NMSU: A Note on the Measurement of Dust Emissions from Moving Sources in Agricultural Field Operations
NMSU branding

Authors: Respectively, College Assistant Professor, Department of Plant and Environmental Sciences, New Mexico State University; Assistant Professor, Department of Environmental Sciences, Louisiana State University; Emeritus Professor, Department of Natural Resources Management and Engineering, University of Connecticut; Professor, Department of Plant and Environmental Sciences, New Mexico State University; Air Resources Engineer, California Air Resources Board; and Associate Professor, Department of Civil and Environmental Engineering, University of Vermont.


Dust emissions from agricultural field operations can cause air quality and human health problems. The objective of this paper is to provide a simple method to measure dust emissions using a single particulate matter (PM) sensor paired with a fast response sonic anemometer for repeated measurements of emission rates (source strength) from a single moving tractor operation in an agriculture field. The method was developed for a pre-plant disking operation in a cotton field with unusually high soil moisture (28%) in the Rio Grande Valley, Las Cruces, NM. The PM10 (particulate matter with diameter ≤ 10 μm) and total suspended particulate (PMTSP) source strengths averaged 70 (±38) g ha-1 and 76 (±42) g ha-1, respectively, over 23 replications. The procedures, calculations, and assumptions of the method are detailed. Weak relationships with atmospheric parameters revealed that source strength is mainly affected by operation type, equipment, and soil conditions. This measurement method can be used for vapors, gases, and other pollutants emitted from ground level if a fast response pollutant-specific sensor is used.


Reduction of air pollutants is primarily achieved by regulating emissions from specific sources. Agricultural operations are under scrutiny as a source of airborne particulate matter (PM), but the wide variety of methods used to research and document the emissions of particulate matter from agricultural field operations leads to disparity in estimated emissions and uncertainty of agriculture's emissions contributions. The lack of a standard method to determine source strength from individual field operations has added considerably to the confusion. The closest to a standard method available is the "integrated horizontal flux method" (Wilson and Shum, 1992). It requires time-averaged measurements of the downwind vertical profile of particulate concentrations, C (g m-3), from fixed sensors, multiplied by the vertical wind speed, U (m s-1), profile. The vertical integration of the CU (g m-2 s-1) profile then provides the horizontal flux through the vertical plane, which contains the measurement profile. Finally, time integration provides the total emission flux when normalized by the field dimensions. Its major problem is the inherent assumption of stationary wind direction for the entire averaging time (i.e., it must move the emitted plume across the sampler array for the entire averaging time). Also, the separation of dust emitted from the field from the background load of dust in the air requires a simultaneous duplicate wind and sampler profile upwind of the field. These problems require a significant investment in multiple sensors and a large number of replications to obtain any statistical accuracy.

Early studies that quantified agricultural particulate emissions represented a very limited number of sites and a low number of replicate samples, and measured mostly total suspended particulate matter or industrial workplace respirable dust (PM4, PM with diameter ≤ 4 μm, in response to federal regulations existing at that time). For development of emission factors for fugitive dust sources, these previous fugitive dust studies employed upwind and downwind arrays (Cowherd et al., 1974; Flocchini et al., 1994) with high volume filtration samplers (Cuscino et al., 1984; Cowherd and Kinsey, 1986) and cascade impactors (Cowherd et al., 1974; Cuscino et al., 1984; Flocchini et al., 1994). Exposure profiling of farm workers (Cowherd et al., 1974; Cuscino et al., 1984; Flocchini et al., 1994) and respirable dust monitor measurements (Snyder and Blackwood, 1977; Clausnitzer and Singer, 1996, 1997) were used to establish farmer exposure but did not quantify emissions to the atmosphere.

Some studies have used field measurements from PM10 (particulate matter with diameter ≤ 10 μm) samplers at a single height (3.3 m) upwind and downwind, and from total suspended particulate (PMTSP) monitors at four heights (3, 5, 7, and 9 m) downwind from the source (Flocchini et al., 1994; Ashbaugh et al., 1996), and then used a box model to estimate emission factors (Ashbaugh et al., 1997). There have also been studies using numerical models to back-calculate the source strength based on the measured atmospheric conditions and the concentrations at downwind locations (e.g., Aylor and Flesh, 2001; Aylor and Qiu, 1996).

All of these fixed upwind/downwind measurement methods suffer from unacceptable errors when using the results for general regulation of the activity. One problem arises because the fixed upwind/downwind methods assume a constant, or at least near constant, mean wind direction for the period of measurement. If the period is long—over a number of hours or days—the turbulent wind direction is highly variable, especially if day and night periods are integrated in the measurements. In this case, the "upwind" and "downwind" measurements are aliased and mean very little. If shorter periods are measured or conditionally sampled in the measurements (i.e., hour averages), the centerline of the plume from the tractor operation is not likely to move over the downwind sensors for a significant part of the hour. Thus, extrapolating the measurements across the average plume is not possible because there is no knowledge of where in the emitted plume the measurements were made. This problem has been solved to date by multiple arrays of samplers, which are at short but varying distances from the tractor path; however, the complications of multiple arrayed samplers limit the number of replications (Paez-Rubio et al., 2006).

A second problem, especially when trying to generalize emissions, is that operation source strength is highly dependent on soil moisture conditions, the particular field operation, and soil texture. To generally define source strengths from field operations, the entire range of these variables must be captured in repeated experiments with each crop, which would require thousands of repetitions over numerous locations and over numerous soil moisture conditions. The "integrated horizontal flux measurement" technique is too difficult to conduct on anything but a research project basis, and the results are too uncertain to use in the above manner.

New mobile lidar technology shows considerable promise to improve the situation because it can obtain accurate spatial distributions of PM10, while point samplers cannot (Holmén et al., 2001a, 2001b; Cassel et al., 2003; Hiscox et al., 2007; Holmén et al., 2007), but the expense and sophistication of this technology makes its general use doubtful in the next decade or two, if ever.

We believe that three improvements can help alleviate the problem of field measurements of source strength with point samplers. The first is the use of the new generation of accurate, fast response point samplers (i.e., the 1-s sampling rate samplers used by Paez-Rubio et al. [2006]). The second is making the samplers mobile to ensure they are in a position where the emitted dust plumes move across them (e.g., Coates, 1996). Thus, single passes across a field can be valid replications that will allow numerous independent measurements in a variety of different conditions in relatively short periods of time. The third is the use of an in-field, fast response 3-D sonic anemometer to measure the short-term wind field dynamics and allow accurate calculations of the near source dispersion and resulting amounts of aerosols crossing the sampler locations.

This paper details the use of these methods to measure the source strength of PM10 and PMTSP for a disking operation at Las Cruces, NM, when the soil had an average moisture content of 28%. The source strengths measured here represent the first calculations of this nature performed on time scales on the order of 1 s, which will improve significantly the accuracy of our dispersion model (Wang et al., 2008) for agricultural field operations.

Materials and Methods

The dust emitted from a source (e.g., tractor with implement) moving across a field can be considered as a series of puffs leaving the implement and traveling downwind (Wang et al., 2008). A puff is defined as a volume of dust produced from the operation when the equipment arrives at a location; it is assumed that puffs were contiguous.

Figure 1 depicts a puff produced during a disking operation and its increase in volume as it is moved downwind. A sensor located a short distance downwind from the tractor path measures the concentration of the puff, and then the source strength can be calculated as


where E is the source strength gs -1), V is the volume (m3) of the puff at the PM sensor's location, C is the average concentration (g m-3) of dust in the puff, and t is the time period over which the puff was created. If we assume the puff is cuboid-shaped, where the width of the puff (X) is the width of the source implement (i.e., disk) in the wind direction, the length of the puff (Y) is the length traveled by the implement in 1 s (Y = tST, where ST is the tractor speed [m s-1]), and the height of the puff (Z) is a function of the air turbulence, then


Particle concentrations, C(z) (g m-3), in the puff dispersed from the surface source decrease with height in the near-field and can be generally described by a logarithmic function


where the constants a and b are functions of the atmospheric turbulence expressed by the aerodynamic friction velocity (u*) and z is the height. The height of the puff (Z) is found by solving Equation 3 for C(z) = 0.

The average concentration in the puff is then


If the concentration sensors are just a few meters away from the source of the puff, the puff will pass over the sensors within just a few seconds when wind speed is greater than about 1 m s-1. In this short time, the spread of the puff in the horizontal directions will be negligible because advection is the primary horizontal transport mechanism, and X and Y can be assumed to be the same as their initial values (X0,Y0). But this does not hold with the height of the plume at the sensors (Z0), which must be estimated as explained below. Now the source strength in g s-1 is


The source strength can be converted to the unit of g m-2 using the field area worked during the period of interest.

Fig. 1: A schematic diagram of puff source, sensor location, and wind speed.

Figure 1. A puff is produced and travels to the PM sensors later during disking operations.

Field Measurements

Agricultural disking operations were conducted at New Mexico State University's Leyendecker Plant Science Research Center near Las Cruces, NM, on March 31, 2005 (32.2° N, 106.8° W; elev. 1,180 m). The details of the disking operations and initial experiments are described in Holmén et al. (2007) and Hiscox et al. (2007). The experimental field was about 250 m by 100 m and was flood irrigated. The soil was a mixture of Armijo clay loam and Harkey loam soil types. Cotton was planted in this field during the previous growing season. The average soil moisture was 28% and the average daytime air relative humidity was 14%.

Dust emissions from the field operations were measured by point samplers (described below) located in the field. The disking equipment (John Deere T0310, Moline, IL) and tractor (John Deere 782) made single passes across the field from the southwest to the northeast. Sampling was conducted during each pass and for several minutes after each pass. At the end of each pass, the tractor stopped at the end of the field and turned off its engine until all sampling was completed and the dust plume generated had moved out of the sampling area. Then another pass across the field was made and the sampling was repeated. This procedure allowed the use of each pass as an independent replicate. Measurements are available from 23 passes during the disking operation.

The samplers were mounted on the 3-point hitch of a second tractor so they could be repositioned easily before each pass to always be near the center of the field and a short distance (3.7–9.8 m) downwind from the tractor's path. The samplers were run continuously during the pass. During each pass, an optical PM10 sensor (GT640A, Met One Instruments, Inc., Grants Pass, OR) was located at 1.7 m height; it measured the 1-min average PM10 concentration (μg m-3). The sensor air flow rate was 4.0 L min-1. A second optical sensor (GT640A) sampled the 1-min average of PMTSP at the same location, height, and sampling frequency as the PM10 sampler.

An electrical low pressure cascade impactor (ELPI, Dekati LTD, Tampere, Finland) made size-distributed particle number measurements at 2-s intervals. It was located at the same downwind location and height as the two GT640As. It operated at 30.0 L min-1.

During the experiments, a 3-D sonic anemometer (CSAT3, Campbell Scientific Inc., Logan, UT) was used to measure the 20 Hz wind component velocities (u, v, w) and air temperature (T) at 1.5 m height at the field edge. The friction velocity (u*, m s-1), Monin-Obukhov Length (L, m), wind speed (U, m s-1), wind direction, and turbulence intensity (I = σw/U) were calculated for each sampling period (each pass). L is calculated after Stull (2001), L = -(ρ cp T u*3)/(k g H), where ρ is the air density, cp is the specific heat of air, T is the air temperature, u* is the friction velocity, k is the von Karman constant (0.4), g is the acceleration due to gravity, and H is the sensible heat flux.

Source Strength Calculations

Initial puff dimensions

We first defined the initial puff size as a cuboid at the location of the samplers. The initial cuboid had a width (X0) of 3.96 m, which was the total width of the disk, a length (Y0), which was the tractor's travel distance during 1 s, and a height (Z0), which varied with wind conditions as explained below (Figure 1).

As the initial cuboid (puff) travels downwind from the tractor to the sampler, turbulent dispersion will expand its size. It was assumed that, in the downwind direction, advection is the major dispersion factor. The travel time of the puff from the tractor to the sampler is calculated based on mean wind speed and direction, averaged over the 3-min time period of the tractor pass. The distance (D) between the sensors and the tractor path (the puff origin location) in the direction of wind ranged from 3.7 to 9.8 m for the various passes (Figure 1). The mean wind speed was from 3.6 to 6.3 m s-1, and the puff traveling time from the puff origin to the sensors ranged from 0.6 to 2.6 s. Since the traveling time was very short, the conditions for Equation 5 were satisfied; that is, the width (X) of the cuboid did not change (X = X0 = 3.96 m) when it arrived at the samplers. The same reasoning applies to the cuboid length, which is the length of travel by the tractor in one second, Y = Y0 = ST x 1 s.

In contrast to the horizontal dimensions, which are advected unchanged, the cuboid height and vertical concentration distribution at the sensors vary due to turbulent atmospheric conditions during the short travel time. The vertical upward dispersion causes the concentration profile to decrease with height. The sampler measurements in this experiment where taken at a single height, and the cuboid height could therefore not be measured directly. The cuboid height and concentration distribution were estimated based on previous vertical profile measurements from a lidar unit by Holmén et al. (2001a). Their experiments measured the downwind vertical concentration profiles of PM10 produced during disking operations in cotton fields. The measurement heights were from 1.7 to 50 m. Their measurements were separated into different wind speed categories. Among them, speeds of 3.53 to 4.62 m s-1 were closest to our experimental wind conditions. Therefore, we used their measured profiles under this wind category. The concentrations were normalized by the concentration value at the 1.7 m point. We used this profile to estimate our cuboid height at the sensors.

We fit a logarithmic curve to their reported measurements (see Figure 2) to yield an equation


where C(z)normalized is the concentration normalized by the z = 1.7 m value.

Setting C(z)normalized to 0 allows prediction of cuboid height for each class. The concentration reached a zero value at 14.8 m. Therefore, 14.8 m was used as cuboid height (Z0) for source strength calculations.

Fig. 2: Line graph showing downwind vertical profile of normalized concentration from cotton field disking operation.

Figure 2. Downwind vertical profile of normalized concentration from cotton field disking operation for wind speed of 3.54 to 4.62 m s-1 based on the lidar data in Holmén et al. (2001).

Concentration in the cuboid

As with the cuboid dimensions, the horizontal concentration distribution was assumed to be constant.

In the vertical dimension, the instantaneous peak concentration values of PM10 (C(z0)max10, µg m-3) and PMTSP (C(z0)maxTSP, µg m-3) at the cuboid passing time were used as the concentration at the measurement height, z0 = 1.7 m, which was then distributed vertically according to Equation 6. C(z0)max10 and C(z0)maxTSP were found using the following procedure.

The cuboid passed the PM sensors in a time period of about 1 s (period = X U-1, where U is the wind speed [m s-1] as demonstrated in Figure 3). U ranged from 3.6 to 6.3 m-1 for the various passes. The ELPI provided 2-s particle concentration numbers, and the values were expressed in the unit of particle number m-3. Accurate mass conversion is difficult without particle density measurements, which were not available. Therefore, to obtain physically realizable units, the data from the PM10 and PMTSP were used to provide the absolute mass concentration values. The PM10 and PMTSP sensors' 1-min average concentrations were scaled by the ELPI time series of 2-s concentrations to determine the peak values of PM10 (C(z0)max10) and PMTSP (C(z0)maxTSP). This scaling was accomplished by determining the ratio (R) of the peak ELPI-measured concentration (C(z0)MaxELPI) to the summation of concentration over 1 min measured by the ELPI during the puff transit time past the ELPI instrument (Figure 3; note that ELPI 1-min data have 30 data points)


where C(z0)ELPI is the mass concentration of ELPI at the measurement height (z0 = 1.7 m). The ELPI-measured number concentrations were converted to mass concentrations based on the ELPI User Manual, version 4. The ratio R is a relative value, and its accuracy is not related to the accuracy of the particulate density, which was not available.


where C(z0)PM10 is the corresponding measured 1-min PMC10 average mass concentration. Note that because R is calculated by 30 data points of 2-s ELPI data, C(z0)PM10 x 30 is the summation of PM10 concentration during each corresponding 1-min period. Similarly,


where C(z0)PMTSP is the corresponding 1-min PMTSP mass concentration.

The technique presented here scales peak concentration when the plume passing through the sensors was calculated by using the available sensor data of ELPI, PM10, and PMTSP. In future experiments, this scaling will not be necessary because faster response optical sensors of PM10 and PMTSP are available and will provide direct measurements of high frequency (1 s) peak concentration.

Fig. 3: Line graph showing example time series of normalized mass concentration from the ELPI cascade impactor.

Figure 3. Example time series of normalized mass concentration from the ELPI cascade impactor at height z0 = 1.7 m (C(z0)ELPI) during pass 3, puff traveling past the PM and ELPI sensors. In this case, Equation 7 equals 0.355. C(z0)MaxELPI = 1 is the normalized peak concentration.

Results and Discussion

Tables 1 and 2 list the micrometeorology and sample data for each of the 23 replicate tractor passes. During the disking operation, u* ranged from 0.27 to 0.53 m s-1 (Table 1), wind speed ranged from 3.6 to 6.3 m s-1, and wind direction varied from 312.5 to 26.9°. All atmospheric conditions were unstable (negative), except in the evening when they became stable. Turbulence intensity was approximately 0.1 during the passes.

Table 1. Micrometeorological Parameters for the 23 Passes

Pass # Time U
1 11:40:40–11:43:55 6.1 340.9 0.44 0.09 -0.6
2 11:47:52–11:50:36 4.9 326.8 0.36 0.11 -0.5
3 11:57:58–12:00:52 6.3 328.0 0.46 0.10 -2.3
4 12:04:49–12:07:43 4.5 344.9 0.33 0.14 -0.7
5 13:48:51–13:51:35 4.9 316.6 0.36 0.10 -1.3
6 14:00:53–14:03:36 3.7 323.3 0.27 0.13 -0.7
7 14:13:29–14:18:35 5.4 343.0 0.39 0.11 -0.7
8 14:34:40–14:37:34 4.9 323.2 0.35 0.12 -1.4
9 14:45:23–14:48:15 5.9 312.5 0.43 0.11 -1.0
10 14:58:30–15:01:00 4.4 327.5 0.32 0.11 -0.9
11 15:13:32–15:16:20 4.4 326.9 0.32 0.11 -0.7
12 15:44:07–15:46:57 4.6 333.8 0.33 0.12 -1.3
13 15:57:10–15:59:50 4.4 344.0 0.32 0.11 -1.1
14 16:18:50–16:21:31 5.4 338.0 0.39 0.10 -1.2
15 16:46:54–16:49:45 5.2 330.9 0.38 0.11 -3.3
16 16:55:37–16:58:30 5.0 329.9 0.36 0.10 -1.9
17 17:18:22–17:21:01 5.0 343.0 0.36 0.10 -3.1
18 17:27:32–17:30:16 4.4 359.9 0.32 0.11 -3.3
19 17:39:40–17:42:27 4.4 339.7 0.32 0.11 -7.1
20 17:49:20–17:52:03 3.6 358.6 0.26 0.11 -3.1
21 17:59:47–18:02:25 4.8 5.2 0.35 0.09 -5.4
22 18:09:13–18:11:53 5.7 9.4 0.41 0.09 -18.3
23 18:21:31–18:24:32 4.4 26.9 0.32 0.09 11.0

Table 2. Dust Concentrations at the Samplers and Source Strengths for the 23 Passes

Pass # C(z0)max10
(µg m-3)
(µg m-3)
(µg m-3)
(µg m-3)
(mg s-1)
(mg s-1)
(g ha-1)
(g ha-1)
1 1,071.1 1,178.2 754 830 32 35 63 69
2 347.7 382.5 245 269 12 13 20 23
3 644.0 708.4 537 590 21 23 38 42
4 1,062.6 1,168.8 687 756 35 39 63 69
5 1,657.5 1,823.2 1,167 1,284 58 64 98 107
6 1,364.5 1,501.0 882 970 48 53 80 88
7 1,232.5 1,355.7 868 955 23 25 73 80
8 601.7 661.8 424 466 20 22 35 39
9 797.1 876.8 561 617 27 29 47 52
10 825.2 907.7 533 587 32 35 49 53
11 3,237.0 3,560.7 2,093 2,302 111 122 191 210
12 2,218.7 2,440.6 1,434 1,578 75 83 131 144
13 1,830.7 2,013.7 1,183 1,302 66 72 108 119
14 785.2 863.7 553 608 28 31 46 51
15 798.5 878.4 562 618 27 30 47 52
16 732.5 805.8 516 567 24 27 43 47
17 414.1 455.5 292 321 15 16 24 27
18 955.0 1,050.5 617 679 33 37 56 62
19 1,305.1 1,435.6 844 928 45 49 77 85
20 1,117.7 1,229.5 723 795 39 43 66 72
21 1,009.7 1,110.7 653 718 37 40 59 65
22 1,765.9 1,942.5 1,243 1,368 63 70 104 114
23 1,288.3 1,417.1 833 916 41 45 76 83
Average 1,176.6 1,294.3 791 871 40 44 70 76
645.0 709.5 412 454 23 25 38 42

The average source strength for PM10 was 40 mg s-1, with a standard deviation (σ) of 23 mg s-1; this equates to 70 g ha-1 with σ = 38 g ha-1. For PMTSP, average source strength was 44 mg s-1, σ = 25 mg s-1 (average = 76 g ha-1, σ = 42 g ha-1). Note that 92% of the PMTSP was smaller than or equal to 10 µm in diameter (PM10). This 92% value was also confirmed by the particle size distribution, which was provided by the mass measurements using the ELPI and a micro-orifice uniform deposit impactor (MOUDI; Model 110, MSP Corporation, Shoreview, MN), located next to and at the same height as the PM10 samplers (details are in Holmén et al. [2008]).

Finally, the effect of atmospheric parameters on source strength was examined. Regression analyses of the source strengths as functions of the various meteorological parameters shown in Table 1 were conducted. The regressions showed that the source strength was essentially independent of the near-field micrometeorology. The strongest correlation with any of the atmospheric parameters was R2 = 0.2, thus reinforcing the idea that source strength from disking operations mainly depends on the equipment and the soil conditions rather than atmospheric variables. Soil drying over the day could have introduced some of the variability, but no trend with time was apparent.

The soil moisture in this study was high. Therefore, when compared to other studies in the literature, these source strengths are very low. Cassel et al. (2003) reported disking operations with PM10 source strength of 78 mg m-2 (780 g ha-2) for a cotton field with soil moisture of 13%. This was about 10 times higher than the source strength (70 g ha-1) in this study. Holmén et al. (2001a, 2001b) reported 8 out of 13 cotton field disking experiments (soil moisture 11.50–18.15%) resulted in PM10 source strengths 5 to 16 times higher than this study, 4 out of the 13 experiments had similar source strengths to this study (36, 76, 92, and 140 g ha-1), and one, at 6 g ha-1, was 12 times lower than source strengths in this study.


This paper reports on the use of in-field PM sensors for repeated measurement of emitted dust source strength from a pre-plant disking operation in an irrigated cotton field. The calculations and assumptions of the method are detailed. For this method to be utilized in future studies, we recommend the following procedures be followed.

  • Pre-knowledge of the vertical profile of relative concentration as a function of wind parameter (wind speed or friction wind velocity) downwind of the tractor should be needed. Therefore, a preliminary experiment with multiple samplers arrayed vertically is highly desirable.
  • A 3-D sonic anemometer, sampling rate 1 Hz, should be set up in the field.
  • Sampler(s) should be set downwind of the tractor path near the middle of the path swath. Rule of thumb: Distance from the path should be < 4 s travel time downwind to satisfy the near-field puff assumption in Equation 5 (i.e., for U = 1 m s-1, sampler(s) should be set < 4 m downwind, for U = 2 m s-1, samplers should be set up to 8 m downwind, etc).
  • Tractor and implement should make single passes across the field.
  • After each pass, samplers should be moved to be the appropriate distance downwind from the next pass/replication.

Based on the experimental data, source strength is mainly affected by the operation type, equipment, and soil conditions rather than local atmospheric parameters. The results from this study and further measurements by these methods are providing source strength input data for our agriculture dust dispersion model (Wang et al., 2008).


This work was supported with funds from the USDA NRI CSREES program under Contract No. 2007-55112-17849 and 2004-35112-14230, the University of Connecticut Storrs Agricultural Experiment Station, and the New Mexico State University Agricultural Experiment Station. The authors are grateful to the staff at the Agricultural Experiment Station at New Mexico State University for their generous cooperation during the field experiments.

Literature Cited

Ashbaugh, L. L., R. Matsumura, T. James, O. Carvacho, and R. Flocchini. 1996. Modeling PM10 dust emissions from field harvest operations. In Proceedings, International Conference on Air Pollution from Agricultural Operations, 155–159.

Ashbaugh, L. L., R. Matsumura, and R. Flocchini. 1997. Calculation of PM source strength using a simple box model and a vertical profiling method. In Emission Inventory: Planning for the Future, Proceedings of the 1997 AWMA/EPA Emission Inventory Conference, October 28–30, Research Triangle Park, NC. Vol 1, 282–288.

Aylor, D. E., and T. K. Flesch. 2001. Estimating spore release rates using a Lagrangian stochastic simulation model. Journal of Applied Meteorology, 40, 1196–1208.

Aylor, D. E., and J. Qiu. 1996. Micrometeorological determination of release rate of Venturia inaequalis ascospores from a ground-level source during rain. Agricultural and Forest Meteorology, 81, 157–178.

Cassel, T., K. Trzepla-Nabaglo, and R. Flocchini. 2003. PM10 emission factors for harvest and tillage of row crops. 12th International Emission Inventory Conference, San Diego, CA, April 29–May 1, 2003. Available at http://www.epa.gov/ttn/chief/conference/ei12/poster/cassel.pdf.

Clausnitzer, H., and M. J. Singer. 1996. Respirable-dust production from agricultural operations in the Sacramento Valley, California. Journal of Environmental Quality, 25, 877–884.

Clausnitzer, H., and M. J. Singer. 1997. Intensive land preparation emits respirable dust. California Agriculture, 51(2), 27–30.

Coates, W. 1996. Particulates generated by five cotton tillage systems. Transactions of the ASAE, 39, 1593–1598.

Cowherd Jr., C., K. Axetall Jr., C. M. Guenther, and G. A. Jutze. 1974. Development of emission factors for fugitive dust sources [NTIS Publication no. PB-238 262]. Kansas City, MO: Midwest Research Institute.

Cowherd Jr., C., and J. S. Kinsey. 1986. Identification, assessment and control of fugitive particulate emissions [EPA/600/8-86/023]. Research Triangle Park, NC: U.S. Environmental Protection Agency.

Cuscino Jr., T. A., J. S. Kinsey, and R. Hackney. 1984. The role of agricultural practices in fugitive dust emissions. Kansas City, MO: Midwest Research Institute.

Etyemezian, V., S. Ahonen, D. Nikolic, J. Gillies, H. Kuhns, D. Gillette, and J. Veranth. 2004. Deposition and removal of fugitive dust in the arid southwestern United States: Measurements and model results. Journal of the Air and Waste Management Association, 54, 1099–1111.

Flocchini, R. G., T. A. Cahill, R. T. Matsumura, O. Carvacho, and Z. Lu. 1994. Study of fugitive PM-10 emissions for selected agricultural practices on selected agricultural soils [SJV Grant File _ 20960]. Davis: University of California.

Hiscox, A. L., D. R. Miller, B. A. Holmén, W. Yang, and J. Wang. 2007. Near field dust exposure from cotton field tilling and harvesting. Journal of Environmental Quality, 37, 551–556.

Holmén, B. A., T. A. James, L. L. Ashbaugh, and R. G. Flocchini. 2001a. Lidar-assisted measurement of PM10 emissions from agricultural tilling in California's San Joaquin Valley - Part I: lidar. Atmospheric Environment, 35, 3251–3264.

Holmén, B. A., T. A. James, L. L. Ashbaugh, and R. G. Flocchini. 2001b. Lidar-assisted measurement of PM10 emissions from agricultural tilling in California's San Joaquin Valley - Part II: emission factors. Atmospheric Environment, 35, 3265–3277.

Holmén, B. A., D. R. Miller, A. L. Hiscox, W. Yang ,J. Wang, T. W. Sammis, and R. Bottoms. 2007. Near-source particulate emissions and plume dynamics from agricultural field operations. Journal of Atmospheric Chemistry, 59, 117–134.

Paez-Rubio, T., X. Hua, J. Anderson, and J. Peccia. 2006. Particulate matter composition and emission rates from the disk incorporation of class B biosolids into soil. Atmospheric Environment, 40, 7034–7045.

Snyder, J. W., and T. R. Blackwood. 1977. Source assessment: Mechanical harvesting of cotton state of the art [EPA/600/2-77-107d]. Research Triangle Park, NC: U.S. Environmental Protection Agency.

Stull, R. B. 2001. An introduction to boundary layer meteorology. Dordrecht, The Netherlands: Kluwer Academic Publishers.

Wang, J., A. L. Hiscox, D. R. Miller, and T. W. Sammis. 2008. A dynamic Lagrangian, field-scale model of dust dispersion from agriculture tilling operations. Transactions of the ASABE, 51, 1763–1774.

Wilson, J. D., and W. K. N. Shum. 1992. A re-examination of the integrated horizontal flux method for estimating volatilization from circular plots. Agricultural and Forest Meteorology, 57, 281–295.

Brand names appearing in publications are for product identification purposes only. No endorsement is intended, nor is criticism implied of similar products not mentioned. Persons using such products assume responsibility for their use in accordance with current label directions of the manufacturer.

To find more resources for your business, home, or family, visit the College of Agriculture and Home Economics on the World Wide Web at aces.nmsu.edu.

Contents of publications may be freely reproduced for educational purposes. All other rights reserved. For permission to use publications for other purposes, contact pubs@nmsu.edu or the authors listed on the publication.

New Mexico State University is an equal opportunity/affirmative action employer and educator. NMSU and the U.S. Department of Agriculture cooperating.

Printed and electronicaly distributed August 2009, Las Cruces, NM.