur et al., 1999) to less than 5 000 ppmv for most of the time. To measure atmospheric concentrations of CO 2 , a particular approach should be implemented. The approach involves the reduction of CO 2 to CH 4 with hydrogen over a nickel catalyst and detection by the FID (Cowper & DeRose, 1983).
In order to obtain the required sensitivity for the quantification of N 2 O at concentrations found in ambient air, the electron capture detector (ECD) is commonly used to measure atmospheric concentrations. Some work has been done with N 2 alone as a carrier gas (Jiang et al., 2007;Arnold et al., 2001;Loftfield et al., 1997) while others were performed with an Ar/CH 4 mix (95/5) as carrier gas or as make-up gas to the detector (Jiang et al., 2007;Heinemeyer & Kaiser, 1996;Sitaula et al., 1992;Weiss, 1981;Mosier & Mack, 1980). It was also noted that impurities in a carrier gas can strongly influence the response of the ECD (Phillips et al., 1979) and that the addition of O 2 in the carrier gas increases the sensitivity of ECD to allow the determination of atmospheric concentrations of CO 2 (Cowper & DeRose, 1983). Even if the dynamic range of the ECD is limited, the range of concentration for CO 2 and N 2 O encountered for most confinement buildings are limited. The typical measured concentrations range from atmospheric concentration (360 ppmv for CO 2 and 0,31 ppmv for N 2 O; Brasseur et al., 1999) up to 5 000 ppmv for CO 2 and up to 10 ppmv for N 2 O.

Sampling and management of gas samples
The system developed for the quantification of gas emissions from the agricultural sector has two main functions, first the collection and management of the sample and second the analysis of the sample. Samples are taken sequentially from several sampling points and continuously transported to the analysis system. CO 2 , CH 4 and N 2 O are analyzed with a gas chromatograph (GC). An example of the gas sample collection system is shown in Fig. 3.
For each sample location, gases are pumped through a membrane filter made of polytetrafluoroethylene (50 mm diameter, 0,2 μm pore size) and routed in a Teflon tube (6,4 mm OD, 0.8 mm wall) of variable length depending on the distance between the sampling location and the analysis system. The Teflon tubes are connected to a rotary valve allowing the sequential sampling and analysis of up to 16 locations. A purge diaphragm pump allows The gas flow from the source to the analyser is provided by the main diaphragm pump which delivers the gas to a stainless steel tee fitting. A small filtering sleeve made of sintered stainless steel (7 μm pore size) placed upstream of the main pump provides extra equipment protection against fine dust. The stainless steel tee allows the diaphragm pump to draw a fraction of the sample to flow continuously through the 1 000 µL sample loop of the GC. Two solenoid valves are used to isolate the sample loop in order to balance the pressure of the sample with atmospheric pressure before injection. The sample excess is exhausted to the atmosphere.
The solenoid valve placed between the 16-position rotary valve and the main pump allow for selecting the sample gas analysed from the rotary valve or a selection of calibration gases including ambient air and pressurized gas cylinders controlled by solenoid valves.
All components of the sample collection system that are in contact with the sample gas are either Teflon (ex.: sample tubes), Teflon coated (ex. : pump diaphragms) or stainless steel 316 (ex.: rotary valve, fittings, etc.).
The Teflon sampling tubes from the sampling locations to the analyzer are placed inside a series of ducts maintained by circulating air at approximately 35 °C. The complete system including the 16-position rotary valve, the GC and the pumps are installed in a temperature controlled mobile laboratory. T-type thermocouples are used to monitor the temperature in the heated ducts and the temperature inside and outside the mobile lab. A data logger controlled by the computer of the GC can acquire and archive various parameters measured during periods of analysis. It also controls the electric actuator of the 16-position rotary valve and the various solenoid activated valves of the gas collection system of the samples.

Chromatographic analysis of greenhouse gases
The strategy for the chromatographic analysis is the separation of the three gases on packed columns filled with the porous polymer Porapak Q 80/100 mesh. A pre-column (3,2 mm OD, 1 m long) connected in series before the analytical column (3,2 mm OD, 3 m long) removes some substances that may be present in the gas samples. These substances, which are retained longer than N 2 O in the pre-column, can include water, NH 3 and some sulphur compounds that may have adverse effects on the detectors or on the columns.
CH 4 is quantified with a FID, while CO 2 and N 2 O are measured with an ECD. However following the initial set-up of the chromatographic analysis, CO 2 was quantified with the FID after reduction with hydrogen over a nickel catalyst. Subsequently, following the chance observation of the detection of CO 2 by the ECD, the quantification of CO 2 is transferred to the ECD. This minimises the gradual loss of the effectiveness of the nickel catalyst and also allows a better separation of CO 2 and N 2 O which is not affected by the disruption of the baseline caused by the actuation of the detector valve. Figure 4 shows a schematic of the tubing configuration of the GC used for the separation and quantification of the greenhouse gases and Fig. 5 shows a picture of the column arrangement inside the GC oven and a picture of the outside of the GC.
The oven of the GC is maintained at 60 °C for the duration of the analysis. The 10-port injection valve and the 6-port detector valve are mounted on top of the GC. The zero grade nitrogen is used as carrier gas and is introduced at the three entry points of the pneumatic system of the GC at an equal flow rate of 25 ml min -1 . The two detectors are maintained at 325 °C and the FID is supplied with 30 ml min -1 of hydrogen UHP and with 300 ml min -1 of zero grade air produced by a commercial generator. All the necessary tubing, fittings and valves installed in the GC are made of stainless steel.
The sequence of the analysis begins when the injection valve is actuated to allow the carrier gas to flow through the sample loop and thus transfer the sample gas into the pre-column and the analytical column. After elution of the N 2 O from the pre-column, the injection valve returns to its original position to allow the back flush of the pre-column and further elution of the analytical column. After detection of CH 4 on the FID, the detector valve is actuated to allow quantification of the CO 2 and N 2 O on the ECD. The retention times are 2,9 min for CO 2 , 1,6 min for CH 4 and 3,8 min for N 2 O. Examples of chromatograms obtained for analysis of standard calibration gases and ambient air are presented in Figs. 6 and 7, respectively. With an analysis time of 5 min for the chromatographic analysis and a turnover of less than 7 min between analyses, the chromatographic system allows continuous acquisition of sufficient data to adequately describe agricultural process.
Normal operation of the GC is provided in part by the control software for the electrical parameters and also by periodic checks of the pressure on low pressure gauges on the www.intechopen.com cylinder gas regulators and on the pressure gauges mounted on three controls of the carrier gas admission in the GC which operate near 234 kPa.

Quality control of the analyses
Chromatographic instrumentation analysis is usually calibrated at the beginning of each period of analysis using standard calibration gases qualified as "Certified Standards". The three GHG are available mixed and diluted with N 2 in a single cylinder and at concentrations similar to those expected in real samples. The response factors of the chromatographic analysis are calculated with a single point calibration for each analyzed gas since the response of the detectors used are linear over the concentration ranges encountered.
To document the long-term performance of the overall system and to control the quality of the data obtained, standard analysis are performed automatically at a specified frequency. The samples used are the ambient air in the mobile laboratory and the standard gases from cylinders used for calibration of the GC. The curves showing the responses of the system over time are used to observe and confirm the periods of normal operation of the system and other statistical calculations on the results are used to estimate the overall accuracy of analysis including management and quantification of the sample. To determine the gas emission rates of an animal housing facility, the air exchange rate of the facility must be measured accurately. In northern climates, HV buildings have become more common since buildings relying on NV during cold weather conditions develop numerous problems. These problems include poor inlet air distribution into the building, poor control of the exchange air within the building, uneven temperature and air velocity distribution throughout the building. Therefore, HV buildings rely on MV during cold weather conditions and on NV during spring, summer and fall weather conditions.
Fan airflow rates can be evaluated using a standardized ventilation conduit developed using the standard ANSI/ASHRAE 41.2-1987 (RA 92) (ASHRAE, 1992). The static pressure difference between the interior of the room and the outside of the building as well as the fan rotational speed of each ventilation stage should be continuously measured during the trials. During measurements within the duct, conditions occurring during each trial are recreated (static pressure and rotation speeds of each stage of ventilation). Collected data make it possible to calculate regression equations for each fan predicting air flow rate based on room static pressure and fan rotation speed. Regression equations are calculated from the supplier information .
For MV buildings, the ventilation related data collected are: daily number of animals and their mass, hourly mean exhaust fan rotational speeds, building static pressures and hourly temperatures inside and outside the building.
Ventilation rates can be predicted using a CO 2 balance. Using data monitored in that barn, the animal CO 2 production values can be calculated from published data like, e.g., values suggested by CIGR (2002).

Natural and hybrid ventilated barn
Determining the ventilation rate of NV or HV buildings remains challenging. With ventilation controls becoming more sophisticated, side curtain and ridge vent openings can be controlled more precisely based on anticipatory logic used to determine upcoming temperatures and wind conditions. Traditionally, NV buildings have only been used for larger animals where changes in the thermal environment are not as critical. With improved controller logic, NV buildings will soon be used more extensively for housing smaller animals. Consequently, the ability to predict airflow in these facilities is very important in controlling the thermal and non-thermal well-being of the confined animals.
A number of researchers have proposed methods to calculate the air exchange rates in NV buildings. These airflow rates are dependent on wind speed at the opening and effectiveness of the opening. Morsing et al. (2002) and Choinière et al. (1988) reported coefficients to be used for agricultural buildings to calculate the wind speed at the opening. Nääs et al. (1998) and Choinière et al. (1988) described an algorithm to determine an opening effectiveness relative to the wind direction. This methodology is described as the ventilation rate due to wind and thermal buoyancy.
For HV buildings, the ventilation related data collected included: daily number of animals and their mass, hourly wind speed and direction, curtain opening areas, operating status of the exhaust fans and hourly temperatures inside and outside the building. As for MV buildings, ventilation rates from NV and HV barns can be predicted by establishing a CO 2 balance. In order to be representative, the CO 2 concentrations should be measured at least at three locations: on each side of the barn close to the curtains and in the center of the barn at the animal level. For HV barns, the CO 2 concentrations should also be measured close to the fans.

Swine production
The literature identifies emission factors from the three types of swine confinement buildings ( Table 2). The swine production system begins with the maternity stage comprising both of gestating sows and farrowing sows with their piglets. Gestating sows are usually reared either in individual stalls or in group pens. Farrowing sows and their piglets are mainly kept in farrowing crates. The swine nursery or post-weaning building rears the weaning piglets brought from the maternity. Piglets remain there until they reach a certain weight, generally between 20 and 25 kg. Feed and ambient conditions are adjusted with pig growth. Then, pigs are transferred to grower/finisher facilities until they leave for slaughter.  Overall, sows produce more CO 2 on an animal basis (5,29 kg d -1 animal -1 ) than weanling piglets (0,55 kg d -1 animal -1 ) or grower/finisher pigs (1,92 kg d -1 animal -1 ) since the CO 2 production increases as the animal weight grows. In fact, the emission ratios between two growing stages correspond approximately to the animal unit ratios.
In the same way, the greater amount of urine and feces excreted by sows favours the establishment of anaerobic conditions and the CH 4 emission (30,1 g d -1 animal -1 ) in comparison with the offspring (2,77 and 5,54 g d -1 animal -1 , respectively for weanling piglets and grower/finisher pigs).
N 2 O emissions from maternity and nursery were relatively close to zero as found in several studies. Grower/finisher pigs emit 0,66 g N 2 O d -1 animal -1 . The non frequent change in protein requirements during the growth stage leads to more N being excreted.

In animal production
Several technologies have been developed to reduce gas and odour emissions from livestock housing. Several of these techniques, including reduction efficiencies for each technology, were inventoried by Godbout et al. (2010) from an exhaustive literature review. The inventory revealed that progress has been made in reducing odour and ammonia emissions, while less concern has been placed on GHG emissions. Three distinct techniques have been recognized, namely under slat separation, air cleaning and nutrient management.

In-barn manure management: under-slat separation
Separating urine and feces beneath the slats and removing both the solid and liquid fractions frequently is a reliable manure management technique to reduce gas and odour emissions from buildings (Andersson, 1995;Arogo et al., 2001;Bernard et al., 2003;Jongebreur, 1981). The aim of this technique is to separate solid and liquid phases of the manure immediately after it falls through the slats to reduce the contact time between both phases. Three major under-slat manure separation systems have been studied over the years: the conveyor net, the V-shape scraper and the conveyor belt.
The conveyor net is composed of a mesh, tensioned under the slats, through which urine can flow while feces are collected. When this separation system is mechanized, the conveyor scrapes the feces to the end of the building while the urine stays in a gutter that is sloped toward a conventional storage pit (Fig. 8a).
(a) (b) (c) Fig. 8. a) Conveyor net ; b) V-shape scraper ; c) Diagram of the conveyor belt (adapted from Lemay et al., 2007) www.intechopen.com In V-shaped scrapers (Fig. 8b), the feces stay on inclined gutter walls while urine is gathered into the bottom of the gutter and continuously drains out of the room by gravity. The solid fraction remaining on the inclined walls is scraped at a certain frequency using mechanically driven scrapers. Godbout et al. (2010) measured CO 2 and CH 4 emissions from a V-shaped scraper and compared these to a pull-plug system (emptied every week). However, even if emission reductions were observed, there were no statistical differences.
Conveyor belts (Fig. 8c) were adapted from poultry to swine production, placing the belt at an angle under the slatted portion of the pens. Its lower edge feeds into a pipe that collects the urine and transports it to the end of the building, thus allowing the separate collection of urine and feces within the hog house (van Kempen et al., 2003). Koger et al. (2003) founded that CH 4 emission from a belt-based housing were reduced between 52 and 83% throughout the grower period studied comparatively to conventional pig houses.
Most of studies evaluating these techniques have been mainly conducted in order to measure the reduction of NH 3 emissions (Voermans and van Asseldonk, 1990;Voermans and van Poppel, 1993;Hendriks and van de Weerdhof, 1999). The study carried out by  found 13% and 19% CO 2 and CH 4 emission reduction, respectively, from a conveyor net compared to a drainage system without separation (emptied once a week). The same gas reduction values could be expected from the other separation techniques since they use the same principle.

Air cleaning
Air cleaning techniques for livestock buildings have the potential to improve air quality. However, efforts have been focused for improving performances on the reduction of dust and the abatement of NH 3 and H 2 S. The air cleaning techniques are classified into two broad categories, physicochemical treatment and biological treatment .
The physical -chemical absorption (scrubbing) is the physicochemical method most widely used for the treatment of air. This is a technology developed for many industrial applications. Gases are absorbed when the air from the barn is in contact with a liquid in which gas become soluble within the solution. The mass transfer from gas to the liquid is achieved by using a filter material within the filter unit (Devinny et al., 1999). The filter material usually has a large porosity, or void volume, and a large specific area (Melse & Ogink, 2005). Water is often used as the liquid solvent and its pH can be adjusted (basic or acid) depending on the pollutant to increase the solubility of gases. Contaminated air is introduced, either horizontally (crosscurrent) or upwards (counter-current), resulting in good contact between air and water, and enabling mass transfer from gas to liquid phase. A fraction of the trickling water is continuously recirculated; another fraction is discharged and replaced by fresh water (Fig. 9) (Melse & Ogink, 2005).
On the other hand, a biological treatment of air is based on the capacity of microorganisms to transform organic and inorganic pollutants into non-toxic compounds and odour free (Devinny et al., 1999;Hartung et al., 2001;Revah & Morgan-Segastume, 2005). Three main types of bioreactors are currently used: biofilters, biotrickling filters and bioscrubbers (Fig.  10). The basic mechanism is the same for all biological treatment systems; the difference is due to the equipment configuration to carry out the transfer between the gas and the liquid, and on the pollutant biodegradation process (Table 4) (Devinny et al. 1999;Revah & Morgan-Segastume, 2005). Removal efficiency for NH 3 and H 2 S emissions with a biological treatment can range from 6 to 100% and 3 to 99%, respectively (Nicolai & Janni 2001;Armeen, 2008;Iranpour et al., 2005). The reduction of odour emission is also widely variable, going from 29 to 100% depending of the operation conditions (Luo, 2001). A first bioreactor prototype developed by Belzile et al. (2010) found that NH 3 emissions from small-scale swine chambers were reduced by 62 to 91% and H 2 S emissions were decreased by 24 to 66% by the biological treatment compared to a drainage system without separation (emptied once a week). However no significant reduction was obtained for CO 2 and CH 4 emissions at this first stage. Fig. 9. A counter-current air scrubber (adapted from Melse & Ogink, 2005) www.intechopen.com The discharged water from a scrubber might be used as nitrogen fertilizer for crops; sometimes the water is added to the liquid manure storage (Melse et al., 2009). The discharge water from a biotrickling filter might be treated in a denitrification process in order to decrease the nitrogen content (Melse et al., 2009;Sakuma et al., 2008

Nutrient management
An additional method to reduce emissions caused by excess nitrogen is the alteration of the ratio of nitrogen excretion in urine versus feces by nutrient management (Mroz et al., 1993). Reduction of dietary protein combined with supplementation of synthetic amino acids in pig diets might reduce total nitrogen excretion by 25 to 40% (Hartung & Phillips, 1994;Kay & Lee, 1997). Additionally, the inclusion of fermentable carbohydrates or non-starch polysaccharides into diets stimulates bacterial fermentation in the hindgut and reduced urinary versus fecal nitrogen ratio by 68% (Canh et al., 1997a).
However, GHG emission reduction is generally not measured or documented when using nutrient management. Principally, studies target reducing NH 3 and other odorant compound emissions (Garry et al., 2007;Le et al., 2006;Lyngbye et al., 2006). GHG emissions measurements (CO 2 , CH 4 and N 2 O) were carried out by Godbout et al. (2010) when protein content is reduced and lysine is increased in the diet. As a result, such diet treatment presented no impact on CO 2 and N 2 O emissions while CH 4 emissions increased by 58% compared to a commercial diet. Therefore, a more thorough analysis should be carried out for a better understanding of dietary management in GHG emission reduction.

Summary and conclusions
Contaminants exhausted from confined animal buildings include various gases, dust particles, micro-organisms and odours. The most important gases are CO 2 , NH 3 , H 2 S, CH 4 , N 2 O and some trace gases (aldehydes, amines, aromatics, organic acids, sulphur compounds, etc.). The main GHG emitted from livestock building are CH 4 , N 2 O (from manure decomposition) and CO 2 (from animal metabolism). Generally, CH 4 emissions are more present in liquid manure management while N 2 O is produced under solid manure management.
The emission is the product of the gas concentration and the air exchange rate. An accurate measurement of these values is very important and is still a challenge today for emissions from agricultural sources. Since the agricultural emissions are very low, the concentration measurement requires equipment having great sensitivity and selectivity like those utilizing optical properties of gas such as FTIR, PAS and NDIR or separation techniques like chromatography with selective detectors. Most of the gases in air which are mainly composed of N 2 , O 2 and Ar with several others gases in trace concentrations like CO 2 , CH 4 and N 2 O can be separated by chromatography and detected by different detectors more or less specific to the target gas. The technique is simple, proven and allows the simultaneous quantification of CO 2 , CH 4 and N 2 O. Compared to other techniques having the required sensitivity, like modern spectroscopic techniques, gas chromatography is known to produce reliable results and can be envisaged as moderate to low cost techniques with easy apparatus implementation.
The air flow measurement is very important and often, a lot of uncertainty is related to this value bringing an error in the emission determination. The measurement techniques are function of the ventilation system. Three main systems exist: MV, NV and HV buildings. Various methods have been developed to measure ventilation rate from animal housing facilities including airborne tracer techniques, diffusion of animal-produced CO 2 or heat, vane anemometers and orifice plates. Each of these methods, however, has limitations.
In the case where is not possible to measure emissions, values from literature can be used for swine, broiler and layer productions. The typical CO 2 emissions for sows, weanling piglets and grower/finisher pigs are 5,29, 0,55 and 1,92 kg d -1 animal -1 , respectively. The greater amount of urine and faeces excreted by sows favours the establishment of anaerobic conditions and the CH 4 emissions (30,1 g d -1 animal -1 ) in comparison with the offspring (2,77 and 5,54 g d -1 animal -1 , respectively for weanling piglets and grower/finisher pigs). N 2 O emissions from maternity and nursery were relatively close to zero as found in several studies. Grower / finisher pigs emit 0,66 g N 2 O d -1 animal -1 . Broiler and layer productions emit similar quantities of CO 2 to the atmosphere (31,5 and 28,2 kg yr -1 animal -1 , respectively). However, the different layer systems using liquid manure management generate a greater emission factor for CH 4 (44,7 g yr -1 animal -1 ) comparatively to broiler systems with litter (12,3 g yr -1 animal -1 ). The N 2 O emissions range from 10,9 to 17,6 g yr -1 head -1 . However, since the gas emissions are influenced by many factors, on-site measurement should be privileged instead of typical values from literature to determine the typical emissions from building.
Several technologies have been developed to reduce odour and gas emissions from swine housing. Two in-barn approaches are encouraging: the under slat separation system and diet manipulation while air cleaning systems have been developed for the exhaust air outlet.
In agreement with the literature, the under slat separation system can reduce around 13% and 19% of CO 2 and CH 4 emissions, respectively. No studies have reported GHG emission reductions using diet manipulation. Many types of air cleaning systems already exist and they have been developed mainly for odour and ammonia emission reductions and the effect on GHG is not clearly shown.