This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
Agriculture and livestock production are key economic, social and environmental activities in the EU and the global context. Intensified agricultural practises are coupled with emissions of nitrogen containing compounds which pose problems of environmental pollution. Pollution from these activities is transboundary in nature and with growing awareness of the finite nature of global resources featuring prominently in legislation and the social consciousnesses addressing the best available techniques for agriculture that are environmentally less damaging is a current and important issue. Ammonia (NH3) is a particularly important trace gas associated with agriculture (Aneja et. al., 2008). The release of NH3 from agricultural practises contributes the harmful effects of high nitrogen loads (Galloway et. al., 2003; Erisman et. al., 2007) which relates to environmental damage to aquatic ecosystems (Kirchner et. al., 1999) soil acidification (Burton and Turner, 2003), woodland (Pitcairn et. al., 1998), air pollution and negative health impacts for humans (Dockery and Pope, 1994) and animal species (Flint et. al., 2000). Monitoring ammonia emissions from agriculture is an important part in mitigating the negative environmental and health effects associated with ammonia volatization. Menzi et. al. (2006) note the underlying processes of NH3 emission are an important research topic due to increasing pressure on agriculture to reduce emissions. Contributing to the auspices of the BAT Farm project this thesis aims to add to existing knowledge relating to ammonia measurement and analysis. BAT Farm is a joint project of Glasgow Caledonian University in Scotland, Teagasc the Irish Agriculture and Food Development Agency, IRSTEA the French National Research Institute of Science and Technology for Environment and Agriculture, INTIA and Neiker a public institute for agricultural research, which is also Spanish, is the lead partner. The project is international in nature as can be seen from the diverse range of partners involved, the over arching aim of the project is to determine a scientifically based protocol for evaluating realistic options for the abatement of environmental impacts in air and water resources and from livestock. BAT stands for best available technique and the name symbolises the work the project is undertaking in evaluating and compiling available agricultural techniques. This thesis is a contribution to the project as it evaluates the best available technique (BAT) for passive ammonia sampling by comparing samplers based on Ferm (1986) and Scholtens (1996); furthermore this thesis also employs and evaluates the analysis method of using an ion probe for the detection of ammonia as well as and in comparison to the standard colorimetric analysis method.
Ammonia (NH3) is part of the group of pollutants which contribute to transboundary air pollution, acid deposition and nitrogen deposition. Sulphur dioxide (SO2) and nitrogen dioxides (NOx), which are also part of the group of pollutants, produce sulphuric acid and nitric acid respectively in the atmosphere. However, ammonia is a base and as such is the main neutralising compound for these substances producing ammonium sulphates and ammonium nitrates in which form much of the long range transportation of ammonia takes place (Sutton et. al. 2004, 796). In terms of chemistry ammonia is a colourless gas which absorbs radiation in the infrared spectrum. NH3 combines readily with a proton H+ to form ammonium ions (NH4+) (McGinn and Janzen 1997, 140). As can be seen from figure x ammonia that enters the troposphere is transferred by this process from the air into acid cloud droplets, atmospheric removal of NH4+ and NH3 takes place via wet and dry deposition. Close to source dry deposition of NH3 is the most important removal process; however, at greater distances from the source wet removal of particulate NH4+ is the most important (Schjoerring et. al. 1992, 13). Animal manure and urine contain between 1 and 4% ammonium which exists in equilibrium with volatile gaseous ammonia. Stored manure when brought into contact with soil or large volumes of air the equilibrium will displace to the ammonium form (Ferm et. al. 2000, 2).
Figure x. 'Wet and dry deposition of NH3'
Source: Sutton et. al. (2004,796).
Ammonia emission is caused by the bacterial decomposition of undigested nitrogen containing compounds in faeces and urine and urea. Urea is rapidly converted into ammonia as a result of urease activity in faecal microbes (Merino et. al. 2008). Ammonia is emitted from the available total ammonical-Nitrogen (TAN) in the excreta of livestock; the amount of available TAN can be increased by the mineralisation of organic Nitrogen during the management of manure. TAN can also be depleted by gaseous emission, losses to leachate and through immobilisation in litter. A proportion of TAN is lost at each stage of manure management predominantly as NH3 (Webb et. al. 2006, 7223). As such, agriculture dominates the UK ammonia emissions with over 89% of the total emissions coming from agriculture, in Scotland specifically 91% of emissions are attributed to agriculture (MacCarthy et. al. 2011). Slurry storage accounted for 13% of the UK's ammonia emissions in 2009; slurry spreading accounted for a further 22% of the UK's total emissions (Misselbrook et. al. 2010). Dairy cattle contribute to ammonia emissions via the nitrogen content of their excreta; Tamminga (1992) estimates that approximately 80% of a dairy cow's nitrogen intake will be excreted through urine and faeces. This is important in terms of the volatilisation to ammonia. As such the monitoring of ammonia in the agricultural setting and on a dairy farm specifically is an important part of reducing and mitigating harmful emissions.
Ammonia and Slurry specific emissions
The majority of livestock derived ammonia lost to the atmosphere comes from slurry; both spread on fields and stored (McGinn and Janzen 1997, 140). Slurry stored in uncovered tanks can contribute towards atmospheric ammonia, Sommer et. al (1993) have shown that uncovered pig and cattle slurry can emit 3-5g NH3-N m-2 d-1. The ammonia and ammonium ions present in slurry derive from the degradation of urea; urea is readily hydrolysed by urease and converted to ammonium after excretion. Free ammonia is formed when a portion of the ammonium ions dissociate, though the amount of free ammonia ions is dependent on temperature and pH (Burton and Turner 2003, 81-82). Land spreading of slurry and solid manure is the largest source of NH3 emissions in the agricultural setting (Jarvis and Pain, 1990; Stadelmann et. al. 1998: Misselbrook et. al., 2000). Reidy et. al. (2008) note that the application of organic livestock waste on fields accounts for approximately one third to one half of all agricultural NH3 emissions. As the study location is an organic farm measuring ammonia emissions are of particular importance. The volatilization of ammonia from slurry application is depended on varied factors such as; the application technique, the composition of the slurry in terms of consistency and dry matter content, the crop cover and soil type of the area the slurry is being applied to and weather conditions (Sommer and Olesen, 1991, 2000; Bussink et al., 1994; Braschkat et al., 1997). Slurry storage contributes to NH3 and GHG emissions Amon et. al. (2006). The way in which slurry is stored has impacts on emission rates as found by Amon et. al. (2006), uncovered storage tanks emit more than tanks with layers of coverage such as straw or a wooden cover. Sommer et. al. (2006) note that composition of slurry, in terms of dry matter content and soil water content, also have effects on ammonia emissions in the field.
Environmental impacts of Ammonia
Kirchmann et. al. (1998, 1) note that ammonia emissions have direct consequences to terrestrial and aquatic ecosystems. Deposited NH3 and NH4+ may be oxidised to form nitrate in the soil which acidifies the soil. Potential acid deposition in ecosystems is in the form of NH3 or NH4+ with NH4+ acting as a vector for acidity from SO2 and NOx also. Furthermore deposition of NH3 and NH4+ in soil leads to further emission of the greenhouse gas nitrous oxide (N2O) (Sutton et. al. 2004, 796-797). In aquatic ecosystems, both fresh water and marine, excesses of NH3 and NH4+ cause eutrophication which contributes to decreased biodiversity (Kirchmann et. al. 1998, 1). Ammonia is toxic to fish species with salmon being particularly sensitive, which has economic as well as biodiversity implications in Scotland. Burton and Turner (2003) note that pollution incidents related to slurry run off are usually visible by high levels of deceased fish. In Scotland ecosystems such as bogs, moorlands and mountain ecosystems are vulnerable to the addition of nitrogen via NH3 and NH4+ deposition as these types of ecosystems have adapted to low nutrient availability. Ground fauna of woodland ecosystems have also been affected through this atmospheric deposition (Sutton et. al. 2004, 797). Impacts on these varied ecosystems through these changes to nutrient cycling have further impacts on biodiversity and bird and animal species through affects to habitats.
Impacts on Agriculture related to Ammonia
In agricultural situations ammonia volatization reduces the nitrogen (N) efficiency of animal slurry as fertiliser and also decreases the degree of certainty in predictions of availability of manure N to crops (Sommer et. al. 2006, 229). This affects the ability of farming practitioners to adequately anticipate amounts of fertiliser needed for crops which can result in the need for extra applications. This has impacts on farming schedules as well as costs related to re-application and time costs of employees. As ammonia is highly water soluble it will to a substantial degree be deposited in close proximity to the source (Ferm et. al. 2000, 2). As such ammonia storage on farms can lead to deposition on farm locations which can cause local toxic effects damaging surrounding vegetation (Kirchmann et. al. 1998, 1). This can have detrimental effects on animal health if toxic vegetation is ingested. Indeed, Flint et. al. (2000) note that animal health may also be affected by problems of air quality resultant from ammonia emissions due to stress which will have financial implications for farmers in veterinary costs. The health of agricultural workers may also be impacted upon by poor air quality arising from ammonia emissions which raises questions of health and safety at work (Flint et. al. 2000, 389).
Legislation relating to Ammonia
In 1979 the United Nations Economic Commission for Europe (UNECE) introduced the Convention on Long-Range Transboundary Air Pollution (CLRTAP) in response to the transboundary environmental issue of acid rain. CLRTAP initially focused on sulphur dioxide (SO2) emissions leading to the need for wider focus resulting in the Gothenburg Protocol in 1999. The Gothenburg Protocol to Abate Acidification, Eutrophication and Ground-level Ozone set national emission ceilings for SO2, NOx, NH3 and volatile organic compounds (VOCs) which had to be met by 2010 onwards. The National Emissions Ceiling Directive 2001 (2001/81/EC) also set emission limits in European law. The Emission Ceilings Directive 2001 came into UK law as the National Emission Ceilings Regulations in 2002 (DEFRA 2011).
The EU Directive on Integrated Pollution Prevention and Control (IPPC) includes the application of Best Available Techniques (BAT) to reduce NH3 emissions from large poultry and pig farms (Sutton et. al. 1998, 807). Furthermore, slurry stores require a rigid cover and lagoons a floating cover from 2007 onwards under IPPC (Misselbrook et. al. 2010, 9). The use of passive samplers for ammonia measurement is supported by the European Union Council Directive 96/62/EC, 27th September 1996, resulting from the framework for preliminary air quality assessment, optimisation of station siting, measurement generalisation and evaluation relating to Air Quality Assessment and Management (Kirchner et. al 1999, 256).
The Thematic Strategy on Air Pollution (TSAP) 2005 expresses environmental objectives for 2020 based on 2000 levels. The objectives aim to reduce the area of forest ecosystems where acid deposition exceeds critical loads for acidification by 74%, and reduce ecosystems area where nitrogen deposition exceeds the critical loads for eutrophication by 43%. Furthermore, TSAP sets the target to reduce live years lost from particulate matter (YOLLs) at 47% compared to the assessment for 2000. Achieving these goals requires significant further emissions reductions of various air pollutants. It has been estimated that emissions of ammonia 80-95% which typically originates from agricultural sources would need to be reduced by 27% in the EU25 by 2020 compared to 2000 levels. Since prior legislation and current projections of livestock change can only account for 4% of the decline, the remaining 23% will have to be met by introduction of specific abatement measures in agriculture (Spranger et. al. 2009).
The Nitrates Directive 1991 (91/676/EEC) aims to protect water quality across Europe from nitrates from agriculture polluting ground and surface waters by promoting good agricultural practices. Good agricultural practice consists of; limits on when fertilizers can be applied to soil in order to allow nitrogen availability only when crops need it, limits on the conditions under which fertilizer can be applied, requirements for minimum storage capacity for livestock manure and recommendations for crop rotation, catch crops and soil winter cover for the prevention of leaching during wet seasons (European Commission 2012).
Aims and Objectives
This projects aims to compare the Scholtens (1996) method of passive ammonia sampling to the older Ferm (1986) method using a real farm setting. It is the aim of this work to use both types of passive sampler to monitor ammonia emissions from a slurry tank containing excreta from dairy cows and to monitor ammonia emissions after this slurry has been spread on a field. The analysis of the ammonia emissions also aims to be two fold; using the commonly used method of colorimetric analysis and the less common ion probe method. This analysis also has the aim of giving comparability between the techniques.
1. Develop a test rig for passive sampling in a real farm setting
2. Create sampling protocol to monitor ammonia emissions using Ferm (1986) and Scholtens (1996) samplers
3. Set up laboratory procedure to analyse sampler's ammonia content
4. Compare results of ammonia collected by each sampler type
5. Compare analysis methods for determining ammonia content of samplers
6. Determine best available technique
Scope of Study
This project seeks to compare passive sampler types between Ferm (1986) and Scholtens (1996), as such, this study does not consider other types of sampler. This study is locationally fixed to Rainton Farm in Dumfries and as such only considers the ammonia emissions of a dairy farm as opposed to other types of farm. However, this study contributes to the wider literature relating to passive sampling as the samplers used can be utilised to monitor gaseous emissions other than ammonia. The samplers can also be utilised in varied other locations and on different types of farm. This study can then be utilised as a basis for future work relating to gas sampling in the agricultural setting in the global context. This study does not consider the climatic conditions that took place during the field experiment, these conditions could be observed for future work as an explanatory tool for the results obtained; however, this study focuses on comparing the results obtained by using Ferm (1986) samplers in comparison to Scholtens (1996) rather than accounting for results. This study does not employ any other sampler type to measure ammonia emissions this study focuses on the comparison of low cost samplers, as such Leuning et. al. (1985) 'shuttles' are not considered, neither are more expensive computer based methods.
Atmospheric emissions of ammonia can have negative effects on ecosystems both natural and managed at the local and global level; as such measurement of ammonia has an important role to play in ammonia abatement strategies. As livestock production contributes highly to ammonia emissions it is imperative that monitoring takes place at all levels of production. This thesis is primarily focused on the manure storage and spreading stages of livestock, due to the high level of emissions associated with these stages, and the best way to passively sample ammonia during these stages. Monitoring of ammonia emissions must be simple and cost effective to encourage more farms to implement such strategies; as such this thesis compares two low cost samplers that require little training to operate. This thesis is a contribution to the EU funded BAT Farm project as it evaluates the best available technique for passive ammonia sampling by comparing the Ferm (1986) method and the Scholtens (1996) method which claims to have advantages over Ferm (1986); furthermore this thesis also employs and evaluates the analysis method of using an ion probe for the detection of ammonia as well as and in comparison to the standard colorimetric analysis method. BAT Farm is a joint project of Glasgow Caledonian University in Scotland, Teagasc the Irish Agriculture and Food Development Agency, IRSTEA the French National Research Institute of Science and Technology for Environment and Agriculture, INTIA and Neiker a public institute for agricultural research, which is also Spanish, is the lead partner. The project is international in nature as can be seen from the diverse range of partners involved, the over arching aim of the project is to determine a scientifically based protocol for evaluating realistic options for the abatement of environmental impacts in air and water resources and from livestock.
A passive sampler is a device that is capable of taking a gas or vapour sample from the atmosphere at a rate controlled by the physical process of gaseous diffusion through the device via a barrier such as a static air layer, porous material or permeation through a membrane. However, for the device to be passive it cannot involve active movement of air through the device such as an air pump. The gas molecules are transported through the sampler by diffusion as a result of air pressure and air temperature. The flux is achieved by the placement of an efficient sorbent for the target gas behind the barrier, the difference between the ambient concentration and the sorbent concentration compared to the ambient concentration is the driving force. The average next flux of the target gas is obtained by analysis of the sorbent (Carmichael et. al., 2003, 1295). The flow rate through the sampler is calculated using the equation:
mNH3 = DNH3 x A
where mNH3 is the passive sampler flow rate measured in cm3 min-1, DNH3 is the gas diffusion coefficient measured in cm2 min-1, A is the effective cross sectional area of the sampler measured in cm2 and Δx is the diffusion resistance of the sampler measured in cm. The ambient concentration of ammonia captured by the sampler depends upon the mass of ammonia captured by the sorbent, the duration of exposure to the gas target and the diffusion coefficient. The mass of ammonia adsorbed by the sorbent (QNH3) is:
QNH3 = (ce - cb)V
where ce is the filter extract coefficient of ammonium measured in μg mL-1, cbis the extract concentration of an exposed blank sampler also measured in μg mL-1 and V is the volume of the extract measured in mL (Puchalski et. al. 2011, 3158).
Ammonia can be measured by passive samplers because the infrared absorption and acid-base properties of NH3 can be utilised for detection purposes. To measure atmospheric NH3 an acid medium is used and the selective, cumulative absorption is measured (McGinn and Janzen 1997, 140). Ammonia and Ammonium readily react with acidic media; hydrochloric acid (Janzen and McGinn 1991), sulphuric acid (Harper et. al. 1983: Sherlock et. al. 1989), boric acid (Kissel et. al. 1977), phosphoric acid (Bussink 1994) and citric acid (Fehsenfeld et. al. 2002). Due to this passive samplers generally feature an acid base to capture atmospheric ammonia.
Types of passive samplers
The techniques for monitoring ammonia emissions fall into two main categories; those that use an acid base to collect ammonia and those that utilise sensors to measure changes in wind direction or concentration (Phillips et. al. (2001, 2). Ferm (1986) and Scholtens (1996) samplers fall into the category of those that use an acid base to monitor ammonia. However Leuning et. al. (1985) constructed and tested a passive ammonia sampler prior to Ferm (1986). Schjoerring et. al. (1992) and Hansen et. al (1998) have used passive samplers based on the Ferm (1986) design rather than adopting the Scholtens (1996) design. Detector tubes (Dräger, 1997), denuders (Ferm, 1979), and 'badge' type samplers (Willems?), chemiluminescence detectors utilise sensors comprising the second category of ammonia emissions samplers. This thesis is concerned with the lower tech options for ammonia sampling and as such will only focus on those options.
Leuning 'shuttles' (1985)
Leuning et. al. (1985) constructed a passive sampler as shown in figure x. The sampler is referred to as a Leuning 'shuttle' due to the shape; it is this shape with its vanes and pivots to keep it pointing upwind that allows the sampler to respond to changes in wind direction. The sampler utilises an oxalic acid base to capture ammonia as it flows through the device. Once the ammonia has been captured the oxalic acid coating is extracted with water and can be analysed using the standard colorimetric method as well as ion chromatography or specific electrode. Phillips et. al. (2001) note that the Leuning et. al. (1985) sampler is limited in that it has a high initial purchase cost, and due to its design is more suited to monitoring ammonia from plots as opposed to slurry stores or livestock buildings.
Figure x. Leuning 'shuttle' schematic
Source: Leuning et. al. (1985, 1118).
Sherlock et. al. (1989) conducted an experiment to test the Leuning et. al. (1985) sampler. The test was designed to check the accuracy of the sampler in measuring ammonia loss from fertilised plots of land. It was found that the sampler functioned adequately in comparison to micrometeorological acid bubbler method used to measure ammonia concentration and cup anemometers used for wind speed. Woods et. al. (2000) also tested the Leuning et. al. (1985) sampler against one based on Schjoerring et. al. (1992), which is based on Ferm (1986), in a farm setting. Like Phillips et. al. (2001) Woods et. al. (2000) noted the high cost associated with construction and usage of Leuning 'shuttles'. Also, Woods et. al. (2000) found that the Ferm (1986) type sampler functioned better than the Leuning et. al. (1985) sampler in terms of accuracy and ease of use.
Ferm 'tubes' (1986)
The Ferm (1986) passive sampler measures horizontal ammonia flux. The sampler consists of two glass tubes, as shown in figure x, A and B which are referred to in this study as Ferm In and Ferm Out for A1 B1 and A2 B2 respectively. One end of each sampler in the pair is fitted with a steel disc with a 1mm diameter hole in the centre to restrict air flow through the tube. The tubes are internally coated with acid for the NH3 to adhere to. The average airflow in the two samplers directed in opposite directions is proportional to the wind speed vector along the tubes, as such; the tube facing the wind collects more of the flux. The emissions from an ammonia source can be obtained using a mass balance approach. The mass balance approach consists of measuring the ammonia transport through four vertical 'walls' which surround the ammonia source. The transport is determined by multiplication of the horizontal flux with the 'wall' area. As the flux varies with height measurement at several heights is needed. The fluxes coming in through the vertical surfaces need to be subtracted from the flux going out; as such the wind direction across the ammonia source need not be constant to yield results (Ferm et. al. 2005, 7108). Flint et. al. (2000) note that Ferm's (1986) original design comprised two sampler tubes because when the incident airflow in a single tube is oblique to the tube inlet then there can be a too rapid or too slow decrease in air speed inside the tube. The decrease in air flow being too fast is resultant from the flow restrictor orifice, the steel disc, and too slow being resultant from the open end of the tube.
Figure x. Ferm (1986) passive sampler
Source: Ferm et. al. (2005, 7107)
The Ferm passive sampler was developed and tested in a wind tunnel under laboratory conditions (Ferm, 1986; Hansen 1997). The sampler was also tested on a plot that provided artificial ammonia emissions (Schojerring et. al., 1992: Sommer et. al., 1996). Phillips et. al. (1998) utilised Ferm (1986) samplers to measure ammonia emissions from livestock housing in the UK. Ferm (1986) passive samplers have also been used in the field in Sweden (Ferm et. al., 2000) and in Poland (Ferm et. al., 2005).
Hansen et. al. (1998) used a passive sampler based on Ferm (1986) which was mounted on a wind vain in order to act in the same fashion as the sampler constructed by Leuning et. al. (1985). Mounting the sampler on a wind vane served to allow the sampler to change direction and thus overcome the limitation of the stationary aspect of Ferm (1986). Phillips et. al. (2001) note that while the sampler used by Hansen et. al. (1998) had the benefit of having a lesser cost and being simpler to operate than Leuning et. al. (1985) it still had the same inherent drawbacks of being less applicable to slurry stores and buildings. Phillips et. al. (2001) have compared Ferm (1986) 'tubes' with other sampling techniques and found that for directly measuring ammonia flux the Ferm (1986) method has the advantages of being simple and cheap to use. However, it was noted by Phillips et. al. (2001) that the Ferm (1986) method required a lengthy period after sampling in the laboratory for analysis. In terms of time a limitation of the Ferm (1986) passive sampler is that two samplers are needed at each sample height; Ferm in and Ferm out. The need for two samplers requires extra time in the preparation and analysis of tubes. Scholtens (1996) devised a sampler that only uses one tube to overcome the limitations of the Ferm method.
Figure x shows the Scholtens (1996) passive sampler. This sampler follows the same principles as the Ferm (1986) sampler but differs in the location of the steel disc to regulate air flow. As can be seen from figure x the steel disc is located in between the two sample chambers so that one tube is needed as opposed to two following Ferm (1986). The main role of locating the steel disc in the centre of the sampler is to regulate the amount of air that flows through it. Locating the disc in the centre of the sampler means that it functions as both sampler and flow meter by measuring the pressure drop across the tubes. The use of two chambers in one tube also means that the sampler can detect the direction of the flux assuming that the sampler is operated below saturation levels of the acid coated inserts inside the chambers (Scholtens et. al. 2003a). Scholtens et. al. (2003a, 97) note that the sampler has four advantages over Ferm (1986); firstly, it only requires one tube for either side of the flux pane as opposed to the two that Ferm (1986) uses. Secondly, the glass fibre inserts used in the tubes mean long term exposure in animal houses is possible. Thirdly, accurate calibrations can be made to the sampler without the presence of an ammonia source and finally designs for application tailored samplers are easily validated.
Figure x. Scholtens (1996) Passive sampler design
Source: Scholtens et. al. (2003a, 96)
The Scholtens (1996) as sampler described by Scholtens et. al. (2003a) was tested by Scholtens et. al. (2003b) under laboratory conditions. Scholtens et. al. (2003b) tested ten variations of the Scholtens (1996) sampler in a wind tunnel; it was found that the original Scholtens (1996) sampler did not function as expected. At expected ideal function the curve of air velocity in the orifice against the angle of incidence would be a pure cosine curve, however, it was found that only a general cosine curve was present. The maximum air velocities in the orifice are not found at 0o or 180o as would be expected, at these angles the velocity was approximately 95% of the maximum values. The maximum values were in fact found at angles of approximately 20o previous to and after 0o and 180o. Around the 90o and 270o angles where the sampler is aligned perpendicular to the air flow instability is noted, furthermore the airflow through the sampler reverses at angles of incidence of; 75o, 95o, 265o and 280o. It was noted that the readings of pressure drop at these angles had a high level of uncertainty, however, between the angles of incidence of 120o and 240o and 300o and 60o there was little difference in the curve for the varying velocities of airflow applied. As such the angle of airflow is an important factor in the functionality of a Scholtens (1996) passive sampler.
Using passive samplers
Tang et. al. (2001) note that passive diffusion samplers are widely used in the monitoring and measurement of ammonia as they are low cost, they are low tech in nature as they do not require a power source, they are also simple and flexible to use and deploy. This type of sampler can also be deployed in virtually any location and large numbers can be used at once which provides vast amounts of spatial and temporal data. In the field these samplers require little maintenance which coupled with the ease of deployment and handling means that operational costs are low. Furthermore, specialist training is not required to for operation, deployment or maintenance of these samplers. Flint et. al. (2000) note that using passive samplers has benefits in utilising more sample areas as opposed to extrapolating results from a limited spatial data set. This has benefits in overcoming turbulent atmospheres and the use of fragile and more expensive equipment such as direct chemiluminescence (Demmers et. al., 1998) filtration packs (Sickles et. al., 1999) denuders (Leuning et. al.1985) or gas scrubbers (Fehsenfeld, 1995).
Limitations of passive samplers in field applications have been noted by Fehsenfeld et. al. (2002); the preparation of tubes, extraction of ammonia from samples collected and the laboratory analysis are time consuming and labour intensive. Fehsenfeld et. al. (2002) also note that when using samplers of this kind large numbers of measurements and monitoring of ammonia that requires continuous measurement are difficult to obtain. Furthermore, this type of sampler requires considerable sample time at low concentrations of ammonia. In passive samplers that use an acid base to absorb ammonia there is the danger of other present alkaline gasses also being captured. Amine gasses will also be captured by the acid base but in comparison to ammonia the concentrations are generally low enough that they are generally negligible, Phillips et. al. (2001) deduce that total global amine emissions are approximately 1% of ammonia emissions.
Field sampling of ammonia has received much attention in the literature, as such it is useful to consider prior works when assessing the contribution this work aims to provide. There have been comprehensive reviews of past work; Phillips et. al. (2001) gave a detailed account of techniques for ammonia measurement. McGinn and Janzen (1998) provided techniques to measure ammonia loss form soil after slurry application. Ammonia has been sampled at manure heaps using passive samplers (Karlsson, 1994; Phillips et. al., 1997; Petersen et. al., 1998; Rohde and Karlsson, 2002) and slurry stores (Ferm and Svensson, 1992; Hess and Hügle, 1994; Karlsson, 1996; Sommer et al., 1996; Phillips et al., 1997), and after slurry spreading (Ferm and Christensen 1997: Ferm et. al., 1999). Passive sampling of ammonia emissions from animal housing has been the subject of much work (Demmers et. al., 1998; Hartung, 1998; Phillips et. al., 1998; Wathes et. al., 1998; Demmers et. al., 1999; Phillips et. al., 2001; Scholtens et. al., 2003). Ferm et. al. (2005) note that buildings create a turbulent air movement close to the building, and can in some cases cause a backward flux towards the building. As such, this thesis did not sample from animal housing due to possible associated problems of sampling buildings. However, future studies could comprise an element of sampling from animal houses in order to compare Ferm (1986) and Scholtens (1996).
Field Work relating to this study
Sommer et. al. (1996) used a known source of ammonia, a tank with NH4HCO3, to estimate the accuracy of passive ammonia samplers following Ferm (1986). It was found that the accuracy varied with the number of measuring heights used, with low levels close to the source they received accuracy better than 10%. Ferm et. al. (2005) note that positioning horizontal flux samplers, such as Ferm (1986) samplers, mounted in a fixed position around the ammonia source to be sampled is a practical technique for measuring ammonia emissions. As such, this study follows both prior works in using the fixed source of a slurry tank with horizontal samplers positioned around it to compare the sampler method between Ferm (1986) and Scholtens (1996).
Ferm et. al. (2005) note that the temperature between the ammonia source and the surrounding air appears to be an important factor in recording ammonia emissions. The thermal convection increases between the ammonia source And the surrounding air which results in an efficient upward transport of the air layer just above the ammonia source which is saturated with ammonia. This upward movement of the saturated ammonia air layer allows an air layer containing less ammonia to form which will displace the ammonia/ammonium equilibrium forming more ammonia available for volatilisation. In terms of slurry spreading Sommer and Hutchings (2001) also note the connection between the ammonia source and air temperature; solar warming increases the air temperature thus increasing the turbulence in the atmosphere which transports ammonia away from the source. Other studies have also noted a positive relationship between ammonia emission after slurry spreading and temperature or solar radiation (Brunke et. al., 1988; Sommer et. al., 1991; Moal et. al., 1995; Braschkat et. al., 1997; Sommer et al., 1997). Due to the upward movement of air layers this study positioned the passive samplers on masts above the ammonia sources so as to take Ferm et. al. (2005) and Sommer and Hutchings (2001) into consideration in terms of experimental design.
When investigating ammonia emissions of slurry stores Ferm and Svensson (1992) compared the Ferm (1986) method with a chamber method of ammonia detection and found that the Ferm (1986) method of measuring horizontal flux gave a 20 % higher emissions reading. The experiment was then recreated by Ferm et. al. (2000) by monitoring the ammonia emissions after band spreading of pig slurry on a wheat crop; in this experiment the Ferm (1986) method recorded a 24% lower emission reading. Indeed, Ferm et. al. (2005) note that it is difficult to rate the accuracy of the Ferm (1986) method in the field due to a lack of a correct reference technique. This study considers the accuracy of Ferm (1986) samplers in relation to the Scholtens (1996) sampler.
Wood et. al. (2000) compared Leuning et. al. (1985) samplers to Schjoerring et. al. (1992) samplers; which were based on Ferm's (1986) original design. The study compared the samplers in a real farm setting using a circular plot of land spread with urea that was consistent with Leuning et. al.'s (1985) original experiment. It was found that the sampler based on Ferm (1986) had the advantage over the Leuning et. al (1985) sampler as it was more cost efficient and easier to handle and showed high correlation efficient for horizontal and vertical flux. The limitations of cost noted by Woods et. al. (2000) in regards to the use of Leuning et. al. (1985) samplers is also applicable to this study. Funding was not available for the use of Leuning et. al. (1985) samplers. Furthermore, comparing samplers that are cost effective gives a more even ground for comparison.
Passive samplers have been used for ammonia detection in real farm settings globally; Asia, Africa and South America (Carmichael et. al., 2003), Poland (Ferm et. al., 2005), Sweden (Ferm et. al., 2000; Rohde et. al. 2006) Switzerland (Thöni et. al. 2004) and the United States (Rabaud et. al., 2001; Roadman et. al., 2003; Puchalski et. al., 2011). Carmichael et. al. (2003) sampled in different locations, however, comparison of sampler types was not part of the study. Ferm et. al.(2005) sampled ammonia emissions from manure storage and manure spreading on commercial farms in Poland. This study is similar to Ferm et. al.(2005) in the respect that manure storage and spreading are sampled and this study also makes use of Ferm (1986) samplers as did Ferm et. al. (2005). The main difference between this study and Ferm et. al. (2005) is that this study is a comparison of the functionality of passive sampler types rather than sampling over all. Ferm et. al. (2000) is similar to this work as that study compared sampler types in Sweden, however, Ferm et. (2000) was limited to testing ammonia emissions after slurry spreading.
Puchalski et. al (2011) describe the use of three types of passive samplers in field tests; the Adapted Low-Cost Passive High-Absorption (ALPHA), the Radiello and the Ogawa. The samplers were tested at different location at different times. The ALPHA sampler was tested in North Carolina over the period May 2007 to May 2008 comprising of 48 sample periods at 20 different locations. The Radiello sampler was utilised by the Ammonia Monitoring Network between 2007 and 2010 at locations across the American mid-West. The Ogawa samplers were deployed at seven sites in New Mexico, Colorado, Oklahoma, Texas, Wisconsin and Ohio in 2008. While Puchalski et. al. (2011) had a longer study period across more varied sites as compared to this study the samplers were not intended to be compared directly to each other, rather they were described to allow different users to select which most met their needs. As such, Puchalski et. al. (2011) lack the direct comparability which this study offers.
The measurement of vertical ammonia fluxes from wind speed-weighted average vertical concentration gradients has been undertaken (Schjoerring, 1995: Hansen et. al., 1998). Genermont et. al. (1998) have undertaken the measurement of ammonia flux after slurry spreading in field conditions using the Mass Balance Method (MBM) and chemiluminescence analyser with acid traps for verification. Chemiluminescence analyser samplers were deployed after slurry spreading at 5 different heights; the flux gradient was measured continuously using the chemiluminescence analyser. Passive samplers in the form of acid traps were also used for verification of results and analysed using colorimetric analysis. Genermont et. al. (1998) found that the MBM gave reliable flux estimates and that ammonia flux could be directly determined in real time using the chemiluminescence analyser. This study uses Ferm (1986) and Scholtens (1996) samplers to consider ammonia emissions from which flux can be calculated in future works.
Gaps in the literature
Having discussed different types of passive samplers, field tests of samplers and tests which comprised of comparisons between samplers it is evident that there has been little if any work comparing Ferm (1986) and Scholtens (1996) under field conditions. Scholtens et. al. (2003) notes the perceived advantages that the Scholtens (1996) sampler has over Ferm (1986) but there is a lack of evidence confirming these perceived advantages. Indeed, there is a lack of literature accounting to the testing of Scholtens (1996) samplers in a non- laboratory setting. Furthermore, there has been global study of ammonia emissions sampling with the majority of work coming from Scandinavia, mainland Europe and America with some studies carried out in Latin America and Asia. Despite the EU legislation relating to ammonia and other sources of air pollution there are few studies carried out in Scotland specifically or the UK as a whole. As such, this thesis is ideally placed to address these gaps, testing Ferm (1986) and Scholtens (1996) in a real farm setting in Scotland and comparing these samplers.
Materials and Methods
For the purpose of this study Rainton Farm in Dumfries was selected as the location for sampling. Rainton was selected as it already had ties with Glasgow Caledonian University and the BAT Farm project. Rainton is an organic dairy farm and its owner's David and Wilma Finlay are committed to sustainability and achieving low environmental impacts in each aspect of farm management (Cream of Galloway, 2012). As such, David and Wilma agreed to work with GCU and the BAT Farm project to address and mitigate their environmental impacts as recommended by the project. The willingness of David and Wilma coupled with the location and type of farm setting Rainton offered is the main reason for selection.
Passive samplers were constructed following Ferm (1986) and Scholtens (1996). 10cm long glass tubes with diameter 1cm were filled with rolled litmus paper and joined together based on either Ferm (1986) or Scholtens (1996). The litmus paper was 1cm by 4cm and was rolled using a rolling device and then inserted into the tube while rolled around a thin glass rod and then pushed gently in place by the aid of a cotton bud. The rolling device used is commonly used to roll cigarettes; however, it was very beneficial to this study as it gave uniformity to the construction process. The rolled paper was centred in the tube. For the samplers following Ferm (1986) the tubes were made by joining two tubes together using shrink wrap and restricting the flow through the sampler by placing a 1cm diameter stainless steel disc with a hole 1mm in diameter in the centre of the disc at one end. The placement of the tubes in the field as Ferm In and Ferm Out correspond to the steel disc being at the left end of the tube for Ferm In and the right side for Ferm Out. The samplers following Scholtens (1996) were constructed in the same manner with the steel disc being located between two glass tubes containing rolled litmus paper. The tubes were coated with 18 molar sulphuric acid so that the litmus paper would form the acid base to capture NH3. The tubes were numbered using red and black permanent markers with numbers in black indicating the end of the sampler that should face the ammonia source. For the first week of the study on the slurry tower approximately 128 samplers were constructed, 96 of the Ferm variety, and 24 of Scholtens as well as extra samplers to function as field blanks. The same amounts of tubes were constructed for the second and third weeks of the study. For the measurement of ammonia emissions after slurry spreading 336 Ferm tubes were constructed, 252 Scholtens were constructed including those for the background mast and extra tubes were also constructed for field blanks. Gloves were worn at all times during construction and tubes were never placed directly onto any surface to reduce any contamination.
Tang et. al (2001) note that passive samplers can be easily contaminated via air and handling in laboratory conditions if proper care is not exercised. Due to this concern the samplers used in this experiment were only handled when wearing gloves in all stages from construction to analysis to reduce contact with ammonium on hands and fingers. In the field the samplers were put directly into zip lock bags for transportation and kept in these bags in air tight boxes until analysis began.
As sampling locations, Rainton's slurry storage tower and a field where slurry would be spread were selected. The slurry tower was chosen as it gave a spatially fixed ammonia source to be monitored following Ferm et. al. (2005). It was also chosen to sample ammonia emissions after slurry spreading due to the large volume of previous work which has been undertaken on the topic.
Figure x depicts the release of ammonia from stored manure. The slurry tank at Rainton follows this principle. In order for the slurry tower to be used masts were erected at the compass points North, East, South and West. Four masts were selected following Sommer et. al. (1996) and Ferm et. al. (2005) in order to allow the collection of as much ammonia as possible due to the air circulation being impacted by the convection of air streams above the slurry tower. The masts were constructed of timber with hose clamps used to hold the samplers situated at heights of 0.2, 0.4, 0.8, 1.2, 2.2 and 3.3 metres from the top rim of the slurry tank to allow air circulation. The Ferm samplers were positioned with Ferm in and Ferm out at each height and a Scholtens situated at the same height for comparison, all samplers were operated simultaneously. The masts were raised and lowered using a pulley system in order for the samplers to be placed and collected in safety as the tower is 6 metres high. The samplers were also positioned onto the mast prior to that mast being raised using the pulley system to ensure that each mast had a continuous sample time. The samplers were installed on January 24th 2012, February 8th 2012 and third date. Once the samplers were put in place they were collected approximately one week later.
Figure x. Biochemical mechanisms of ammonia release from livestock manure.
Source: Burton and Turner (2003, 83)
Field Spreading of Slurry
The field samplers were deployed after slurry had been spread using a splashplate system. Splash plate slurry spreading broadcasts a thin film of slurry over fields by expelling slurry from a tanker under pressure. This method is the most commonly used in Europe as it is the cheapest method of slurry application (Burton et. al. 1996). The field has an area of 12 acres, with 10 acres of cuttable area. The mast was situated on the Northern edge, a third of the way across the field from the eastern edge. The samplers were situated at heights of 0.2, 0.4, 0.8, 1.2, 2.2 and 3.3 metres with one Scholtens sampler and two Ferm samplers, Ferm In and Ferm out, at each height, a background mast with samplers at heights of 0.2, 0.8 and 2.2 was also used. The samplers were angled at compass directions; North to South and East to West. The samplers were deployed at seven time intervals; 60 minutes, 120 minutes, 180 minutes, 1080 minutes, 1440 minutes, 2880 minutes and 4320 minutes. The background mast results were not used for this study as the concern here is a direct method comparison between Ferm (1986) and Scholtens (1996) samplers not the actual amount of ammonia emitted, as such background measurements were not necessary. The mast was put in place so that for future work the ammonia flux of the samplers can be calculated.
The samples were prepared for analysis by careful deconstruction of the tubes into their former two barrel format. The separated tubes were then sealed with stoppers and placed in racks while 3.5ml of de-ionised water was added with the use of a Gibson pipette into each and a stopper used to the seal the top of the tube. The tubes were left for a minimum of 16 and a maximum of 24 hours and then decanted into sealable vials, sealed and labelled with the number the tube had previously been given. The vials were then stored ready for analysis. Gloves were worn at all times and changed as necessary.
Analysis of Samples
The samples were analysed using colorimetric and ion probe methods to ascertain the ammonia concentration. For each type of analysis a calibration curve, as shown in figure x and x below, was made to extrapolate the volume of ammonia in each sample.
Figure x. Calibration curve ion probe
Figure x. Calibration curve colorimetric
Each calibration curve was made by creating a 100mg/L stock solution of Ammonium Chloride (NH4+Cl) by measuring 0.2865g of anhydrous ammonium chloride using a mass balance. The ammonium chloride was then dissolved in a volume of approximately 400 to 500ml 15 â„¦ deionised water, the solution was then made up to 1000ml by adding more 15 â„¦ deionised water to form the stock solution. The solution was then diluted in volumetric flasks to give the following concentrations; 1mg/L, 2mg/L, 5mg/L, 10 mg/L, 20 mg/L, 30 mg/L. 40 mg/L, 50 mg/L, 60mg/L, 70 mg/L, 80 mg/L and 90 mg/L. Each solution was then analysed using the colorimetric and ion probe method which described as follows. As can be seen from figure x and x the R2 values for the ion probe method and colorimetric method were 0.992 and 0.9982 respectively, which show a high correlation factor between ammonia concentration and the analysis method results obtained.
In colorimetric analysis the modified Bertholot method supplied by Fortress Diagnostics was used in order to give the samples colour. The principle of this method is that the ammonium ions in the sample react with salicylate and hypochlorite to form an indophenols complex which shows in the green spectrum. To achieve this process 0.1 ml of the sample was pippetted, using a 1ml Labnet autoclave pipette, into a 2 sided cuvett supplied by Fisher Scientific, UK, which had been marked with the sample number. 1ml of reagent 1, which is made up of 62 mmol/l Sodium Salicylate and 5 mmol/l Sodium Nitroprusside, was then added to each sample. Finally 0.5 ml of Reagent 2, Sodium Hypochlorite, was added to each sample. The same pipette was used for every sample to ensure consistency and to reduce error. A development time of exactly 10 minutes was allowed to elapse before the samples were analysed, all samplers were analysed within approximately 1 hour 30 minutes of this 10 minute maturation window. The samples were analysed using a Jenway 6505 UV/Vis. Spectrophotometer with the wavelength set at 600, the zero value was calibrated using deionised water. The samples were run in duplicate and an average taken to adjust for any human error.
An ammonium ion-selective electrode from ELIT was utilised to measure the NH4+ in the samples. This method was chosen as the ELIT electrode measures the NH4+ ion directly in the solution, this means that there is no need for pre-treatment of the solution and no noxious gasses are liberated during analysis. For analysis the probe was calibrated using 1000 parts per million NH4Cl, a time frame of between 5 and 10 minutes was observed to allow the probe to reach a constant reading. Once this constant was achieved the probe was washed for 1 minute in de-ionised water being stirred at 100 turns per minute. The reading for the de-ionised water was recorded each time and only used if the reading was below 270 mV, on any occasion where the reading was greater than 270 mV the water was changed and the minute repeated until a lower reading was attained. The probe was then carefully dried and inserted into the sample for 1 minute, at the end of this minute the reading was taken as it was shown to ensure each sample was given the same time frame. Once the reading was taken the probe was returned to fresh de-ionised water for 1 minute for washing. This process was repeated for all samples.
Calculating Ammonia content of samples
An R2 value of 0.9992, as shown in figure x for the ion probe calibration curve, shows the high level of correlation between the NH4Cl solution concentration and the mV reading of the ion probe for solution. The R2 value was achieved by taking a natural log of the NH4Cl solutions. This was corrected in the results by using an excel function to compensate by returning the natural log to the original value. To give the ammonia content for the ion probe readings after adjusting for the natural log the excel sheet was programmed to multiply the sample value if it was above the zero point of 262.62 by 0.046, then minus 11.145 and multiply this by 3.5 so as to account for the volume of the sample. These values are taken from the calibration curve in figure x.
To calculate the ammonia content of the samples according to the colorimetric method no natural log was used, however this does not affect the comparability of results as the calculation for the ion probe method was composed to adjust for this. The ammonia content was calculated from the colorimetric data by programming the spreadsheet to multiply the sample value if it was above the zero point of 0.26801 by 66.255, then minus 17.757 and multiply the total by 3.5 to account for the volume of the sample. The values are taken from the calibration curve in figure x. As noted previously the high R2 values attained for the calibration curve shows that there is a high correlation between the ammonia concentration and method readings meaning that there is a high confidence level in the results.
Calculating ammonia flux
The ammonia flux for the field after slurry spreading can be calculated using the following equation:
= Integrated VF of sample mast - Integrated VF of background mast
Where x is the fetch length in metres from the sample mast to the edge of the plot or the distance of slurry that was measured. In the current study the fetch was 15m. This study does not consider ammonia flux but for future work this is the method that would be used.
In the first week of the experiment the numbers were coated in clear nitrocellulose polyvinyl coating (nail polish) to protect the numbers from the elements. However, this coating did offer adequate protection to the samplers and shrink wrap was used as the coating in the subsequent weeks of the experiment. There were no problems encountered with the shrink wrap. In week one of the experiment the numbers of the samplers located on the North mast above the slurry tower were lost and as such these results have not been included as the placement of the tubes could not be verified in order for comparison.
During the analysis a fault with the ELIT ion probe was observed. The probe began to malfunction, in that the reading time for deionised water became extended. Following the manufacturers guideline the probe was then cleaned with ethanol and left in a concentrated solution of NH4Cl for upwards of 24 hours to recalibrate. After cleaning the probe functioned normally and a cleaning schedule was put in place so as to avoid any further complications. However, the probe suffered a breakage and had to be replaced after the initial cleaning problem. Once the new probe arrived it was recalibrated and the calibration curve was redone to check the accuracy of the new probe. Once it was verified the analysis continued without further setback.