Evaluation of Particulate Matter Abatement Strategies for Almond Harvest

ABSTRACT

Almond harvest accounts for substantial PM₁₀ (particulate matter [PM] ≤10 μm in nominal aerodynamic diameter) emissions in California each harvest season. This paper evaluates the effects of using reduced-pass sweepers and lower harvester separation fan speeds (930 rpm) on lowering PM emissions from almond harvesting operations. In-canopy measurements of PM concentrations were collected along with PM concentration measurements at the orchard boundary; these were used in conjunction with on-site meteorological data and inverse dispersion modeling to back-calculate emission rates from the measured concentrations. The harvester discharge plume was measured as a function of visible plume opacity during conditioning operations. Reduced-pass sweeping showed the potential for reducing PM emissions, but results were confounded because of differences in orchard maturity and irrigation methods. Fuel consumption and sweeping time per unit area were reduced when comparing a reduced-pass sweeper to a conventional sweeper. Reducing the separation fan speed from 1080 to 930 rpm led to reductions in PM emissions. In general, foreign matter levels within harvested product were nominally affected by separation fan speed in the south (less mature) orchard; however, in samples conditioned using the lower fan speed from the north (more mature) orchard, these levels were unacceptable.

IMPLICATIONS

The results of this research indicate that PM emissions from almond sweeping operations may be reduced by use of reduced-pass sweepers. Additionally, increased efficiencies in fuel consumption and time required for sweeping may be realized by use of reduced-pass sweepers. Reducing harvester separation fan speeds results in lower emissions from nut conditioning, but foreign matter levels in conditioned samples from more mature orchards were unacceptable.

INTRODUCTION

California almond farmers produce over 75% of the world's almonds. In 2009, approximately 630 Gg of almonds were harvested in California on approximately 254,000 bearing hectares, with a total value of $1.8 billion.1 Over 80% (209,000 ha) of the bearing crop is located within the San Joaquin Valley Air Pollution Control District, which is aggressively working to reduce PM₁₀ and PM_2.5 (particulate matter [PM] ≤10 and ≤2.5 μm in nominal aerodynamic diameter, respectively) emissions from all sources, including agricultural industries. The current emission factor applied to all almond harvesting operations is 4570 kg PM₁₀/km²·yr, accounting for 12 Gg of PM₁₀ each year.2

For the past several years, the Almond Board of California has been investigating abatement strategies to reduce PM emissions from almond harvest operations. Goodrich et al.3 compared conventional almond sweeping (using three blower-passes per harvested row) and reduced-pass almond sweeping (using one blower-pass per harvested row). The authors reported that reducing the number of blower-passes from three to one lowered the average PM₁₀ emissions by 49%.

Sweeping practices may affect emissions from pickup operations because sweeping practices that increase the amount of soil material in the windrow may increase PM emissions during nut pickup. Some sweeper operators set the sweeper head lower than recommended by manufacturers in an attempt to decrease the number of unhar-vested nuts left on the orchard floor. Downey et al.4 tested the change in opacity of the harvester PM plume due to setting the sweeper head to induce 1.27 cm (0.5 in.) of ground interference relative to the same sweeper set according to manufacturer specifications. They reported a 32% increase in opacity measurements in the plume from nut pickup operations harvesting windrows of nuts after using improper sweeper head adjustments compared with PM emitted from harvest operations of windrows formed with proper settings.

In an effort to determine the implications of the work of Downey et al.4 for regulated pollutants, Faulkner and Capareda5 measured concentrations of total suspended particulate (TSP), PM₁₀, and PM_2.5 at the edge of an orchard and used inverse dispersion modeling to back-calculate emission rates to test the effect of sweeping depth on PM emissions from both operations. They reported no differences in emissions of PM₁₀ or PM_2.5 during sweeping as a function of sweeper depth, but, like Downey et al.,4 Faulkner and Capareda5 reported that PM emissions during pickup of windrows formed using proper sweeper settings trended lower (by ∼50%) than those formed using improper sweeper settings.

Downey et al.4 also tested the effect of reducing harvester ground speed on opacity measurements in the exhaust plume of almond pickup machines. They found that reducing harvester ground speed without reducing the power take-off speed of the tractor led to lower opacity measurements in the plume relative to emissions from harvesters operating at typical speeds. In a follow-up study, Faulkner et al.6 again used measured ambient PM concentrations and inverse dispersion modeling from nut pickup operations to test the effect of reduced harvester ground speed on PM emissions. In line with the findings of Downey et al.,4 Faulkner et al.6 reported that TSP emission rates were lower for the slower harvester speed, but no differences were detected in PM₁₀ or PM_2.5 emission rates as a result of changing harvester ground speed.

Little work has been published considering modifications to machine design for the purpose of reducing PM emissions. Southard et al.7 analyzed the effect of cleaning chain length on dust generation, and Whitelock et al.8 developed a prototype cyclone system for removing PM from the exhaust stream of a pecan harvester. The device developed by Whitelock et al.8 has shown promise in pecan orchards, but the large size of the system makes application in almond orchards problematic.

Ponpesh et al.9 tested the effects of decreasing harvester separation fan speed on relative PM emissions. The authors reported that decreasing harvester airflow by 18% resulted in a decrease in plume opacity and reduced in-orchard TSP and PM₁₀ concentrations by 62 and 77%, respectively. The results of Ponpesh et al.9 indicate that decreasing separation fan speeds may result in decreased emissions from the harvester, but they do not quantify differences at the edge of the orchard, where ambient air quality standards are often applied by state air pollution regulatory agencies.

To save on labor costs and harvesting time, several manufacturers of orchard harvesting equipment have introduced reduced-pass nut sweepers with auxiliary sweeping units that allow a single harvester to cover a wider width in one pass (Figure 1). These machines reduce the number of passes and total time required for sweeping an orchard. As such, there is speculation that these machines may also reduce the total emissions of PM₁₀ and/or PM_2.5 during sweeping operations relative to conventional sweepers.

Figure 1. Reduced-pass sweeper.

The objectives of this study were as follows:

•	Evaluate the effectiveness of reduced-pass sweepers for reducing PM emissions from almond sweeping operations relative to conventional sweepers
•	Evaluate differences in in-orchard time and fuel use between reduced-pass and conventional sweepers
•	Evaluate the effect of reducing harvester separation fan speeds on PM emissions from almond conditioning operations
•	Identify changes in composition of windrowed materials and conditioned almonds on the basis of sweeper treatment and harvester separation fan speed.

METHODS

Sampling was conducted in two orchards in the Central Sacramento Valley near Arbuckle, CA, in August 2009. Both orchards were planted in Hillgate loam that was 18.8% clay. Trees in the north orchard were 11 yr old and irrigated with aboveground irrigation whereas trees in the south orchard were 9 yr old and were irrigated using subsurface drip irrigation. It should be noted that the 2 yr difference in orchard age was visibly noticeable with regards to tree size. In both orchards, trees were planted in approximately 400-m (0.25-mi) rows oriented in a north-south direction with 6.7 m (22 ft) between rows and 5.5 m (18 ft) between trees in the same row.

Each plot consisted of 10 tree rows. Almond growers commonly plant a combination of almond varieties in a given orchard to achieve cross-pollination. The usual combination is a nonpareil variety with a “pollinator” variety or a nonpareil with two “pollinator” varieties, such as carmel and butte, in each orchard. In newer orchards, including those in which sampling was conducted, the nonpareil varieties are normally planted every other row with the other varieties planted on an alternating basis. However, during the harvesting of nonpareils, all windrows are used for pickup and conditioning operations, virtually using the whole area for the harvest process. Therefore, although each plot consisted of 10 tree rows and 10 windrows were created, only the nuts from 5 tree rows were harvested during the tested harvest operations. Sampling was conducted during sweeping of all plots. Nuts were then allowed to air-dry in windrows before sampling was again conducted on the same plots during windrow conditioning.

Sweeping Trials

Conventional and reduced-pass sweeping tests were conducted using Flory 77 Series sweepers (Flory Industries) with and without an auxiliary sweeper unit, respectively. Windrows were prepared using a conventional sweeper (two blow-passes and two sweeping passes per tree row) or a reduced-pass sweeper (two passes while simultaneously blowing and sweeping per tree row). The conventional and reduced-pass sweeper used similar engines (60 kW at 2500 rpm and displacement of 4.5 L). Because of constraints from the cooperating grower, all conventional sweeping trials were conducted in the north (more mature, aboveground drip irrigated) orchard, whereas reduced-pass trials were conducted in the south (less mature, subsurface drip irrigated) orchard (Figure 2). Therefore, analyses of the effects of sweeping treatments on PM emissions and the composition of windrows may be confounded by effects of orchard age, differences in soil structure (although the soil type was consistent between orchards), and effects of irrigation methods. Eight trials were conducted using each sweeping treatment.

Figure 2. Orchard layout for sweeping and windrow conditioning tests (not to scale).

During sweeping operations, collocated, low-volume TSP and Federal Reference Method (FRM) PM₁₀ samplers (model PQ100 Inlet; BGI, Inc.) were placed nominally upwind and downwind of each plot to measure the change in ambient PM concentrations due to sweeping. One collocated set of samplers was located upwind of each plot and four collocated sets of samplers were spaced evenly along the width of each plot approximately 15 m (50 ft) downwind from the northern or southern edge of the plot (Figure 3).

Figure 3. Sampler configuration (not to scale).

Because of the errors associated with FRM sampling in agricultural environments identified by Buser et al.,10 TSP concentrations were measured alongside FRM PM₁₀ samplers. TSP measurements were conducted with samplers designed by Wanjura et al.11 to reduce variations in sampler flow rate that lead to high uncertainty in FRM TSP concentration measurements. PM₁₀ measurements were conducted using the same airflow control unit as the TSP samplers and an FRM PM₁₀ sampling inlet. Tests from each plot lasted approximately 1 hr. Data collection and analysis were conducted using the methods described by Faulkner et al.6 In summary, the particle size distribution (PSD) of PM collected on TSP filters having more than 200 μg of PM were analyzed using a particle size analyzer (Mastersizer 2000, Malvern Instruments, Inc.). The PSD (described by a log-normal mass distribution) of each sample was determined and characterized by a mass median diameter (MMD) and geometric standard deviation (GSD).12 The MMDs were converted from equivalent spherical diameter (ESD) to aerodynamic equivalent diameter (AED) using a particle density (ρ_p) of 2.6 g/cm³ (on the basis of soil particle samples taken from each orchard and analyzed using a pycnometer [AccuPyc 1330, Micro-metrics]) and a shape factor of 1.05 (eq 1). This shape factor was a departure from earlier work described by Faulkner et al..6 In the study presented here, a particle shape factor of 1.05 was used given the slightly aspherical shape of soil particles collected from the orchards during sampling as seen using a scanning electron microscope (Figure 4).

Figure 4. Scanning electron microscope image of collected PM particle.

(1)

where ρ_p is in grams per cubic centimeter and χ is the shape factor.

The resulting PSD was then used to determine the true percentage of PM₁₀ and PM_2.5 on each filter according to eq 2.

(2)

where C _i is the concentration of PM smaller than or equal to size i, C _TSP is the concentration of TSP, i is the indicator size (10 μm for PM₁₀ and 2.5 μm for PM_2.5), and f(x) is the probability density function of PSD of the PM.

During concentration measurements, the following instruments were used to collect on-site meteorological data in an open field 20 m north of the sampled orchard:

•	A two-dimensional sonic anemometer (Wind-Sonic1, Gill Instruments, Ltd.) was used to measure the wind speed (accuracy ±2%) and direction (accuracy ±3°) 3 m above the ground surface at a frequency of 4 Hz.
•	A three-dimensional sonic anemometer (model 81000, R.M. Young, Co.) was used to collect data for use in defining the stability of the surface layer (accuracy: wind speed ±1% root mean square (rms); wind direction ±2°) at 2 m above the ground at a sampling frequency of 4 Hz.
•	A barometric pressure sensor (model 278, Setra Systems, Inc.; accuracy: ±0.25%) recorded every 5 min.
•	A temperature (accuracy ±0.5 °C) and relative humidity (accuracy: ±1.5%) probe mounted in a solar radiation shield at 2 m (HMP50, Campbell Scientific, Inc.) recorded every 5 min.
•	Two pyranometers, one mounted face up (CMP 22, Kipp and Zonen; accuracy: ±3 W/m2) and one mounted face down (CMP 6, Kipp and Zonen; accuracy: ±4 W/m2) were used to measure net solar radiation at a sampling frequency of 5 min.

Additional meteorological parameters were calculated according to U.S. Environmental Protection Agency guidance.13 The dimensions of each test plot and corresponding meteorological data were then used with the American Meteorological Society/Environmental Protection Agency Regulatory Model (AERMOD) to determine fluxes (μg/m²·sec) from each of the downwind samplers for each sampling period according to the protocol described by Faulkner et al.6 Each of the four sampler sets used at each plot provided an independent measurement of concentration, leading to four independent estimates of the emissions flux for each plot. These four fluxes were considered replicated measurements of emissions for a given plot, such that eight average fluxes (4 TSP; 4 FRM PM₁₀) were used to determine the emissions for each sweeping treatment.

In-orchard concentrations of PM₁₀ and TSP were also collected during tests. Mini-Vol gravimetric air samplers (Mini-Vol Portable Air Sampler, Airmetrics, Inc.) were placed in the middle of each replicate test block during all sweeping and windrow conditioning tests. The samplers, which have been described elsewhere,9 operate at a flow rate of 5 L min⁻¹ and use manufacturer-supplied heads with embedded filters for gravimetrically collecting TSP and PM₁₀ during field tests. Data from these measurements are presented as mass collected during the field tests. Air samplers were suspended approximately 1.5 m above the ground surface and were aligned with the lower portion of tree canopies (Figure 5).

Figure 5. Windrow conditioning while measuring in-orchard TSP/PM₁₀ and opacity.

Before sweeping tests, nuts were collected beneath five trees from separate rows within each plot. Because of the large mass for each sample, five 0.5-kg sub-samples were obtained from each primary sample. These subsamples were weighed, and the number of nuts was determined. Results allowed transformation back to the total number of nuts per sample area on the basis of total mass of the primary samples. Figure 6 shows a typical sample area for nut collection. Nuts remaining after the sweeping operation were counted from the previously sampled areas to establish sweeper efficiency estimates.

Figure 6. Example sample area for pre- and postsweep nut counts (not to scale).

After sweeping operations, five samples were collected from separate windrows within each plot for material separation analysis (described later) and to determine if differences existed between windrows on the basis of sweeper type. Additionally, fuel and time-in-orchard efficiency estimates were determined during sweeping operations. Sweeper time in orchard was determined for each test block; sweeper fuel consumption efficiency was determined from elapsed engine hours (neglecting idle engine hours required for replicated test setups) and diesel fuel volume consumed throughout the test period.

Conditioning Trials

Conditioning trials were conducted using a Flory Industries model 8500 self-propelled harvester operated at a constant ground speed averaging 5.1 km/hr for both orchards. Emissions were measured using the standard fan speed for almonds of 1080 rpm (control) and an experimental fan speed of 930 rpm, which was achieved by replacing the drive belt and sheaves such that ground speed was not altered. Although conventional and reduced-pass sweeping tests were carried out in separate orchards, harvester separation fan speed tests during windrow conditioning were randomized through both orchards. The result was that four conditioning trials at each fan speed were conducted in the north orchard (conventional sweeper) and four were conducted in the south orchard (reduced-pass sweeper). Sampling and data analysis for conditioning trials using FRM and Mini-Vol samplers were conducted in the same manner as described for sweeping trials.

Separation fan exhaust was also measured using an opacity monitoring device (model FW300; Sick Maihak, GmbH). The monitor and field staging of the device has been described elsewhere4,9; a brief description follows. The device is a two-step direct transmissometer; visible light (650 nm) is transmitted across a measuring path to a reflector and back to a photodiode detector. Light attenuation due to PM-laden air is transformed from an electrical signal to percent opacity in which 0% opacity represents clean air. An external purge air blower (supplied by the manufacturer) provides clean, PM-free air for maintaining optical integrity. Opacity data were captured at

1-sec intervals and retrieved from a laptop computer after field testing using the manufacturer's software. Field staging of the opacity monitor is shown in Figure 5. The measurement device was staged on a small trailer and aligned along one tree gap, two rows downwind from the harvester, to sample the PM plume as the machine moved through the orchard. Six and eight test blocks during standard and reduced fan speed windrow conditioning were measured, respectively, with the opacity monitor. Three replicate opacity measurements were obtained for each measurement event within each test block. Each replicated measurement (one pass of the harvester during windrow conditioning) required repositioning the trailer (i.e., following the harvester through the test block) to capture data.

Size Fractionation

After windrow conditioning, five windrow samples were collected from each plot in the different orchards coinciding with the different separation fan speed tests. Five subsamples (0.5 kg each) were collected from each primary sample for sieve analysis (size separation). Each sub-sample was placed in a sieve series and mechanically separated using conditions reported previously.9 Retained materials on the separate sieves were collected and weighed to establish if differences existed from different fan speed settings (or sweeper type, as discussed earlier). The following size ranges were used: particle sizes greater than 18.850 mm represent nuts and twigs; particle sizes between 9.423 and 18.850 mm represent leaves, small nuts, and twigs; particle sizes between 5.6 and 9.423 mm represent leaves and grass; particle sizes between 2 and 5.6 mm represent grass; and particle sizes of 2 mm or less represent soil.

Windrow samples were averaged based on their location and experimental treatment; that is, results from wind-rows after sweeping within the north orchard were averaged separate from windrow materials within the south orchard. Results from windrow conditioning samples were averaged based on fan speed and orchard location. Average mass fractions of the sieve separations were analyzed with two tests from the Statistical Analysis Software—analysis of variance (ANOVA) and Duncan's new multiple range test (α = 0.05). All other in-orchard field data results are reported as averages with standard errors on the basis of orchard location and/or machine type.

RESULTS AND DISCUSSION

Meteorological conditions during sampling are shown in Table 1. Average characteristics of PSDs measured during sweeping and conditioning trials are shown in Table 2. As expected, the PSD of PM emitted during harvest operations is not dependent on sweeping or conditioning treatments but on the nature of the parent material, which did not differ between tests. No statistical differences in PSDs were detected between sweeping treatments (P = 0.575 for MMD; P = 0.917 for GSD). Similarly, no statistical differences in MMDs or GSDs measured during conditioning trials were detected among separation fan speeds (P = 0.659 for MMD; P = 0.591 for GSD), sweeping treatment (P = 0.581 for MMD; P = 0.624 for GSD), or fan-speed–sweeper interactions (P = 0.175 for MMD; P = 0.712 for GSD).

Table 1. Meteorological parameters measured on-site during sampling

Parameter	Sweeping						Conditioning
	Conventional			Reduced Pass			1080 rpm			930 rpm
	Min	Max	Average	Min	Max	Average	Min	Max	Average	Min	Max	Average
Albedo	0.15	0.31	0.18	0.16	0.33	0.20	0.15	0.32	0.18	0.16	0.17	0.16
Bowen ratio	0.10	0.57	0.23	0.20	0.32	0.24	0.15	0.21	0.19	0.10	0.22	0.15
Relative humidity (%)	47	61	54	35	47	40	39	51	45	21	30	26
Temperature (°C)	25	29	27	31	32	32	26	28	27	30	31	30
Solar radiation (W/m^2)	506	753	647	551	734	652	601	749	686	613	725	671
Wind speed (m/sec)	0.9	1.8	1.4	1.0	1.6	1.3	1.0	1.7	1.4	1.2	1.6	1.4

Table 2. PSD parameters from TSP filters

Parameter	Sweeping		Conditioning
	Conventional	Reduced Pass	1080 rpm	930 rpm
MMD (μm AED)	12.2 x	12.0 x	12.1 x	12.4 x
GSD	1.97 x	1.96 x	2.03 x	2.07 x
Average % PM10	38.5	39.5	39.5	38.3
Average % PM2.5	1.1	1.2	1.5	1.6
Notes: No statistical differences were detected in means in the same row followed by the same letter (α = 0.05).

Sweeping Trials

Average in-orchard TSP and PM₁₀ concentrations from sweeping trials are shown in Table 3 along with emissions estimates derived from ambient concentrations, inverse dispersion modeling, and particle size analysis. Reduced-pass sweeping resulted in approximately 33% less TSP and PM₁₀ within the orchard canopy (although not significantly different at the 95% confidence level), whereas 66 and 48% reductions were seen in TSP and PM₁₀ emissions measured using an FRM sampler, respectively. However, the different sweeping operations were carried out in two separate orchards—conventional sweeping was done within a more mature orchard (north orchard) with microemitter surface irrigation whereas reduced-pass sweeping was done within a less mature orchard (south orchard) with subsurface irrigation.

Table 3. Mass and emissions from sweeping treatments

Sweeping Operation	In-Orchard Mass Measurements (mg)		Measured Emissions (kg/km^2)
	TSP	PM10	TSP	True PM10	True PM2.5	FRM PM10
Conventional	0.56 (0.07)x	0.15 (0.01) x	333 (65) x	128 (25) x	4 (0.7) x	170 (40) x
Reduced pass	0.38 (0.06) x	0.10 (0.01) y	112 (23) y	44 (9) y	1 (0.3) y	89 (27) y
Notes: No statistical differences were detected in means in the same column followed by the same letter (α = 0.05). Standard errors are shown in parentheses. All conventional sweeping was conducted in the north orchard; all reduced-pass sweeping was conducted in the south orchard.

Nut count measurements for determining the efficiency from the different sweeping operations are given in Table 4. Results indicate that the average tree within the south orchard produced approximately 60% less product than the north orchard. Additionally, similar numbers of nuts were left within each orchard (end row effects of nuts left after sweeping were not determined). Results indicate that both sweepers recovered at least 99.7% of the nuts from the orchard floor.

Table 4. Average number of nuts before and after sweeping operations

Sweeping Operation	Before Sweeping	After Sweeping	Nut Recovery (%)
Conventional (north orchard)	4898 (91)	6 (0.5)	99.9
Reduced pass (south orchard)	1914 (80)	5 (0.8)	99.7
Notes: Standard errors shown in parentheses.

Sweeper fuel use and time-in-orchard estimates are given in Table 5. Results show that the conventional sweeper was approximately 27% more fuel efficient per engine hour; however, time to sweep the test block was 51% longer compared with the reduced-pass sweeper. This result is not unexpected because the conventional sweeper required two more passes per windrow compared with the reduced-pass sweeper. However, considering the test block area (∼2.6 ha), the reduced-pass sweeper was more efficient per unit area compared with the conventional sweeper.

Table 5. Sweeper time-in-orchard and fuel use results

			Fuel Consumed
Sweeping Operation	Ground Speed (km/hr)	Engine Hours per Block (hr)	Per Engine Hour (L/hr)	Per Unit Area (L/ha)
Conventional (north orchard)	5.4 (0.26)	1.42 (0.11)	5.8	3.2
Reduced pass (south orchard)	4.3 (0.07)	0.94 (0.03)	7.9	2.8
Notes: Standard errors shown in parentheses.

Results from size separation analysis of windrows after sweeping within the respective orchards are given in Table 6. An ANOVA using a one-way classification for determining the influence of sweepers within the north and south orchards found that the two largest size ranges of materials were significantly different within the separate orchards. Multiple range tests for the mass fraction size ranges were also analyzed. Results indicate that conventional sweeping within the north orchard produced more product in the largest size range, whereas the south orchard produced more material within the next lowest size range. This result also indicates the maturity difference of the orchards as indicated from nut count estimates. More small nuts were produced in the south orchard compared with the north orchard. Both orchards were similar in the amount of material represented within the smallest size ranges (leaves, grass, and soil).

Table 6. Size separation results of windrow samples

	Mass Fraction (%)
Size Separation Range	Conventional	Reduced Pass
Nuts and twigs	75.3 (0.4) x	69.4 (1.0) y
Leaves, small nuts, and twigs	8.1 (0.6) x	13.4 (0.5) y
Leaves and grass	5.1 (0.2) x	4.5 (0.2) x
Grass	6.7 (0.3) x	6.8 (0.3) x
Soil	4.6 (0.2) x	5.9 (0.3) x
Notes: No statistical differences were detected in means in the same row followed by the same letter (α = 0.05). Standard errors are shown in parentheses.

Conditioning Trials

Although conventional and reduced-pass sweeping tests were carried out in separate orchards, harvester separation fan speed tests during windrow conditioning were randomized through both orchards. Average masses of TSP and PM₁₀ collected in the orchard during conditioning trials are shown in Table 7 and indicate that a slower separation fan speed resulted in 26% less TSP and 42% more PM₁₀ within the canopy during conditioning within the north orchard. Slower fan speeds within the south orchard resulted in 33% less within-canopy TSP than the higher separation fan speed, whereas PM₁₀ mea surements were similar. In all cases the south (less mature) orchard in-canopy measurements of TSP and PM₁₀ were less than the north orchard. Opacity measurements and the duration (“time span”) of dust plumes emitted during conditioning were of similar magnitude within each orchard when comparing the separation fan speeds; however, when comparing the north (more established, surface irrigated) orchard to the south orchard, opacity measurements were close to 50% higher and time spans of the emitted fan exhaust plume were 30% longer (results from in-canopy measurements were not significantly different between treatments). The time span from dust plumes is a concern within fields adjacent to urban areas because it indicates the duration with which a plume may persist in a given area.

Table 7. Within-canopy results from conditioning treatments

Conditioning Treatment	North Orchard				South Orchard
	TSP (mg)	PM10 (mg)	Opacity (%)	Time Span (sec)	TSP (mg)	PM10 (mg)	Opacity (%)	Time Span (sec)
1080 rpm	0.46 (0.12)	0.07 (0.03)	4.4 (0.8)	26 (2)	0.39 (0.16)	0.05 (0.02)	2.2 (0.3)	20 (2)
930 rpm	0.34 (0.14)	0.10 (0.04)	4.0 (0.8)	28 (6)	0.26 (0.06)	0.04 (0.01)	1.8 (0.3)	19 (3)
Notes: Standard errors shown in parentheses.

Emissions estimates derived from ambient concentrations, inverse dispersion modeling, and particle size analysis are shown in Table 8. Significant differences were detected in conditioning emissions as a function of fan speed (P = 0.002) but not sweeping method (P = 0.397) or sweeping method-fan speed interactions (P = 0.592). Reducing separation fan speed resulted in substantially lower emissions of TSP and PM₁₀. From Table 8, reducing the fan speed by 14% led to a reduction in PM emissions of over 65%. Although there was likely a reduction in emissions with decreased separation fan speed (lower in-canopy masses, lower opacity values, and lower concentrations at the orchard edge were recorded at the lower fan speed), it is unlikely that emissions were reduced by 65%. Faulkner et al.14 demonstrated that AERMOD is increasingly sensitive to changes in wind speed at values below 3 m/sec. The maximum wind speed during all conditioning trials was 1.7 m/sec, so all dispersion modeling of these tests occurred within the wind speed range at which AERMOD is particularly sensitive to changes in the input parameter.

Table 8. Emissions from conditioning treatments (kg/km²)

Conditioning Treatment	TSP	True PM10	True PM2.5	FRM PM10
1080 rpm	4,018 (489) x	1,585 (193) x	60 (7) x	1,863 (278) x
930 rpm	1,111 (658) y	425 (252) y	18 (10) y	615 (385) y
Notes: No statistical differences were detected in means in the same row followed by the same letter (α = 0.05). Standard errors are shown in parentheses.

Results from the size separation analyses after wind-row conditioning for different separation fan speeds are given in Table 9. Results show that a larger percentage of the desired product (i.e., that within the largest size range) remained within windrows at the faster separation fan speed in the north orchard. Additionally, the slower separation fan speed resulted in retention of at least 60% more of the three smaller size ranges (undesirable product) within the north orchard. Comparing separation fan speeds within the south orchard showed an opposite effect with respect to the smaller size range of material within windrows. Here the slower separation fan speed resulted in approximately 45% less grass within the conditioned windrows, whereas retention of the larger sizes (representing desirable product) was similar. From these results and knowledge of the maturity and irrigation systems within these two orchards, the results imply that the standard separation fan speed should be used for windrow conditioning within mature orchards whereas a reduced fan separation speed may lead to comparable foreign material within harvested product in younger orchards.

Table 9. Size separation results after windrow conditioning

Size Separation Range	Mass Fraction (%)
	North Orchard		South Orchard
	1080 rpm	930 rpm	1080 rpm	930 rpm
Nuts and twigs	84.7 (1.2) x	66.6 (2.0) z	72.4 (2.0) y,z	73.8 (1.0) y
Leaves, small nuts, and twigs	10.5 (0.9) x	19.1 (0.9) y	21.4 (1.6) y	21.5 (0.7) y
Leaves and grass	1.6 (0.3) x	4.1 (0.4) y	1.7 (0.2) x	1.8 (0.2) x
Grass	1.3 (0.2) x	5.5 (0.6) z	2.5 (0.4) y	1.4 (0.1) x
Soil	1.9 (0.2) x	4.7 (0.5) y	2.0 (0.3) x	1.5 (0.4) x
Notes: No statistical differences were detected in means in the same row followed by the same letter (α = 0.05). Standard errors are shown in parentheses.

Additional evaluation of the data using an ANOVA analysis with one-way factorial design with the test blocks randomized across the north and south orchards found there were no significant differences between the size ranges of materials on the basis of harvester separation fan speed. Multiple range tests found that the effects of fan speeds were similar. However, the one-way factorial design also found relatively high rms errors similar in magnitude to the average mass fractions of materials within the respective size ranges. This indicates large variations in the data (understandable with respect to the age of orchards and differences in product yields reported earlier) and the need for larger sample sizes to include product yield as a factor for further analysis.

CONCLUSIONS

The effects of using reduced-pass sweepers (vs. conventional sweepers) and lower separation fan speeds (930 vs. 1080 rpm) on PM emissions from almond harvesting operations were evaluated. Differences in sweeping time and fuel use were concurrently evaluated, and effects of these potential mitigation measures on the composition of windrows and conditioned material were considered. No differences were detected in the PSD characteristics of PM emitted from each operation. Differences in in-canopy masses of TSP or PM₁₀ and/or opacity did not correspond to differences in emissions as determined using concentration measurements at the orchard boundary coupled with inverse dispersion modeling.

Reduced-pass sweeping showed the potential for reducing emissions, but results were confounded by differences in orchard maturity and irrigation methods and were therefore inconclusive. Reduced-pass sweeping demonstrated comparable nut recovery to conventional sweeping (although the conventionally swept orchard produced 60% more product than the orchard swept using the reduced-pass sweeper), but fuel use and time required per unit area were reduced by the reduced-pass sweeper.

Reducing the separation fan speed from 1080 to 930 rpm led to reductions in PM emissions. Generally the slower fan speed during windrow conditioning resulted in less measured mass of in-canopy TSP. In all cases the south orchard resulted in less in-canopy TSP and PM₁₀ versus the north (more mature) orchard. Opacity and time spans of dust plumes were 50% higher and 30% longer, respectively, within the north orchard. Although foreign matter levels were only nominally affected by separation fan speed in the south (less mature) orchard, foreign matter levels in samples conditioned using the lower fan speed from the north (more mature) orchard were unacceptable. The results of these tests imply that the standard separation fan speed should be used for windrow conditioning within mature orchards whereas a reduced fan separation speed may lead to comparable foreign material within harvested product in younger orchards.

ACKNOWLEDGMENTS

The researchers acknowledge the Almond Board of California and the National Research Initiative Competitive Grant no. 2009-55112-05217 from the U.S. Department of Agriculture Cooperative State Research, Education, and Extension Service Air Quality Program for support of this project.