Ammonia emissions from an in-ground finisher hog manure tank

Atmospheric Environment 190 (2018) 43–52

Contents lists available at ScienceDirect

Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv

Ammonia emissions from an in-ground finisher hog manure tank ∗

T

Richard H. Grant , Matthew T. Boehm Department of Agronomy, Purdue University, USA

A R T I C LE I N FO

A B S T R A C T

Keywords: Livestock emissions Ammonia emissions

A large fraction of emitted and ultimately deposited ammonia (NH3) originates from swine waste storage. Understanding the factors that influence these emissions is important in determining how to mitigate NH3 loading of the atmosphere. Hog manure is stored in slurry pits, tanks, or lagoons. Ammonia emissions from a ground-level mid-western hog finisher manure tank collecting manure from a mean of 3508 animals was measured for 8–20 d each quarter of the year for two years. Forty of the 164 measurement days had sufficient measurements to represent entire days. These daily emissions averaged 0.44 g NH3 m−2h−1 (53.1 g d−1 AU−1, 7.2 g d−1 hd−1: AU = 500 kg animal mass, hd = 1 animal) with maximum emissions of 1.62 g NH3 m−2h−1 (194.2 g d−1 AU−1, 26.4 g d−1 hd−1). Emissions from the tank were greater on an area basis but comparable on an animal basis relative to emissions from much larger anaerobic lagoons. Emissions were correlated with air temperature and manure composition, but not wind speed or friction velocity- probably due to the turbulence created by the tank structure under all winds. Crusting of the manure surface in the tank corresponded with a non-significant 10% increase in NH3 emissions. A model of the tank emissions, based on nominal nitrogen loading rates from the literature and taking into account the configuration of the circular tank, had a mean bias error of 0.01 g NH3 m−2h−1 (0.9 g d−1 AU−1, 0.1 g d−1 hd−1) and a root mean square error of 0.26 g NH3m−2h−1 (31.6 d−1 AU−1, 4.3 g d−1 hd−1). Additional emissions measurements from such ground-level manure storage tanks with greater documentation of the manure liquid composition are needed to verify the modeling approach.

1. Introduction Animal agriculture is a significant source of ammonia (NH3) emitted into the atmosphere. The emission of NH3 contributes to long-term global warming through rapid deposition and subsequent production of N2O (USEPA, 2011). Ammonia emissions also contribute to the formation of PM2.5 (USEPA, 2011). Agricultural operations are required to report emission that exceed 220 kg d−1 in compliance with the Emergency Planning and Community Right-To-Know Act in the United States (Centner and Patel, 2010). Emissions from manure storage facilities are a significant fraction of the total farm NH3 emissions. Manure from hog operations is typically stored in a lagoon, a pit/tank, an above-ground tank, or a below ground concrete tank. Manure storages are designed, in part, on the basis of manure loading over the maximum storage period (ASAE, 1998; USDA, 2016). Since the manure loading depends on the transfer of excreted manure, the number and mass of the hogs excreting

is important in defining the necessary size of the storage. While anaerobic lagoons are sized to account for the manure loading and sludge storage for many years, dilution water to encourage bacterial activity, and storage of precipitation less evaporation; manure storage tanks need only handle the manure loading and sludge storage for less than one year, and storage of precipitation (less evaporation) (Worley, 2015). Consequently, the storage of the same manure loading requires much less volume for a tank than for an anaerobic lagoon. Szögi et al. (2005) measured emissions from an anaerobic hog finishing waste lagoon of 0.156 g m−2h−1 (9.75 g NH3 hd−1 d−1) during nine measurement days distributed across a year. Shores et al. (2005) measured NH3 emissions of 0.60 g m−2h−1 (44.1 g hd−1 d−1) from an anaerobic lagoon on one day during July at a finishing farm in North Carolina. Zahn et al. (2001) measured average emissions of 0.66 g m−2h−1 (22.7 g NH3 hd−1 d−1 from an anaerobic lagoon at a Missouri finishing operation during a 14-day measurement campaign in

Abbreviations: a, air over tank; AU, 500 kg animal mass; b, laminar air; BG, background; bLS, backward Lagrangian stochastic; F, fraction of loading; g, conductance; H, Henry's coefficient; hd, 1 animal; K, bulk exchange coefficient; LM, live mass; L, volumetric loading rate; LIQ, liquid; MDL, minimum detection limit; NAEMS, National Air Emissions Monitoring Study; OP, optical path; RPM, Radial plume mapping; s, surface; SA, surface area; Sc, Schmidt number; SD, standard deviation; TDLAS, tunable diode laser absorption spectrometer; TS, total solids; u*, friction velocity; V, volume; β, Van't Hoff temperature coefficient ∗ Corresponding author. 915 W. State St, West Lafayette, IN, USA. E-mail address: [email protected] (R.H. Grant). https://doi.org/10.1016/j.atmosenv.2018.07.009 Received 4 March 2018; Received in revised form 5 July 2018; Accepted 7 July 2018 Available online 09 July 2018 1352-2310/ © 2018 Elsevier Ltd. All rights reserved.

Atmospheric Environment 190 (2018) 43–52

R.H. Grant, M.T. Boehm

Fig. 1. Configuration of study site. Locations of the five retroreflectors on each cardinal direction from the tank are indicated with a prefix of direction from the tank. Reroreflectors 3, 4, and 5 were mounted on towers while 1 and 2 were on tripods. Scanning TDLAS units are indicated by ‘TS’. Image from GoogleEarth®.

measurements of NH3 emissions from tank manure storages have been reported (Sommer, 1997; Gay et al., 2003). Gay et al. (2003) measured the NH3 emissions from a hog finishing farm manure tank using a low speed wind tunnel on the liquid surface on four days and found emissions ranging from 0.306 g m−2h−1 to 2.43 g m−2h−1 with an average of 1.44 g m−2h−1. Sommer (1997) found emissions of up to 1.5 g m−2h−1 using the integrated horizontal flux measurement method for an uncovered slurry tank in Denmark with much higher total N content than typical anaerobic lagoons (e.g. Grant et al., 2013a, 2013b). These two tank emission studies suggest emissions from tanks are much larger than from anaerobic lagoons on an area basis. Sommer (1997) suggested that the relatively high NH3 emission rates from tanks in the field may be due to either a greater NH3 concentration gradient between the liquid and air overlying the liquid or an enhanced transfer rate. Unfortunately, while N content of the manure was reported, there was no information to express the emissions on an animal basis. It is also possible that the exposure of the tank at the ground surface may contribute to the emissions in a similar manner that the configuration of stomatal pores on a leaf influences mass transfer between the leaf and the atmosphere (Parlange and Waggoner, 1970); Montieth and Unsworth, 1990). Crusting of the surface of the stored manure commonly occurs as the DM increases to up to 6% (Wood et al., 2012). Increases in DM and the formation of crusts can occur as increased air temperatures increase evaporation of the stored manure (Misselbrook et al., 2005; Smith et al., 2007). Deeper tanks, with smaller SA:V ratio were found to increase incidence of crusting compared to shallow tanks through providing greater DM under the exposed surface (Smith et al., 2007). Crusting has been shown to decrease NH3 emissions (Sommer et al., 1993; Olesen and Sommer, 1993; Misselbrook et al., 2005; Wood et al., 2012). It is presumed that the decrease in emissions is largely a result of increased resistance to diffusive transport from the liquid to the overlying air (Sommer et al., 1993, 2000; Olesen and Sommer, 1993). However nitrification, denitrification, NO3− leaching may also contribute to the

late summer and early fall. Grant et al. (2016) measured emissions from anaerobic lagoons at finishing farms in two states, finding average daily emissions of 0.18 g m−2h−1 (104 g AU−1 d−1, 16.1 g hd−1 d−1) at a farm in Oklahoma based on 59 d of measurements across two years and of 0.08 g m−2h−1 (36 g AU−1 d−1, 4.4 g hd−1 d−1) at a farm in North Carolina based on 16 d of measurements across two years. The type of the manure storage facility may contribute to the differences in emissions. A laboratory study showed that NH3 emissions decreased as the surface area (SA) to volume (V) ratio of a constant volume of manure decreased from 0.99 to 0.03 m2m-3 (Sievers et al., 2000). Sievers et al. (2000) attributed this to reduced NH3 desorption and not initial nitrogen content. Anaerobic lagoons that were part of the NAEMS had SA:V ratios of 0.25–0.4 m2m-3 (Grant et al., 2016) while storage tanks have a SA:V ratio of 0.1 m2m-3 (Grant and Boehm, 2010a, 2010b) to 0.3 m2m-3 (Sommer, 1997). Muck and Steenuis (1982) modeled the emissions and found that decreasing SA:V results in decreased emissions for a fixed desorption rate because the depth of manure added each day to a tank increases which increases the distance within the manure that the NH3 must diffuse to be emitted at the surface. Consequently manure in tanks would be expected to emit less NH3 than that in lagoons given identical manure composition. The composition of the manure influences NH3 emissions. Total ammoniacal nitrogen (TAN = NH3 + NH4+) in the storage tank increases due to microbial mineralization of organic N in the manure and decreases through immobilization with organic matter, nitrification and denitrification (Sommer et al., 2006). Ni et al. (2010) found in a laboratory study that diluting manure from an average of 6.4% dry matter (DM) to 3.4% DM by increasing manure volume and not changing the total N content increased NH3 emissions. Anaerobic lagoons typically have less than 1% DM while slurry storage tanks have up to 9% DM (USDA, 2008). Tank storage should have reduced NH3 desorption due to the low SA:V (Sievers et al., 2000) and reduced emissions from high DM manure (Ni et al., 2010) compared to lagoons. This is not what has been observed in the field. Only a few 44

Atmospheric Environment 190 (2018) 43–52

R.H. Grant, M.T. Boehm

The interference influenced both the magnitude of the response and the probability that the response was NH3, resulting in an unknown mix of underestimated PIC with high probability of NH3 and underestimated PIC with low probability of NH3 (and therefore not a valid measurement). The low water vapor levels during the 19 December 2007 through 15 January 2008 period could have resulted in relatively low PIC underestimates. Consequently, no correction for the bias was applied to valid measurements. The scanning TDLAS instruments (TDLAS/scanner) were mounted 1 m above the tank rim (abl) to the northeast and southwest of the tank (Fig. 1). The northeast TDLAS/scanner was 12 m northwest of northwest corner of northernmost barn. Optical paths (OP) surrounding the tank were defined by the TDLAS and retro-reflector arrays at 1 m abl. Additional retro-reflector arrays were located at heights of 6 m and 14 m on towers to the northwest and southeast of the tank. Meteorological measurements (barometric pressure, air temperature, relative humidity and solar radiation) were made 10 m to the southwest of the tank and 66 m west of the barns (Fig. 1). The meteorological data were collected on a data logger. The 3D sonic anemometers were located on the meteorological tower at 2.3 m agl, and on the northwest corner tower (Fig. 1) at 4 m and 15.4 m agl (Fig. 1). The 2-m anemometer had a prevailing wind upwind fetch of 1:36 (rise: distance) and 1:33 for NW and S winds, respectively; and 1:22 for E winds coming from the barns. The 4-m anemometer fetch was greater than 1:100 in all directions except for E winds with a 1:78 fetch. Consequently, wind measurements were relatively unaffected by upwind conditions when the fetch did not include the barns (wind directions between 145° and 15°, moving clockwise) (Fig. 1). Calibration checks of the TDLAS and sonic anemometers were made at the beginning and end of each measurement period. The TDLAS NH3 concentration accuracy and precision were maintained at 10% at 50 ppm-m and ± 10% respectively. The sonic anemometers were intercompared with the requirement that the vector components of the wind be within 0.1 ms−1. Calibrations of the TDLAS (seven level calibration with calibration gas) and sonic anemometers (inter-comparison with three standard anemometers) were conducted semi-annually. Calibration corrections were applied to the TDLAS. No corrections were applied to the measured wind components. The NH3 emissions for the tank were computed at half-hour intervals using both the RPM and bLS models (Hashmonay et al., 2008; Flesch et al., 2004; Grant et al., 2016). While the RPM model used the minimum PIC value for the CBG, the bLS model derived the CBG using a Singular Value Decomposition algorithm of the twelve estimates of emissions associated with each of the twelve PIC values (Grant et al., 2013a, 2013b). Emissions were considered representative of the tank for all wind directions except when winds were from 15° to 145° (Fig. 2B and C). Additional validation criteria for emissions based on the bLS model included 1) friction velocity (u*) greater than 0.15 ms−1, 2) absolute value of Monin-Obukhov length (L) greater than 2, 3) touchdown fraction greater than 0.1, 4) standard deviation of wind direction less than 30°, and 5) CBG less than 0.1 μmol mol −1. Additional validation criteria for emissions based on the RPM model included: 1) the topmost PIC was less than or equal to 90% of the midlevel PIC, 2) the Concordance Correlation Factor was greater than 0.8, and 3) the correction factor (A) was less than 0.9 (Hashmonay et al., 2008; Grant et al., 2013a, 2013b). A comparison between the NH3 emissions determined from the RPM and bLS models was conducted according to the USEPA Method 301 ‘Field Validation of Pollutant Measurement Method’ (Grant et al., 2013a). For the comparison of methods, emissions were considered representative of the tank except when winds were from 15° to 145° (Fig. 2B and C). Although the fetch for the nearest OP is better than 1:3 and the distance to the beam line is ten times the diameter of the fan, the upwind concentration measurements were considered influenced by the exhaust fans (NNE to SSE of tank; Fig. 1) (Fig. 2A). Daily mean emissions were estimated for days in which at least 25

decreased emissions (Sommer et al., 2000; Smith et al., 2007). The formed crust of dairy manure had higher total N but lower NH4+ concentrations than the liquid slurry (Smith et al., 2007). A wet crust has less rigidity than a dry crust (Smith et al., 2007) and presumably has less denitrification and less resistance to diffusive transport of NH3. It is hypothesized that the commonly low SA:V and high DM of manure tanks will result in reduced emissions relative to reported anaerobic lagoons on an animal but greater than anaerobic lagoons and area basis. The study reported here was part of the National Air Emissions Monitoring Study (NAEMS); conducted to assist in characterizing NH3 emissions of differing livestock farm management systems across the USA (Grant and Boehm, 2010a). 2. Material and methods Ammonia emissions were measured from a hog finisher farm manure concrete tank located in Iowa. The farm consisted of four barns with the manure tank to the west (Fig. 1). The facility had a capacity of 3840 finishers (average weight 68 kg) in the four units. Manure from 2ft deep pits in each of the four barns was transferred to the tank by pulling a plug to the tank approximately once every ten weeks. The open circular concrete tank had a diameter of 55 m with a side extending from approximately 0.5 m above (agl) to 2 m below ground level, yielding an emission surface area of 2376 m2 and manure capacity of 5763 m3. A fence rose an additional 1 m above the top of the tank. The animal population in the barns varied from 3111 to 3873, with a mean population of 3508 animals. The mean mass of the finisher hogs was 68 kg resulting in a mean of 477 AU (1AU = 500 kg) on the farm. The volatile solids loading rate of the tank was estimated at 2047 kg d−1 (ASAE, 2005). Nitrogen inputs were estimated at 54 g N d−1hd−1 (0.40 kg N −1 d AU−1) based on producer-reported feed composition and mean hog mass, and typical feeding rates of hogs (NRC, 1998). Assuming a nitrogen excretion rate of 0.28 kg N d−1 AU−1 (ASAE, 2005), the N loading rate of the tank was 146 kg d−1. The AU-based loading rate was comparable to those reported for finisher hog operations storing manure in anaerobic lagoons. Given the tank dimensions, this corresponds to a loading rate of 62 g N m−2 d−1, at least six times that of the anaerobic lagoons (Grant et al., 2016). Manure was surface-applied yearly, alternating between adjacent fields and remote fields (located up to 2 mi from the tank). The storage tank was stirred prior to removal. While the tank was circular, the measurements are described below in terms of the north, east, south and west “sides” and “corners” of the tank, which refers to measurements made to the north, east, south and west. 2.1. Measurements The NH3 emissions from the tank were monitored for 8–20 d each quarter of the year for two years (Table 1). Seasons were defined according to climatology (e.g. December, January and February was winter). Measurements were made using scanning Tunable Diode Laser Absorption Spectrometer (TDLAS) open-path instruments, three 3-dimensional (3D) sonic anemometers, and a variety of meteorological instruments. Emissions were determined from these measurements using two emissions models: vertical radial plume mapping (RPM), and backward Lagrangian Stochastic (bLS; Windtrax®, Thunder Beach Scientific, http://www.thunderbeachscientific.com, Edmonton, Alberta, Canada). The path-integrated concentrations (PICs) of NH3 were measured by TDLAS (GasFinder 2®, Boreal Laser, Inc., Edmonton, Canada) along optical paths (OP) defined by TDLAS/scanner systems and retro-reflectors. A TDLAS unit used from 19 December 2007 through 15 January 2008 was later determined to have a water vapor interference that could result in an underestimate of the NH3 (PIC) by 30%–70% for dew point temperatures between -2 °C and 20 °C (Grant et al., 2016). 45

Atmospheric Environment 190 (2018) 43–52

R.H. Grant, M.T. Boehm

Table 1 Activities and status during measurement periods. Period

# days

Producer activity

Date, appearance of tank (color/crust)

Animal inventory

2: 8/30 - 9/26/2007

27 27

8/30, 31–100% crusted 9/17,18 - brown/100% crust 9/ 26/2007 - black/40% crust 12/19, 20 - black/90% crust 1/14, 15 - brown/100% crust 5/14, 5/15, 6/4 - black/no crust 6/4, 5, 12, 23, 24 - black/no crust

3592

3: 12/19 - 1/15/ 2008 4: 5/14 - 6/4/2008 5: 6/4 - 6/25/2008

Pit agitation and pump out to tank on 9/17/07, 9/19/07, and 9/ 21/07 No events

6: 11/13-11/25/ 2008 7: 11/25 - 12/16/ 2008 8: 4/8-4/23/2009

12

3873

9: 7/28 - 8/17/2009

20

11/12, 13, 14 - black/no crust 11/25/2008 - dark/100% frozen 11/25 - dark/100% frozen 12/15, 16 - white - 100% frozen 4/8, 9/2009 - black/no crust 4/22, 23/2009 - brown/ light scum 7/27, 28, 29 - brown/100% crust 8/18 - brown/100% crust

21 21

21 15

No events Pulled plugs in building to drain manure into pits 6/18/08 and 6/ 25/08 No events Drained manure to tank from 2 north barns 12/6/08 and the 2 south barns 12/8/08 No events Pit agitation and pump out to tank 8/12/09 12pm-6pm 8/13 - 8/ 14/09 8am-6pm 8/17/09 8am-12pm

3564 3547 3111

3800 3201 3374

2.2. Theory

half-hour NH3 emissions measurements (both RPM and adjusted bLS measurements) were valid (Grant et al., 2016). Emissions were normalized by source area, mean animal inventory (hd), and animal mass units (AU = 500 kg animal mass).

Emissions from the tank were considered in the context of a conceptual equilibrium two-film theory. The live-mass specific emissions were modeled according to Grant et al. (2016) as:

Fig. 2. Influence of wind direction on NH3 concentrations and tank emissions. The NH3 concentration along the eastern (closed triangle) and western OP (open triangle) (panel A), calculated background concentration (closed diamond), bLS-calculated background concentration (CBG; panel B), and the adjusted bLS-based emissions (open circle) and RPM-based emissions (closed circle) (panel C) are indicated. The dashed lines demarcate a region where bLS emissions estimates were often invalid as a result of the high CBG. 46

Atmospheric Environment 190 (2018) 43–52

R.H. Grant, M.T. Boehm

Table 2 Characteristics of liquid manure samples from the tank. Sampling Date

Surface crusting?

Weather conditions

Manure sample characteristics

Mean Air temperature

n

o

6/4/2008 4/9/2009 4/23/2009a 7/29/2009 8/18/2009 a

No crust No crust No crust Crusted (dry) Crusted (wet)

C

18.1 3.9 16.1 18.0 15.8

3 3 3 3 3

Mean pH

Mean Total Solids

Mean N −1

Mean NH3

SU

% (wet weight basis)

gL

mmol mol

7.54 7.62 7.44 6.91 7.11

1.3 1.8 1.8 4.2 4.4

3.7 4.0 4.7 4.8

4.8 5.1 6.0 6.2

−1

g L−1

mmol mol−1

3.0 3.0 3.6 3.3

2.5 3.2 3.2 3.8 3.5

Estimated date.

QLM = K Sc [CBG − HCLIQ],

Because of the expected strong influence of air temperature on hog manure NH3 emissions (Grant et al., 2016), the influence of friction velocity (representing the influence of mechanical shear across the terrain on turbulent mixing) and wind speed (including the influence of both mechanical shear and buoyancy associated with the terrain on turbulent mixing) and on daily emissions was estimated with a constrained mean daily temperature of 20 °C ± 3 °C. Tank manure surface appearance was recorded while on-site and manure was sampled at the beginning or end of every measurement period when the tank surface was not frozen (Table 1). Liquid samples were collected at three random points around the tank within 1 m of the tank wall at 0.3 m depth. Liquid samples were analyzed by Midwest Laboratories (Omaha, NE) for liquid NH3 (total ammoniacal nitrogen: TAN, Method 350.2; USEPA, 1974) concentration (CLIQ; Eqs. (1) and (2)), pH (Method 9040C; USEPA, 2004), moisture content, total (Keldahl) nitrogen (Method 1687; USEPA, 2001a), and total solids (TS; Method 1684; USEPA, 2001b).

(1)

where the subscript LM refers to the live mass (total live mass on farm = 477 AU), K is the bulk exchange coefficient, Sc is the molecular Schmidt number for NH3 (1.31), CBG is the background NH3 air concentration, H is the NH3 gas solubility in the manure liquid (0.000673; Sander, 2015), and CLIQ is the manure liquid NH3 concentration defined as: 1

1

HCLIQ = LFNH 3 e−β ⎡⎣ T − 298.15 ⎤⎦

(2)

were L is the volumetric N loading rate estimated at 0.280 g N AU−1 d−1 (ASAE, 2005) and 0.19 m3 manure AU−1 d−1 (Chastain et al., 1999), and FNH3 is the fraction of N loading as NH3. The fraction of N loading that was not adsorbed onto organic matter (FNH3) was estimated based on the mean measured lagoon liquid pH (Table 2) and estimated slurry temperature (T) according to DeVisscher et al. (2002), with the dissociation of NH3 and NH4+ of stored pig slurry based on Ni (1999). The temperature influence on H (last exponential term in Eq. (2)) was estimated from Van't Hoff's equation with the value for β determined by non-linear regression using TableCurve® 2D (SYSTAT Software Inc., Richmond, CA). While T in Eq. (2) represents the temperature of the liquid, we used the measured mean daily air temperature as a proxy for the tank liquid temperatures since the liquid temperature was not measured (Grant et al., 2016). The transfer of NH3 from the tank is similar to that from a pore, with flux from the manure surface through the air above the surface and within the tank (pore) walls then diverging out to the atmosphere above the top of the tank (pore). As the height to width ratio decreases, the NH3 within the tank walls approaches that in the open air over flat terrain (Hall et al., 2012). The bulk exchange coefficient K (Eq. (1)) is the sum of multiple conductances: transport through the liquid in the tank (gLIQ), through the tank liquid surface (gs), through the laminar air layer over the surface (gb), and through the height of the tank air space and into the air over the tank (ga) to be diffused downwind.

K = gLIQ + gs + gb + ga.

3. Results and discussion Emissions were measured on 164 days during eight periods across two years (Table 1). Forty-seven days had emissions measurements (1417 one-half hour measurement periods) based on the bLS method. Forty-nine days had emissions measurements (761 one-half hour measurement periods) based on the RPM method. Of the days with measurements, only 23 and 10 of the days had at least 50% of the day with valid measurements using the bLS method and RPM method respectively. Only five days had both a RPM and bLS emissions estimate. The composition of the manure in the tank was measured on five occasions (Table 2). Nitrogen concentrations were comparable to those for anaerobic lagoons (Grant et al., 2016). 3.1. Emissions measurement method comparison A comparison between the NH3 emissions calculated using the RPM and bLS methods was made based on 624 half-hour measurement periods. The mean bLS emission was 0.432 gs−1 (SD 0.384 gs−1) and the mean RPM emission was 0.662 gs−1 (SD 0.210 gs−1) during all halfhourly intervals for which both RMP and bLS emissions were valid. The bLS emissions did not have a significantly different precision (F = 0.70, critical F 1.0) but did have a significant bias over the RPM emissions (t = −20.5, t0.2 = 1.29). This difference between the RPM and bLS emissions methods contrasted with the equivalence of the methods found for more open lagoon emissions measurements (Grant et al., 2013a, 2013b).

(3)

Mean daily conductance of the air over the tank was determined by conversion of the eddy exchange coefficient calculated from the 2.3 m and 4 m wind measurements into a conductance associated with a ‘pore’ (Parlange and Waggoner, 1970; Montieth and Unsworth, 1990) as:

ga =

(u∗ )2 ⎡ πrtan k + 4ztan k ⎤ z anem (Δu/ Δz anem) ⎢ ⎣ πrtan k ztan k ⎥ ⎦

(4)

where u (measured) and u* (calculated) is determined from the sonic anemometer measurements at z = 2.3 and 4 m, (u∗ ) is for the mean anemometer height (z anem ), and the tank depth (ztank) and radius (rtank) assumes the top of the tank opens to a flat surface rather than 0.5 m above the ground. The first term in the right-hand side product defined the conductance for a flat plane based on Schäfer et al. (2012) and the second term describes the correction to a flat plate conductance resulting from the in-ground tank structure.

3.2. Half-hour emissions The only clear difference between the site configurations for the lagoon emission measurements in Grant et al. (2013a); 2013b and the tank emissions at this farm was the relatively close proximity of the 47

Atmospheric Environment 190 (2018) 43–52

R.H. Grant, M.T. Boehm

barns and their exhaust fans to the tank at this farm. While western OP often had higher concentrations than eastern OP, eastern OP were higher than western OP when the winds were from the east where the buildings were located (Figs. 1 and 2A). Given that the typical path length around the tank was 50 m, a concentration of 0.1 μmol mol−1 corresponds to a PIC of approximately 5 ppm-m. This is approximately twice the measured MDL of the TDLAS instruments of 2 ppm-m and therefore represents a near-zero background. Measurements with winds from the direction of the buildings was explicitly excluded from analysis (Fig. 2C). This exclusion appeared to largely exclude the influence of the building exhaust fans (except for wind from the south) since bLSdetermined NH3 CBG was generally less than 0.1 μmol mol −1 with occasional higher values for all valid wind directions and generally higher values for south winds (Fig. 2B). In contrast to the bLS emissions method, the RPM emission method used an upwind PIC for the CBG and eight additional PICs elevated from the near-ground 1-m PICs and hence further from the exhaust fans (Grant et al., 2013a, 2013b). After validation criteria were applied, the half-hourly bLS and RPM emissions were not strongly influenced by wind direction (Fig. 2C). However a comparison of the minimum OP concentration (commonly from an elevated PIC) and the CBG (determined by the bLS model) showed greater variability when the winds were light- as would be expected if the exhaust fans were contributing to the NH3 around the tank (Fig. 3). Since the influence of downwind obstructions on the mean flow is less than that of the flow turbulence, the RPM model was believed to be less influenced by the configuration of the site. Consequently it was assumed that the RPM method was a more accurate emissions measurement method than the bLS method for this farm. The bLS emissions estimates were divided by the Method 301 correction factor (0.63) for mean equivalency to the emissions determined by the RPM method. Since the RPM and bLS methods had equal precision, valid RPM and bLS emissions were both used to estimate the actual emissions. If both RPM and adjusted bLS emissions were valid for the same half-hour time period, the two estimates were averaged for the emissions during that period. All bLS-calculated emissions estimates reported below represent adjusted values unless specified.

Fig. 4. Annual variation in daily mean wind speed, air temperature and NH3 emissions. The variation in wind speed at 2.3 m agl (open circle, panel A), air temperature (open diamond, panel B) and NH3 emissions (closed circle, panel C) are indicated. Daily values based on all measurements (combined adjusted bLS-calculated and RPM-calculated) in the day. The shaded band (panel C) represents the estimated error of the measurement.

in air temperature (Fig. 4B and C). Winds were more variable during the spring than the summer (Fig. 4A). Emissions calculated using potentially water-vapor-affected PIC measurements (discussed above), made during 2007–2008 winter, averaged 0.06 g NH3 m−2h−1 (SD 0.06 g NH3 m−2h−1) A comparison of these measurements with those made with un-affected TDLs during the 2008–2009 winter (mean = 0.07 g NH3 m−2h−1, SD 0.08 g NH3 m−2h−1) showed no apparent influence of the water vapor interference for the 2007–2008 winter measurements. On an area basis, daily emissions averaged 0.45 g NH3 m−2h−1 with maximum daily average emissions of 1.62 g NH3 m−2h−1. The areabased values are comparable to emissions from 26 to 28 m diameter slurry manure tanks in Denmark (Sommer, 1997), but much lower than those measured at the surface of a manure tank in Minnesota (Gay et al., 2003). The mean emission was much higher than measured at lagoons at finisher hog farms in North Carolina and Oklahoma (Grant et al., 2016). The hypothesis that the low SA:V and high DM of the manure tank would result in enhanced emissions relative to anaerobic lagoons on an area basis was supported. On an animal basis, daily emissions averaged 7.3 g d−1 hd−1 (53.9 g d−1 AU−1) with maximum daily average emissions of 26.4 g d−1 hd−1 (194.2 g d−1 AU−1). The mean emissions on an animal basis were less those that reported for most anaerobic lagoons (22.7 g NH3 hd−1 d−1 for a Missouri farm according to Zahn et al., 2001; 16.1 g hd−1 d−1 for an Oklahoma farm according to Grant et al., 2016; 9.8 g NH3 hd−1 d−1 for a Missouri farm according to Szögi et al., 2005) and more than those reported for a North Carolina farm (4.4 g hd−1 d−1; Grant et al., 2016). The hypothesis that NH3 emissions from the manure tank would have lower emissions than the anaerobic lagoon storages on an animal basis was not supported.

3.3. Daily emissions As expected, emissions were highest during the summer and very low during the winter, following the general trend in air temperatures (Fig. 4C). Variability in emissions was also high during the spring and summer, although this variability was clearly associated with variation

3.3.1. Influence of temperature Higher mean daily air temperatures, used as a proxy for liquid surface temperature, corresponded with higher daily emissions (Fig. 5). The temperature influence on daily emissions on a unit area basis was

Fig. 3. Comparison of minimum OP NH3 concentration and the bLS-calculated background concentration (CBG) as a function of wind speed for all NH3 flux measurements with CBG < 0.1 μmol mol−1. 48

Atmospheric Environment 190 (2018) 43–52

R.H. Grant, M.T. Boehm

Fig. 5. Relationship of daily environmental conditions to mean daily NH3 emissions. The relationship between daily mean emissions and air temperature (panel A; r2 = 0.53) and friction velocity at 2.3 m (panel B; r2 = 0.01, P(nonzero slope) = 0.36) is indicated. The solid lines illustrates the nonlinear relationship between air temperature and emissions according to Eq. (2) (panel A) and the linear relationship between friction velocity and emissions (panel B).

modeled according to Eq. (2) (Fig. 5A). The value for β was determined to be 6060 K, with the resultant model accounting for 53% of the emissions variability. The empirically-derived temperature influence (β) was only 6% higher than that determined for the anaerobic lagoon of an Oklahoma finisher farm (Grant et al., 2013a, 2013b). Since the temperature influence on the tank emissions was similar to that for lagoon emissions, the difference in mean emissions was likely a result of the tank area (A) and the mass transfer coefficient (K). Prolonged periods below freezing resulted in freezing of the manure surface. When the surface was frozen, as determined based on the observations (Table 1) and daily air temperatures, the daily NH3 emissions were very low (Fig. 7) with a mean of 0.07 g m−2 h−1 (SD 0.07 g m−2 h−1) (Table 4). We used this variability as an estimate of the minimum detection limit (MDL) of the measurement method (3SD = 0.21 g m−2 h−1; 3.4 g d−1 hd−1, 25 g d−1 AU−1).

Table 4 Daily emissions associated with liquid manure surface condition.

3.3.2. Influence of wind Daily NH3 emissions were correlated with neither wind speed (negative correlation, r2 = 0.08, P(non-zero slope) = 0.001) nor friction velocity at 2.3 m (negative correlation, r2 = 0.007, P(non-zero slope) = 0.36; Fig. 5B). Although a correlation between emissions and winds was predicted based on the manure pit model of Zhang et al. (1994) and prior daily measurements of anaerobic lagoons (Grant el at, 2016), lack of correlation was also observed for NH3 emissions by Sommer (1997). The lack of correlation between emissions and wind conditions or depth to manure surface was probably due to the dominant influence of a pressure gradient across the upwind and downwind tank wall generating significant mechanical turbulence regardless of wind flow across the surrounding terrain. The lack of influence of wind conditions on emissions implied that ga (Eq. (2)) was relatively invariant and gb (Eq. (2)) was not a strong function of u* or U, as is often assumed (DeVisscher et al., 2002; Ro and Hunt, 2006; Grant et al., 2013a).

derived from Grant and Boehm, 2010b). This was counter-intuitive since the manure storage N-loading rate for the NC lagoon was 4.4 kg m−3d−1 (Grant and Boehm, 2016) compared to 18.6 kg m−3d−1 for this storage tank. The N in the manure is distributed as organic N, NH4+ and NH3. The mean NH3 fraction of total N was 0.88. The NH3 fraction of the manure liquid N content was inversely correlated with liquid pH (r2 = 0.67): the relatively weak correlation was likely a result of few values available for correlation (Table 2). A comparison of daily emissions within three days of the manure sampling indicated daily tank emissions were weakly correlated with TAN (r2 = 0.54) (Fig. 6). The stored manure TS was higher when the manure surface was crusted than when it was not (Table 2). This difference in TS was likely partly influenced by a positive contamination of the liquid sample as it was lifted through gaps in the surface crust. A comparison of daily emissions with liquid TS content was also highly correlated with emissions (positive linear correlation with r2 = 0.78). The stored manure TAN was also higher when the manure surface was crusted than when it was not (Table 2). If we assume contamination of the sample with crust material, the TAN in the crust was greater than that in the equal mass of liquid. If we assume negligible contamination of the sample, this suggests that the crusting may have enhanced TAN near

Tank manure surface

No crust Frozen Crust All

Valid measured days1

14 17 9 40

Air temperature (C)

NH3 emissions (g m−2 h−1)

Normalizeda NH3 emissions (g m−2 h−1)

Mean

Mean

Standard Deviation

Mean

12.6 −4.5 19.3 6.8

0.55 0.07 1.03 0.45

0.41 0.07 0.29 0.47

0.80 0.47 1.09 0.72

1: days with more than 24 measurements. a Emissions adjusted to equivalent at 20 °C based on influence of temperature.

3.3.3. Influence of manure composition The mean N in the tank manure of 5.5 mmol mol−1 (4.3 g l−1) was equivalent to that of an anaerobic lagoon in NC (5.4 mmol mol−1; Table 3 Mean seasonal and annual temperatures and NH3 emissions. Season

Measurement days

Mean Air temperature (C)

NH3 emissions Cumulative (kg m

Winter Spring Summer Fall Annual

33 17 37 28 115

−5.8 12.7 21.2 11.3 9.9

0.18 1.18 2.19 0.35 3.90

49

−2

)

Mean (g m−2 h−1)

(g d−1 hd−1)

(g d−1 AU−1)

0.08 0.54 1.00 0.16 0.44

1.2 8.8 16.3 2.6 7.2

9.0 64.4 119.6 19.2 53.1

Atmospheric Environment 190 (2018) 43–52

R.H. Grant, M.T. Boehm

3.4. Modeled daily emissions Daily NH3 emissions were modeled according to Eqs. (1)–(4) for days with more than 24 valid measurements. CBG in Eq. (1) was set to the minimum concentration of the daily mean concentrations for each optical path (Fig. 1). The combined conductance of gs and gb was approximated based on previously reported values (Olesen and Sommer, 1993) as 0.056 m s−1 when there was no crust on the surface and 0.008 m s−1 when the surface was crusted. Assuming no convective mixing of the liquid in the tank and diffusion from the sampling depth of 0.3 m, the mean gL was estimated at 4.64 × 10−6 ms−1 for molar ratios of less than 0.02 mol mol−1 (Frank et al., 1996). Based on the lack of influence of wind velocity on the emissions, ga was assumed to be not influenced by wind speed or turbulence and was set to the calculated median conductance for northerly winds with the tank upwind (0.086 ms−1) corrected by a ‘pore’ flow factor of 0.67 according to Eq. [4]. This conductance was comparable to the modeled conductance near the bottom of the canyon in large eddy simulations of two-dimensional urban canyons (Jeong and Andrews, 2001). The lack of influence of depth to the tank manure surface in Kaharabata et al. (1998) was likely a result of the tank being entirely above ground such that the turbulent exchange from their tank was due to the pressure gradients induced by the flow around the tank and not into the tank. The total conductance (0.16 ms−1) based on Eq. [3] was held constant for all measurements. The annual variation in the emissions model error was evenly distributed with a mean bias error (MBE) of all measurements of only 0.01 g NH3 m−2h−1 (0.9 g d−1 AU−1, 0.1 g d−1 hd−1) (Table 5, Fig. 7B). Estimates of the NH3 emissions when the liquid surface was crustfreewere much better (MBE -0.03 to −0.07 g NH3 m−2h−1; Table 5). This error was in the same range as the error of the generalized open-air lagoon emissions model applied to two finisher operations (OK3 and NC3) in Grant et al. (2016), but greater than the 0.003 g NH3 m−2h−1 MBE of Bajwa and coworkers' equilibrium chamber-based emissions model of eleven hog lagoons (Bajwa et al., 2006). The emissions model root mean square error (RMSE) averaged 0.26 g NH3m−2h−1 (31.6 g d−1 AU−1, 4.3 g d−1 hd−1) (Table 5); comparable to the generalized hog operation model errors for open-air emissions from OK and NC finisher operations. This corresponded to a normalized RMSE of 57%; much lower than the value of 201% for the equilibrium model of Bajwa et al. (2006) for chamber emissions at eleven North Carolina hog lagoons. This is counterintuitive, since the model RMSE would be expected to be better for measurements made under the more controlled environment of the chamber measurements than micrometeorological open-air measurements. The emissions model generally underestimated the emissions when the liquid surface was crusted (Fig. 7B), with a MBE of 0.21 g NH3 m−2h−1 and a relative MBE of 21% (Table 5). This category of tank manure surface resulted in the largest model errors (Table 5, Fig. 7), suggesting that the temperature influence on the emissions was too low

Fig. 6. Relationship between tank liquid manure composition and mean daily NH3 emissions within three days of sampling. Solid line is linear regression fit (r2 = 0.54). Labels indicate mean wind speeds at 15.4 m agl on the days surrounding manure sampling.

the top of the liquid under the crust. Since the high TS corresponded with crusting, and the crusting on these dates were wet crusts, TAN may have been convected from depth in the tank as water evaporated from the crust surface (Sommer et al., 1993). Crusting of the manure storage tank due to lack of mixing might also be expected to influence NH3 emissions. Theoretically a dry crust should decrease gs and K resulting in a reduction of emissions for similar concentration gradients (Sommer et al., 1993; Olesen and Sommer, 1993). The daily emissions however were greater when the tank liquid surface was crusted than when it was not crusted (Table 4), in contrast to this theory but consistent with the relationship of higher manure TAN when the tanks was crusted. Since crusting typically occurred when the air temperatures (and presumably evaporation rates) were high, emissions were normalized to an air temperature of 20° C using the temperature function (Eq. (2)) with β of 6060 K. The mean normalized emissions still indicated 10% higher emission when the surface was crusted than not crusted (Table 4), however the difference was not statistically significant. Similar non-significant effects of crusting were found for dairy manure (Grant and Boehm, 2015), while Wood et al. (2012) noted enhanced NH3 emissions from open-air tanks of dairy manure with high crusting (however the manure surface was never entirely crusted). Sommer et al. (1993) noted that NH3 emissions from hog manure were likely rate-limited by the depletion of NH3 in the air over the crusted surface. In their study this air was uniform and relatively steady while in ours we suspect the air was quite turbulent; providing efficient flushing of the NH3 away from the crusted surface. Wood et al. (2012) suggested that the enhanced NH3 emissions in their partially-crusted tanks may have been due to NH3 convection in the tank due to ebullition. Furthermore, if the crusts were wet, evaporation from the surface may have induced convection of TAN from depth in the tank (Sommer et al., 1993). Assuming complete crust cover conditions occurred between observation periods of this study, enhanced emissions may have occurred under crust conditions due to the convection of TAN and congregation of such bubbles with enhanced NH3 under the crust; thereby increasing the NH3 concentration gradient across the crust and offsetting the effect of decreased total conductance with the presence of the crust. Unfortunately, there were no measurements of crust moisture, no continuous profile measurements of pH, N, or NH3 content of the liquid manure and no profile measurements of NH3 within the tank to evaluate this hypothesis.

Table 5 Modeled daily NH3 emissions associated with liquid manure surface condition. Tank manure surface

Area based MBEa −2

No crust Frozen Crust All a b c

50

Live mass based RMSEb −2

MBE

(g m h−1)

(g m h−1)

(g AU d−1)

−0.03 −0.07 0.21 0.01

0.32 0.11 0.35 0.26

−3.6 −8.3 25.3 0.9

Relative MBEc

RMSE −1

(g AU−1 d−1)

(%)

38.8 13.0 42.1 31.6

−6 −104 21 2

Mean Bias Error. Root Mean Square Error. Relative Mean Bias Error = MBE/measured meanFigures.

Atmospheric Environment 190 (2018) 43–52

R.H. Grant, M.T. Boehm

4. Conclusions Emissions from the tank were lower than documented NH3 emissions from anaerobic lagoons when calculated on an animal mass basis but much higher than lagoons when stated on an area basis. The hypothesis that the configuration of the tank manure storage facility would result in reduced emissions on an animal basis relative to anaerobic lagoons was not supported by the results. The hypothesis that the configuration of the tank manure storage facility would result in higher emissions on an area basis than anaerobic lagoons was supported by the results. This was likely due to: 1) the tank configuration providing a more efficient turbulent NH3 transport from the manure surface than anaerobic lagoons as proposed by Sommer (1997) and 2) the liquid manure surface having a higher concentration of NH3 than anaerobic lagoons. Unfortunately insufficient measurements of the liquid NH3 concentration were made to support the latter contention. Modeling the emissions from the tank differed from recent anaerobic lagoon models in the accounting for the sheltering of the tank liquid surface by the tank walls in the surface turbulent transport processes. Although the greater solids in the tank liquid and crusting of the surface corresponded to increased emissions, the increase was not significantly different from the emissions from the non-crusted manure surface when normalized to 20 °C. Additional emissions measurements from other ground-level manure storage tanks with greater documentation of the manure liquid composition, crust characteristics and gradients of NH3 in the liquid manure and airspace of the tank are needed to verify the transport mechanisms and modeling approach.

Fig. 7. Measured and modeled daily NH3 emissions. The daily emissions based on valid 30-min measurements over more than 50% of the day for the RPM model (solid square) and the bLS model (open square) are indicated in panel A. The difference between measured daily and modeled mean daily emissions (combined RPM and adjusted bLS measurements) are indicated for days with more than 50% (large solid circles). Shaded region in panel A represents the estimated emissions MDL.

and/or the conductance estimated for the crust was too small. Accounting for the tank configuration in the modeling of the mass exchange between the manure surface and the open air (Eq. (4)) was a significant factor in the emissions model. Excluding the ‘pore’ flow factor correction to ga increased the overall MBE to −0.133 g NH3m−2h−1 (−15.9 g AU−1 d−1).

Acknowledgements Thanks go to Alfred Lawrence, Ben Evans, and Scott Cortus for their careful field measurements. Funding for the study was provided by livestock producers, the Agricultural Air Research Council, Inc. and the National Pork Board. Additional assistance was provided by the US Environmental Protection Agency and Purdue University Agricultural Experiment Station.

3.5. Annual emissions Annual tank NH3 emissions, estimated by equally weighting the season mean emissions, were 3.9 kg m−2 (Table 3). These emissions were much higher than the estimated annual emissions from finisher anaerobic lagoons in NC (1.4 kg m −2,Szögi et al., 2005; 0.7 kg m −2, Grant et al., 2016) and OK (1.6 kg m −2, Grant et al., 2016). The daily NH3 emissions were highest during spring and summer and decreased during the winter to less than 0.1 g NH3m−2h−1 (Table 3; Fig. 4). The winter emissions may be underestimated, given the potential for atmospheric moisture interference during measurement period 3 (Table 1). The wide variability during the summer in combination with the limited number of ‘complete’ days of measurements limits confidence in the magnitude of the emissions. Annual tank NH3 emission in terms of the farm inventory was 19 kg AU−1. These emission values were lower than that estimated from a finisher hog anaerobic lagoon in NC (28 kg AU−1,Szögi et al., 2005) but higher than that measured at another finisher hog anaerobic lagoon in NC (13 kg AU−1, Grant et al., 2016). The corresponding NH3 emissions for a finisher hog anaerobic lagoon in OK was 32 kg AU−1 (Grant et al., 2016). Given the assumed nitrogen excretion rate, the mean annual NH3 emission accounted for 19% of the excreted N. This was similar to the total emission of NH3 from finishing hog farms in the Netherlands (Velthof et al., 2012) where covered slurry tanks are extensively used (Eurostat, 2013). Given the producer-supplied feed composition and assumed feeding rate, the N in the feed was 145 kg N AU−1 or 2 kg N hd−1, corresponding to annual mean N emissions as NH3 represented 13% of the N feed inputs. This emission was comparable to that estimated by Hayes et al. (2004) from animal housing for similar crude protein diets (11%–13%).

References ASAE, 1998. Design of anaerobic lagoons for animal waste management. Standard EP403.2. American Society of Agricultural and Biological Engineers Standards 656–659. ASAE, 2005. Manure production and characteristics. Standard D384.2. American Society of Agricultural and Biological Engineers Standards 1–19. Bajwa, K.S., Aneja, V.P., Arya, S.P., 2006. Measurement and estimation of ammonia emissions from lagoon-atmosphere interface using a coupled mass transfer and chemical reactions model, and an equilibrium model. Atmos. Environ. 40, S275–S286. Centner, T.J., Patel, P.G., 2010. Reporting air emissions from animal production activities in the United States. Environ. Int. 36, 237–242. Chastain, J.P., Camberato, J.J., Albrecht, J.E., Adam, J., 1999. Swine Manure Production and Nutrient Content. South Carolina Confined Animal Manure Managers Certification Program, vol. 3. Clemson University, SC, pp. 1–17. DeVisscher, A., Harper, L.A., Westerman, P.W., Liang, Z., Arogo, J., Sharpe, R.R., VanCleemput, O., 2002. Ammonia emissions from anaerobic swine lagoons: model development. J. Appl. Meteorol. 41, 426–433. Eurostat, 2013. Agri-environmental indicator- manure storage. Statistical Office, European Union, Luxembourg. http://ec.europa.eu/eurostat/statistics-explained/ index.php/Agri-environmental_indicator_-_manure_storage, Accessed date: 9 January 2018. Flesch, T.K., Wilson, J.D., Harper, L.A., Crenna, B.P., Sharpe, R.P., 2004. Deducing ground-to-air emissions from observed trace gas concentrations: a field trial. J. Appl. Meteorol. 43, 487–502. Frank, M.J.W., Kuipers, J.A.M., van Swaaij, W.P.M., 1996. Diffusion coefficients and viscosities of CO2+H2O, CO2+CH3OH, NH3+H2O, and NH3+CH3OH liquid mixtures. J. Chem. Eng. Data 41, 297–302. Gay, S.W., Schmidt, D.R., Clanton, C.J., Janni, K.A., Jacobson, L.D., Weiberg, S., 2003. Odor, total reduced sulfur, and ammonia emissions from animal housing facilities and manure storage units in Minnesota. Appl. Eng. Agric. 19, 347–360. Grant, R.H., Boehm, M.T., 2010a. National Air Emissions Monitoring Study: Data from the Midwestern US Pork Production Facility IA3A. Final Report to the Agricultural Air Research Council. Purdue University, West Lafayette 137p. https://archive.epa.gov/ airquality/afo2012/web/pdf/ia3asummaryreport.pdf. Grant, R.H., Boehm, M.T., 2010b. National Air Emissions Monitoring Study: Data from the southeastern US Pork Production Facility NC3A. Final Report to the Agricultural

51

Atmospheric Environment 190 (2018) 43–52

R.H. Grant, M.T. Boehm

Moines, Iowa, USA. Smith, K., Cumby, T., Lapworth, J., Misselbrook, T., Williams, A., 2007. Natural crusting of slurry storage as an abatement measure for ammonia emissions on dairy farms. Biosyst. Eng. 97, 464–471. Sommer, S.G., 1997. Ammonia volatilization from farm tanks containing anaerobically digested animal slurry. Atmos. Environ. 31, 863–868. Sommer, S.G., Christensen, B.T., Nielsen, N.E., Schjorring, J.K., 1993. Ammonia volatilization during storage of cattle and pig slurry: effect of surface cover. J. Agric. Sci. 121, 63–71. Sommer, S.G., Petersen, S.O., Moller, H.B., 2000. Greenhouse gas emissions from stored livestock slurry. J. Environ. Qual. 29, 744–751. Sommer, S.G., Zhang, G.Q., Bannick, A., Chadwick, D., Misselbrook, T., Harrison, R., Hutchings, N.J., Menzi, H., Monteny, G.J., Ni, J.Q., Oenema, O., Webb, J., 2006. Algorithms determining ammonia emissions from buildings housing cattle and pigs and from manure stores. Adv. Agron. 89, 261–335. Szögi, A.A., Vanotti, M.B., Stansbery, A.E., 2005. Reduction of ammonia emissions from treated anaerobic swine lagoon. In: Havenstein, G. (Ed.), Proceedings of the 2005 Animal Waste Management Symposium, Research Triangle Park, NC, USA, 5–7 October 2005. North Carolina State University, NC, USA. USDA, 2008. Agricultural waste management field Handbook, Part 651. US Department of agriculture Natural Resources Conservation service online at. https://www.nrcs. usda.gov/wps/portal/nrcs/detailfull/national/water/?&cid=stelprdb1045935. USDA, 2016. Waste storage facility. Conservation practice standard, Code 313. US Department of agriculture Natural Resources Conservation service 10p, online at. https://www.nrcs.usda.gov/Internet/FSE_DOCUMENTS/nrcs143_026465.pdf. USEPA, 2001a. Method 1684 – Total, Fixed, and Volatile Solids in Water, Solids, and Biosolids. EPA-821/R-01–015 (CD). US Environmental Protection Agency Office of Water, Office of Science and Technology, Washington, DC. USEPA, 2004. pH Electrometric measurement. Method 9040C (Rev. 3). US environmental protection Agency Online at Accessed 3/7/2006" > . www.epa.gov/sw-846/pdfs/ 9040c.pdf, Accessed date: 3 July 2006. USEPA, 1974. Method 350.2. Nitrogen, ammonia (Colorimetric, Titrimetric, PotentiometricDistillation procedure). US environmental protection Agency Online at. www.umass.edu/tei/mwwp/acrobat/epa350_2NH3titration.pdf, Accessed date: 19 January 2006. USEPA, 2001b. Method 1687. Total Kjeldahl Nitrogen in Water and Biosolids by Automated Colorimetry with Preliminary Distillation/Digestion (Draft). EPA-821/R01–004 (CD). US Environmental Protection Agency Office of Water. Office of Science and Technology, Washington, DC. USEPA, 2011. Reactive Nitrogen in the United States: an Analysis of Inputs, Flows, Consequences, and Management Options. Report of the Science Advisory Board EPAsab-11–013. US Environmental Protection Agency Science Advisory Board, Washington, DC. Velthof, G.L., van Bruggen, C., Groenestein, C.M., de Haan, B.J., Hoogeveen, M.W., Huijsmans, J.F.M., 2012. A model for inventory of ammonia emissions from agriculture in The Netherlands. Atmos. Environ. 46, 248–255. Wood, J.D., Gordon, R.J., Wagner-Riddle, C., Dunfield, K.E., Mandani, A., 2012. Relationships between dairy slurry total solids, gas emissions, and surface crusts. J. Environ. Qual. 41, 674–704. Worley, J.W., 2015. Manure storage and treatment systems, Chapter 3. In: Bass, T.M., Risse, L.M. (Eds.), Small Farm Nutrient Management Primer: for Un-permitted Animal Feeding Operations. 12p., Univ. Georgia Extension, Bulletin 1293. Reviewed February 2015, . https://secure.caes.uga.edu/extension/publications/files/pdf/B %201293_5.PDF, Accessed date: 16 September 2017. Zahn, J.A., Tung, A.E., Roberts, B.A., Hatfield, J.L., 2001. Abatement of ammonia and hydrogen sulfide emissions from a swine lagoon using a polymer biocover. J. Air Waste Manag. Assoc. 51, 562–573. Zhang, R.R., Day, D.L., Christenson, L.L., Jepson, W.P., 1994. A computer model for predicting ammonia release rates from swine manure pits. J. Agric. Eng. Res. 58, 223–229.

Air Research Council. Purdue University, West Lafayette 124p. https://archive.epa. gov/airquality/afo2012/web/pdf/nc3asummaryreport.pdf. Grant, R.H., Boehm, M.T., Lawrence, A.F., Heber, A.J., 2013a. Ammonia emissions from anaerobic treatment lagoons at sow and finishing farms in Oklahoma. Agric. For. Meteorol. 180, 203–210. https://doi.org/10.1016/j.agrformet.2013.06.006. Grant, R.H., Boehm, M.T., Lawrence, A.F., 2013b. Comparison of a backward-Lagrangian Stochastic and Vertical Radial Plume Mapping methods for estimating animal waste lagoon emissions. Agric. For. Meteorol. 180, 236–248. https://doi.org/10.1016/j. agrformet.2013.06.013. Grant, R.H., Boehm, M.T., 2015. Manure NH3 and H2S emissions from a western dairy storage basin. J. Environ. Qual. 44, 127–136. https://doi.org/10.2134/jeq2014.05. 0196. Grant, R.H., Boehm, M.T., Heber, A.J., 2016. Ammonia emissions from anaerobic waste lagoons at pork production operations: influence of climate. Agric. For. Meteorol. 228, 73–84. https://doi.org/10.1016/j.agrformet.2016.06.018. Hall, T.C., Britter, R.E., Norford, L.K., 2012. Predicting velocities and turbulent momentum exchange in isolated street canyons. Atmos. Environ. 59, 75–85. Hashmonay, R.A., Varma, R.M., Modrak, M.T., Kagman, R.H., Segall, R.R., Sullivan, P.D., 2008. Radial Plume Mapping: a US EPA Test Method for Area and Fugitive source emission monitoring using optical remote sensing. 21–36. In: Kim, Y.J., Platt, U. (Eds.), Advanced Environmental Monitoring. Springer-Verlag Press. Hayes, E.T., Leek, A.B.G., Curran, T.P., Dodd, V.A., Carton, O.T., Beattie, V.E., O'Doherty, J.V., 2004. The influence of diet crude protein level on odour and ammonia emissions from finishing pig houses. Bioresour. Technol. 91, 309–315. Jeong, S.J., Andrews, M.J., 2001. Application of the k-e turbulence model to the high Reynolds number skimming flow field of a urban street canyon. Atmos. Environ. 36, 1137–1145. Kaharabata, S.K., Schuepp, P.H., Desjardins, R.L., 1998. Methane emissions from aboveground open slurry tanks. Global Biogeochem. Cycles 12, 545–554. Misselbrook, T.H., Brookman, S.K.E., Smith, K.A., Cumby, T., Williams, A.G., McCrory, D.F., 2005. Crusting of stored dairy slurry to abate ammonia emissions: pilot-scale studies. J. Environ. Qual. 34, 411–419. Montieth, J.L., Unsworth, M., 1990. Principles of Environmental Physics, second ed. Edward Arnold, London 291p. Muck, R.E., Steenuis, T.S., 1982. Nitrogen losses from manure storages. Agric. Wastes 4, 41–54. NRC, 1998. Nutrient Requirements of Swine, tenth ed. National Academy Press, Washington, DC. Ni, J.-Q., 1999. Mechanistic models of ammonia release from liquid manure: a review. J. Agric. Eng. Res. 72, 1–17. Ni, J.-Q., Heber, A.J., Sutton, A.L., Kelly, D.T., Patterson, J.A., Kim, S.-T., 2010. Effect of swine manure dilution on ammonia, hydrogen sulfide, carbon dioxide, and sulfur dioxide releases. Sci. Total Environ. 408, 5917–5923. Olesen, J.E., Sommer, S.G., 1993. Modelling effects of wind speed and surface cover on ammonia volatilization from stored pig slurry. Atmos. Environ. 27A, 2567–2574. Parlange, J.-Y., Waggoner, P.E., 1970. Stomatal dimensions and resistance to diffusion. Plant Physiol. 46, 337–342. Ro, K.S., Hunt, P.G., 2006. A new unified equation for wind-driven surficial oxygen transfer into stationary water bosdies. Trans. ASABE (Am. Soc. Agric. Biol. Eng.) 49, 1615–1622. Sander, R., 2015. Compilation of Henry's law constants (version 4.0) for water as solvent. Atmos. Chem. Phys. 15, 4399–4981. Schäfer, K., Grant, R.H., Emeis, S., Raabe, A., von der Heide, C., Schmid, H.P., 2012. Areal-averaged trace gas emission rates from long-range open-path measurements in stable boundary layer conditions. Atmos. Meas. Tech. 5, 1–13. Shores, R.C., Harris, D.B., Thompson, E.L., Vogel, C.A., Natschke, D., Hashmonay, R.A., Wagoner, K.R., Modrak, M., 2005. Plane-integrated open path Fourier transform infrared spectrometry methodology for anaerobic swine lagoon emission measurements. Appl. Eng. Agric. 21, 487–492. Sievers, D.M., Fulhage, C.D., Hoenhe, J., 2000. Animal, agricultural and food processing wastes. In: Proceedings of the Eighth International Symposium, pp. 664–671 Des

52