1

This is author version of article published as: 1

Hudson, Neale A. and Ayoko, Godwin A. (2008) Odour sampling. 2. Comparison 2 of physical and aerodynamic characteristics of sampling devices: A review . 3 Bioresource Technology 99(10):pp. 3993-4007.4

Copyright 2008 Elsevier 5

Odour sampling 2. Comparison of physical and 6

aerodynamic characteristics of sampling devices. 7

A review 8

N. Hudsona,b* and G. A. Ayokoa 9

aInternational Laboratory for Air Quality and Health, School of Physical and Chemical Sciences, Queensland University of 10

Technology, GPO 2434, Brisbane Queensland 4001, Australia and bDepartment of Primary Industries & Fisheries Queensland, PO 11

Box 102, Toowoomba, Queensland 4350, Australia. 12

Abstract 13

Sampling devices differing greatly in shape, size and operating condition have been used to collect air 14

samples to determine rates of emission of volatile substances, including odour. However, physical 15

chemistry principles, in particular the partitioning of volatile substances between two phases as explained 16

by Henrys Law and the relationship between wind velocity and emission rate, suggests that different 17

devices cannot be expected to provide equivalent emission rate estimates. Thus several problems are 18

associated with the use of static and dynamic emission chambers, but the more turbulent devices such as 19

wind tunnels do not appear to be subject to these problems. In general, the ability to relate emission rate 20

estimates obtained from wind tunnel measurements to those derived from device-independent techniques 21

* Corresponding author. E-mail address: neale.hudson@dpi.qld.gov.au Telephone +61-7-4688 1519, Fax +61-

7-4688 1192

2

supports the use of wind tunnels to determine emission rates that can be used as input data for dispersion 1

models. 2

Keywords: odour; emission; wind tunnel; chamber; dynamic; flux. 3

1. Introduction 4

Estimation of rates of emission of volatile materials from area sources such as 5

anaerobic treatment ponds, feedlot pads, compost windrows and municipal wastewater 6

works is a complicated process. Conceptually, two basic processes may be used: 7

1. Device-independent micro-meteorological techniques, where the emission 8

rate is calculated from concentrations measured across the plume of emitted 9

material and local meteorological data, specifically wind velocity profile 10

data [e.g. (Yamulki et al., 1996; Christensen et al., 1996; Denmead et al., 11

2000; Magliulo et al., 2004; Kim et al., 2005)], and 12

2. A sampling device, where a chamber, hood or wind tunnel, is deployed on 13

an emitting surface. The device may be static (sealed or vented) or flushed 14

with contaminant-free carrier at a known velocity or flow rate. The 15

emission rate is calculated as the product of concentration and airflow 16

through the device [e.g. (Raich et al., 1990)(Eklund et al., 1985; Gholson et 17

al., 1991; Fukui and Doskey, 1996; Conen and Smith, 1998; Peu et al., 18

1999)]. 19

In principle, similar techniques may be used to estimate odour emission rates. The 20

entire process comprises: collection of representative odour samples; measurement of 21

flushing rates within the sampling device (if used) or measurement of ambient 22

meteorological conditions (for device-independent methods); determination of odour 23

3

concentration in the sample using dynamic olfactometry, and calculation of odour 1

emission rates using the concentration and flushing rate data or meteorological 2

conditions at the time of sampling. In the case of odour, however, the high cost of 3

odour assessment limits the number of samples that can be realistically analysed. This 4

effectively precludes micrometeorological measurement techniques, which necessitate 5

the collection of a substantial number of samples across the downwind odour profile. 6

The entire odour sampling and assessment process was recently reviewed 7

(Environmental Protection Agency, 2001; Gostelow et al., 2003). In the former review 8

the benefits and disadvantages of micrometeorological, wind tunnel, static- and 9

dynamic-hood sampling procedures were tabulated in some detail, together with the 10

potential application of all techniques. While the technical and methodological 11

difficulties associated with wind tunnels appeared to be least, the authors held back 12

from recommending the use of any specific device for sample collection. It should be 13

noted that the use of any device to collect a sample of odorous air is likely to disturb the 14

emitting surface and thereby the true emission rate. 15

In Australia, a range of wind tunnels and emission chambers have been used to 16

collect odour samples from area sources at intensive livestock farming operations. 17

Regulatory agencies in Australia have developed odour management policies for these 18

facilities based on data derived from a number of these devices (Streeten and McGahan, 19

2000; Skerman, 2000; McGahan et al., 2000; Environment Protection Authority, 20

2001). Accordingly, separation distances and odour impact criteria vary according to 21

the regulatory jurisdiction. This situation arose in part from the very different numeric 22

values obtained from use of the various emission chambers or wind tunnels. This 23

creates difficulties for industries that operate across State lines in Australia, who may 24

4

have quite different license conditions for equivalent operations in different States. 1

These apparent anomalies are difficult to comprehend by the general community, 2

thereby creating difficulties for the various regulatory agencies and producers. Recently 3

the Australian beef cattle feedlot industry facilitated an investigation to reconcile the use 4

of these two sampling devices for odour sampling (Meat & Livestock Australia (MLA) 5

and the Australian Feedlot Producers Association (ALFA), 2002; Nicholas et al., 6

2004). The results of this research confirmed that two sampling devices commonly 7

used in Australia for estimating odour emission rates from area sources provide quite 8

different results. 9

The situation is even more confusing if the scientific literature is consulted. While 10

mass transfer processes have been extensively investigated and described, little 11

guidance is provided in the selection and operation of sampling devices to obtain 12

meaningful emission rate estimates. Review of the literature reveals a wide range of 13

devices that may be used to collect samples of volatile chemicals from liquid and soil 14

surfaces. With the possible exception of measurements of emission rates for mercury 15

(Gustin and Lindberg, 2000) and major atmospheric gases (Wanninkhof et al., 1985; 16

Clark et al., 1995; Crusius and Wanninkhof, 2003), limited comparison of data derived 17

from different sampling devices has taken place. Accordingly, limited data and 18

information exists to guide practitioners in the selection and operation of sampling 19

equipment. Currently, even less information exists to guide practitioners in the 20

selection and use of the most appropriate odour sampling equipment. 21

This review identified some of the devices used to collect samples of volatile 22

material emitted from solid and liquid surfaces. The physical dimensions and typical 23

operating conditions of these devices were summarized, providing an overview of the 24

5

wide range of sizes and operating conditions. Examination of a few representative 1

examples from the literature in more detail identified some of the practical issues and 2

limitations posed by each device. Consideration of the fundamental principles 3

underlying the mass transfer process provided guidance regarding selection of sampling 4

equipment. Finally, information from the literature was used to justify the selection of 5

wind tunnels for estimating odour emission rates. 6

2. Odour as an “analyte” 7

This aspect was recently discussed (Hudson and Ayoko, Submitted for 8

publication), where it was established that “odour” is a complex mixture of many 9

organic and some inorganic chemicals. For example, up to 330 different chemicals 10

belonging to a range of chemical classes including volatile fatty acids, aldehydes and 11

ketones, nitrogen heterocycles, reduced sulphur compounds and phenols, were 12

identified in odour samples derived from piggeries and beef feedlot operations (O'Neill 13

and Phillips, 1992; Schiffman et al., 2001; Wright et al., 2005). As a sample material, 14

odour therefore represents an uncontrolled mixture of chemicals of different classes and 15

quite different physical and chemical properties. Caution is therefore required when 16

selecting a sampling technique to ensure that discrimination is minimized and that the 17

sample is truly representative of the ambient odour composition. This is particularly 18

important when the key odorants associate with intensive livestock-raising are 19

considered. Zahn et al. (2001) generated an artificial piggery odour using 19 specific 20

odorants. It was possible to develop an odour prediction model using nine of these 21

compounds – acetic, butyric, isobutyric, valeric and heptanoic acids, phenol, 4-22

methylphenol 4-ethylphenol and 3-methylindole. These odorants all have relatively 23

small values of non-dimensional Henry constant. Wright et al. (2005) identified 4-24

6

methylphenol, 2’-aminoacetophenone, isovaleric acid and 4-ethylphenol as the most 1

significant odorants downwind from a major piggery. Neither Zahn et al. (2001) or 2

Wright et al. (2005) identified hydrogen sulphide, mercaptans or other reduced sulphur 3

compounds as significant odorants associated with intensive livestock facilities. 4

Reduced sulphur compounds typically have relatively large values of non-dimensional 5

Henry constant. 6

3. Mass transfer processes 7

Hudson and Ayoko recently reviewed mass transfer processes in the context of 8

odour sampling (Submitted for publication). While the major force determining the 9

direction and rate of emission of a solute is the difference in concentration between the 10

two phases, it was established that the magnitude of the dimensionless Henry law 11

constant ( H ′ ) was very significant in determining the air-water transfer velocity of a 12

solute. It was also shown that liquid turbulence was likely to be significant for highly 13

volatile materials with large ( H ′ ) values, while air turbulence (i.e. wind velocity) was 14

very significant to air-water transfer velocities for all solutes, but particularly those with 15

small ( H ′ ) values. The majority of odorants have small ( H ′ ) values. The interaction 16

between the magnitude of ( H ′ ) of individual odorants and the hydrodynamic 17

characteristics of the odour sampling device should therefore be considered carefully 18

before collecting a sample of odorous air. 19

4. Physical characteristics of odour sampling devices 20

A review of the scientific literature revealed that a very wide array of chambers, 21

hoods and tunnels have been used to collect samples of volatile materials. Closer 22

examination revealed that the conditions under which they were operated also varied 23

7

greatly. In view of the uses to which emission rate data is put, it is difficult to believe 1

only very limited comparison of rates of emission of different devices and solutes has 2

taken place. Some representative data included in part 1 of this series (Hudson and 3

Ayoko, Submitted for publication), clearly showed that estimates of odour emission rate 4

are strongly influenced by the selection of sampling device. 5

The wide range of physical dimensions and operating conditions made it difficult 6

to describe and compare the actual conditions within these devices during the sample 7

collection process. To address this situation, the results of a review of the literature 8

were summarised and an attempt made to compare sampling conditions. 9

4.1. Review of the literature 10

Emphasis was given to identifying devices used to collect odour and other volatile 11

substances, particularly volatile substances with characteristics similar to odorants, 12

associated with intensive livestock operations. Representative examples of devices used 13

to collect soil gases were also selected owing to the large number of publications 14

describing their use. 15

Sample collection devices were identified as static or dynamic devices. Static 16

devices were defined as those where air was not actively passed through the device at a 17

measured rate. This included a number of sealed and vented chambers. Dynamic 18

devices included a number of chambers and wind tunnels. In some cases, local wind 19

conditions during the sample collection period determined the actual flushing rates of 20

the wind tunnel devices. The devices were further subdivided into three groups 21

according to their physical shape. Cylindrical, rectangular and a combination of these 22

two basic shapes were identified. 23

8

4.2. Development of criteria for comparison of sampling devices 1

The basic physical characteristics of each device were identified, from which the 2

internal volume and area of the device footprint were calculated. Typical operating 3

conditions were also identified, specifically flushing rates. Flushing rates were related 4

to the footprint area and volume within the device. Ratios of flushing rate to surface 5

area and device volume provided an indication of relative “loading rate”, allowing 6

objective comparison between devices. Where appropriate, horizontal and vertical 7

turbulence was assessed following calculation of the Reynolds number for each 8

dimension of each device (Zhang et al., 2002): 9

υc

eV

LR = Equation 1 10

where eR is the Reynolds number, L is the length of the flushing path (L), cV is 11

average flushing air velocity (L.t-1) and υ is the kinematic velocity of the air (L2.t-1). 12

cV may be estimated from: 13

aQVc = Equation 2 14

where Q is flushing flow rate (L3.t-1) and a is the cross-sectional area of the 15

device (L2) (Gao and Yates, 1998b). Under well-developed boundary layer conditions, 16

the vertical and horizontal components ensure that molecular diffusion and convective 17

mass transfer occur (Bliss et al., 1995). Under these conditions, emission rate becomes 18

a function of wind velocity [e.g. (Smith and Watts, 1994; Jiang and Kaye, 1996)]. 19

Using typical operating conditions for the various devices identified in the relevant 20

publication, boundary layer conditions in either the vertical (height) or horizontal 21

(width) plane of the sampling device were estimated. These estimates provided 22

9

information regarding the relationship between air velocity, turbulence intensity and 1

turbulence within the device and the physical dimensions of the device. This 2

information is summarised in Table 1 and Table 2. These Tables identify boundary 3

layer “compliance”. Compliance indicated that fully developed boundary layer 4

conditions were likely to exist within the device, i.e. rates of emission of volatile 5

materials were likely to be a function of wind velocity within the device. Statistics 6

describing the basic physical characteristics and operating conditions were also 7

calculated for the various classes of device; these are summarised in Table 3 to Table 6. 8

5. Characteristics of sampling devices 9

5.1. Static chambers 10

Static chambers are more likely to be circular than rectangular. Eight of 27 11

cylindrical devices were static, while only five of 65 rectangular devices were static. 12

This was probably due to the ready availability of pipe rather than consideration of mass 13

transfer processes. The basic physical dimensions and operating parameters of the 14

cylindrical and rectangular static chambers listed in Table 1 and Table 2 are compared 15

in Table 3 and Table 5 respectively. While it was difficult to make a meaningful 16

comparison of rectangular and cylindrical shapes, obvious differences existed in 17

chamber heights, volumes, area:volume ratios and height:width or height:radius ratios. 18

Median chamber heights, volumes and height:width ratios were more than 30% greater 19

for rectangular devices than equivalent median values for cylindrical static chambers. 20

The area:volume ratios (loading ratio) of rectangular static chambers were about 35% 21

smaller than those of cylindrical static chambers. The larger loading ratio of cylindrical 22

chambers implies that a larger mass of material would be released into a cylindrical 23

10

static chamber per unit time than into a rectangular chamber. In practical terms, the 1

response time of the two chamber shapes would therefore be different. A case could be 2

made to adjust emission rates provided by chambers of different size and shape to 3

normalise measured emission rates. 4

5.2. Dynamic chambers 5

Dynamic devices are far more likely to have a rectangular than cylindrical 6

footprint. Nineteen of 76 dynamic devices were cylindrical, whereas 57 were 7

rectangular. This probably reflects that it is simpler to flush a box shape continuously 8

by passing an air stream from one side to the other (effectively minimising stagnant 9

zones), than a cylindrical shape. 10

Median chamber volumes and area:volume ratios (loading ratios) differed by less 11

than 15% for cylindrical and rectangular shaped devices. The median surface areas of 12

rectangular chambers surveyed were about 38% larger than those of cylindrical 13

chambers. In contrast, median flushing rates, area related sweep rate and air exchange 14

rates were 625%, 480% and 1200% larger for the rectangular devices than for 15

cylindrical devices respectively. These differences implied: 16

• Rectangular chambers were potentially more representative of emitting surfaces 17

owing to the larger area covered; 18

• Samples derived from rectangular chambers would be less concentrated than those 19

derived from cylindrical chambers owing to the dilution by the much larger flushing 20

rates; 21

11

• Conversely, increases in headspace concentrations of volatile compounds were 1

unlikely to occur with rectangular chambers – suppression of emission rate would 2

therefore be unlikely. 3

6. Selected case studies from the literature 4

Bekku et al. (1997) utilised static, dynamic and flow-through chambers to measure 5

rates of emission of carbon dioxide from soils. They confirmed that emission rates 6

could be depressed by increases in concentrations of carbon dioxide in the headspace of 7

static chambers. They were able to minimise the impact on measured carbon dioxide 8

emission rates by limiting the chamber placement period. 9

Conen and Smith (1998) compared emission rates obtained with conventional 10

closed chambers with those obtained from vented chambers. It was proposed that a vent 11

tube would minimise the influence of pressure on emission rate estimates. On relatively 12

permeable soils, vented chambers provided estimates of emission rate up to five times 13

those of sealed chambers. Emission rates were also strongly dependent on wind 14

velocity. This difference was attributed to a venturi effect that caused soil air to enter 15

the chamber. On heavier (less permeable) soils, vented chambers provided emission 16

rate estimates that were about 88% those of sealed chambers. In these circumstances, 17

the venturi tube was thought to cause loss of material following diffusion. Their 18

research identified that measured emission rates were strongly dependent on chamber 19

operation, ambient conditions and the characteristics of the emitting substrate. 20

Closed and dynamic chambers were used by Gao and Yates (1998a) to estimate 21

soil fumigant emission rates. Closed chambers underestimated emission rates, with 22

underestimation increasing with chamber placement time. Dynamic chambers also 23

12

depressed emission rates, if the flushing rates were too low. Underestimation by both 1

chamber types was attributed to increasing concentrations of volatile material in the 2

chamber headspace, which in turn depressed the concentration gradient across the soil-3

air interface. It was also demonstrated that overestimation of emission rate was possible 4

if the flushing rate in a dynamic chamber was too high, particularly for vacuum driven 5

systems operated on sandy permeable soils. This was attributed to the creation of a 6

pressure deficit within the chamber. It was recommended that airflow rates be 7

optimised (increased) to minimise increases in headspace concentrations of emitted 8

materials, while avoiding pressure deficits by ensuring that flow rates were not too high. 9

They concluded that: “the ideal combination of airflow rate and pressure deficit is 10

probably case specific and difficult to assess”. 11

Gao et al. (1997) constructed a shallow, flow-through emission chamber to 12

measure soil gas emission rates. The chamber incorporated flow straightening and 13

contraction sections at the inlet and exit respectively to minimise stagnant zones and 14

equalise flows across the chamber width. Measured emission rates agreed quite closely 15

with those derived from micrometeorological methods. It was recommended that 16

dynamic emission measurement devices be utilised more widely because they produced 17

a uniform horizontal air stream that was easy to model. In addition, the wind velocity 18

within the device could be matched to that outside to simulate ambient conditions. 19

Gillis and Miller (2000) developed a dynamic emission chamber to measure 20

mercury emissions from soils. A range of flushing rates were applied to the device – 21

measured emission rates were strongly dependent on flow rate. It was also established 22

that the measured emission rate was dependent on ambient wind velocity and 23

orientation of the device relative to the wind direction. Differences in mercury emission 24

13

rates previously measured using dynamic chamber devices and micrometeorological 1

methods were attributed to both chamber wind velocity and external wind effects. 2

The US EPA dynamic emission chamber (“flux chamber”) was developed to create 3

a standardised method for measuring the emission of industrial chemicals from 4

contaminated soils and liquid storage facilities (Kienbusch, 1986; Gholson et al., 1989). 5

It was selected in preference to laboratory-scale emission measurements or 6

micrometeorological methods because of ease of use and reduced costs. It was 7

acknowledged that the device did not account fully for all the variables that controlled 8

emission, e.g. the volatility of individual chemical species, ambient meteorological 9

conditions, surface roughness, soil moisture and permeability. It was felt however that 10

management actions and treatment-induced impacts on the emission process were quite 11

significant and might dominate the processes under some circumstances. Use of a 12

device within which consistent, reproducible conditions could be established was 13

therefore the most important criterion – a flux chamber provided this capability. 14

During development, the device was extensively tested to assess the impact of 15

external factors such as solar radiation as well as operational factors such as flushing 16

rate and placement depth. Measured emission rates for three representative chemicals 17

were compared with those provided by a wind tunnel device. Recoveries of 38 18

industrial contaminants were also measured. This extensive testing is often cited by 19

users of the device as justification for selection. However, careful consideration of the 20

results of these tests indicate that they do not really have much relevance to odour 21

sampling. The average recovery rates reported for the 40 test industrial chemicals was 22

104% (Kienbusch, 1986). These recoveries were determined by introducing a 23

calibration gas containing known concentrations of standard chemicals into a chamber 24

14

placed on a Teflon™ surface with a central inlet port. The recovery rate was therefore a 1

measure of loss of material through absorption or leakage from the system, not to 2

recovery from a liquid or soil surface relative to a “true” emission rate. 3

Rates of emission of three selected chemicals were compared with those obtained 4

from a wind-tunnel. The differences in emission rates were reported as bias and are 5

listed in Table 7. In the original research, no explanation was offered for the bias 6

observed. Consideration of the properties of the chemicals listed in Table 7 provides 7

some insight regarding the cause of the difference in measured emission rate. In Part 1 8

of this series (Hudson and Ayoko, Submitted for publication), we highlighted that rates 9

of emission are strongly influenced by turbulence in liquid and/or gas phase. Rathbun 10

and Tai (1986; 1987) and Chiou et al. (1980; 1983) and Lee et al. (2004) demonstrated 11

that emission rates for a range of chemicals increased as surface wind velocity 12

increased. The magnitude of the dimensionless Henry law constant ( H ′ ) provides an 13

indication of the phase in which mass transfer rates will be controlled. If H ′ greatly 14

exceeds about 0.001, volatilisation tends to be dominated by liquid phase turbulence. If 15

H ′ is much smaller than about 0.001, volatilisation will tend to be dominated by gas 16

phase turbulence, for which wind velocity serves as a surrogate. For compounds with 17

intermediate values of H ′ (0.0005 – 0.05), both liquid and gas phase turbulence 18

determines emission rates. Of the chemicals listed in Table 7, two have intermediate 19

values of H ′ , while the other two have values of H ′ that imply that liquid phase 20

turbulence will control emission rates. The bias in emission rates observed for the test 21

chemicals are quite consistent with the values of H ′ . The difference in measured 22

emission rate is likely to be largest for the compounds with small values of H ′ . The 23

increased wind velocity within the wind tunnel increased the emission rate of the two 24

15

chemicals with smallest H ′ values relative to the more quiescent emission chamber, 1

reported originally as a larger bias. 2

Ryden et al. (1985) described the use of a wind tunnel to measure ammonia 3

emission rates from grassland relative to a micrometeorological method. A fan allowed 4

the flushing rate within the tunnel to be adjusted over a range of wind velocities from 0 5

to 3.0 m/s. During use, the tunnel wind velocity was adjusted to be similar to that of the 6

outside wind velocity. When the tunnel wind velocity was maintained at 1 m/s, the 7

tunnel emission rate differed from values obtained by micrometeorological mass 8

balance by a factor of two to five. When the tunnel wind velocity was maintained as 9

close to ambient wind velocity as possible, no significant difference in emission rate 10

derived from the two techniques was observed. 11

During an investigation of odour emissions from feedlots, Smith & Watts (1994) 12

compared emission rates obtained from two wind tunnels (Table 2). Odour emission 13

rate to velocity relationships were very similar for both tunnels and differences were 14

explained in terms of the different tunnel heights. Differences in absolute odour 15

emission rate were observed. These were attributed to the relative height differences 16

between the tunnels, which influenced the velocity profiles within the tunnels. 17

Smith (1993) also investigated downwind, back-calculation methods for assessing 18

odour emission rates from feedlot surfaces (1995). He was able to demonstrate that 19

odour emission rates estimated from samples derived from wind tunnels were consistent 20

with those derived from a simple Gaussian dispersion model. This was a significant 21

finding – it was the first time a device-independent method had been used to validate 22

odour emission rates derived from a wind tunnel. 23

16

More recently Hudson et al. (2004) and Galvin et al. (2004) were able to confirm 1

the findings of Smith. The same Gaussian dispersion model was used to estimate odour 2

emission rates from samples collected downwind from anaerobic ponds treating piggery 3

waste. A wind tunnel was used to collect odour samples directly from the pond surface, 4

from which emission rates were calculated. Emission rates were calculated for up to 5

nine discrete anaerobic ponds over three seasons. The overall relationship between the 6

direct and downwind derived odour emission rates had an r2 value of 0.76 (n = 27), 7

indicating a reasonable correlation between the two methods. 8

The development of a wind tunnel by researchers at the University of New South 9

Wales (UNSW) arose from the requirement to create a device with well defined 10

aerodynamic characteristics and where the relationship between tunnel wind velocity 11

and odour emission rate was consistent and manageable (Bliss et al., 1995; Jiang et al., 12

1995; Wang et al., 2001b). It was demonstrated that ammonia emission rates were a 13

function of wind velocity to the power of 0.5. This result was reasonably consistent 14

with that of Smith and Watts (1994), who observed a relation to the power of 0.63. The 15

difference in magnitude of exponent was explained by differences in the surface 16

roughness of the emitting surfaces. 17

The relationship between odour emission rate and wind velocity is important in the 18

context of odour impact assessment, where convective mass transfer and dispersion 19

strongly influence the likelihood and severity of impact. Smith and Watts (1994) 20

concluded that their results “confirm the need to operate any wind tunnel at a wind 21

speed equivalent to the ambient wind speed, if the purpose of the sampling is to 22

determine “real” emission rates”. 23

17

This conclusion was in part validated by Sattler et al. (2002), who collected 1

hydrogen sulphide samples from a municipal wastewater treatment plant using an 2

emission chamber based on the principles of the US EPA dynamic emission chamber. 3

The emission rates obtained from these measurements were used as input to a dispersion 4

model. Predicted hydrogen sulphide concentrations were compared with odour 5

complaint data. It was clearly evident that the predicted hydrogen sulphide 6

concentrations were too low, i.e. the emission rate values derived from the emission 7

chamber were too low. The authors noted “the effect of wind speed on dispersing 8

pollutants is already accounted for in traditional dispersion modelling methodology; 9

the effect of wind speed on release or generation of emissions in the first place is not”. 10

They were able to better match the hydrogen sulphide concentrations with odour 11

observation records by applying a wind velocity correction factor derived from Jiang et 12

al. (1995): 13

5.0

1

212 ⎟⎟

⎠

⎞⎜⎜⎝

⎛=

vvEE Equation 3 14

where 1E and 2E are emission rate at wind velocity one and two respectively ( 1v 15

and 2v ). Although the wind velocity adjustment yielded credible results based on flux 16

chamber emission rate estimates, Sattler and McDonald (2002) concluded that “wind 17

effects would be best accounted for by using a wind tunnel to estimate emission rates 18

and then apply a fetch adjustment”. It is difficult to correct emission rate estimates 19

made with a flux chamber to those likely for a wind tunnel because of the absence of a 20

clear relationship between these values, evident from our results (Hudson and Ayoko, 21

Submitted for publication) and the data from Jiang and Kaye cited by Sattler and 22

McDonald (2002). 23

18

7. Implications of device dimensions, operating characteristics for selection of 1

odour sampling equipment 2

7.1. Device dimension and flushing rate considerations 3

By definition, static chambers are not flushed. Concentrations of volatile 4

substances will increase in the headspace of the device as a function of placement time, 5

decreasing the chemical driving force and thereby the rate of emission of volatile 6

material over the chamber placement period time. Successful use of these devices 7

requires short placement periods and collection of a series of samples of small volume 8

over a short period commencing immediately following chamber placement. The large 9

size of sample (40 to 120 L) required for odour measurement alone renders static 10

chambers totally unsuitable for odour sampling if meaningful emission rates are 11

required. 12

Despite the dynamic nature of the US EPA emission chamber, the large sample 13

volume and long chamber placement time recommended calls into question the use of 14

this device for odour sampling. Sampling rates must be kept small to avoid creating a 15

vacuum within the chamber, leading to chamber placement periods of 60 to 80 minutes. 16

It was observed that odour concentrations in samples derived from these devices were 17

very high (Hudson and Ayoko, Submitted for publication; Sattler and McDonald, 2002). 18

Depression of odour emission is quite probable under these conditions. Therefore, in 19

addition to being impractical, the long chamber placement period may directly lead to 20

alteration of the sample composition. This issue has never been addressed. 21

It is not possible to determine the velocity at which flushing air passes over the 22

emitting surface of a circular chamber. While the chamber volume may be well mixed 23

19

(but no evidence has been published to verify this assumption), it is likely that areas of 1

the chamber will experience higher flushing air velocities than others. Our research 2

(Hudson et al., Unpublished data) has shown that four distinct zones of higher velocity 3

exist within the US EPA flux chamber. These arise from the number and orientation of 4

holes delivering flushing air into the chamber. Once again, the impact of the resulting 5

velocity gradients on odour sample composition has not been investigated. 6

The interaction of flushing air velocities and the magnitude of the dimensionless 7

Henry law constant ( H ′ ) may influence the rates of emission of volatile substances and 8

thereby the composition of the sample. Many odorants derived from livestock wastes 9

have small H ′ values, making the composition and concentrations of odour samples 10

particularly vulnerable to air turbulence and velocity. Generally, measured odour 11

concentrations are used to calculate odour emission rates, used as input to dispersion 12

models. Eliminating or reducing the effects of turbulence and velocity may have the 13

potential to discriminate against certain odorants (leading to unnaturally low 14

representation), or selection of others (increasing concentrations relative to natural 15

conditions). This may have quite significant implications for the odour concentrations 16

measured and thereby the processes currently used to predict adequate buffer distances. 17

It was possible to validate odour emission rates derived from wind tunnel samples 18

using device-independent methods. It was also demonstrated that emission chamber 19

hydrogen sulphide results could be correlated with ambient concentration data (using 20

odour complaint information) provided a wind velocity/turbulence correction was made. 21

Other investigations regarding rates of emission of mercury, ammonia, chlorinated 22

organic chemicals and moderately polar, water soluble chemicals demonstrated that 23

emission rates derived from dynamic sampling devices (wind tunnels) could be related 24

20

to those obtained using micrometeorological methods. This provides compelling 1

evidence to support the view that wind tunnel devices should be selected for odour 2

sample collection above dynamic, chamber-type devices that do not allow wind velocity 3

and turbulence conditions to be managed. 4

8. Conclusions 5

Principles well founded in physical chemistry predict that rates of emission of 6

volatile compounds will be influenced by wind velocity (or atmospheric turbulence), 7

liquid turbulence and the properties of the volatile material. The value of the Henry law 8

coefficient provides a reasonable indication as to whether liquid or atmospheric 9

turbulence is likely to dominate the emission process. 10

Consideration of the large variety of shapes and sizes of devices used to collect 11

samples of volatile material, coupled with the range of conditions imposed by the 12

operating conditions indicates that comparable estimates of emission rate should not be 13

anticipated from different devices. A limited number of comparative studies 14

demonstrated that odour emission rate estimates varied markedly between relatively 15

turbulent devices such as wind tunnels and quiescent devices such as closed or dynamic 16

flux chambers. 17

Numerous studies involving use of closed or dynamic emission chambers 18

confirmed that emission rates measured with these devices are subject to external 19

influences, including ambient wind velocity, direction and the permeability of the 20

emitting surface. These influences may be transient and are outside of the control of 21

experimental design or sampling practice. As such, they are likely to introduce errors 22

into the emission rate measurement process. 23

21

The clear relationship between wind velocity (atmospheric turbulence) and 1

emission rate indicated that dynamic or turbulent sampling devices such as wind tunnels 2

were more likely to model natural emission processes. This was of particular 3

importance if the emission rate estimates were to be used as input to a dispersion model. 4

The fact that odour emission rates estimated from samples derived from wind 5

tunnels could be validated using device-independent techniques such as downwind 6

sampling following by back-calculation, provided further evidence that wind tunnels 7

should be preferred to estimate odour emission rates. 8

Acknowledgement 9

This work was funded in part by the Department of Primary Industries & Fisheries, 10

Queensland. 11

References 12

Allaire, S., Yates, S., Ernst, F. and Gan, J. 2002. A dynamic two-dimensional system 13

for measuring volatile organic compound volatilization and movement in soils. 14

Journal of Environmental Quality 31 1079-1087. 15

Baek, B-H, Koziel, JA, and Parker, DB. Comparison of ammonia and hydrogen 16

sulphide mass emission rates from artificial cattle feedlot surfaces using two designs 17

of isolation flux chambers. 493-501p. 18

Baek, B-H, Koziel, JA, Spinhirne, JD, Parker, DB, and Cole, NA. 2003. Estimations of 19

ammonia and hydrogen sulphide fluxes from cattle feedlot surfaces in Texas high 20

plains. ASAE 701P1403, 123-130p. 3rd International conference on air pollution 21

from agricultural operations, Research Triangle Park, North Carolina. ASAE: North 22

22

Carolina. 1

Bekku, Y., Koizumi, H., Oikawa, T. and Iwaki, H. 1997. Examination of four methods 2

for measuring soil respiration. Applied Soil Ecology 5(3), 247-254. 3

Bliss, P., Jiang, K. and Schulz, T. 1995. The development of a sampling system for 4

determining odor emission rates from areal surfaces: Part II. Mathematical model. 5

Journal of Air & Waste Management Association 45 989-994. 6

Bockreis, A. and Steinberg, I. 2005. Measurement of odour with focus on sampling 7

techniques: Assessment and Abatement of Odourous Emmissions. Waste 8

Management 25(9), 859-863. 9

Capareda, SC, Boriack, CN, Mukhtar, S, Mutlu, A, and Shaw, BW. 2004. Recovery of 10

gaseous emission from ground level area sources of ammonia and hydrogen sulphide 11

using dynamic isolation flux chambers. 044013p. 2004 ASAE/CSAE Annual 12

International Meeting, Ottawa, Ontario, Canada. ASAE: St Joseph, Michigan. 13

Carpi, A. and Lindberg, S.E. 1998. Application of a teflon(TM) dynamic flux chamber 14

for quantifying soil mercury flux: Tests and results over background soil: 15

Atmospheric Transport, Chemistry and Deposition of Mercury. Atmospheric 16

Environment 32(5), 873-882. 17

Chiou, C.T., Kohnert, R.L., Freed, V.H. and Tonkyn, R.G. 1983. Predictions of 18

evaporative loss rates of solutes in stagnant and turbulent waters in relation to rates of 19

reference materials. Environment International 9(1), 13-17. 20

Chiou, C.T., Freed, V.H., Peters, L.J. and Kohnert, R.L. 1980. Evaporation of solutes 21

23

from water. Environment International 3(3), 231-236. 1

Christensen, S., Ambus, P., Arah, J., Clayton, H., Galle, B., Griffith, D., Hargreaves, K., 2

Klemedtsson, L., Lind, A.-M., Maag, M., Scott, A., Skiba, U., Smith, K., Welling, M. 3

and Wienhold, F. 1996. Nitrous oxide emission from an agricultural field: 4

comparison between measurements by flux chamber and micrometeorological 5

techniques. Atmospheric Environment 30(24), 4183-4190. 6

Clark, J.F., Schlosser, P., Wanninkhof, R., Simpson, H.J., Schuster, W.S.F. and Ho, 7

D.T. 1995. Gas transfer velocities for SF6 and 3He in a small pond at low wind 8

speeds. Geophysical Research Letters 22(2), 93-96. 9

Conen, F. and Smith, K. 1998. A re-examination of closed flux chamber methods for 10

the measurement of trace gas emissions from soils to the atmosphere. European 11

Journal of Soil Science 49 701-707. 12

Crusius, J. and Wanninkhof, R. 2003. Gas transfer velocities measured at low wind 13

speed over a lake. Limnology and Oceanography 48(3), 1010-1017. 14

de Bortoli, M., Ghezzi, E., Knoppel, H. and Vissers, H. 1999. A new test chamber to 15

measure material emissions under controlled air velocity. Environmental Science & 16

Technology 33 1760-1765. 17

de Mello, W. and Hines, M. 1994. Application of static and dynamic enclosures for 18

determining dimethyl sulphide and carbonyl sulphide exchange in Sphagnum 19

peatlands: Implications for the magnitude and direction of flux. Journal of 20

Geophysical Research 99(D7), 14,601-14,607. 21

24

Denmead, O.T., Leuning, R., Jamie, I. and Griffith, D.W.T. 2000. Nitrous oxide 1

emissions from grazed pastures: measurements at different scales. Chemosphere - 2

Global Change Science 2(3-4), 301-312. 3

Di Francesco, F., Ferrara, R. and Mazzolai, B. 1998. Two ways of using a chamber for 4

mercury flux measurement - a simple mathematical approach. The Science of the 5

Total Environment 213 33-41. 6

Dillon, H.K. and Rumph, P.F. 1998. An environmental chamber for investigating the 7

evaporation of volatile chemicals. American Industrial Hygiene Association Journal 8

59(3), 200 9

Duchemin, E., Lucotte, M. and Canuel, R. 1999. Comparison of static chamber and 10

thin boundary equation methods for measuring greenhouse gas emissions from large 11

water bodies. Environmental Science & Technology 33 350-357. 12

Eklund, B., Balfour, W. and Schmidt, C. 1985. Measurement of fugitive volatile 13

organic emission rates. Environmental Progress 4(3), 199-202. 14

Environment Protection Authority, N.S.W. 2001. Draft policy - Assessment and 15

management of odour from stationary sources in NSW (& technical note 8). Sydney, 16

NSW: NSW EPA. 17

Environmental Protection Agency Environmental Protection Agency, (Ed.) 2001. 18

Odour impacts and odour emission control measures for intensive agriculture. R&D 19

Report Series No. 14, OdourNet UK: Bath, UK. 20

Ferrara, R. and Mazzolai, B. 1998. A dynamic flux chamber to measure mercury 21

25

emission from aquatic systems. The Science of the Total Environment 245 51-57. 1

Frechen, F.-B., Frey, M., Wett, M. and Loser, C. 2004. Aerodynamic performance of a 2

low speed wind tunnel. Water Science and Technology 50(4), 57-64. 3

Fukui, Y. and Doskey, P. 1996. An enclosure technique for measuring nonmethane 4

organic compound emissions from grasslands. Journal of Environmental Quality 25 5

601-610. 6

Galvin, G., Lowe, S. and Smith, R. 2004. The Validation Of A Simple Gaussian 7

Dispersion Model For Determining Odour Emission Rates From Area Sources. 8

Developments in Chemical Engineering & Mineral Processing 12(5/6), 1-14. 9

Gao, F and Yates, SR. 1997. Analysis of flux chambers for measuring VOC emissions 10

at soil and water surface. Special conference on control of emissions of odor and 11

volatile organic compounds, Alexandria. ,Water Environment Federation: 12

Alexandria. 13

Gao, F. and Yates, S. 1998a. Laboratory study of closed and dynamic flux chambers: 14

Experimental results and implications for field application. Journal of Geophysical 15

Research 103(D20), 26,115-26,125. 16

Gao, F. and Yates, S. 1998b. Simulation of enclosure-based methods for measuring 17

gas emissions from soil to the atmosphere. Journal of Geophysical Research 18

103(D20), 26,127-26,136. 19

Gao, F., Yates, S., Yates, M., Gan, J. and Ernst, F. 1997. Design, fabrication and 20

application of a dynamic chamber for measuring gas emissions from soil. 21

26

Environmental Science & Technology 31(1), 148-153. 1

Gholson, A., Albritton, J. and Jayanty, R. 1989. Evaluation of the flux chamber 2

method for measuring volatile organic emissions from surface impoundments. 3

EPA/600/3-89/008, US EPA: Research Triangle Park, North Carolina. 4

Gholson, A., Albritton, J. and Jayanty, R. 1991. Evaluation of an enclosure method for 5

measuring emissions of volatile organic compounds from quiescent liquid surfaces. 6

Environmental Science & Technology 25(3), 519-524. 7

Gillis, A. and Miller, D.R. 2000. Some potential errors in the measurement of mercury 8

gas exchange at the soil surface using a dynamic flux chamber. The Science of The 9

Total Environment 260(1-3), 181-189. 10

Gostelow, P., Longhurst, P., Parsons, S. and Stuetz, R. School of Water Sciences, C.U., 11

(Ed.) 2003. Review of odour sampling and measurement. Final report for EB 12

Milton Keynes Limited, Cranfield University: Cranfield University. 13

Gustin, M. and Lindberg, S. 2000. Assessing the contribution of natural sources to the 14

global mercury cycle: The importance of intercomparing dynamic flux 15

measurements. Fresenius Journal Analytical Chemistry 366 417-422. 16

Haghighat, F. and Zhang, Y. 1999. Modelling of emission of volatile organic 17

compounds from building materials - estimation of gas-phase mass transfer 18

coefficient. Building and Environment 34 377-389. 19

Hansen, M., Henriksen, K. and Sommer, S. In press. Observations of production and 20

emission of greenhouse gases and ammonia during storage of solids separated from 21

27

pig slurry: Effects of covering. Atmospheric Environment 1

Heber, A., Ni, J.-Q. and Lim, T. 2002. Odor flux measurements at a facultative swine 2

lagoon stratified by surface aeration. Applied Engineering in Agriculture 18(5), 3

593-602. 4

Hudson, N. and Ayoko, G. Submitted for publication. Odour sampling 1. Physical 5

chemistry considerations. Bioresource Technology 6

Hudson, N., Duperouzel, D., Dunlop, M. and Gallagher, E. Unpublished data. Odour 7

concentration and emission rate data. Unpublished data, Queensland Department of 8

Primary Industries and Fisheries 9

Hudson, N., Galvin, G. and Lowe, S. 2004. The effect of loading rate and spatial 10

variability on pond odour emission, Final report for Australian Pork Limited, Project 11

1628. Queensland Department of Primary Industries & Fisheries: Toowoomba, 12

Queensland. 13

Husted, S. 1993. An open chamber technique for determination of methane emission 14

from stored livestock manure. Atmospheric Environment 27(11), 1635-1642. 15

Jiang, K., Bliss, P. and Schulz, T. 1995. The development of a sampling system for 16

determining odor emission rates from areal surfaces: Part I. Aerodynamic 17

performance. Journal of Air & Waste Management Association 45 917-922. 18

Jiang, K. and Kaye, R. 1996. Comparison study on portable wind tunnel system and 19

isolation chamber for determination of VOC's from areal sources. Water Science and 20

Technology 34(3-4), 583-589. 21

28

Kienbusch, M. 1986. Measurement of gaseous emission rates from land surfaces using 1

an emission isolation flux chamber. EPA/600/8-86/008, US EPA: Environmental 2

Monitoring Systems Laboratory, Las Vegas,Nevada. 3

Kim, K.-H., Kim, M.-Y. and Kim, J. 2005. Application of micrometeorological 4

approaches to measure methane exchange in a dry paddy field in the western caost of 5

Korea. Chemosphere 59 1613-1624. 6

Lague, C, Gaudet, E, Agnew, J, and Fonstad, TA. 2004. Greenhouse gas and odour 7

emissions from liquid swine manure storage facilities in Saskatchewan. 044157p. 8

2004 ASAE/CSAE Annual International Meeting, Ottawa, Ontario, Canada. ASAE: 9

St Joseph, Michigan. 10

Lee, J.-F., Chao, H.-P., Chiou, C. and Manes, M. 2004. Turbulence effects on 11

volatilization rates of liquids and solutes. Environmental Science & Technology 38 12

4327-4333. 13

Leyris, C., Guillot, J.-M., Fanlo, J.-L. and Pourtier, L. 2005. Comparison and 14

development of dynamic flux chambers to determine odorous compound emission 15

rates from area sources. Chemosphere 59(3), 415-421. 16

Lindberg, S., Zhang, H., Vette, A., Gustin, M., Barnett, M. and Kuiken, T. 2002. 17

Dynamic flux chamber measurement of gaseous mercury emission fluxes over soils. 18

Part 2 - effect of flushing flow rate and verification of a two-resistance exchange 19

interface simulation model. Atmospheric Environment 36 847-859. 20

Lindvall, T., Noren, O. and Thyselius, L. 1974. Odor reduction for liquid manure 21

29

systems. Transactions of the American Society of Agricultural Engineers 24 508-1

512. 2

Liu, Q., Bundy, D. and Hoff, S. 1995. A study on the air flow and odor emission rate 3

from a simplified open manure storage tank. Transactions of the American Society of 4

Agricultural Engineers 38(6), 1881-1886. 5

Loubet, B., Cellier, P., Flura, D. and Genermont, S. 1999a. An evaluation of the wind 6

tunnel technique for estimating ammonia volatilization from land: Part 1. Analysis 7

and improvement of accuracy. Journal of Agricultural Engineering Research 72 71-8

81. 9

Loubet, B., Cellier, P., Flura, D. and Genermont, S. 1999b. An evaluation of the wind 10

tunnel technique for estimating ammonia volatilization from land: Part 2. Influence 11

of the tunnel on transfer process. Journal of Agricultural Engineering Research 72 12

83-92. 13

Magliulo, V., Alterio, G. and Peressotti, A. 2004. Experimental and numerical test of 14

the micrometeorological mass difference technique for the measurment of trace gas 15

emissions from small plots. Environmental Science & Technology 38 2693-2700. 16

McGahan, E., Skerman, A., Casey, K. and Streeten, T. 2000. Separation guidelines for 17

Queensland piggeries. Queensland Department of Primary Industries: Toowoomba. 18

Meat & Livestock Australia (MLA) and the Australian Feedlot Producers Association 19

(ALFA). 2002. Development of odour performance criteria for the Australian 20

feedlot industry. Sydney: MLA and ALFA. 21

30

Meisinger, J., Lefcourt, A. and Thompson, R. 2001. Construction and validation of 1

small mobile wind tunnels for studying ammonia volatilization. Applied Engineering 2

in Agriculture 17(3), 375-381. 3

Navaratnasamy, M, Feddes, J, and Edeogu, I. 2004. Comparison of wind tunnel and 4

vented flux chamber in measuring odour emission rates. Paper no. 044184. 5

ASAE/CSAE Annual international meeting, Ottawa, Ontario . ASAE: St Joseph, 6

Michigan. 7

Nicholas, P., Watts, P., Heinrich, N., Hudson, N. and Bell, K. 2004. Development of 8

odour performance criteria for the Australian feedlot industry. Report prepared for 9

MLA, project FLOT.323. Toowoomba, Queensland: FSA Consulting. 10

O'Neill, D. and Phillips, V. 1992. A review of the control of odour nuisance from 11

livestock buildings: Part 3, Properties of the odorous substances which have been 12

identified in livestock wastes or in the air around them. Journal of Agricultural 13

Engineering Research 53 23-50. 14

Parrish, D., Williams, E., Fahey, D., Liu, S. and Fehsenfeld, F. 1987. Measurement of 15

nitrogen oxide fluxes from soils: Intercomparison of enclosure and gradient 16

measurement techniques. Journal of Geophysical Research 92(D2), 2165-2171. 17

Pelletier, F, Godbout, S, Marquis, A, Savard, L-O, Larouche, J-P, Lemay, SP, and 18

Joncas, R. 2004. Greenhouse gas and odour emissions from liquid swine manure 19

storage and manure treatment facilities in Quebec. 044158p. 2004 ASAE/CSAE 20

Annual International Meeting, Ottawa, Ontario, Canada. ASAE: St Joseph, 21

Michigan. 22

31

Peu, P., Beline, F. and Martinez, J. 1999. A floating chamber for estimating nitrous 1

oxide emissions from farm scale treatment units for livestock wastes. Journal of 2

Agricultural Engineering Research 73 101-104. 3

Poach, M.E., Hunt, P.G., Reddy, G.B., Stone, K.C., Matheny, T.A., Johnson, M.H. and 4

Sadler, E.J. 2004. Ammonia Volatilization from Marsh-Pond-Marsh Constructed 5

Wetlands Treating Swine Wastewater. Journal of Environmental Quality 33(3), 6

844-851. 7

Poissant, L. and Casimir, A. 1998. Water-soil and soil-air exchange rate of total 8

gaseous mercury measured at background sites . Atmospheric Environment 32(5), 9

883-893. 10

Raich, J., Bowden, R. and Steudler, P. 1990. Comparison of two static chamber 11

techniques for determining carbon dioxide efflux from forest soils. Soil Science 12

Society of America Journal 54 1754-1757. 13

Rathbun, R. and Tai, D. 1986. Gas-film coefficients for the volatilization of ethylene 14

dibromide from water. Environmental Science & Technology 20(9), 949-952. 15

Rathbun, R. and Tai, D. 1987. Volatilization of ethylene dibromide from water. 16

Environmental Science & Technology 21(3), 248-252. 17

Ravikrishna, R., Valsaraj, K.T., Yost, S., Price, C.B. and Brannon, J.M. 1998. Air 18

emissions from exposed, contaminated sediment and dredged materials: 2. Diffusion 19

from laboratory-spiked and aged field sediments. Journal of Hazardous Materials 20

60(1), 89-104. 21

32

Reichman, R. and Rolston, D.E. 2002. Design and Performance of a Dynamic Gas 1

Flux Chamber. Journal of Environmental Quality 31(6), 1774-1781. 2

Rochette, P., Ellert, B., Gregorich, E., Desjardins, R., Pattey, E., Lessard, R. and 3

Johnson . 1997. Description of dynamic closed chamber for measuring soil 4

respiration and its comparison with other techniques. Canadian Journal of Soil 5

Science 77(2), 195-203. 6

Roelle, P.A. and Aneja, V.P. 2002. Characterization of ammonia emissions from soils 7

in the upper coastal plain, North Carolina. Atmospheric Environment 36(6), 1087-8

1097. 9

Ryden, J. and Lockyer, D. 1985. Evaluation of a system of wind tunnels for field 10

studies of ammonia loass from grassland through volatilisation. Journal of Science of 11

Food and Agriculture 36 781-788. 12

Sadek, S., Lee, T., Smith, J. and Gebel, R. 1998. Flux chamber measurements of mass 13

transfer rates. Journal of Environmental Engineering 124(2), 111-121. 14

Sarwar, G., Corsi, R.L., Kinney, K.A., Banks, J.A., Torres, V.M. and Schmidt, C. 2005. 15

Measurements of ammonia emissions from oak and pine forests and development of a 16

non-industrial ammonia emissions inventory in texas. Atmospheric Environment 17

39(37), 7137-7153. 18

Sattler, ML and McDonald, ET. 2002. Time, temperature and wind speed: the 19

importance of varying emission rates in odor modelling. Odors and Toxics Air 20

Emissions 2002, Albuquerque, New Mexico, USA. Albuquerque, New Mexico, USA. 21

33

Schiffman, S., Bennett, J. and Raymer, J. 2001. Quantification of odors and odorants 1

from swine operations in North Carolina. Agricultural and Forest Meteorology 108 2

213-240. 3

Schmidt, D. R., Bicudo, J. R., and Janni, K. 1999. Determining odor emissions from 4

manure surfaces using a wind tunnel. Paper No. 994033p. ASAE Annual 5

International Meeting, ASAE: St Joseph, Michigan. 6

Sironi, S., Capelli, L., Centola, P., Del Rosso, R. and Il Grande, M. 2005. Odour 7

emission factors for assessment and prediction of Italian MSW landfills odour 8

impact. Atmospheric Environment 39(29), 5387-5394. 9

Skerman, A. 2000. Reference manual for the establishment and operation of beef cattle 10

feedlots in Queensland. Brisbane, Queensland: Department of Primary Industries & 11

Fisheries. 12

Skiba, U., DiMarco, C., Hargreaves, K., Sneath, R. and McCartney, L. 2006. Nitrous 13

oxide emissions from a dung heap measured by chambers and plume methods. 14

Agriculture, Ecosystems & Environment 112(2-3), 135-139. 15

Smith, R. 1995. A gaussian model for estimating odour emissions from area sources. 16

Mathematical Computer Modelling 22(9), 23-29. 17

Smith, R.J. 1993. Dispersion of odours from ground level agricultural sources. Journal 18

of Agricultural Engineering Research 54 187-200. 19

Smith, R. and Watts, P. 1994. Determination of odour emission rates from cattle 20

feedlots: Part 2, Evaluation of two wind tunnels of different size. Journal of 21

34

Agricultural Engineering Research 58 231-240. 1

Sohn, J.H., Smith, R.J., Hudson, N.A. and Choi, H.L. 2005. Gas Sampling Efficiencies 2

and Aerodynamic Characteristics of a Laboratory Wind Tunnel for Odour 3

Measurement. Biosystems Engineering 92(1), 37-46. 4

Sommer, S., McGinn, S., Hao, X. and Larney, F. 2004. Techniques for measuring gas 5

emissions from a composting stockpile of cattle manure. Atmospheric Environment 6

38 4643-4652. 7

Staudinger, J. and Roberts, P. 1996. A critical review of Henry's law constants for 8

environmental applications. Critical Reviews in Environmental Science and 9

Technology 26(3), 205-297. 10

Streeten, T. and McGahan, E. 2000. Environmental Code of Practice for Queensland 11

Piggeries. Brisbane, Queensland: Department of Primary Industries. 12

Taha, M., Pollard, S., Sarkar, U. and Longhurst, P. 2005. Estimating fugitive 13

bioaerosol releases from static compost windrows: Feasibility of a portable wind 14

tunnel approach. Waste Management 25 445-450. 15

Thornton, F., Shurpali, N., Bock, B. and Reddy, K. 1998. N2O and NO emissions from 16

poultry litter and urea applications to Bermuda grass. Atmospheric Environment 17

32(9), 1623-1630. 18

Valsaraj, K.T., Choy, B., Ravikrishna, R., Reible, D.D., Thibodeaux, L.J., Price, C.B., 19

Brannon, J.M. and Myers, T.E. 1997. Air emissions from exposed, contaminated 20

sediments and dredged materials 1. Experimental data in laboratory microcosms and 21

35

mathematical modelling. Journal of Hazardous Materials 54(1-2), 65-87. 1

Wang, D., Yates, S., Ernst, F., Knuteson, J. and Brown, G. 2001a. Volatilization of 2

1,3-dichloropropene under different application methods. Water, Air and Soil 3

Pollution 127 109-123. 4

Wang, X, Jiang, J, and Kaye, R. 2001b. Improvement of wind tunnel sampling system 5

for odour and VOC's. 147-154 p. 1st IWA International Conference on Odour and 6

VOCs: Measurement, Regulation and Control Techniques, University of New South 7

Wales, Sydney. 8

Wanninkhof, R., Ledwell, J. and Broecker, W. 1985. Gas exchange-wind speed 9

relation measured with sulphur hexafluoride on a lake. Science 227(4691), 1224-10

1226. 11

Wright, D., Eaton, D., Nielsen, L., Kuhrt, F., Koziel, J., Spinhirne, J. and Parker, D. 12

2005. Multidimensional gas chromatography-olfactometry for identification and 13

prioritization of malodors from confined animal feeding operations. Journal of 14

Agricultural and Food Chemistry 53 8663-8672. 15

Xiao, Z.F., Munthe, J., Schroeder, W.H. and Lindqvist, O. 1991. Vertical fluxes of 16

volatile mercury over forest soil and lake surfaces in Sweden. Tellus B 43(3), 267-17

279. 18

Yamulki, S., Goulding, K., Webster, C. and Harrison, R. 1995. Studies on NO and 19

N2O fluxes from a wheat field. Atmospheric Environment 29(14), 1627-1635. 20

Yamulki, S., Harrison, R. and Goulding, K. 1996. Ammonia surface-exchange above 21

36

an agricultural field in southeast England. Atmospheric Environment 30(1), 109-1

118. 2

Zahn, J., DiSpirito, A., Do, Y., Brooks, B., Cooper, E. and Hatfield, J. 2001. 3

Correlation of human olfactometry responses to airborne concentrations of 4

malodorous volatile organic compounds emitted from swine effluent. Journal of 5

Environmental Quality 30 624-634. 6

Zhang, H., Lindberg, S., Barnett, M., Vette, A. and Gustin, M. 2002. Dynamic flux 7

chamber measurement of gaseous mercury emission fluxes over soils. Part 1: 8

simulation of gaseous mercury emissions from soils using a two-resistance exchange 9

interface model. Atmospheric Environment 36 835-846. 10

Zhang, H., Lindberg, S., Marsik, F. and Keeler, G. 2001. Mercury air/surface exchange 11

kinetics of background soils of the Tahquamenon River watershed in the Michigan 12

upper peninsula. Water, Air and Soil Pollution 126 151-169. 13

Zhang, J.S., Shaw, C.Y., Kanabus-Kaminska, J.M., MacDonald, R.A., Magee, R.J., 14

Lusztyk, E. and Weichert, H.J. 1996. Study of air velocity and turbulence effects on 15

organic compound emissions from building materials/furnishings using a new small 16

test chamber. ASTM Special Technical Publication 184-199.17

37

1

Tables 2

3

Table 1. Physical characteristics and typical operating conditions of representative cylindrical sampling devices. 4

5

Shapea Radius

(m) Height

(m) Base area

(m2) Chamber volume

(m3) A/V ratio

(loading ratio) H/R ratio Typical flushing rate

(m3/s) Area-related sweep rate

(m3/m2/hour) Exchange rate

(/hour) Exchange period

(sec) Reference

C 0.076 0.12 0.018 0.0022 8.182 1.579 0 0 0 N/A (Sommer et al., 2004) C 0.0805 N/Ab 0.02 N/A N/A N/A 0 0 N/A N/A (Bockreis and Steinberg, 2005) C 0.1375 0.2 0.059 0.0118 5 1.455 0 0 0 N/A (Raich et al., 1990) C 0.1435 0.083 0.065 0.0054 12.037 0.578 0 0 0 N/A (Raich et al., 1990) C 0.2 0.13 0.126 0.0164 7.683 0.65 0 0 0 N/A (Conen and Smith, 1998) C 0.2 0.13 0.126 0.0164 7.683 0.65 0 0 0 N/A (Conen and Smith, 1998) C 0.2 0.1 0.126 0.0126 10 0.5 0 0 0 N/A (Skiba et al., 2006) C 0.325 0.2 0.332 0.0664 5 0.615 0 0 0 N/A (Rochette et al., 1997)

CS 0.203 0.203 0.13 0.0175 7.429 1 0.00003 0.72 5.16 698 (Poissant and Casimir, 1998) CS 0.205 0.18 0.132 0.03 4.4 0.878 0.00003 0.72 4.02 896 (Gholson et al., 1991) C 0.1875 N/A 0.11 0.055 2 N/A 0.00004 1.44 2.7 1333 (Sadek et al., 1998) C 0.2 0.1 0.126 0.0126 10 0.5 0.00005 1.44 14.28 252 (Skiba et al., 2006)

CS 0.205 0.18 0.132 0.03 4.4 0.878 0.00005 1.44 6 600 (Sarwar et al., 2005) CS 0.1325 0.42 0.055 0.0231 2.381 3.17 0.00008 5.4 12.96 278 (Baek et al., ) CS 0.13275 0.472 0.055 0.026 2.115 3.556 0.00008 5.4 11.52 312 (Roelle and Aneja, 2002) CS 0.2032 0.1778 0.13 0.03 4.333 0.875 0.00008 2.16 10.02 359 (Kienbusch, 1986) CS 0.205 0.18 0.132 0.03 4.4 0.878 0.00008 2.16 10.02 359 (Gholson et al., 1991) CS 0.205 0.18 0.132 0.03 4.4 0.878 0.00008 2.16 10.02 359 (Sarwar et al., 2005)

38

Shapea Radius

(m) Height

(m) Base area

(m2) Chamber volume

(m3) A/V ratio

(loading ratio) H/R ratio Typical flushing rate

(m3/s) Area-related sweep rate

(m3/m2/hour) Exchange rate

(/hour) Exchange period

(sec) Reference

CS 0.2475 0.65 0.193 0.0645 2.992 2.626 0.00008 1.44 4.68 769 (Baek et al., ) C 0.3 1 0.283 0.283 1 3.333 0.00009 1.08 1.2 3000 (Lague et al., 2004) C 0.1275 0.472 0.051 0.0241 2.116 3.702 0.00011 7.92 16.18 222 (Baek et al., 2003)

CS 0.2475 0.309 0.193 0.0645 2.992 1.248 0.00012 2.16 6.51 553 (Capareda et al., 2004) CS 0.205 0.18 0.132 0.03 4.4 0.878 0.00017 4.68 19.98 180 (Gholson et al., 1991) C 0.087 0.11 0.024 0.0022 10.909 1.264 0.00025 37.44 409.08 9 (Lindberg et al., 2002) C 0.096 0.035 0.029 0.001 29 0.365 0.00025 30.96 900 4 (Lindberg et al., 2002) C 0.24 0.09 0.045 0.0022 20.455 0.375 0.00033 26.28 545.46 7 (Lindberg et al., 2002) C 0.28 0.18 0.067 0.0112 5.982 0.643 0.0005 27 160.74 22 (Lindberg et al., 2002)

a C indicates a cylindrical footprint, CS indicates a cylindrical footprint with a domed top 1

b N/A indicates that this data was not available from the original publication 2

3

4

5

6

39

1

Table 2. Physical characteristics and typical operating conditions of representative rectangular sampling devices. 2

3 BLa

Compliance? Length (m)

Width (m)

Height (m)

Base area (m2)

Face area (m2)

Chamber volume

(m3)

A/V ratio (loading ratio)

(m2/m3) L/W ratio

(m/m) H/R ratio

(m/m)

Typical flushing

rate (m3/s)

Linear velocity (m/s)

Area-related sweep rate

(m3/m2/hour)

Exchange rate

(/hour)

Exchange period

(s) Hb Vc Reference

0.1 0.03 0.002 0.003 0.00006 0.00001 300 3.333 0.02 0.00002 0.3333 24.12 6000 1 Yes No (Valsaraj et al., 1997)

0.1 0.1 0.15 0.01 0.015 0.0015 6.667 1 1.5 0 0 0 0 N/Ad N/A N/A (Gao and Yates, 1998a)12

0.15 0.25 0.002 0.038 0.0005 0.0001 380 0.6 0.013 0.00003 0.06 2.88 1020 4 Yes No (Ravikrishna et al., 1998) 0.198 0.2 0.05 0.04 0.01 0.002 20 0.99 0.253 0.00035 0.035 31.68 630 6 Yes Yes (Gao and Yates, 1997) 0.198 0.2 0.05 0.04 0.01 0.002 20 0.99 0.253 0.00005 0.005 4.68 90 40 No No (Gao and Yates, 1998a) 0.198 0.2 0.05 0.04 0.01 0.002 20 0.99 0.253 0.0022 0.22 198 3960 1 Yes Yes (Gao and Yates, 1998a)

0.3 0.1 0.1 0.03 0.01 0.003 10 3 0.333 0.00025 0.025 29.88 300 12 Yes Yes (Lindberg et al., 2002) 0.3 0.3 0.5 0.09 0.15 0.045 2 1 1.667 0 0 0 0 N/A N/A N/A (Duchemin et al., 1999)

0.3 0.3 0.3 0.09 0.09 0.027 3.333 1 1 0 0 0 0 N/A N/A N/A (de Mello and Hines, 1994)

0.3 0.3 0.3 0.09 0.09 0.027 3.333 1 1 0.00005 0.0006 2.16 6.66 541 No No (de Mello and Hines, 1994)

0.3 0.3 0.31 0.09 0.093 0.0279 3.226 1 1.033 0.00007 0.0008 2.88 8.58 420 No No (Parrish et al., 1987) 0.36 0.36 0.1 0.13 0.036 0.013 10 1 0.278 0.00003 0.0008 0.72 9.24 390 No No (Yamulki et al., 1996)

0.4 0.35 0.05 0.14 0.018 0.007 20 1.143 0.125 0.00097 0.0539 24.84 497.16 7 Yes Yes (Haghighat and Zhang, 1999)

0.4 0.4 0.115 0.16 0.046 0.0184 8.696 1 0.288 0.00017 0.0037 3.96 32.58 111 No No (Reichman and Rolston, 2002)

0.405 0.402 0.095 0.163 0.038 0.0155 10.516 1.007 0.235 0.00041 0.0108 9 94.86 38 Yes No (Gao et al., 1997) 0.415 0.115 0.08 0.048 0.009 0.0038 12.632 3.609 0.193 0.00013 0.0144 9.72 120 30 Yes No (Wang et al., 2001a) 0.415 0.415 0.08 0.172 0.033 0.0138 12.464 1 0.193 0.0005 0.0152 10.44 130.44 28 Yes No (Wang et al., 2001a) 0.43 0.29 0.16 0.125 0.046 0.02 6.25 1.483 0.372 0.00023 0.005 6.48 42 86 No No (de Bortoli et al., 1999)

0.5 0.1 0.035 0.05 0.004 0.0018 27.778 5 0.07 0.00001 0.0025 0.72 23.33 154 No No (Di Francesco et al., 1998)

0.5 0.2 0.075 0.1 0.015 0.0075 13.333 2.5 0.15 0.00003 0.002 1.08 16.02 225 No No (Yamulki et al., 1995) 0.5 0.2 0.075 0.1 0.015 0.0075 13.333 2.5 0.15 0 0 0 0 N/A N/A N/A (Yamulki et al., 1995) 0.5 0.25 0.08 0.125 0.02 0.01 12.5 2 0.16 0.006 0.3 172.8 2160 2 Yes Yes (Sironi et al., 2005) 0.55 0.6 0.08 0.33 0.048 0.0264 12.5 0.917 0.145 0.00317 0.066 34.56 431.82 8 Yes Yes (Zhang et al., 1996)

40

BLa Compliance? Length

(m) Width

(m) Height

(m)

Base area (m2)

Face area (m2)

Chamber volume

(m3)

A/V ratio (loading ratio)

(m2/m3) L/W ratio

(m/m) H/R ratio

(m/m)

Typical flushing

rate (m3/s)

Linear velocity (m/s)

Area-related sweep rate

(m3/m2/hour)

Exchange rate

(/hour)

Exchange period

(s) Hb Vc Reference

0.6 0.025 0.05 0.015 0.001 0.0008 18.75 24 0.083 0.000003 0.003 0.72 11.25 320 No No (Allaire et al., 2002)

0.6 0.2 0.2 0.12 0.04 0.024 5 3 0.333 0.0001 0.0025 2.88 15 240 No No (Ferrara and Mazzolai, 1998)

0.6 0.2 0.2 0.12 0.04 0.024 5 3 0.333 0.0005 0.0125 15.12 75 48 Yes Yes (Lindberg et al., 2002)

0.6 0.2 0.2 0.12 0.04 0.024 5 3 0.333 0.00008 0.002 2.52 12.48 288 No No (Carpi and Lindberg, 1998)

0.6 0.2 0.2 0.12 0.04 0.024 5 3 0.333 0.0001 0.0025 2.88 15 240 No No (Zhang et al., 2001) 0.6 0.3 0.3 0.18 0.09 0.054 3.333 2 0.5 0.00006 0.0007 1.08 4.02 896 No No (Gillis and Miller, 2000) 0.6 0.3 0.3 0.18 0.09 0.054 3.333 2 0.5 0.00018 0.002 3.6 12 300 No No (Gillis and Miller, 2000) 0.6 0.3 0.2 0.18 0.06 0.036 5 2 0.333 0.00005 0.0008 1.08 4.98 723 No No (Gillis and Miller, 2000) 0.6 0.3 0.2 0.18 0.06 0.036 5 2 0.333 0.0002 0.0033 3.96 19.98 180 No No (Gillis and Miller, 2000) 0.6 0.3 0.15 0.18 0.045 0.027 6.667 2 0.25 0.00006 0.0013 1.08 7.56 476 No No (Gillis and Miller, 2000) 0.6 0.3 0.15 0.18 0.045 0.027 6.667 2 0.25 0.00019 0.0042 3.96 24.9 145 No No (Gillis and Miller, 2000) 0.6 0.4 0.375 0.24 0.15 0.0845 2.84 1.5 0.625 0.00003 0.0002 0.36 1.08 3333 No No (Fukui and Doskey, 1996) 0.6 0.4 0.21 0.24 0.084 0.0504 4.762 1.5 0.35 0.00042 0.005 6.48 29.76 121 Yes No (Peu et al., 1999) 0.7 0.2 0.214 0.14 0.043 0.03 4.667 3.5 0.306 0.00003 0.0007 0.72 3 1200 No No (Christensen et al., 1996) 0.7 0.2 0.3 0.14 0.06 0.042 3.333 3.5 0.429 0.00002 0.0003 0.36 1.44 2500 No No (Husted, 1993) 0.7 0.55 0.72 0.385 0.396 0.2772 1.389 1.273 1.029 0.00042 0.0011 3.96 5.4 667 No No (Dillon and Rumph, 1998) 0.76 0.45 0.2 0.342 0.09 0.0684 5 1.689 0.263 0 0 0 0 N/A N/A N/A (Thornton et al., 1998) 0.8 0.2 0.2 0.16 0.04 0.032 5 4 0.25 0.00003 0.0008 0.72 2.82 1277 No No (Xiao et al., 1991) 0.8 0.4 0.25 0.32 0.1 0.08 4 2 0.313 0.03333 0.3333 375.12 1500 2 Yes Yes (Jiang et al., 1995)

0.8 0.4 0.25 0.32 0.08 4 2 0.313 0.015 N/A 168.84 675 5 N/A N/A (Navaratnasamy et al., 2004)

0.8 0.4 0.25 0.32 0.1 0.08 4 2 0.313 0.03 0.3 337.68 1350 3 Yes Yes (Navaratnasamy et al., 2004)

0.8 0.4 0.25 0.32 0.1 0.08 4 2 0.313 0.03 0.3 337.68 1350 3 Yes Yes (Schmidt et al., 1999) 0.8 0.4 0.25 0.32 0.1 0.08 4 2 0.313 0.019 0.19 213.84 855 4 Yes Yes (Taha et al., 2005) 1 0.1 0.2 0.1 0.02 0.02 5 10 0.2 0.06 3 2160 10800 0 Yes Yes (Leyris et al., 2005) 1 0.25 0.2 0.25 0.05 0.05 5 4 0.2 0.05 1 720 3600 1 Yes Yes (Smith and Watts, 1994) 1 0.33 0.08 0.33 0.026 0.0264 12.5 3.03 0.08 0.00068 0.0262 7.56 93.18 39 Yes Yes (Frechen et al., 2004) 1 0.5 0.5 0.5 0.25 0.25 2 2 0.5 0.025 0.1 180 360 10 Yes Yes (Sohn et al., 2005)

1.2 0.15 0.18 0.18 0.027 0.0324 5.556 8 0.15 0.081 3 1620 9000 0 Yes Yes (Leyris et al., 2005)

41

BLa Compliance? Length

(m) Width

(m) Height

(m)

Base area (m2)

Face area (m2)

Chamber volume

(m3)

A/V ratio (loading ratio)

(m2/m3) L/W ratio

(m/m) H/R ratio

(m/m)

Typical flushing

rate (m3/s)

Linear velocity (m/s)

Area-related sweep rate

(m3/m2/hour)

Exchange rate

(/hour)

Exchange period

(s) Hb Vc Reference

1.5 1 0.4 1.5 0.4 0.6 2.5 1.5 0.267 0.03833 0.0958 92.16 229.98 16 Yes Yes (Lindvall et al., 1974) 1.52 0.25 0.115 0.38 0.029 0.0437 8.696 6.08 0.076 0.00067 0.0231 6.48 54.9 66 Yes Yes (Pelletier et al., 2004) 1.6 1.3 0.51 2.08 0.663 1.0608 1.961 1.231 0.319 0.1326 0.2 229.68 450 8 Yes Yes (Gholson et al., 1991) 1.6 1.3 0.51 2.08 0.663 1.0608 1.961 1.231 0.319 0.3315 0.5 573.84 1125 3 Yes Yes (Gholson et al., 1991) 2.44 0.305 0.25 0.744 0.076 0.1861 3.998 8 0.102 0.07667 1.0088 371.16 1483.08 2.4 Yes Yes (Heber et al., 2002) 2.9 2.22 1.23 6.438 2.731 7.9187 0.813 1.306 0.424 3.8 1.3914 2124.72 1727.58 2 Yes Yes (Liu et al., 1995) 4 1 2.5 4 2.5 10 0.4 4 0.625 3.75 1.5 3375 1350 3 Yes Yes (Poach et al., 2004)

15.6 7.3 4.5 113.88 32.85 512.46 0.222 2.137 0.288 3.75 0.1142 118.44 26.34 137 Yes Yes (Hansen et al., In press)

0 0 0 0 0 0 N/A N/A N/A 0 N/A N/A N/A N/A N/A N/A (Sattler and McDonald, 2002)

2 0.5 0.45 1 0.225 0.45 2.222 4 0.225 0.1125 0.5 405 900 4 Yes Yes (Meisinger et al., 2001)

2 0.5 0.45 1 0.225 0.45 2.222 4 0.225 0.225 1 810 1800 2 Yes Yes (Ryden and Lockyer, 1985)

2 0.5 0.45 1 0.225 0.45 2.222 4 0.225 0.225 1 810 1800 2 Yes Yes (Smith and Watts, 1994) 2 0.5 0.45 1 0.225 0.45 2.222 4 0.225 0.125 0.5556 450 1000.02 4 Yes Yes (Loubet et al., 1999a) 2 0.5 0.45 1 0.225 0.45 2.222 4 0.225 0.25 1.1111 900 1999.98 2 Yes Yes (Loubet et al., 1999b)

a BL = Boundary layer; compliance indicates that well-developed boundary layer conditions are likely under typical operating conditions, where 1 mass transfer is likely to be a function of wind velocity within the device 2

b H = Horizontal plane 3 c V = Vertical plane 4

d N/A data not available from the original publication 5

6

42

1

2

Table 3. Comparison of physical characteristics of static cylindrical sampling devices (n = 8) 3

4

Statistic Radius (m) Height (m) Base area (m2)

Chamber volume (m3)

A/V ratio (loading ratio)

(m2/m3) H/R ratio

(m/m)

Typical flushing

rate (m3/s)

Linear velocity (m/s)

Area-related sweep rate (m3/m2/hour)

Exchange rate(/hour)

Exchange period

(s)

Average 0.17 0.138 0.109 0.019 7.941 0.861 N/A N/A N/A N/A N/A

Median 0.172 0.13 0.0955 0.0126 7.683 0.65 N/A N/A N/A N/A N/A

Minimum 0.076 0.083 0.018 0.0022 5 0.5 N/A N/A N/A N/A N/A

Maximum 0.325 0.2 0.332 0.0664 12.037 1.579 N/A N/A N/A N/A N/A

Ratio, maximum: minimum

4.3 2.4 18.4 30.2 2.4 3.2 N/A N/A N/A N/A N/A

5

43

1

Table 4. Comparison of physical characteristics of dynamic cylindrical sampling devices (n = 19) 2

3

Statistic Radius (m) Height (m) Base area (m2)

Chamber volume (m3)

A/V ratio (loading

ratio) (m2/m3)

H/R ratio (m/m)

Typical flushing rate

(m3/s)

Linear velocity

(m/s)

Area-related sweep rate

(m3/m2/hour)

Exchange rate (/hour)

Exchange period

(s)

Average 0.2 0.3 0.1 0.04 6.6 1.5 0.0001 N/A 8.5 113.2 537.5 Median 0.205 0.18 0.13 0.03 4.4 0.878 0.00008 N/A 2.16 10.02 359

Minimum 0.087 0.035 0.024 0.001 1 0.365 0.00003 N/A 0.72 1.2 4 Maximum 0.3 1 0.283 0.283 29 3.702 0.0005 N/A 37.44 900 3000

Ratio, maximum: minimum

3.4 28.6 11.8 283 29 10.1 16.7 N/A 52 750 750

4

5

44

1

Table 5. Comparison of physical characteristics of static rectangular sampling devices (n = 5) 2

3

Device description

Length (m)

Width(m)

Height (m)

Base area (m2)

Face area (m2)

Chamber volume

(m3)

A/V ratio

(loading ratio)

(m2/m3)

L/W ratio

(m/m)

H/R ratio

(m/m)

Typical flushing

rate (m3/s)

Linear velocity (m/s)

Area-related sweep rate

(m3/m2/hour)

Exchange rate

(/hour)

Exchange period

(s)

Average 0.327 0.225 0.204 0.105 0.06 0.025 6.067 1.438 0.916 N/A N/A N/A N/A N/A

Median 0.3 0.25 0.175 0.09 0.0525 0.01725 5 1 1 N/A N/A N/A N/A N/A

Minimum 0.1 0.1 0.075 0.01 0.015 0.0015 2 1 0.15 N/A N/A N/A N/A N/A

Maximum 0.76 0.45 0.5 0.342 0.15 0.0684 13.333 2.5 1.667 N/A N/A N/A N/A N/A

Ratio, maximum: minimum

7.6 4.5 6.7 34.2 10 45.6 6.7 2.5 11.1 N/A N/A N/A N/A N/A

4

5

45

1

Table 6. Comparison of physical characteristics of dynamic rectangular sampling devices (n = 57) 2

3

Device description

Length (m)

Width(m)

Height (m)

Base area (m2)

Face area (m2)

Chamber volume

(m3)

A/V ratio

(loading ratio)

(m2/m3)

L/W ratio

(m/m)

H/R ratio

(m/m)

Typical flushing

rate (m3/s)

Linear velocity (m/s)

Area-related sweep rate

(m3/m2/hour)

Exchange rate

(/hour)

Exchange period

(s)

Average 1.175 0.528 0.362 2.523 0.78 9.428 13.28 2.649 0.317 0.231 0.323 297.973 935.74 273.735

Median 0.6 0.3 0.2 0.18 0.049 0.0324 5 2 0.288 0.0005 0.02405 10.44 120 38

Minimum 0.15 0.1 0.002 0.03 0.0005 0.0001 0.222 0.6 0.013 0.00001 0.0002 0.36 1.08 1

Maximum 15.6 7.3 4.5 113.88 32.85 512.46 380 10 1.033 3.8 3 3375 10800 3333

Ratio, maximum: minimum

104 73 2250 3796 65700 5124600 1711.7 16.7 79.5 380000 15000 9375 10000 3333

4

5

6

7

1

1

Table 7. Bias in measured emission rate reported for the US EPA 2

dynamic emission chamber 3

4

Chemical Dimensionless Henry constant H’ (Staudinger and Roberts, 1996)

Solubility,

g/100 g

Bias in emission

rate relative to wind

tunnel, %

Reference

Acetone 0.00136 Infinite -82 (Jiang and Kaye, 1996)

Methyl ethyl ketone

0.00153 35 -68 (Gholson et al., 1989)

Toluene 0.273 0.05 -38 (Gholson et al., 1989)

1,1,1-trichloroethylene

0.318 0.1 -21 (Gholson et al., 1989)

5

6

7