Prediction of Agricultural Contaminant Concentrations in Ambient Air

*Steven Cryer and Ian van Wesenbeeck*

## **Abstract**

Monitoring ambient air to assess environmental exposure and risk for volatile agricultural chemicals requires extensive resources and logistical effort. The cost and technical limitations of monitoring can be mitigated using a validated air dispersion model to simulate concentrations of volatile organic chemicals in ambient air. The SOil Fumigant Exposure Assessment (SOFEA) model was developed to explore volatile pesticide exposure and bystander risk. SOFEA assembles sources and source strengths, uses weather data from the region of interest, and executes an air dispersion model (AERMOD, ISCST3) to simulate pesticide concentrations at user defined receptors that can be used in exposure and risk assessment. This work highlights SOFEA development from inception and modifications over the last 1.5 decades, to the current delivery within the public domain. Various examples for the soil fumigant 1,3-dichloropropene are provided.

**Keywords:** air dispersion modeling, SOil Fumigant Exposure Assessment tool (SOFEA), AMS/EPA Regulatory Model (AERMOD), 1,3-dichropropene, Gaussian plume

## **1. Introduction**

The development of a numerical modeling tool for the soil fumigant 1,3-D started several decades ago using the Industrial Source Complex Short Term (ISCST3) air dispersion model [1]. Early work was extended by incorporating a soil fate modeling tool, the Pesticide Root Zone Model (PRZM3), to simulate the source strength used in ISCST3 air dispersion calculations [2]. This initial work was the forerunner of the SOil Fumigant Exposure Assessment system (SOFEA), a stochastic numerical modeling tool developed by Corteva Agriscience as a regulatory tool to evaluate and manage human inhalation exposure potential associated with the use of soil fumigants and other semi-volatile or volatile compounds [3]. There are no existing models for predicting pesticide exposure that can easily incorporate multiple fields throughout the year that mimic use rates and volatility that ultimately govern exposure. Even today, SOFEA has more attributes and functionality when addressing exposure risk from the use of volatile (or semi-volatile) pesticides than other agricultural models.

SOFEA calculates fumigant concentrations in air arising from volatility losses from treated agricultural fields for entire agricultural regions using multiple transient source terms (treated fields), Geographic Information System (GIS) information, agronomic specific variables, user specified buffer zones, and field

re-entry intervals. The United States Environmental Protection Agency (USEPA) Gaussian plume models ISCST3 [4] and/or the AMS/EPA Regulatory Model (AERMOD) [5–6] are used for air dispersion calculations. Examples of ISCST3 simulations include modeling vehicle exhausts in urban areas [7] and modeling the ambient air concentrations at specified buffer distances for single fields in agriculture [8]. AERMOD simulations include estimates of mercury levels in air [9]), NO2 and SO2 predictions in Thailand [10], and air concentrations resulting from emissions from agricultural fields treated with the soil fumigant 1,3-dichloropropene [11].

This work provides a summary for SOFEA development and use in agriculture over the past decade. SOFEA uses field observed (or numerically generated) fumigant flux profiles from soil as transient source terms for both agricultural shank injection and drip-irrigation applications of a soil fumigant. Measured reference flux observations are scaled based upon depth of incorporation into soil and the time of year, to map the complete flux response surface from field/numerical observations, however a soil physics model can also be used to estimate flux from soil (e.g., HYDRUS [12], CHAIN\_2D [13], etc.). Weather information, field size, application date, application rate, application type, depth of soil incorporation, pesticide degradation rates in air, tarp presence at the soil surface, ag-capable land, field re-treatment from 1 year to the next, buffer setbacks, and other sensitive parameters are varied stochastically using Monte Carlo techniques to mimic region and crop specific agronomic practices. Agricultural regions up to 49,210 km<sup>2</sup> (19,000 mi<sup>2</sup> ) can be simulated for temporal periods ranging from 1 day to more than 70 years for the purpose of addressing acute (24 h), short-term (72 h), sub-chronic (28 or 90 days) or chronic exposure. Multi-year, multiple field simulations are conducted using either random field placement in all agricultural capable areas, by selectively placing fields in historical or prospective areas, and/or placing fields of a specific size and spatial location if this information is available. Regional land cover, elevation, and population information cavn be used to refine source placement (treated fields), dispersion calculations, and exposure assessments. Both current and anticipated/forecasted fumigant scenarios can be simulated to provide risk managers with the necessary information to make sound regulatory decisions.

SOFEA has been used for regulatory decision making in California, was reviewed in the 2004 USEPA Scientific Advisory Panel (SAP) meeting [14], and is currently being used in the registration review process for 1,3-dicholopropene (1,3-D) with USEPA. Algorithms used by SOFEA to refine exposure predictions and manage acute, sub-chronic, and chronic risk associated with the use of soil fumigants on a local or regional basis are presented. Although SOFEA was originally designed specifically to describe air concentrations for the soil fumigant 1,3-D, the model is readily adaptable to generically describe the post-fumigation air concentrations of other soil fumigants and semi-volatile organic chemicals. SOFEA can now be executed in "retrospective" mode which allows the user to specify specific field locations where known fumigant applications are made and specific receptor locations where fumigant concentrations are measured which makes the model predictions available for comparison to field monitoring observations. Other enhancements to SOFEA include the option of importing flux (emissions) from soil simulated by a soil physics model (HYDRUS, CHAIN\_2D) in lieu of flux obtained from field experiments. SOFEA will soon be in the public domain and can be obtained from Exponent Inc. (U.S. based consulting company) for use with all volatile or semi-volatile pesticides if parameterized appropriately. This manuscript

**17**

**Figure 2.**

*Prediction of Agricultural Contaminant Concentrations in Ambient Air*

from inception to release of the model to the public domain.

summarizes SOFEA capabilities in both in prospective and retrospective mode,

A generic methodology to determine fumigant concentrations in ambient air in large and diverse air sheds has been developed (**Figure 1**). Directionally averaged air concentrations within entire air sheds are determined using a multiple source Gaussian dispersion model that has been modified to include Monte Carlo sampling techniques, ties to GIS databases, and agronomic practices. Time averaged transient air concentrations simulated via a numerical model can be used to assess exposure and risk for an unlimited number of

SOFEA enables the determination of "area-wide" concentration profiles for user

specified distances that account for multiple field applications. Thus, the effect of fields "off-gassing" at different points in time and space are accounted for by SOFEA. The user can evaluate the impact of the buffer on the acute exposure for residents and by-standers by specifying a buffer distance from the edge of treated fields. The user can also determine the chronic exposure to individuals residing in the use area by specifying the total mass applied on an annual basis and running SOFEA for a full year or multiple years. SOFEA inputs and outputs are easily exported to other file formats or programs. Concentrations of soil fumigants in air

*SOFEA is an intelligent input file generator and output repository for agronomic use of the USEPA Gaussian* 

*DOI: http://dx.doi.org/10.5772/intechopen.86091*

**2. Background**

scenarios.

**Figure 1.**

*plume model AERMOD (and formerly ISCST3).*

*SOil Fumigant Exposure Assessment Model overview.*

summarizes SOFEA capabilities in both in prospective and retrospective mode, from inception to release of the model to the public domain.

## **2. Background**

*Atmospheric Air Pollution and Monitoring*

1,3-dichloropropene [11].

practices. Agricultural regions up to 49,210 km<sup>2</sup>

re-entry intervals. The United States Environmental Protection Agency (USEPA) Gaussian plume models ISCST3 [4] and/or the AMS/EPA Regulatory Model (AERMOD) [5–6] are used for air dispersion calculations. Examples of ISCST3 simulations include modeling vehicle exhausts in urban areas [7] and modeling the ambient air concentrations at specified buffer distances for single fields in agriculture [8]. AERMOD simulations include estimates of mercury levels in air [9]), NO2 and SO2 predictions in Thailand [10], and air concentrations resulting from emissions from agricultural fields treated with the soil fumigant

This work provides a summary for SOFEA development and use in agriculture over the past decade. SOFEA uses field observed (or numerically generated) fumigant flux profiles from soil as transient source terms for both agricultural shank injection and drip-irrigation applications of a soil fumigant. Measured reference flux observations are scaled based upon depth of incorporation into soil and the time of year, to map the complete flux response surface from field/numerical observations, however a soil physics model can also be used to estimate flux from soil (e.g., HYDRUS [12], CHAIN\_2D [13], etc.). Weather information, field size, application date, application rate, application type, depth of soil incorporation, pesticide degradation rates in air, tarp presence at the soil surface, ag-capable land, field re-treatment from 1 year to the next, buffer setbacks, and other sensitive parameters are varied stochastically using Monte Carlo techniques to mimic region and crop specific agronomic

for temporal periods ranging from 1 day to more than 70 years for the purpose of addressing acute (24 h), short-term (72 h), sub-chronic (28 or 90 days) or chronic exposure. Multi-year, multiple field simulations are conducted using either random field placement in all agricultural capable areas, by selectively placing fields in historical or prospective areas, and/or placing fields of a specific size and spatial location if this information is available. Regional land cover, elevation, and population information cavn be used to refine source placement (treated fields), dispersion calculations, and exposure assessments. Both current and anticipated/forecasted fumigant scenarios can be simulated to provide risk managers with the necessary information to make sound regulatory

SOFEA has been used for regulatory decision making in California, was reviewed in the 2004 USEPA Scientific Advisory Panel (SAP) meeting [14], and is currently being used in the registration review process for 1,3-dicholopropene (1,3-D) with USEPA. Algorithms used by SOFEA to refine exposure predictions and manage acute, sub-chronic, and chronic risk associated with the use of soil fumigants on a local or regional basis are presented. Although SOFEA was originally designed specifically to describe air concentrations for the soil fumigant 1,3-D, the model is readily adaptable to generically describe the post-fumigation air concentrations of other soil fumigants and semi-volatile organic chemicals. SOFEA can now be executed in "retrospective" mode which allows the user to specify specific field locations where known fumigant applications are made and specific receptor locations where fumigant concentrations are measured which makes the model predictions available for comparison to field monitoring observations. Other enhancements to SOFEA include the option of importing flux (emissions) from soil simulated by a soil physics model (HYDRUS, CHAIN\_2D) in lieu of flux obtained from field experiments. SOFEA will soon be in the public domain and can be obtained from Exponent Inc. (U.S. based consulting company) for use with all volatile or semi-volatile pesticides if parameterized appropriately. This manuscript

(19,000 mi<sup>2</sup>

) can be simulated

**16**

decisions.

A generic methodology to determine fumigant concentrations in ambient air in large and diverse air sheds has been developed (**Figure 1**). Directionally averaged air concentrations within entire air sheds are determined using a multiple source Gaussian dispersion model that has been modified to include Monte Carlo sampling techniques, ties to GIS databases, and agronomic practices. Time averaged transient air concentrations simulated via a numerical model can be used to assess exposure and risk for an unlimited number of scenarios.

SOFEA enables the determination of "area-wide" concentration profiles for user specified distances that account for multiple field applications. Thus, the effect of fields "off-gassing" at different points in time and space are accounted for by SOFEA. The user can evaluate the impact of the buffer on the acute exposure for residents and by-standers by specifying a buffer distance from the edge of treated fields. The user can also determine the chronic exposure to individuals residing in the use area by specifying the total mass applied on an annual basis and running SOFEA for a full year or multiple years. SOFEA inputs and outputs are easily exported to other file formats or programs. Concentrations of soil fumigants in air

**Figure 1.**

*SOFEA is an intelligent input file generator and output repository for agronomic use of the USEPA Gaussian plume model AERMOD (and formerly ISCST3).*

#### **Figure 2.**

*SOil Fumigant Exposure Assessment Model overview.*

are associated with *x*, *y*, *z* co-ordinates and associated with proximity to treated fields as well as human populations (if population census data is available). The SOFEA model is readily adaptable to generically describe the post-fumigation air concentration of other organic contaminants, **Figure 2**.

## **3. SOFEA refinements and use, past and present**

## **3.1 Air dispersion model**

The ISCST3 [4] and AERMOD [5, 6] models were developed by the USEPA as regulatory tools for predicting concentrations of air contaminants in diverse air sheds. Both are Gaussian plume models useful for estimating air quality surrounding contaminant release sites. AERMOD replaced ISCST3 by USEPA, although ISCST3 is still in use by some researchers and regulatory authorities. ISCST3 has been widely used to simulate probability distribution functions (PDFs) of fumigant concentrations in air within townships in California to estimate acute and chronic bystander exposure [15, 16].

## **3.2 Parameter representation**

The complex terrain algorithms of ISCST3/AERMOD can account for the effects of elevation changes within specific regions should this information be available. Population information (if provided or known) can also be used in population-based risk assessments. The 2010 U.S. census data lists population densities by census blocks and is a good choice for population information. A township is defined

**19**

*Prediction of Agricultural Contaminant Concentrations in Ambient Air*

according to the Public Land Survey System (PLSS) and is nominally 6 × 6 mi (9.66 × 9.66 km) in area, **Figure 3**. The spatial locations for receptors placed uniformly or weighted in a central township are user specified, and appropriate land cover, elevation, and population data from GIS data bases are necessary inputs. Township information must include land cover such that ag-capable land can be quantified. Elevation and population information are optional. The impact of sources external to the central township domain will depend on the persistence and drift characteristics of the pesticide being simulated. However, a model evaluation for 1,3-D showed that 3 × 3 townships was adequate to account for edge effects [17].

An air shed is defined as a volume of air overlying a square surface area, where source terms (i.e., treated fields) throughout the air shed can contribute to overall air concentrations at specific locations. Although historical SOFEA simulations focused on air concentrations in either a single township or a 3 × 3 township domain, the model can be used to simulate concentrations across much larger airsheds. The complex terrain algorithms of ISCST3 or AERMOD can take advantage of elevation changes within specific regions. Receptors can be placed uniformly in a central township or over a multi-township domain. Source terms can be placed anywhere in the simulation domain, which can include up to 23 × 23 townships

specify the annual pesticide mass applied to any township within a 23 × 23 township domain, appropriate GIS information, receptor spacing and heights, as well as

Concentrations of a soil fumigant in air resulting from transient agricultural source terms are also dependent upon meteorological conditions, application timing, and other agronomic properties. A mechanism was required that could propagate parametric uncertainty in sensitive model inputs to air concentration predictions. Monte Carlo (MC) methods provide a straightforward technique to propagate such uncertainty in independent parameters to dependent output variables [18, 19]. PDFs can include fumigated field sizes, application rates, application dates etc. SOFEA can also be used in retrospective mode, where exact treated field locations and application parameters (mass applied, date applied, etc.), and receptor locations are known. Variability in input is described by PDFs that are randomly sampled to generate input parameter sequences. Stochastic variables for SOFEA include application rate, date, and hour of day initiated, pesticide depth of incorporation, presence of a tarp at the soil surface, application type (shank injection or drip irrigation), field size, weather year, and pesticide properties such as degradation rates in air. Output predictions are no longer deterministic, but rather a discrete distribution is generated from which exceedance probabilities and return frequencies can be calculated (e.g., 1-in-100-year exposure potential, and so on). Air quality modeling work is in accordance with the policy established by the U.S. EPA for Air Quality Models and

appropriate PDFs characterizing agronomic practices within the region.

follows the guidelines set forth by U.S. EPA for Monte Carlo Analysis [20].

The original version of SOFEA required the MS Excel add-on program Crystal Ball™ (Decisioneering, Inc.) to transform ISCST3 from a deterministic model into a stochastic/deterministic system; however, subsequent versions have been modified to include Visual Basic Applications (VBA) algorithms that obviate the need for Crystal Ball™. In SOFEA2, an ISCST3 input file is exported from Excel that is based upon appropriate selections from user defined PDFs that are derived from

). When running in prospective mode, the user need only

*DOI: http://dx.doi.org/10.5772/intechopen.86091*

**3.3 Air shed simulation domain**

= 19,000 mi2

**3.4 Stochastic portrayal**

(49,210 km2

*Prediction of Agricultural Contaminant Concentrations in Ambient Air DOI: http://dx.doi.org/10.5772/intechopen.86091*

according to the Public Land Survey System (PLSS) and is nominally 6 × 6 mi (9.66 × 9.66 km) in area, **Figure 3**. The spatial locations for receptors placed uniformly or weighted in a central township are user specified, and appropriate land cover, elevation, and population data from GIS data bases are necessary inputs. Township information must include land cover such that ag-capable land can be quantified. Elevation and population information are optional. The impact of sources external to the central township domain will depend on the persistence and drift characteristics of the pesticide being simulated. However, a model evaluation for 1,3-D showed that 3 × 3 townships was adequate to account for edge effects [17].

#### **3.3 Air shed simulation domain**

*Atmospheric Air Pollution and Monitoring*

**3.1 Air dispersion model**

**3.2 Parameter representation**

concentration of other organic contaminants, **Figure 2**.

**3. SOFEA refinements and use, past and present**

are associated with *x*, *y*, *z* co-ordinates and associated with proximity to treated fields as well as human populations (if population census data is available). The SOFEA model is readily adaptable to generically describe the post-fumigation air

The ISCST3 [4] and AERMOD [5, 6] models were developed by the USEPA as regulatory tools for predicting concentrations of air contaminants in diverse air sheds. Both are Gaussian plume models useful for estimating air quality surrounding contaminant release sites. AERMOD replaced ISCST3 by USEPA, although ISCST3 is still in use by some researchers and regulatory authorities. ISCST3 has been widely used to simulate probability distribution functions (PDFs) of fumigant concentrations in air within townships in California to estimate acute and chronic bystander exposure [15, 16].

The complex terrain algorithms of ISCST3/AERMOD can account for the effects of elevation changes within specific regions should this information be available. Population information (if provided or known) can also be used in population-based risk assessments. The 2010 U.S. census data lists population densities by census blocks and is a good choice for population information. A township is defined

*Example of a 3 × 3 township air shed for SOFEA simulation from the state of California.*

**18**

**Figure 3.**

An air shed is defined as a volume of air overlying a square surface area, where source terms (i.e., treated fields) throughout the air shed can contribute to overall air concentrations at specific locations. Although historical SOFEA simulations focused on air concentrations in either a single township or a 3 × 3 township domain, the model can be used to simulate concentrations across much larger airsheds. The complex terrain algorithms of ISCST3 or AERMOD can take advantage of elevation changes within specific regions. Receptors can be placed uniformly in a central township or over a multi-township domain. Source terms can be placed anywhere in the simulation domain, which can include up to 23 × 23 townships (49,210 km2 = 19,000 mi2 ). When running in prospective mode, the user need only specify the annual pesticide mass applied to any township within a 23 × 23 township domain, appropriate GIS information, receptor spacing and heights, as well as appropriate PDFs characterizing agronomic practices within the region.

### **3.4 Stochastic portrayal**

Concentrations of a soil fumigant in air resulting from transient agricultural source terms are also dependent upon meteorological conditions, application timing, and other agronomic properties. A mechanism was required that could propagate parametric uncertainty in sensitive model inputs to air concentration predictions. Monte Carlo (MC) methods provide a straightforward technique to propagate such uncertainty in independent parameters to dependent output variables [18, 19]. PDFs can include fumigated field sizes, application rates, application dates etc. SOFEA can also be used in retrospective mode, where exact treated field locations and application parameters (mass applied, date applied, etc.), and receptor locations are known. Variability in input is described by PDFs that are randomly sampled to generate input parameter sequences. Stochastic variables for SOFEA include application rate, date, and hour of day initiated, pesticide depth of incorporation, presence of a tarp at the soil surface, application type (shank injection or drip irrigation), field size, weather year, and pesticide properties such as degradation rates in air. Output predictions are no longer deterministic, but rather a discrete distribution is generated from which exceedance probabilities and return frequencies can be calculated (e.g., 1-in-100-year exposure potential, and so on). Air quality modeling work is in accordance with the policy established by the U.S. EPA for Air Quality Models and follows the guidelines set forth by U.S. EPA for Monte Carlo Analysis [20].

The original version of SOFEA required the MS Excel add-on program Crystal Ball™ (Decisioneering, Inc.) to transform ISCST3 from a deterministic model into a stochastic/deterministic system; however, subsequent versions have been modified to include Visual Basic Applications (VBA) algorithms that obviate the need for Crystal Ball™. In SOFEA2, an ISCST3 input file is exported from Excel that is based upon appropriate selections from user defined PDFs that are derived from

actual agronomic data. Excel, ISCST3, and VBA programs were coupled to allow the transparent integration of the Monte Carlo component in SOFEA3, which used, but ISCST3 had changes in mixing height calculations under calm conditions such that simulation with parametric uncertainty more closed matched monitoring observations [21]. The latest version of SOFEA (SOFEA4) contains identical functionality as the original version but was rewritten in C++ and Qt to replace VBA programming, ISCST3 replaced by AERMOD, and results using SOFEA4 use are found elsewhere [11]. SOFEA4 provides automation, transparency, a Graphical User Interface (GUI), the use of AERMOD, and maintainability for future support.

#### **3.5 Crop selection and simulation domain**

Fumigants are used on a variety of high valued agricultural commodities. Each commodity/crop is potentially unique, with different agronomic management practices. The crops chosen can be based upon current or future forecasted fumigant uses, and currently up to five different crop types can be considered. Predominant crops where soil fumigants are used include tree and vine (TV), field crops (FC), nursery crops (NC), strawberries (SB) and post-plant vines (PP). The contributions of a soil fumigant to air quality from each crop are easily extractable by keeping the crop types/parameters unique during simulation. This aids in determining appropriate Best Management Practices (BMP's) by crop type. SOFEA uses the supplied PDFs to generate the agronomic variables (e.g., field size, application rate, etc.) for each crop type. Thus, if a region is dominated by one crop type, all five crop types in SOFEA can be parameterized with the same data (if desired) to minimize computer memory requirements.

#### **3.6 Receptors**

Receptors are specific (*x*, *y*, *z*) locations in the simulation domain where air concentrations are calculated. These receptors can be uniformly placed within the township for chronic exposure predictions, or at specific setback distances around treated fields if acute exposure assessment is required, **Figure 4**. Historical SOFEA simulations in CA have assumed a rectangular grid of 36 equally spaced receptors per township section (i.e., a 1 mi<sup>2</sup> area) which yields 1296 receptors per township at a spacing of 268.2 m (880 feet) [11, 21]. Receptors are typically placed at 1.5 m above the ground to mimic the breathing height on adult. Ultimately, any desired

**Figure 4.** *Receptor placement at user specified buffer setback near treated fields (left) or uniform grid placement (right).*

**21**

*Prediction of Agricultural Contaminant Concentrations in Ambient Air*

receptor location and density, and height can be specified by the user, allowing

Many data bases and GIS software programs exist to extract appropriate information for use in SOFEA, **Figure 5**. SOFEA is not a GIS tool but rather uses GIS information that has been assembled using software such as ArcView™ (ESRI, Inc.). Land cover information is obtained by Landsat Thematic Mapper images (30-m resolution) that contain 21 unique land classifications [available from the National Land Cover Data (NLCD) database at http://landcover.usgs. gov/natllandcover.htm]. Elevation information is obtained from the USGS Digital Elevation Models (DEM) data at 1:24,000 scales. Population

information is given by census blocks and populated with data from the 2010 US Census, and GIS information is used to parameterize the air shed for ISCST3/AERMOD simulations (e.g., ag-capable land where fields

A single location for weather data is used in SOFEA to represent weather conditions from the region of interest. Meteorological information includes hourly air stability class, wind speed, air temperature, wind direction and mixing and ceiling height for ISCST3, along with the Monin-Obukhov stability length for AERMOD. Wind speed and direction are critical parameters, and for larger simulation domains with potentially greater surface roughness length (*z*) (due to trees, buildings, fence rows, etc.), wind speed is preferable measured at a height of 10-m. A rule of thumb for determining the minimum height of the wind sensor is 7\**z* [22]. Flat fallow fields typically have a roughness length *z* < 0.1-m, and therefore an anemometer height of 2-m is adequate. The user creates a weather library for each year of weather and this library is assigned a uniform distribution when SOFEA is executed in prospective mode, or actual weather information for a specific time frame when running in retrospective mode. Weather data is available from public sources such as the California Irrigation Management System (CIMIS), the National Oceanic and Atmospheric Administration (NOAA), or

individual receptors to be placed anywhere in the simulation domain.

*Example of the continually updated public domain data bases available to the user of SOFEA.*

*DOI: http://dx.doi.org/10.5772/intechopen.86091*

**3.7 GIS data layers**

**Figure 5.**

can be placed, etc.).

**3.8 Meteorological data**

*Prediction of Agricultural Contaminant Concentrations in Ambient Air DOI: http://dx.doi.org/10.5772/intechopen.86091*

**Figure 5.** *Example of the continually updated public domain data bases available to the user of SOFEA.*

receptor location and density, and height can be specified by the user, allowing individual receptors to be placed anywhere in the simulation domain.

## **3.7 GIS data layers**

*Atmospheric Air Pollution and Monitoring*

actual agronomic data. Excel, ISCST3, and VBA programs were coupled to allow the transparent integration of the Monte Carlo component in SOFEA3, which used, but ISCST3 had changes in mixing height calculations under calm conditions such that simulation with parametric uncertainty more closed matched monitoring observations [21]. The latest version of SOFEA (SOFEA4) contains identical functionality as the original version but was rewritten in C++ and Qt to replace VBA programming, ISCST3 replaced by AERMOD, and results using SOFEA4 use are found elsewhere [11]. SOFEA4 provides automation, transparency, a Graphical User Interface (GUI),

Fumigants are used on a variety of high valued agricultural commodities. Each commodity/crop is potentially unique, with different agronomic management practices. The crops chosen can be based upon current or future forecasted fumigant uses, and currently up to five different crop types can be considered. Predominant crops where soil fumigants are used include tree and vine (TV), field crops (FC), nursery crops (NC), strawberries (SB) and post-plant vines (PP). The contributions of a soil fumigant to air quality from each crop are easily extractable by keeping the crop types/parameters unique during simulation. This aids in determining appropriate Best Management Practices (BMP's) by crop type. SOFEA uses the supplied PDFs to generate the agronomic variables (e.g., field size, application rate, etc.) for each crop type. Thus, if a region is dominated by one crop type, all five crop types in SOFEA can be parameterized with the same data (if desired) to minimize computer memory requirements.

Receptors are specific (*x*, *y*, *z*) locations in the simulation domain where air concentrations are calculated. These receptors can be uniformly placed within the township for chronic exposure predictions, or at specific setback distances around treated fields if acute exposure assessment is required, **Figure 4**. Historical SOFEA simulations in CA have assumed a rectangular grid of 36 equally spaced receptors

at a spacing of 268.2 m (880 feet) [11, 21]. Receptors are typically placed at 1.5 m above the ground to mimic the breathing height on adult. Ultimately, any desired

*Receptor placement at user specified buffer setback near treated fields (left) or uniform grid placement (right).*

area) which yields 1296 receptors per township

the use of AERMOD, and maintainability for future support.

**3.5 Crop selection and simulation domain**

**20**

**Figure 4.**

**3.6 Receptors**

per township section (i.e., a 1 mi<sup>2</sup>

Many data bases and GIS software programs exist to extract appropriate information for use in SOFEA, **Figure 5**. SOFEA is not a GIS tool but rather uses GIS information that has been assembled using software such as ArcView™ (ESRI, Inc.). Land cover information is obtained by Landsat Thematic Mapper images (30-m resolution) that contain 21 unique land classifications [available from the National Land Cover Data (NLCD) database at http://landcover.usgs. gov/natllandcover.htm]. Elevation information is obtained from the USGS Digital Elevation Models (DEM) data at 1:24,000 scales. Population information is given by census blocks and populated with data from the 2010 US Census, and GIS information is used to parameterize the air shed for ISCST3/AERMOD simulations (e.g., ag-capable land where fields can be placed, etc.).

## **3.8 Meteorological data**

A single location for weather data is used in SOFEA to represent weather conditions from the region of interest. Meteorological information includes hourly air stability class, wind speed, air temperature, wind direction and mixing and ceiling height for ISCST3, along with the Monin-Obukhov stability length for AERMOD. Wind speed and direction are critical parameters, and for larger simulation domains with potentially greater surface roughness length (*z*) (due to trees, buildings, fence rows, etc.), wind speed is preferable measured at a height of 10-m. A rule of thumb for determining the minimum height of the wind sensor is 7\**z* [22]. Flat fallow fields typically have a roughness length *z* < 0.1-m, and therefore an anemometer height of 2-m is adequate. The user creates a weather library for each year of weather and this library is assigned a uniform distribution when SOFEA is executed in prospective mode, or actual weather information for a specific time frame when running in retrospective mode. Weather data is available from public sources such as the California Irrigation Management System (CIMIS), the National Oceanic and Atmospheric Administration (NOAA), or

the Florida Automated Weather Network (FAWN), or could be collected by a dedicated weather station installed in the simulation domain. The weather station should collect, at minimum, hourly precipitation, solar radiation, air temperature, and wind speed and direction (SOFEA requirements). Weather data must be pre-processed using PCRAMMET if the ISCST model is used, or the AERMET pre-processor if AERMOD is used. Pre-processing is conducted outside of the SOFEA model framework.

## **3.9 Source placement**

In prospective mode, sources (treated fields) can be placed randomly or weighted to specific township locations, **Figure 6**. All ag-capable land (all land excluding urban areas, water bodies, barren, rock, quarries, and wetlands) is used and placement is based on a uniform probability of occurrence (known as random field placement). However, there are situations in high pesticide product use regions where treated field locations are known, and section weighting can be used to ensure that product use spatially represents historical needs. A township section is 1/36 of the township area and the user can specify the probability that these sections are locations where fields are placed. Receptors in these regions will register higher chronic soil fumigant air concentrations due to the increased field (i.e., source) density. Section-weighting probabilities can be based on expert judgment and/or historical product use records. When sections "fill up" and can no longer contain another treated field, a "spill-over" algorithm is introduced in SOFEA where the fields are then placed in sections surrounding the section that is "filled."

## **3.10 Township allocation of fumigant mass**

In California, the amount of 1,3-D applied annually cannot exceed a mandated township allocation which is set based on acceptable levels of chronic exposure. Each township is assigned an allocation amount based on CA permit conditions (or some other a user-supplied amount), so this system can and has been applied in

#### **Figure 6.**

*Use of section weighted placement of treated fields within a township based upon specific sections having high fumigant use. Each field is assigned input based upon user specified PDFs.*

**23**

**Figure 7.**

*impact of depth of pesticide incorporation (right).*

*Prediction of Agricultural Contaminant Concentrations in Ambient Air*

is used where the chemical flux from soil to the air is defined as.

h<sup>−</sup><sup>1</sup>

can also be used to develop flux profiles for different conditions.

other fumigant use areas across the United States. The amount of pesticide used in a given township is thus given as a fraction of this user specified township allocation.

The California Department of Pesticide Regulations (CDPR) uses a simple procedure to account for seasonal and incorporation depth variability on pesticide volatility losses to represent the complete flux response profile. Volatility losses are sensitive to temperature and depth of soil incorporation [2] and a simple expression

Fluxi = R × Fri × Sincorp × Syr (1)

where, Fluxi = scaled hourly flux loss from soil into air for hour "i" based upon

Fri = observed flux rate (reference profile based on a field experiment, or modeling), Sincorp = scaling factor for depth of incorporation (dimensionless), and Syr = scaling

Although the CDPR approach only uses a single flux profile for each application type, these profiles are modified by soil incorporation depth and time of year. Also, models such as HYDRUS [23], STANMOD [24], CHAIN\_2D [13] and PRZM3 [25]

California is sectioned into warm and cool seasons where increased emission to the atmosphere occurs under warm conditions and is arbitrarily increased by a factor of 1.6× by CDPR. Therefore, Syr is assigned a value of 1.6 to account for gross seasonal temperature effects during the warm season. This warm season can be a specific time of the year (as for CDPR) or the SOFEA user can use a continuous sinusoidal function, where the amplitude and frequencies are daily average air temperatures based upon what day within the year a pesticide application is made (**Figure 7**, left). In addition, several constraints on the depth of incorporation are used for CA and are given in **Figure 7** (right). The user can specify how the pesticide incorporation depth in soil can alter the cumulative mass loss from the soil surface (linear, exponential, CDPR) by selecting from options in the drop-down

*Example of CDPR application factor or sinusoidal modeling for percent of applied volatilized (left) and* 

), R = pesticide application rate (kg ha<sup>−</sup><sup>1</sup>

),

*DOI: http://dx.doi.org/10.5772/intechopen.86091*

**3.11 Application scaling factor (CA only)**

an actual field trial, Fri (kg ha<sup>−</sup><sup>1</sup>

factor for time of year (dimensionless).

**3.12 Temporal representation Syr (CA only)**

other fumigant use areas across the United States. The amount of pesticide used in a given township is thus given as a fraction of this user specified township allocation.

## **3.11 Application scaling factor (CA only)**

*Atmospheric Air Pollution and Monitoring*

SOFEA model framework.

the section that is "filled."

**3.10 Township allocation of fumigant mass**

**3.9 Source placement**

the Florida Automated Weather Network (FAWN), or could be collected by a dedicated weather station installed in the simulation domain. The weather station should collect, at minimum, hourly precipitation, solar radiation, air temperature, and wind speed and direction (SOFEA requirements). Weather data must be pre-processed using PCRAMMET if the ISCST model is used, or the AERMET pre-processor if AERMOD is used. Pre-processing is conducted outside of the

In prospective mode, sources (treated fields) can be placed randomly or weighted to specific township locations, **Figure 6**. All ag-capable land (all land excluding urban areas, water bodies, barren, rock, quarries, and wetlands) is used and placement is based on a uniform probability of occurrence (known as random field placement). However, there are situations in high pesticide product use regions where treated field locations are known, and section weighting can be used to ensure that product use spatially represents historical needs. A township section is 1/36 of the township area and the user can specify the probability that these sections are locations where fields are placed. Receptors in these regions will register higher chronic soil fumigant air concentrations due to the increased field (i.e., source) density. Section-weighting probabilities can be based on expert judgment and/or historical product use records. When sections "fill up" and can no longer contain another treated field, a "spill-over" algorithm is introduced in SOFEA where the fields are then placed in sections surrounding

In California, the amount of 1,3-D applied annually cannot exceed a mandated township allocation which is set based on acceptable levels of chronic exposure. Each township is assigned an allocation amount based on CA permit conditions (or some other a user-supplied amount), so this system can and has been applied in

*Use of section weighted placement of treated fields within a township based upon specific sections having high* 

*fumigant use. Each field is assigned input based upon user specified PDFs.*

**22**

**Figure 6.**

The California Department of Pesticide Regulations (CDPR) uses a simple procedure to account for seasonal and incorporation depth variability on pesticide volatility losses to represent the complete flux response profile. Volatility losses are sensitive to temperature and depth of soil incorporation [2] and a simple expression is used where the chemical flux from soil to the air is defined as.

$$\mathbf{Flux}\_{\mathbf{i}} = \mathbf{R} \times \mathbf{F r}\_{\mathbf{i}} \times \mathbf{S}\_{\text{incorrp}} \times \mathbf{S}\_{\text{yr}} \tag{1}$$

where, Fluxi = scaled hourly flux loss from soil into air for hour "i" based upon an actual field trial, Fri (kg ha<sup>−</sup><sup>1</sup> h<sup>−</sup><sup>1</sup> ), R = pesticide application rate (kg ha<sup>−</sup><sup>1</sup> ), Fri = observed flux rate (reference profile based on a field experiment, or modeling), Sincorp = scaling factor for depth of incorporation (dimensionless), and Syr = scaling factor for time of year (dimensionless).

Although the CDPR approach only uses a single flux profile for each application type, these profiles are modified by soil incorporation depth and time of year. Also, models such as HYDRUS [23], STANMOD [24], CHAIN\_2D [13] and PRZM3 [25] can also be used to develop flux profiles for different conditions.

## **3.12 Temporal representation Syr (CA only)**

California is sectioned into warm and cool seasons where increased emission to the atmosphere occurs under warm conditions and is arbitrarily increased by a factor of 1.6× by CDPR. Therefore, Syr is assigned a value of 1.6 to account for gross seasonal temperature effects during the warm season. This warm season can be a specific time of the year (as for CDPR) or the SOFEA user can use a continuous sinusoidal function, where the amplitude and frequencies are daily average air temperatures based upon what day within the year a pesticide application is made (**Figure 7**, left). In addition, several constraints on the depth of incorporation are used for CA and are given in **Figure 7** (right). The user can specify how the pesticide incorporation depth in soil can alter the cumulative mass loss from the soil surface (linear, exponential, CDPR) by selecting from options in the drop-down

#### **Figure 7.**

*Example of CDPR application factor or sinusoidal modeling for percent of applied volatilized (left) and impact of depth of pesticide incorporation (right).*

**Figure 8.** *Obtaining soil volatility flux estimates using a model such as CHAIN\_2D or HYDRUS.*

menu in the SOFEA GUI. The type of seasonal scaling (CDPR or sinusoidal) can also be selected by the user in the SOFEA GUI.

## **3.13 Model output characterization**

SOFEA is used to execute the air dispersion models ISCST3 (historical) and/or AERMOD (recent). Hourly output from these models can be analyzed according to user selections for post processing output concentrations (e.g., 1-h, 1-day, 3-day, 15-day, annual, and so forth). Functionality for the current version of SOFEA is somewhat different than in the earlier versions, but the bulk of functionality for SOFEA are the same and found elsewhere [3].

In earlier versions of SOFEA, input and output were facilitated via a VBA interface that utilizes EXCEL spreadsheets containing user supplied PDF's of application parameters. Users could create inputs based on actual field data and pesticide use information, or generate hypothetical distributions of use parameters such as field sizes, application rates and timing, depth of injection, etc. Over the years, SOFEA has evolved from a VBA model using only ISCST3 to a C++ interface that can drive AERMOD simulations (e.g., SOFEA4). A user guide for the most recent version of SOFEA4 is currently in preparation and should be available sometime in 2019.

Chemical flux estimates can be obtained from a variety of different experimental sources but can also be estimated from soil physics models such as HYDRUS [23], STANMOD [24] and CHAIN\_2D [13], **Figure 8**. Such models can and have been used to simulate both volatility from soil along with movement into the soil profile. Advantages of using a soil physics model to estimate pesticide flux loss deal with low cost and the semi-infinite parameter space that can be explored with simulation techniques. Field studies in five different states that explored atmospheric flux loss for chloropicrin and 1,3-dichloropropene were validated with CHAIN\_2D and indicate the model can correctly capture both peak and cumulative emissions effectively for these two soil fumigants [17, 26]. Thus, CHAIN\_2D and similar models are useful

**25**

the tarps.

**Figure 9.**

*within soil for biological efficacy.*

*Prediction of Agricultural Contaminant Concentrations in Ambient Air*

tools for extrapolating flux predictions to diverse scenarios where experimental observations are unavailable [27–29]. Examples of various mitigations strategies that can be explored using soil physics models include the use of agricultural films, increased soil injection depth for the fumigant, and under a near semi-infinite param-

*Example use of soil physics model CHAIN\_2D where exploration of multiple mitigations strategies such as water sealing, tarps, depth of incorporation, etc. can be simulated, as well as concentration by time (C × T)* 

eter combinations of meteorological, soil and agronomic properties, **Figure 9**.

Both cumulative and 1-h maximum air concentrations were simulated and compared against field observations with good success (the best observations and simulations results were between 6 and 8%). In addition, both ISCST3 and CALPUFF air dispersion models showed similar order of magnitude output predictions [17]. Chloropicrin and 1,3-D emissions through Totally Impermeable Film (TIF) were compared using HYRUS where the fumigant flux was simulated within a factor of ~2, though the timing of the peak was over-predicted by the model [29]. The authors suggest that field-based calibration should be conducted when tarps are used because of the lack of representative field effective permeability data for

Soil physics models should first be validated with field observations before being used to extrapolate to a variety of different conditions. Cryer and van Wesenbeeck [26] used CHAIN\_2D to validate against field observations, and then coupled

CHAIN\_2D to several USEPA air dispersion models (ISCST3 [4] and CALPUFF [30]).

Most inputs can be specified as either discrete values, or as PDFs. If possible, PDFs should be used to maximize Monte Carlo capabilities of the SOFEA modeling system and encompass uncertainties and variability in model inputs. SOFEA can generate fumigant concentrations for each receptor in the simulation domain (up to 11,664 receptors have been simulated in a nine-township air shed), averaged over specific time intervals (24-h and yearly) or periods specified by the user. For example, the user could specify the output of 24-h average, 60-day average, and

*DOI: http://dx.doi.org/10.5772/intechopen.86091*

*Prediction of Agricultural Contaminant Concentrations in Ambient Air DOI: http://dx.doi.org/10.5772/intechopen.86091*

**Figure 9.**

*Atmospheric Air Pollution and Monitoring*

menu in the SOFEA GUI. The type of seasonal scaling (CDPR or sinusoidal) can

*Obtaining soil volatility flux estimates using a model such as CHAIN\_2D or HYDRUS.*

SOFEA is used to execute the air dispersion models ISCST3 (historical) and/or AERMOD (recent). Hourly output from these models can be analyzed according to user selections for post processing output concentrations (e.g., 1-h, 1-day, 3-day, 15-day, annual, and so forth). Functionality for the current version of SOFEA is somewhat different than in the earlier versions, but the bulk of functionality for

In earlier versions of SOFEA, input and output were facilitated via a VBA interface that utilizes EXCEL spreadsheets containing user supplied PDF's of application parameters. Users could create inputs based on actual field data and pesticide use information, or generate hypothetical distributions of use parameters such as field sizes, application rates and timing, depth of injection, etc. Over the years, SOFEA has evolved from a VBA model using only ISCST3 to a C++ interface that can drive AERMOD simulations (e.g., SOFEA4). A user guide for the most recent version of SOFEA4 is currently in preparation and should be available sometime in 2019.

Chemical flux estimates can be obtained from a variety of different experimental sources but can also be estimated from soil physics models such as HYDRUS [23], STANMOD [24] and CHAIN\_2D [13], **Figure 8**. Such models can and have been used to simulate both volatility from soil along with movement into the soil profile. Advantages of using a soil physics model to estimate pesticide flux loss deal with low cost and the semi-infinite parameter space that can be explored with simulation techniques. Field studies in five different states that explored atmospheric flux loss for chloropicrin and 1,3-dichloropropene were validated with CHAIN\_2D and indicate the model can correctly capture both peak and cumulative emissions effectively for these two soil fumigants [17, 26]. Thus, CHAIN\_2D and similar models are useful

also be selected by the user in the SOFEA GUI.

SOFEA are the same and found elsewhere [3].

**3.13 Model output characterization**

**Figure 8.**

**24**

*Example use of soil physics model CHAIN\_2D where exploration of multiple mitigations strategies such as water sealing, tarps, depth of incorporation, etc. can be simulated, as well as concentration by time (C × T) within soil for biological efficacy.*

tools for extrapolating flux predictions to diverse scenarios where experimental observations are unavailable [27–29]. Examples of various mitigations strategies that can be explored using soil physics models include the use of agricultural films, increased soil injection depth for the fumigant, and under a near semi-infinite parameter combinations of meteorological, soil and agronomic properties, **Figure 9**.

Soil physics models should first be validated with field observations before being used to extrapolate to a variety of different conditions. Cryer and van Wesenbeeck [26] used CHAIN\_2D to validate against field observations, and then coupled CHAIN\_2D to several USEPA air dispersion models (ISCST3 [4] and CALPUFF [30]). Both cumulative and 1-h maximum air concentrations were simulated and compared against field observations with good success (the best observations and simulations results were between 6 and 8%). In addition, both ISCST3 and CALPUFF air dispersion models showed similar order of magnitude output predictions [17]. Chloropicrin and 1,3-D emissions through Totally Impermeable Film (TIF) were compared using HYRUS where the fumigant flux was simulated within a factor of ~2, though the timing of the peak was over-predicted by the model [29]. The authors suggest that field-based calibration should be conducted when tarps are used because of the lack of representative field effective permeability data for the tarps.

Most inputs can be specified as either discrete values, or as PDFs. If possible, PDFs should be used to maximize Monte Carlo capabilities of the SOFEA modeling system and encompass uncertainties and variability in model inputs. SOFEA can generate fumigant concentrations for each receptor in the simulation domain (up to 11,664 receptors have been simulated in a nine-township air shed), averaged over specific time intervals (24-h and yearly) or periods specified by the user. For example, the user could specify the output of 24-h average, 60-day average, and

annual average concentration PDFs, for assessing acute, sub chronic and chronic risk to exposed populations.

### **3.14 SOFEA sensitivity analysis**

A sensitivity analysis was conducted with earlier versions of SOFEA to determine which variables had the greatest impact on model predicted concentrations [3]. The dependent variable endpoint in the sensitivity analysis for Kern County, CA was the 15-D multi-direction average air concentration at 30.5 m buffer. Sensitive parameters were the crop percentage, application rate, application date and weather year, in addition to the amount of 1,3-D mass applied in a township and the proximity of a treated field to a monitoring location. Additional parametric sensitivity analysis for CHAIN\_2D/ISCST3 showed several soil and irrigation parameters as consistently sensitive, including depth of incorporation into soil, tarp material, and initial soil water content [17].

### **3.15 Historical uses of SOFEA**

A moderate overprediction in air concentrations was made by SOFEA when predicting regional air concentrations for Ventura and Merced counties in California [31] which included 25 contiguous townships and treated at 1.5 times the current township allocation using 1,3-D (or at maximum levels of 1,3-D used between 1999 and 2006). However, this work provided an example of how SOFEA could be used using actual agronomic practices to manage the use of soil fumigant products and long-term exposure and risk to residents located in high-use regions. This publication also discussed how high-use rural areas leading to the highest predicted air concentrations could be used in a formalized risk assessment. The observation that the high concentrations were surrounding the downwind locations around treated fields was first predicted by Cryer and van Wesenbeeck [1] before the SOFEA modeling tool was fully developed.

SOFEA was improved with the release of SOFEA2 which eliminated the need for the third-party software Crystal Ball™ while also incorporating the ability to specify unique agronomic fields and air monitoring receptor locations. Further refinement includes post-processing hourly concentration predictions for precise starting intervals, the capacity to incorporate specific field flux loss from soil physics modeling for each field in the simulation domain, and the ability to accommodate drip applications made to vineyards. SOFEA and SOFEA2 generated the same output distributions when identically parameterized (unpublished work of Corteva Agriscience).

A 1,3-D air monitoring study was conducted in a high fumigant use area in Merced, CA, where 3-day average air concentrations were measured continuously at the approximate center at each of nine townships over a 14½ month period [21]. This monitoring study was designed specifically for validating SOFEA. Although SOFEA2 predicted the general pattern and correct order of magnitude for 1,3-D air concentrations as a function of time, it failed to recover the highest observed 1,3-D concentrations of the monitoring study which typically occurred in December. It was found the atmospheric mixing height was a significant parameter affecting the modeled 1,3-D concentrations. An algorithm that adjusted the PCRAMMET mixing height based on measured wind speed and air temperature was found to improve the simulated concentrations significantly, however the inclusion of AERMOD in SOFEA3 improved the model fit to observed data without requiring any mixing height adjustment. Comparison of the output probability density functions (PDFs) for 72 h 1,3-D concentrations between monitoring observations and SOFEA4

**27**

tion can be obtained.

**Figure 10.**

*Prediction of Agricultural Contaminant Concentrations in Ambient Air*

simulation indicate that slight under-prediction of concentrations above the 99th percentile was off-set by slight over-prediction of the 1,3-D concentration distribution below the 99th percentile, resulting in the annual average 1,3-D concentration for the nine-receptor monitoring domain being slightly over predicted (<2%). This suggests that without further refinement, based upon field validation observations, SOFEA2 results are representative but conservative estimates of exposure for 1,3-D if border township contributions and mixing height (MH) adjustments for calm periods are considered. SOFEA2 proved a useful tool for estimating airborne levels of 1,3-D but showed some weakness when incorporating ISCST3 [21] and was

*Exceedance probability for Merced CA study from measured and simulated conditions when AERMOD* 

AERMOD was used in conjunction with SOFEA2 after 2016 (now named SOFEA3), following the knowledge that MH was one of the most sensitive variables and that ISCST3 and its associated meteorological pre-processer over-estimated MH during stable air (calm conditions). Analysis showed SOFEA3, when using AERMOD in lieu of ISCST3 as the air dispersion model, improves the predictions of observed 1,3-D concentrations, and obviates the need for adjustments of the MHs in the processed weather file, as was required with SOFEA2. Improvements are a result of the refined algorithms in AERMOD for prediction of MH during calm conditions, based on updated understanding of Planetary Boundary Layer (BPL) dynamics, the use of the Monin-Obukhov length scale (*L*), and the calculation of a convective and mechanical MH, the latter which is used only for stable conditions (when *L* > 0). **Figure 10** shows SOFEA3 results compared to the 2010–2011 California (Merced) monitoring study, while **Figure 11** represents the same simulation predictions at a much finer resolution (11,664 receptors per township) such that contour plots for air concentra-

SOFEA3 was rewritten using C++ and Qt to replace VBA, and now this latest

version is denoted SOFEA4. SOFEA4 was used to simulate 1,3-dichloropropene (1,3-D) concentrations in ambient air in three agricultural areas in the USA where soil fumigation is a critical aspect of pest management and crop production. The regions explored by van Wesenbeeck et al. [11] are the Pacific Northwest, the mid-Atlantic coast, and the Southeast coastal plain, **Figure 12**. The Merced, CA monitoring study served to represent the southwest region of the U.S. SOFEA4 is the latest modeling tool of SOFEA that has been modified to use AERMOD, the EPA's recommended regulatory air dispersion model, to predict short-, medium- and long-term pesticide concentrations in air resulting

*DOI: http://dx.doi.org/10.5772/intechopen.86091*

renamed SOFEA3 when AERMOD was included.

*replaced ISCST3 in SOFEA3 as the dispersion model for predictions.*

*Prediction of Agricultural Contaminant Concentrations in Ambient Air DOI: http://dx.doi.org/10.5772/intechopen.86091*

**Figure 10.**

*Atmospheric Air Pollution and Monitoring*

risk to exposed populations.

**3.14 SOFEA sensitivity analysis**

material, and initial soil water content [17].

**3.15 Historical uses of SOFEA**

modeling tool was fully developed.

Corteva Agriscience).

annual average concentration PDFs, for assessing acute, sub chronic and chronic

A sensitivity analysis was conducted with earlier versions of SOFEA to determine which variables had the greatest impact on model predicted concentrations [3]. The dependent variable endpoint in the sensitivity analysis for Kern County, CA was the 15-D multi-direction average air concentration at 30.5 m buffer. Sensitive parameters were the crop percentage, application rate, application date and weather year, in addition to the amount of 1,3-D mass applied in a township and the proximity of a treated field to a monitoring location. Additional parametric sensitivity analysis for CHAIN\_2D/ISCST3 showed several soil and irrigation parameters as consistently sensitive, including depth of incorporation into soil, tarp

A moderate overprediction in air concentrations was made by SOFEA when predicting regional air concentrations for Ventura and Merced counties in California [31] which included 25 contiguous townships and treated at 1.5 times the current township allocation using 1,3-D (or at maximum levels of 1,3-D used between 1999 and 2006). However, this work provided an example of how SOFEA could be used using actual agronomic practices to manage the use of soil fumigant products and long-term exposure and risk to residents located in high-use regions. This publication also discussed how high-use rural areas leading to the highest predicted air concentrations could be used in a formalized risk assessment. The observation that the high concentrations were surrounding the downwind locations around treated fields was first predicted by Cryer and van Wesenbeeck [1] before the SOFEA

SOFEA was improved with the release of SOFEA2 which eliminated the need for the third-party software Crystal Ball™ while also incorporating the ability to specify unique agronomic fields and air monitoring receptor locations. Further refinement includes post-processing hourly concentration predictions for precise starting intervals, the capacity to incorporate specific field flux loss from soil physics modeling for each field in the simulation domain, and the ability to accommodate drip applications made to vineyards. SOFEA and SOFEA2 generated the same output distributions when identically parameterized (unpublished work of

A 1,3-D air monitoring study was conducted in a high fumigant use area in Merced, CA, where 3-day average air concentrations were measured continuously at the approximate center at each of nine townships over a 14½ month period [21]. This monitoring study was designed specifically for validating SOFEA. Although SOFEA2 predicted the general pattern and correct order of magnitude for 1,3-D air concentrations as a function of time, it failed to recover the highest observed 1,3-D concentrations of the monitoring study which typically occurred in December. It was found the atmospheric mixing height was a significant parameter affecting the modeled 1,3-D concentrations. An algorithm that adjusted the PCRAMMET mixing height based on measured wind speed and air temperature was found to improve the simulated concentrations significantly, however the inclusion of AERMOD in SOFEA3 improved the model fit to observed data without requiring any mixing height adjustment. Comparison of the output probability density functions (PDFs) for 72 h 1,3-D concentrations between monitoring observations and SOFEA4

**26**

*Exceedance probability for Merced CA study from measured and simulated conditions when AERMOD replaced ISCST3 in SOFEA3 as the dispersion model for predictions.*

simulation indicate that slight under-prediction of concentrations above the 99th percentile was off-set by slight over-prediction of the 1,3-D concentration distribution below the 99th percentile, resulting in the annual average 1,3-D concentration for the nine-receptor monitoring domain being slightly over predicted (<2%). This suggests that without further refinement, based upon field validation observations, SOFEA2 results are representative but conservative estimates of exposure for 1,3-D if border township contributions and mixing height (MH) adjustments for calm periods are considered. SOFEA2 proved a useful tool for estimating airborne levels of 1,3-D but showed some weakness when incorporating ISCST3 [21] and was renamed SOFEA3 when AERMOD was included.

AERMOD was used in conjunction with SOFEA2 after 2016 (now named SOFEA3), following the knowledge that MH was one of the most sensitive variables and that ISCST3 and its associated meteorological pre-processer over-estimated MH during stable air (calm conditions). Analysis showed SOFEA3, when using AERMOD in lieu of ISCST3 as the air dispersion model, improves the predictions of observed 1,3-D concentrations, and obviates the need for adjustments of the MHs in the processed weather file, as was required with SOFEA2. Improvements are a result of the refined algorithms in AERMOD for prediction of MH during calm conditions, based on updated understanding of Planetary Boundary Layer (BPL) dynamics, the use of the Monin-Obukhov length scale (*L*), and the calculation of a convective and mechanical MH, the latter which is used only for stable conditions (when *L* > 0). **Figure 10** shows SOFEA3 results compared to the 2010–2011 California (Merced) monitoring study, while **Figure 11** represents the same simulation predictions at a much finer resolution (11,664 receptors per township) such that contour plots for air concentration can be obtained.

SOFEA3 was rewritten using C++ and Qt to replace VBA, and now this latest version is denoted SOFEA4. SOFEA4 was used to simulate 1,3-dichloropropene (1,3-D) concentrations in ambient air in three agricultural areas in the USA where soil fumigation is a critical aspect of pest management and crop production. The regions explored by van Wesenbeeck et al. [11] are the Pacific Northwest, the mid-Atlantic coast, and the Southeast coastal plain, **Figure 12**. The Merced, CA monitoring study served to represent the southwest region of the U.S. SOFEA4 is the latest modeling tool of SOFEA that has been modified to use AERMOD, the EPA's recommended regulatory air dispersion model, to predict short-, medium- and long-term pesticide concentrations in air resulting

#### **Figure 11.**

*Example of simulated SOFEA3 (i.e., AERMOD used) chronic air concentration predictions over a threetownship area of Merced CA where field location, application rate and date are known and air concentrations at a central location in each township was monitored over a 14-month period.*

#### **Figure 12.**

*Geographic region for SOFEA4 simulations for high 1,3-D use areas of the United States [11]. Results for the Southwest are found elsewhere [21].*

from representative agronomic practices and large air sheds. SOFEA3/4 with AERMOD improved model predictions over what SOFEA2 with ISCST3 produced, due to the more realistic mixing height (MH) calculated by the weather pre-processor that subsequently resulted in higher concentrations during calm (stable air) periods. Advantages of using SOFEA4 to model fumigant concentrations over monitoring approaches for fumigant concentrations in air, include the ability to predict concentrations at a much greater temporal frequency (and spatial locations) than could be accomplished by monitoring alone. 1,3-D application data obtained from growers along with local weather data was used to parameterize SOFEA4, and it was found the Human Equivalent Concentrations

**29**

*Prediction of Agricultural Contaminant Concentrations in Ambient Air*

not exceeded for four study areas with intense 1,3-D use [11, 21].

(HECs) for acute, short-term, subchronic, and chronic exposure for 1,3-D were

The SOil Fumigant Exposure Assessment (SOFEA) model was originally developed in 2005 [3], expanded/refined multiple times over the last 14 years to explore volatile pesticide exposure and bystander risk (most recent version being SOFEA4 where C++ replaces VBA, and now with AERMOD the principle model used). Multiple publications using SOFEA have been documented since SOFEA was first developed [3, 11, 14, 17, 21, 31], with the most recent manuscript on SOFEA4 use for high use regions in the United States currently undergoing a journal review process for publication [11]. SOFEA assembles sources (agricultural fields), various management practices, source strengths (pesticide flux rates), weather data from the region of interest, and executes an air dispersion model [AERMOD, ISCST3 (historical)] to simulate pesticide concentrations at user defined receptors whose concentration predictions can be used in exposure and risk assessment procedures. This book chapter describes the historical development of SOFEA up to the latest version (SOFEA4) including all attributes that have been described over the years. SOFEA now uses AERMOD, the officially sanctioned USEPA regulatory air dispersion model, in lieu of ISCST3 for air dispersion simulations. Recent SOFEA simulation results were compared to the ambient air monitoring data collected in an intensive 1,3-D fumigation field trial in Merced, CA specifically designed to validate SOFEA against monitoring information. SOFEA4 (using AERMOD) was shown to improve the prediction of high concentrations (and thus the annual average concentration) compared to SOFEA2 (using ISCST3) and earlier versions of SOFEA. Better comparison against field observations using SOFEA4 was attributed to the improved characterization of the Planetary Boundary Layer (PBL) during calm period conditions (low wind), and more realistic Mixing Height (MH)

calculations that are employed by AERMOD compared to ISCST3 [11, 21].

local growers in each study area [11]).

The validated SOFEA4 model was further used to simulate 1,3-D air concentrations in three study areas across the US having significant 1,3-D use [11]. These areas include Quincy, WA (representing the Pacific Northwest), Wilson, NC (representing the Atlantic coastal plain), and St. John's, FL (representing GA/ FL). Including Merced, CA results [21], these four agricultural regions represent land areas that account for ~95% of the 1,3-D sold in the USA, **Figure 12**. The most recent publication (under review) of SOFEA deals with actual field locations and 1,3-D application parameters used for two annual product use cycles (2015–2016, and 2016–2017) in multiple high use regions in the United States (as documented by

SOFEA has exceptional attributes and functionality compared to other similar modeling tools for addressing the exposure and risk from the use of volatile (or semi-volatile) pesticides. In a collaborative effort between Corteva Agriscience and Exponent, Inc. (an Engineering and Scientific Consulting company), SOFEA3 was upgraded to modern software engineering standards (renamed SOFEA4), and a new graphical interface was developed with C++ and Qt to provide users an Integrated Development Environment-like (IDE) experience in creating new simulation projects. Summary statistics, moving averages and quantiles can be calculated efficiently over millions of data points, comprising hourly concentrations over several years and thousands of receptors. A beta version of SOFEA4 is currently being tested and evaluated for release, and will be made publicly available online via Exponent, Inc. in mid-2019 for use with other volatile or semi-volatile chemicals for large-scale environmental and

*DOI: http://dx.doi.org/10.5772/intechopen.86091*

**4. Discussion**

(HECs) for acute, short-term, subchronic, and chronic exposure for 1,3-D were not exceeded for four study areas with intense 1,3-D use [11, 21].

## **4. Discussion**

*Atmospheric Air Pollution and Monitoring*

**28**

**Figure 12.**

**Figure 11.**

*Southwest are found elsewhere [21].*

from representative agronomic practices and large air sheds. SOFEA3/4 with AERMOD improved model predictions over what SOFEA2 with ISCST3 produced, due to the more realistic mixing height (MH) calculated by the weather pre-processor that subsequently resulted in higher concentrations during calm (stable air) periods. Advantages of using SOFEA4 to model fumigant concentrations over monitoring approaches for fumigant concentrations in air, include the ability to predict concentrations at a much greater temporal frequency (and spatial locations) than could be accomplished by monitoring alone. 1,3-D application data obtained from growers along with local weather data was used to parameterize SOFEA4, and it was found the Human Equivalent Concentrations

*Geographic region for SOFEA4 simulations for high 1,3-D use areas of the United States [11]. Results for the* 

*Example of simulated SOFEA3 (i.e., AERMOD used) chronic air concentration predictions over a threetownship area of Merced CA where field location, application rate and date are known and air concentrations* 

*at a central location in each township was monitored over a 14-month period.*

The SOil Fumigant Exposure Assessment (SOFEA) model was originally developed in 2005 [3], expanded/refined multiple times over the last 14 years to explore volatile pesticide exposure and bystander risk (most recent version being SOFEA4 where C++ replaces VBA, and now with AERMOD the principle model used). Multiple publications using SOFEA have been documented since SOFEA was first developed [3, 11, 14, 17, 21, 31], with the most recent manuscript on SOFEA4 use for high use regions in the United States currently undergoing a journal review process for publication [11]. SOFEA assembles sources (agricultural fields), various management practices, source strengths (pesticide flux rates), weather data from the region of interest, and executes an air dispersion model [AERMOD, ISCST3 (historical)] to simulate pesticide concentrations at user defined receptors whose concentration predictions can be used in exposure and risk assessment procedures. This book chapter describes the historical development of SOFEA up to the latest version (SOFEA4) including all attributes that have been described over the years. SOFEA now uses AERMOD, the officially sanctioned USEPA regulatory air dispersion model, in lieu of ISCST3 for air dispersion simulations. Recent SOFEA simulation results were compared to the ambient air monitoring data collected in an intensive 1,3-D fumigation field trial in Merced, CA specifically designed to validate SOFEA against monitoring information. SOFEA4 (using AERMOD) was shown to improve the prediction of high concentrations (and thus the annual average concentration) compared to SOFEA2 (using ISCST3) and earlier versions of SOFEA. Better comparison against field observations using SOFEA4 was attributed to the improved characterization of the Planetary Boundary Layer (PBL) during calm period conditions (low wind), and more realistic Mixing Height (MH) calculations that are employed by AERMOD compared to ISCST3 [11, 21].

The validated SOFEA4 model was further used to simulate 1,3-D air concentrations in three study areas across the US having significant 1,3-D use [11]. These areas include Quincy, WA (representing the Pacific Northwest), Wilson, NC (representing the Atlantic coastal plain), and St. John's, FL (representing GA/ FL). Including Merced, CA results [21], these four agricultural regions represent land areas that account for ~95% of the 1,3-D sold in the USA, **Figure 12**. The most recent publication (under review) of SOFEA deals with actual field locations and 1,3-D application parameters used for two annual product use cycles (2015–2016, and 2016–2017) in multiple high use regions in the United States (as documented by local growers in each study area [11]).

SOFEA has exceptional attributes and functionality compared to other similar modeling tools for addressing the exposure and risk from the use of volatile (or semi-volatile) pesticides. In a collaborative effort between Corteva Agriscience and Exponent, Inc. (an Engineering and Scientific Consulting company), SOFEA3 was upgraded to modern software engineering standards (renamed SOFEA4), and a new graphical interface was developed with C++ and Qt to provide users an Integrated Development Environment-like (IDE) experience in creating new simulation projects. Summary statistics, moving averages and quantiles can be calculated efficiently over millions of data points, comprising hourly concentrations over several years and thousands of receptors. A beta version of SOFEA4 is currently being tested and evaluated for release, and will be made publicly available online via Exponent, Inc. in mid-2019 for use with other volatile or semi-volatile chemicals for large-scale environmental and risk assessment procedures. Parties interested in using SOFEA4 can contact Exponent Inc. directly to obtain both the user's guide and the latest version of the model.

## **5. Conclusions**

SOFEA is used to predict soil fumigant and semi-volatile pesticide air concentrations under actual or projected use. SOFEA, first released in 2005 [3], has over a decade of development and refinement and is a comprehensive numerical tool that has been validated against many field trials and monitoring studies using 1,3-dichloropropene as summarized in this manuscript. Both the timing and magnitudes for more than 450 1,3-D treated fields in a 5 × 5 township domain were followed over a 1.5-year period [11, 21] in a field trial in CA specifically designed for SOFEA validation.

Examples of a soil physics model, CHAIN\_2D, for flux predictions has also been used for source strength predictions in SOFEA, in addition to using field observations that are scaled by depth of soil incorporation and the time of the year when the application is made. Comparison of SOFEA predictions against other field studies has been good. SOFEA is a powerful tool to simulate air concentrations for regional agronomic conditions when multiple fields are simulated under typical agronomic conditions. SOFEA is especially useful for regulatory risk managers and product stewards who often are required to make decisions when only limited or incomplete data is available.

Chemical exposure information generated by SOFEA can be and has been used in a formalized risk assessment where risk to human populations is addressed. Understanding how various agronomic BMPs affect acute, sub chronic, and chronic exposure is an essential requirement for proper stewardship of volatile and semivolatile pesticides, and SOFEA can be used for validation purposes or when limited, or no experimental evidence is available. SOFEA4 is being released to the public domain later in 2019 such that any user wanting to simulate air concentrations for volatile and semi-volatile chemicals in large and diverse airsheds can be incorporated, in support of (or in lieu of) monitoring trials.


**31**

**Author details**

Corteva Agriscience, USA

provided the original work is properly cited.

Steven Cryer\* and Ian van Wesenbeeck

\*Address all correspondence to: sacryer@dow.com

© 2019 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/3.0), which permits unrestricted use, distribution, and reproduction in any medium,

*Prediction of Agricultural Contaminant Concentrations in Ambient Air*

with a graphical user interface

USEPA United States Environmental Protection Agency

SOFEA SOil Fumigant Exposure Assessment system

Qt cross-platform framework for developing application software

*DOI: http://dx.doi.org/10.5772/intechopen.86091*

PLSS public land survey system

TIF totally impermeable film

VBA visual basic for applications

PRZM3 USEPA Pesticide Root Zone Model

SAP USEPA Scientific Advisory Panel

## **List of abbreviations**

*Prediction of Agricultural Contaminant Concentrations in Ambient Air DOI: http://dx.doi.org/10.5772/intechopen.86091*

*Atmospheric Air Pollution and Monitoring*

**5. Conclusions**

SOFEA validation.

incomplete data is available.

**List of abbreviations**

1,3-D 1,3-dichloropropene

BMP best management practice

solute transport C++ general purpose program language

Exponent Inc. US-based consulting company GIS geographic information system

software)

PDF probability density function

rated, in support of (or in lieu of) monitoring trials.

AERMOD USEPA Regulatory Model (Gaussian plume) ArcGIS GIS for geographic information by ESRI

CDPR California Department of Pesticide Regulations

ESRI Environmental Systems Research Institute

ArcView Entry level of ArcGIS Desktop, a GIS software by ESRI

CHAIN\_2D computer program for 2-D variability saturated water flow, heat,

Crystal Ball™ software add-on for MS Excel developed by Decisioneering, Inc.

Hydrus water, heat, solute transport in saturated porous media (modeling

ISCST3 USEPA Industrial Source Complex Model (Gaussian plume)

risk assessment procedures. Parties interested in using SOFEA4 can contact Exponent

SOFEA is used to predict soil fumigant and semi-volatile pesticide air concentrations under actual or projected use. SOFEA, first released in 2005 [3], has over a decade of development and refinement and is a comprehensive numerical tool that has been validated against many field trials and monitoring studies using 1,3-dichloropropene as summarized in this manuscript. Both the timing and magnitudes for more than 450 1,3-D treated fields in a 5 × 5 township domain were followed over a 1.5-year period [11, 21] in a field trial in CA specifically designed for

Examples of a soil physics model, CHAIN\_2D, for flux predictions has also been used for source strength predictions in SOFEA, in addition to using field observations that are scaled by depth of soil incorporation and the time of the year when the application is made. Comparison of SOFEA predictions against other field studies has been good. SOFEA is a powerful tool to simulate air concentrations for regional agronomic conditions when multiple fields are simulated under typical agronomic conditions. SOFEA is especially useful for regulatory risk managers and product stewards who often are required to make decisions when only limited or

Chemical exposure information generated by SOFEA can be and has been used

in a formalized risk assessment where risk to human populations is addressed. Understanding how various agronomic BMPs affect acute, sub chronic, and chronic exposure is an essential requirement for proper stewardship of volatile and semivolatile pesticides, and SOFEA can be used for validation purposes or when limited, or no experimental evidence is available. SOFEA4 is being released to the public domain later in 2019 such that any user wanting to simulate air concentrations for volatile and semi-volatile chemicals in large and diverse airsheds can be incorpo-

Inc. directly to obtain both the user's guide and the latest version of the model.

**30**


## **Author details**

Steven Cryer\* and Ian van Wesenbeeck Corteva Agriscience, USA

\*Address all correspondence to: sacryer@dow.com

© 2019 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

## **References**

[1] Cryer SA, van Wesenbeeck IJ. Predicted 1,3-dichloropropene air concentrations resulting from tree and vine applications in California. Journal of Environmental Quality. 2001;**20**:1887-1895

[2] Cryer SA, van Wesenbeeck IJ, Knuteson JA. Predicting regional emissions and near-field air concentrations of soil fumigants using modest numerical algorithms: A case study using 1,3-dichloropropene. Journal of Agricultural and Food Chemistry. 2003;**51**:3401-3409

[3] Cryer SA. Predicting soil fumigant acute, sub-chronic, and chronic air concentrations under diverse agronomic practices. Journal of Environmental Quality. 2005;**34**:2197-2207

[4] USEPA. Industrial Source Complex (ISC3) Dispersion Model User's Guide; Volumes I and II. EPA-454/B-95-003a and b. Research Triangle Park, North Carolina: U.S. Environmental Protection Agency; 1995

[5] AERMOD, United States Environmental Protection Agency. AERMOD: Description of Model Formulation. EPA-454/R-03-004. Research Triable Park, North Carolina: U.S. Environmental Protection Agency; 2004. Available from: http://www. epa.gov/scram001/7thconf/aermod/ aermod\_mfd.pdf

[6] Cimorelli AJ, Perry SG, Venkatram A, Weil JC, Paine RJ, Wilson RB, et al. AERMOD: Description of Model Formulations. EPR-454/R-03-004. Research Triangle Park, North Carolina: U.S. Enviromental Protection Agency; 2004

[7] Hao J, He D, Wu Y, Fu L, He K. A study of the emission and concentration distribution of vehicular pollutants in the urban area of Beijing. Atmospheric Environment. 1999;**34**:454-465

[8] Sullivan DA, Holdworth MT, Hinka DJ. A Monte Carlo-based dispersion modeling of off-gassing releases from the fumigant metamsodium for determining distances to exposure endpoints. Atmospheric Environment. 2004;**38**:2741-2481

[9] Heckel PF, LeMasters GK. The use of AERMOD air pollution dispersion models to estimate residual ambient concentrations of elemental mercury. Water, Air, and Soil Pollution. 2011;**219**(1-4):377-388

[10] Jittra N, Pinthong N, Thepanondh S. Performance evaluation of AERMOD and CALPUFF air dispersion models in industrial complex area. Air, Soil and Water Research. 2015;**8**:87-95. DOI: 10.4137/ASWR.S32781

[11] de Cirugeda Helle O, van Wesenbeeck IJ, Cryer SA. SOFEA modeling of 1,3-Dichloropropene concentrations in ambient air in high fumigant use areas of the United States. American Chemical Society (ACS). 255th National meeting. Boston, MA. August 19-23, 2018

[12] Yu C, Zheng C. HYDRUS: Software for flow and transport modeling in variably saturated media. Software Spotlight, Ground Water. 2010;**48**(6):787-791

[13] Šimůnek J, van Genuchten MTh. The CHAIN\_2D Code for Simulating Two-Dimensional Movement of Water, Heat, and Multiple Solutes in Variably-Saturated Porous Media, Version 1.1. Research Report No. 136; Riverside, California: U.S. Salinity Laboratory, USDA, ARS; 1994

[14] USEPA Agency Scientific Advisory Panel 2004. Fumigant Bystander Exposure Model Review: Soil Fumigant Exposure Assessment System (SOFEA) Using Telone as a Case Study. Available

**33**

1997

*Prediction of Agricultural Contaminant Concentrations in Ambient Air*

Quality Assessment Division, Research Triangle Park, NC: EPA-454/R-18-003 U.S. Environmental Protection Agency. 2018. Available from: https://www3. epa.gov/ttn/scram/models/aermod/

[23] Šimůnek J, Van Genuchten MTh, Sejna M. The HYDRUS software package for simulating two- and three-dimensional movement of water, heat and multiple solutes in variablesaturated media. Technical manual, version 3.0, PC-Progress, Prague, Czech

[24] van Genuchten MT, Šimůnek J, Leij FL, Toride N, Šejna M. STANMOD: Model use, calibration and validation, special issue standard/engineering procedures for model calibration and validation. Transactions of the ASABE.

[25] PRZM-3. A model for Predicting Pesticide and Nitrogen Fate in the Crop Root and Unsaturated Soil Zones: User's Manual for Release 3.12.2.

EPA/600/R-05/111. Ecosystems Research Division, National Exposure Research Laboratory, Athens, GA, 30605-2700, U.S. Environmental Protection Agency, Office of Research and Development,

aermod\_mfed.pdf

Republic, 2018. p. 274

2012;**5**(4):1353-1366

Washington, DC; 2005

against chloropicrin and

agwat.2014.04.007

Environmental Modeling and Assessment. 2010;**15**:309-318

[27] Wang J, Huang G, Zhan H, Mohanty BP, Zheng J, Huang Q, et al. Evaluation of soil water dynamics and crop yield under furrow irrigation with a two-dimensional flow and crop growth coupled model. Agricultural Water Management. 2014;**141**(10-22):2014. DOI: 10.1016/j.

[26] Cryer SA, van Wesenbeeck IJ. Estimating field volatility of Soil Fumigants Using CHAIN\_2D: Mitigation methods and comparison

1,3-dichloropropene field observations.

*DOI: http://dx.doi.org/10.5772/intechopen.86091*

from: https://archive.epa.gov/scipoly/ sap/meetings/web/html/090904\_mtg. html [Accessed: 10 September 2004]

[15] Ross LJ, Johnson B, Kim KD, Hsu J. Prediction of methyl bromide flux from area sources using the ISCST model. Journal of Environmental Quality.

[16] Yates SR, Ashworth DJ, Zheng W, Knuteson J, van Wesenbeeck IJ. Effect of deep injection on field-scale emissions of 1,3-dichloropropene and chloropicrin

from bare soil. Atmospheric Environment. 2016;**137**:135-145

[17] Cryer SA, van Wesenbeeck IJ. Coupling field observations, soil modeling, and air dispersion algorithms

to estimate 1,3-dichloropropene and chloropicrin flux and exposure. Journal of Environmental Quality. 2011;**2011**(40):1450-1461. (Invited Paper for Special Submissions: Agricultural Air Quality)

[18] Rubinstein RY. Simulation and the Monte Carlo Method. New York, NY, USA: John Wiley & Sons, Inc.; 1981.

[20] USEPA. Guiding Principles for Monte Carlo Analysis. EPA/630/R-97/001. Risk Assessment Forum. Washington, DC: U.S. Environmental Protection Agency;

[21] van Wesenbeeck IJ, Cryer SA, de Cirugeda Helle O, Li C, Driver J. Comparison of regional air dispersion

monitoring data for the soil fumigant 1,3-dichloropropene. Science of the Total Environment. 2016:569-570 603-610

Formulation and Evaluation. Office of Air Quality Planning and Standards, Air

simulation and ambient air

[22] USEPA. AERMOD Model

[19] Yakowitz S. Computational Probability and Simulation. Reading, MA: Addison-Wesley Publishers; 1977

ISBN: 0471089176

1996;**25**:885-891

*Prediction of Agricultural Contaminant Concentrations in Ambient Air DOI: http://dx.doi.org/10.5772/intechopen.86091*

from: https://archive.epa.gov/scipoly/ sap/meetings/web/html/090904\_mtg. html [Accessed: 10 September 2004]

[15] Ross LJ, Johnson B, Kim KD, Hsu J. Prediction of methyl bromide flux from area sources using the ISCST model. Journal of Environmental Quality. 1996;**25**:885-891

[16] Yates SR, Ashworth DJ, Zheng W, Knuteson J, van Wesenbeeck IJ. Effect of deep injection on field-scale emissions of 1,3-dichloropropene and chloropicrin from bare soil. Atmospheric Environment. 2016;**137**:135-145

[17] Cryer SA, van Wesenbeeck IJ. Coupling field observations, soil modeling, and air dispersion algorithms to estimate 1,3-dichloropropene and chloropicrin flux and exposure. Journal of Environmental Quality. 2011;**2011**(40):1450-1461. (Invited Paper for Special Submissions: Agricultural Air Quality)

[18] Rubinstein RY. Simulation and the Monte Carlo Method. New York, NY, USA: John Wiley & Sons, Inc.; 1981. ISBN: 0471089176

[19] Yakowitz S. Computational Probability and Simulation. Reading, MA: Addison-Wesley Publishers; 1977

[20] USEPA. Guiding Principles for Monte Carlo Analysis. EPA/630/R-97/001. Risk Assessment Forum. Washington, DC: U.S. Environmental Protection Agency; 1997

[21] van Wesenbeeck IJ, Cryer SA, de Cirugeda Helle O, Li C, Driver J. Comparison of regional air dispersion simulation and ambient air monitoring data for the soil fumigant 1,3-dichloropropene. Science of the Total Environment. 2016:569-570 603-610

[22] USEPA. AERMOD Model Formulation and Evaluation. Office of Air Quality Planning and Standards, Air Quality Assessment Division, Research Triangle Park, NC: EPA-454/R-18-003 U.S. Environmental Protection Agency. 2018. Available from: https://www3. epa.gov/ttn/scram/models/aermod/ aermod\_mfed.pdf

[23] Šimůnek J, Van Genuchten MTh, Sejna M. The HYDRUS software package for simulating two- and three-dimensional movement of water, heat and multiple solutes in variablesaturated media. Technical manual, version 3.0, PC-Progress, Prague, Czech Republic, 2018. p. 274

[24] van Genuchten MT, Šimůnek J, Leij FL, Toride N, Šejna M. STANMOD: Model use, calibration and validation, special issue standard/engineering procedures for model calibration and validation. Transactions of the ASABE. 2012;**5**(4):1353-1366

[25] PRZM-3. A model for Predicting Pesticide and Nitrogen Fate in the Crop Root and Unsaturated Soil Zones: User's Manual for Release 3.12.2. EPA/600/R-05/111. Ecosystems Research Division, National Exposure Research Laboratory, Athens, GA, 30605-2700, U.S. Environmental Protection Agency, Office of Research and Development, Washington, DC; 2005

[26] Cryer SA, van Wesenbeeck IJ. Estimating field volatility of Soil Fumigants Using CHAIN\_2D: Mitigation methods and comparison against chloropicrin and 1,3-dichloropropene field observations. Environmental Modeling and Assessment. 2010;**15**:309-318

[27] Wang J, Huang G, Zhan H, Mohanty BP, Zheng J, Huang Q, et al. Evaluation of soil water dynamics and crop yield under furrow irrigation with a two-dimensional flow and crop growth coupled model. Agricultural Water Management. 2014;**141**(10-22):2014. DOI: 10.1016/j. agwat.2014.04.007

**32**

*Atmospheric Air Pollution and Monitoring*

[1] Cryer SA, van Wesenbeeck IJ. Predicted 1,3-dichloropropene air concentrations resulting from tree and vine applications in California. Journal of Environmental Quality.

[8] Sullivan DA, Holdworth MT, Hinka DJ. A Monte Carlo-based dispersion modeling of off-gassing releases from the fumigant metamsodium for determining distances to exposure endpoints. Atmospheric Environment. 2004;**38**:2741-2481

[9] Heckel PF, LeMasters GK. The use of AERMOD air pollution dispersion models to estimate residual ambient concentrations of elemental mercury.

[10] Jittra N, Pinthong N, Thepanondh S. Performance evaluation of AERMOD and CALPUFF air dispersion models in industrial complex area. Air, Soil and Water Research. 2015;**8**:87-95. DOI:

Water, Air, and Soil Pollution.

2011;**219**(1-4):377-388

10.4137/ASWR.S32781

August 19-23, 2018

2010;**48**(6):787-791

USDA, ARS; 1994

[11] de Cirugeda Helle O, van Wesenbeeck IJ, Cryer SA. SOFEA modeling of 1,3-Dichloropropene concentrations in ambient air in high fumigant use areas of the United States. American Chemical Society (ACS). 255th National meeting. Boston, MA.

[12] Yu C, Zheng C. HYDRUS: Software for flow and transport modeling in variably saturated media. Software Spotlight, Ground Water.

[13] Šimůnek J, van Genuchten MTh. The CHAIN\_2D Code for Simulating Two-Dimensional Movement of Water, Heat, and Multiple Solutes in Variably-Saturated Porous Media, Version 1.1. Research Report No. 136; Riverside, California: U.S. Salinity Laboratory,

[14] USEPA Agency Scientific Advisory Panel 2004. Fumigant Bystander Exposure Model Review: Soil Fumigant Exposure Assessment System (SOFEA) Using Telone as a Case Study. Available

[2] Cryer SA, van Wesenbeeck IJ, Knuteson JA. Predicting regional emissions and near-field air

concentrations of soil fumigants using modest numerical algorithms: A case study using 1,3-dichloropropene. Journal of Agricultural and Food Chemistry. 2003;**51**:3401-3409

[3] Cryer SA. Predicting soil fumigant acute, sub-chronic, and chronic air concentrations under diverse agronomic practices. Journal of Environmental

[4] USEPA. Industrial Source Complex (ISC3) Dispersion Model User's Guide; Volumes I and II. EPA-454/B-95-003a and b. Research Triangle Park, North Carolina: U.S. Environmental Protection

[6] Cimorelli AJ, Perry SG, Venkatram A, Weil JC, Paine RJ, Wilson RB, et al. AERMOD: Description of Model Formulations. EPR-454/R-03-004. Research Triangle Park, North Carolina: U.S. Enviromental Protection Agency; 2004

[7] Hao J, He D, Wu Y, Fu L, He K. A study of the emission and concentration distribution of vehicular pollutants in the urban area of Beijing. Atmospheric

Environment. 1999;**34**:454-465

Quality. 2005;**34**:2197-2207

[5] AERMOD, United States Environmental Protection Agency. AERMOD: Description of Model Formulation. EPA-454/R-03-004. Research Triable Park, North Carolina: U.S. Environmental Protection Agency; 2004. Available from: http://www. epa.gov/scram001/7thconf/aermod/

Agency; 1995

aermod\_mfd.pdf

2001;**20**:1887-1895

**References**

#### *Atmospheric Air Pollution and Monitoring*

[28] Abdou HM, Flury M. Simulation of water flow and solute transport in free-drainage lysimeters and field soils with heterogeneous structures. European Journal of Soil Science. 2004;**55**(2):229-241

**Chapter 4**

**Abstract**

Zaria-Nigeria

**35**

**1. Introduction**

Atmospheric Air Pollution in

Nigeria: A Correlation between

Vehicular Traffic and Criteria

Pollutant Levels

*and Mohammed Ahmed Bichi*

observed outdoor pollutant measurements.

*Yahaya Abbas Aliyu, Joel Ondego Botai,*

*Aliyu Zailani Abubakar,Terwase Tosin Youngu,*

*Jimoh Olanrewaju Sule, Mohammed Wachin Shebe*

In Nigeria, the rising levels of used/poorly maintained vehicles are contributing to most urban air pollution with possible repercussion on the general public health. This study evaluates the inferences of vehicular traffic surge on outdoor pollutant measurement using Zaria, northern Nigeria, as a case study. The study collected a 1-year time-series dataset for the vehicular count and the respective outdoor criteria pollutant measurements over 19 study sites. The vehicular traffic was categorized into motorcycles (2-W), tricycles (3-W), cars, buses, light-duty vehicles (LDV) and heavy-duty vehicles (HDV). The outdoor pollutants that were measured include carbon monoxide (CO), sulfur dioxide (SO2) and particulate matter (PM2.5/PM10). We utilized validated portable monitors (CW-HAT200 particulate counter and the MSA Altair 5x multigas sensor) for the outdoor measurements during December 2015–November 2016. The observed measurements for the validation procedure were normally distributed [kurtosis (0.301); skewness (0.334)] and coefficient of determination (R2 ≥ 0.808). The time-series analysis of particulate matter (PM) measurements displayed alarming concentrations levels. Combined vehicular traffic density analysis revealed significant contribution (R ≥ 0.619) to the population exposed outdoor pollutant measurements. The 2-W (motorcycle) was found to be the vehicular category that attributed the most significant relationship with

**Keywords:** urban air quality, vehicular traffic, portable sensors, criteria pollutants,

In most developing countries, atmospheric pollution continues to affect exposed population health [1–3]. In Africa, air quality studies are devising alternative and reliable means to obtain pollutant measurements for research. The approach

[29] Spurlock F, Johnson B, Tuli A, Gao S, Tao J, Sartori F, et al. Simulation of fumigant transport and volatilization from tarped broadcast applications. Vadose Zone Journal. 2013;**12**(3):1-10

[30] Scire JS, Strimaitis DG, Yamartino RJ. A User's Guide for The CALPUFF Dispersion Model. Concord, MA: Earth Tech., Inc.; 2002

[31] van Wesenbeeck IJ, Cryer SA, Havens PL, Houtman BA. Use of SOFEA to predict 1,3-D concentrations in air in high use regions of California. Journal of Environmental Quality. 2011;**40**:1462-1469

## **Chapter 4**

*Atmospheric Air Pollution and Monitoring*

[28] Abdou HM, Flury M. Simulation of water flow and solute transport in free-drainage lysimeters and field soils with heterogeneous structures. European Journal of Soil Science.

[29] Spurlock F, Johnson B, Tuli A, Gao S, Tao J, Sartori F, et al. Simulation of fumigant transport and volatilization from tarped broadcast applications. Vadose Zone Journal. 2013;**12**(3):1-10

[30] Scire JS, Strimaitis DG, Yamartino RJ. A User's Guide for The CALPUFF Dispersion Model. Concord, MA: Earth

[31] van Wesenbeeck IJ, Cryer SA, Havens PL, Houtman BA. Use of SOFEA to predict 1,3-D concentrations in air in high use regions of California. Journal of Environmental Quality.

2004;**55**(2):229-241

Tech., Inc.; 2002

2011;**40**:1462-1469

**34**
