AGDISP and AGDISPpro: A Comprehensive Guide to Predicting Pesticide Spray Drift in Air

Stella Jenkins Dec 02, 2025 375

This article provides a detailed exploration of the AGDISP model, a critical tool for researchers and environmental assessors predicting off-target pesticide spray drift.

AGDISP and AGDISPpro: A Comprehensive Guide to Predicting Pesticide Spray Drift in Air

Abstract

This article provides a detailed exploration of the AGDISP model, a critical tool for researchers and environmental assessors predicting off-target pesticide spray drift. It covers the model's foundational principles, from its Lagrangian-based mechanics for near-field predictions to its Gaussian extension for far-field transport. The scope includes practical guidance on model application and parameterization for various equipment, including emerging remotely piloted aerial systems (RPAAS). It further addresses model limitations, optimization strategies, and a synthesis of recent validation studies that benchmark the model's performance against field data, offering a complete resource for ecological risk assessment and application planning.

Understanding AGDISP: The Core Principles of Spray Drift Modeling

AGDISP (AGricultural DISPersal) is a mechanistic, Lagrangian-based model developed by the US Forest Service that predicts the off-target drift and deposition of spray applications [1]. As a "first-principles" science-based model, it employs fundamental physical equations to simulate the release, dispersion, and deposition of spray droplets downwind of application areas [1] [2]. Originally designed for optimizing aerial spraying operations, the model incorporates detailed algorithms that account for complex interactions between spraying equipment, release parameters, droplet physics, and meteorological conditions [1] [2].

The model has evolved significantly from its initial focus on fixed-wing aircraft to encompass a wide range of application technologies. AGDISP's development represents a continuous effort to address the spray drift prediction needs of regulatory bodies, applicators, and researchers [3] [4]. The United States Environmental Protection Agency (USEPA) now recognizes AGDISP and its derivative AgDRIFT as key tools for pesticide risk assessment, particularly for estimating downwind deposition from aerial, ground boom, and orchard/vineyard applications [1].

Table: AGDISP Model Versions and Primary Applications

Model Version Primary Application Methods Key Features Regulatory Status
AGDISP (v8.26) Aerial, ground boom, forestry, adulticide/mosquitocide [1] First-principles mechanistic model; ~36 adjustable inputs [2] Used by EPA for drift estimation [1]
AgDRIFT (v2.1.1) Aerial, ground boom, orchard/vineyard airblast [1] Modified version of AGDISP for screening-level assessment [1] Used in EPA risk assessments and label generation [1] [5]
AGDISPpro Remotely Piloted Aerial Application Systems (RPAAS/spray drones) [3] [4] Combines atmospheric transport with multi-rotor aerodynamic models [3] Under evaluation by regulatory bodies [3]

Model Evolution and Technological Expansion

From Fixed-Wing Aircraft to Modern Application Systems

AGDISP's initial development focused on conventional fixed-wing agricultural aircraft, but its framework has proven adaptable to diverse spraying technologies. The model now includes capabilities for forestry applications, mosquito control operations, and ground boom sprayers [1]. This expansion reflects the model's robust fundamental architecture, which can accommodate different release scenarios by modifying input parameters that define equipment characteristics and operational settings.

The model's transition to encompass new technologies is particularly evident in its incorporation of orchard airblast spraying applications. While initially designed for aerial application, researchers have worked to adapt AGDISP's mechanistic approach to airblast scenarios, though this requires accounting for additional complexities such as canopy characteristics and sprayer type [2]. The continuous development efforts have positioned AGDISP as a versatile platform that can be extended to various spraying technologies through appropriate parameterization and validation.

Incorporation of Spray Drone (RPAAS) Modeling

The most recent expansion of AGDISP addresses the growing use of Remotely Piloted Aerial Application Systems (RPAAS), commonly known as spray drones [3] [4]. AGDISPpro represents a significant advancement by incorporating aerodynamic models specific to multi-rotor systems, combining atmospheric transport algorithms with detailed characterizations of drone-induced airflow patterns [3]. This development fills a critical gap in regulatory modeling frameworks, as spray drones present unique operational characteristics compared to conventional aircraft.

Current versions of AGDISPpro include at least nine different RPAAS models, including quadcopter (e.g., PV22) and hexacopter (e.g., PV35X) configurations [3]. The validation efforts for these drone models have demonstrated promising results, with indices of agreement between predictions and field observations ranging from 0.47 to 0.94 across different spray quality scenarios [3] [4]. This expansion maintains AGDISP's relevance amid technological shifts in application methods while providing regulators with tools to assess emerging application technologies.

G Original Original AGDISP (Fixed-Wing Aircraft) Expansion1 Ground Boom Sprayers Original->Expansion1 Expansion2 Orchard Airblast Sprayers Original->Expansion2 Expansion3 Forestry Applications Original->Expansion3 Current AGDISPpro (Spray Drones/RPAAS) Expansion1->Current Expansion2->Current Expansion3->Current

AGDISP Model Evolution Pathway

AGDISPpro Validation and Performance Metrics

Field Validation Methodology for Spray Drift Assessment

The validation of AGDISPpro for spray drone applications followed rigorous experimental protocols involving field studies with commercial unmanned aerial systems [3] [4]. These studies employed two primary experimental designs: single-swath applications using Medium and Extremely Coarse spray nozzles, and four-swath applications using Fine and Ultra Coarse spray nozzles [4]. The research collected both in-swath and downwind deposition measurements using artificial samplers, which were subsequently analyzed to generate quantitative drift deposition data comparable to model predictions.

Meteorological conditions during validation trials were meticulously recorded, as wind speed, wind direction, and atmospheric pressure have been identified as significant factors accounting for variability in drift measurements [3] [2]. The field experiments were conducted under conditions representative of typical operational scenarios, with applications made at standard flight heights and forward speeds for the respective drone systems. This approach ensured that validation data reflected real-world conditions that the model would need to accurately simulate.

Quantitative Performance Assessment

AGDISPpro demonstrates varying performance levels depending on spray quality characteristics. For Extremely Coarse spray droplet spectra, the model showed indices of agreement ranging from 0.61 to 0.94 (n=12), indicating reasonably accurate predictions [3] [4]. With Medium droplet spectra, performance was more variable, with indices of agreement ranging from 0.47 to 0.92 (n=9) [4]. In multiple-swath scenarios, the model performed better with Fine sprays (indices of agreement: 0.86-0.93, n=3) compared to Ultra Coarse sprays (indices of agreement: 0.48-0.55) [3] [4].

A consistent observation across validation studies was AGDISPpro's tendency to slightly under-predict off-target deposition values compared to field measurements, as indicated by negative mean bias error values [4]. Researchers identified uncertainty in swath width and swath displacement parameters as significant contributors to this discrepancy, as these factors substantially affect the magnitude of peak deposition and the position of the spray deposition plume [3] [4]. This systematic under-prediction highlights areas for future model refinement while confirming the model's fundamental utility for spray drift estimation.

Table: AGDISPpro Performance Across Spray Qualities

Spray Quality Application Type Index of Agreement Range Sample Size (n) Systematic Bias
Medium Single-swath 0.47 - 0.92 9 Under-prediction [4]
Extremely Coarse Single-swath 0.61 - 0.94 12 Under-prediction [3] [4]
Fine Four-swath 0.86 - 0.93 3 Under-prediction [3] [4]
Ultra Coarse Four-swath 0.48 - 0.55 Not specified Under-prediction [3] [4]

Experimental Protocols for Spray Drift Measurement

Field Setup and Sampler Configuration

Spray drift validation studies require meticulous field setup following EPA-approved protocols [2]. The standard approach involves establishing a primary sampling line perpendicular to the wind direction, with artificial samplers distributed at increasing distances from the application area edge. Typical configurations place samplers from immediately downwind of the treatment area to distances exceeding 500 meters, depending on the application method and expected drift potential [2]. In almond orchard studies, for example, samplers were deployed up to 183 meters from the orchard edge, with terminal drift detected as far as 531 meters for artificial foliage samplers [2].

The sampling apparatus generally includes multiple sampler types to capture different drift phenomena: Mylar cards or Petri dishes for deposition measurement, artificial foliage to simulate plant interception, and horizontal strings to assess airborne droplet capture [2] [5]. Water-sensitive cards are frequently employed to quantify spray coverage patterns and droplet density [5]. All samplers are typically positioned at standardized heights relevant to potential exposure scenarios, often matching canopy height for agricultural crops or relevant inhalation heights for human exposure assessment.

Tracer Application and Meteorological Data Collection

Validation studies commonly use fluorescent dye tracers (such as pyranine) mixed with water to simulate pesticide sprays [2]. This enables precise quantification of deposition without the complications of active pesticide ingredients. Applications are conducted at representative operating parameters, including appropriate application rates (e.g., 935.4 L/ha for orchard applications), nozzle types, pressure settings, and vehicle speeds [2]. For drone applications, multiple flight swaths are often performed to establish characteristic deposition patterns.

Comprehensive meteorological monitoring is essential throughout testing, requiring instrumentation that measures wind speed, wind direction, air temperature, relative humidity, and atmospheric stability at multiple heights [2] [5]. Data should be collected at high frequency (e.g., every 1-5 seconds) throughout each replication, with averages and standard deviations calculated for appropriate intervals. EPA protocols specify that applications should only proceed when wind direction deviates less than 30° from the established sampling line and within defined wind speed ranges (typically 3-15 km/h) to ensure data quality [5].

G P1 1. Site Selection & Setup P2 2. Sampler Deployment P1->P2 Sub1 Establish transects perpendicular to wind P3 3. Meteorological Instrumentation P2->P3 Sub2 Position Mylar cards, WSC, artificial foliage P4 4. Spray Application P3->P4 Sub3 Install weather stations at application height P5 5. Sample Collection & Analysis P4->P5 Sub4 Apply fluorescent tracer with calibrated equipment P6 6. Data Validation P5->P6 Sub5 Retrieve samplers; fluorometry analysis Sub6 Compare with model predictions

Spray Drift Validation Workflow

Comparative Drift Analysis Across Application Methods

Drift Potential of Different Application Technologies

Research consistently demonstrates significant variation in drift potential between application methods. Comparative studies of ground and aerial applications using the herbicide florpyrauxifen-benzyl found that aerial applications resulted in 5.0- to 8.6-fold greater downwind drift compared to ground boom systems when both applied Coarse sprays in 13 km/h average wind conditions [5]. This increased drift potential translated to more extensive downwind crop injury, with aerial applications causing nearly 100% reduction in soybean reproductive structures at 61 meters downwind, compared to approximately 25% reduction for ground applications at 30.5 meters [5].

The primary factors contributing to these differences include release height, droplet size spectrum, and equipment-induced airflow. Ground applications typically release spray at heights of 0.5-1 meter above the crop canopy, while aerial applications occur at 3-5 meters, resulting in significantly longer droplet flight time and greater drift potential [5]. Additionally, aircraft-generated turbulence and rotor downwash from spray drones create complex airflow patterns that influence droplet trajectory and evaporation rates, creating distinctive drift profiles that differ from ground-based applications [3] [5].

Droplet Size and Meteorological Influences

Droplet size characteristics represent one of the most significant controllable factors affecting drift potential across all application methods. Research shows that finer spray qualities (producing smaller droplets) consistently result in greater downwind drift compared to coarser sprays [3] [5]. AGDISP incorporates detailed droplet size distribution data, typically characterized using the American Society of Agricultural and Biological Engineers (ASABE) S572.1 standard, which defines categories ranging from Fine to Ultra Coarse [3] [6].

Meteorological conditions during application substantially influence drift outcomes, with wind speed, temperature, relative humidity, and atmospheric stability all contributing to variability in measured drift [2] [5]. Statistical analyses of field data indicate that wind direction, wind speed, and atmospheric pressure are particularly significant in accounting for drift variability between spray trials [2]. These relationships are mechanistically represented in AGDISP through physical equations that model droplet evaporation, transport, and deposition processes in response to environmental conditions.

Table: Key Drift-Influencing Factors Modeled in AGDISP

Factor Category Specific Parameters Impact on Drift Potential AGDISP Representation
Equipment Parameters Release height, nozzle type, airspeed, boom configuration [5] Higher release height increases drift; finer sprays increase drift [5] Detailed equipment libraries and nozzle characteristics [2]
Spray Formulation Droplet size spectrum, evaporation rate, formulation type [5] Smaller droplets drift farther; rapid evaporation increases drift [5] Lagrangian droplet tracking with evaporation algorithms [2]
Meteorological Conditions Wind speed, temperature, relative humidity, atmospheric stability [2] [5] Higher wind speed increases drift; low humidity increases evaporation [2] Physical equations modeling droplet transport and fate [2]
Operational Practices Swath width, swath displacement, forward speed, upwind adjustments [3] [6] Proper swath displacement reduces downwind drift [6] Application scenario parameterization [3]

The Scientist's Toolkit: Research Reagent Solutions

Table: Essential Materials for Spray Drift Research

Research Reagent/Material Function in Spray Drift Studies Application Example Citation
Fluorescent Dyes (e.g., Pyranine) Tracer compound for quantifying spray deposition and drift Applied in water solution as pesticide surrogate; enables fluorometry analysis [2]
Artificial Deposition Samplers Collect spray droplets for quantitative analysis Mylar cards, Petri dishes positioned at varying distances and heights [2] [5]
Water-Sensitive Cards (WSC) Qualitatively assess spray coverage and droplet density Visualize spray patterns through color change reaction [5]
Artificial Foliage Simulate plant interception of spray droplets Measure deposit characteristics on realistic surface types [2]
Horizontal String Samplers Capture airborne spray droplets Assess potential inhalation exposure and airborne drift [2]
Meteorological Instrumentation Characterize environmental conditions during application Weather stations measuring wind speed/direction, temperature, RH, pressure [2] [5]
Fluorometer Quantify dye concentration on deposition samplers Precisely measure tracer deposition for comparison with model predictions [2]

Regulatory Implementation and Future Directions

AGDISP and its derivative models play increasingly important roles in regulatory decision-making processes. The EPA employs these models in pesticide registration reviews to assess potential drift risks to human health and the environment [1] [6]. Recent regulatory developments have emphasized shifting from Tier 1 (screening-level) to Tier 3 (higher-tier) modeling approaches in AgDRIFT, which incorporate more realistic application parameters specific to modern equipment and practices [6]. This transition provides more accurate risk assessments that reflect actual industry practices rather than conservative worst-case scenarios.

Future model development will likely focus on reducing uncertainties in current predictions, particularly regarding swath width and displacement parameters for drone applications [3] [4]. Additionally, integration with emerging regulatory frameworks such as the Endangered Species Act (ESA) consultation processes and the Vulnerable Species Pilot Project (VSPP) will require enhanced spatial modeling capabilities [6]. The ongoing validation of AGDISPpro for spray drone applications represents a critical step in ensuring that regulatory decisions for these emerging technologies reflect their actual environmental profiles rather than extrapolations from conventional application methods [3] [4]. As application technologies continue to evolve, AGDISP's mechanistic foundation provides a adaptable framework for predicting and mitigating spray drift across diverse application scenarios.

Lagrangian-based models represent a fundamental approach in fluid dynamics for simulating the transport and deposition of discrete particles, such as pesticide spray droplets. Unlike Eulerian methods that observe flow properties at fixed points in space, the Lagrangian framework tracks individual droplets or "parcels" as they move through time and space, providing a natural and intuitive method for predicting spray drift. This particle-tracking methodology is particularly well-suited for modeling pesticide application as it directly simulates the physical processes governing droplet motion, including gravitational settling, aerodynamic drag, and turbulent dispersion.

The AGDISP model (Agricultural DISPersal model), initially developed with support from NASA and subsequently enhanced over several decades, represents the most comprehensive implementation of Lagrangian mechanics for aerial spray applications [7] [8]. Its core formulation solves the equations of motion for each droplet within a stochastic framework, accounting for the complex interplay of aircraft wake vortices, atmospheric turbulence, and droplet evaporation [7]. This approach has proven sufficiently robust to form the basis for regulatory assessments conducted by agencies such as the U.S. Environmental Protection Agency [8].

Fundamental Mathematical Framework

Core Equations of Motion

The Lagrangian approach calculates droplet trajectories by solving Newton's second law of motion for each droplet, accounting for all significant external forces. The primary equation governing droplet motion is:

Equation 1: Lagrangian Particle Motion Equation

Where:

  • m_p = mass of the droplet
  • v_i = velocity component of the droplet
  • t = time
  • F_drag = aerodynamic drag force
  • F_gravity = gravitational force (including buoyancy effects)
  • F_virtual mass = force required to accelerate the surrounding fluid

The drag force, typically the dominant force after droplet release, is calculated as:

Equation 2: Aerodynamic Drag Force

Where:

  • C_D = drag coefficient (function of droplet Reynolds number)
  • d = droplet diameter
  • ρ = air density
  • u = air velocity vector
  • v = droplet velocity vector

The gravitational settling term accounts for the density difference between the droplet and the surrounding air:

Where m_f is the mass of the fluid displaced by the droplet and g_i is the gravitational acceleration component [9].

Turbulence Modeling

Atmospheric turbulence significantly influences spray dispersion, particularly in the downwind direction. Lagrangian models incorporate this effect through stochastic tracking methods, where turbulent velocity fluctuations are modeled as random processes based on local turbulence characteristics. The discrete form of the trajectory equation incorporates a random component:

Where u_i' represents the random velocity fluctuation derived from Gaussian probability distributions with variances based on turbulent kinetic energy [7] [10].

Table 1: Fundamental Forces in Lagrangian Spray Models

Force Component Mathematical Expression Physical Significance Dominant Influence Factors
Aerodynamic Drag F_drag = C_D * (πd²/8) * ρ * |u - v| * (u_i - v_i) Controls deceleration and cross-wind transport Droplet diameter, relative velocity, air density
Gravitational Settling F_gravity = (m_p - m_f) * g_i Determines vertical deposition rate Droplet mass, density difference (buoyancy)
Virtual Mass F_vm = C_vm * ρ * V_p * (du_i/dt - dv_i/dt) Accounts for fluid acceleration around droplet Acceleration rate, fluid density
Pressure Gradient F_p = -V_p * (∂p/∂x_i) Responds to atmospheric pressure fields Pressure spatial distribution

AGDISP Model Architecture and Implementation

Near-Field Wake Vortex Dynamics

A critical innovation in AGDISP is its detailed treatment of aircraft wake vortices, which dominate near-field droplet transport (typically within 0-800 meters downwind) [7] [8]. The model represents the aircraft wake as a pair of counter-rotating wing-tip vortices that entrain and transport droplets in their flow field. These vortices create a complex velocity field that redistributes droplets downward between the vortices and upward outside them, significantly influencing the initial deposition pattern [7].

The implementation in AGDISP uses an analytical vortex model coupled with the Lagrangian droplet tracking, enabling efficient computation of the vortex decay and its effect on droplet trajectories. This approach captures the essential physics observed in computational fluid dynamics (CFD) simulations, including the asymmetric droplet trajectories between port and starboard vortices in crosswind conditions [7].

Far-Field Gaussian Extension

For far-field drift predictions beyond 800 meters (up to 20 kilometers), AGDISP employs a Gaussian plume model extension [8]. This hybrid approach recognizes that while Lagrangian mechanics accurately captures near-field physics where individual droplet trajectories matter, the statistical approach of Gaussian models becomes more computationally efficient for long-range transport where turbulent diffusion dominates.

The Gaussian extension models the spray cloud concentration distribution assuming Gaussian distributions in both vertical and cross-wind directions, with dispersion parameters based on atmospheric stability classes [8] [11]. This linkage creates a comprehensive modeling system capable of predicting drift from the application site to ecologically relevant distances.

G AGDISP Model Architecture cluster_inputs Input Parameters cluster_core AGDISP Core Engine cluster_outputs Output Modules Aircraft Aircraft Configuration (wingspan, weight) Lagrangian Lagrangian Particle Tracking (Equations of Motion) Aircraft->Lagrangian Nozzles Nozzle Properties (type, position, flow rate) Nozzles->Lagrangian Droplets Droplet Spectrum (size distribution, velocity) Droplets->Lagrangian Atmosphere Atmospheric Conditions (wind, turbulence, stability) Atmosphere->Lagrangian WakeVortex Wake Vortex Model (tip vortices decay) Lagrangian->WakeVortex Evaporation Evaporation Module (droplet size reduction) WakeVortex->Evaporation NearField Near-Field Predictions (0-800 m deposition) Evaporation->NearField GaussianExt Gaussian Extension (far-field to 20 km) NearField->GaussianExt DriftRisk Drift Risk Assessment GaussianExt->DriftRisk

Experimental Protocols for Model Validation

Near-Field Deposition Measurement

Purpose: To validate AGDISP predictions of spray deposition patterns in the immediate application area and near-field drift (0-800 m).

Materials and Equipment:

  • Spray application equipment (aircraft or ground rig)
  • Petri dishes or chromatography paper deposition collectors
  • Meteorological station (wind speed, direction, temperature, humidity)
  • Laser diffraction droplet size analyzer
  • Fluorescent tracer dye and fluorometer
  • GPS unit for spatial referencing

Procedure:

  • Pre-Application Setup:
    • Arrange deposition collectors in a grid pattern extending from the application swath to 800 m downwind
    • Measure meteorological conditions at multiple heights (2 m, 10 m)
    • Characterize droplet spectrum using laser diffraction at release point

  • Application:

    • Prepare spray mixture with fluorescent tracer at known concentration
    • Conduct application at predetermined height, speed, and swath width
    • Record actual flight parameters using GPS

  • Post-Application:

    • Collect deposition samples at predetermined time intervals
    • Extract tracer from collectors using standardized solvent volume
    • Analyze tracer concentration using fluorometer
    • Calculate deposition mass per unit area

  • Data Analysis:

    • Compare measured deposition pattern with AGDISP predictions
    • Calculate statistical measures of agreement (R², RMSE)
    • Perform sensitivity analysis on critical input parameters [7] [3]

Far-Field Drift Validation

Purpose: To validate the Gaussian extension of AGDISP for long-range drift predictions (0.8-20 km).

Materials and Equipment:

  • High-volume air samplers with filter arrays
  • Rotorod samplers or similar aerosol collection devices
  • Tethered balloons or sodar for atmospheric profiling
  • Radioactive or stable isotope tracers (for sensitive detection)
  • Gamma spectrometers or mass spectrometers

Procedure:

  • Site Selection:
    • Choose relatively flat, uniform terrain to minimize topographic effects
    • Establish sampling lines at multiple downwind distances (1, 2, 5, 10, 20 km)

  • Atmospheric Characterization:

    • Measure vertical profiles of wind speed, direction, and temperature
    • Classify atmospheric stability using Pasquill-Gifford categories
    • Record continuous meteorological data throughout experiment

  • Tracer Application and Sampling:

    • Apply tracer material using standardized aerial application methods
    • Operate air samplers for predetermined intervals (typically 1-2 hours)
    • Collect aerosol samples at multiple heights where feasible

  • Laboratory Analysis:

    • Process filter samples using appropriate analytical methods
    • Quantify tracer mass collected at each sampling location
    • Calculate airborne concentration and ground deposition values

  • Model Comparison:

    • Run AGDISP with Gaussian extension using measured input parameters
    • Compare predicted and measured concentration/deposition patterns
    • Evaluate model performance across different stability classes [8]

Table 2: Key Parameters for AGDISP Model Validation Experiments

Parameter Category Specific Measurements Measurement Equipment Validation Metric
Droplet Characteristics Size distribution, velocity spectrum, evaporation rate Laser diffraction, high-speed photography, weighing Deposition spatial pattern, mass balance
Atmospheric Conditions Wind speed/direction profile, temperature, humidity, turbulence Sonic anemometers, weather stations, sodar Downwind drift profile, deposition curve
Application Parameters Release height, aircraft speed/swath, nozzle type/position GPS, flight data recorder, pressure gauges Near-field deposition, swath uniformity
Deposition/Concentration Ground deposition mass, airborne concentration Fluorometry, chromatography, mass spectrometry R², RMSE, agreement with predicted values

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 3: Essential Research Reagents and Materials for Lagrangian Spray Studies

Reagent/Material Function Application Protocol Key Characteristics
Fluorescent Tracers (e.g., Rhodamine WT, Fluorescein) Quantify deposition and drift through fluorescence detection Mix in spray tank at known concentration; analyze collectors with fluorometer High fluorescence yield, photostability, water solubility
Evaporation Retardants (e.g., glycerol, oil-based emulsions) Modify droplet evaporation rate for volatility studies Add to spray mixture at varying concentrations; compare evaporation rates Low vapor pressure, compatible with spray equipment
Tracer Microspheres (fluorescent or radioactive) Model droplet behavior without evaporation effects Suspend in spray mixture; detect using specialized instrumentation Narrow size distribution, density similar to spray droplets
Wind Tunnel Facilities Controlled environment for parameter isolation Conduct scaled experiments with reproducible turbulence conditions Adjustable wind speed, turbulence generation, sampling ports
Droplet Size Analyzers (laser diffraction) Characterize initial droplet spectrum Measure droplet size distribution at nozzle exit plane Wide dynamic range, rapid sampling capability
Computational Resources (HPC clusters) Execute complex Lagrangian simulations Run multiple scenarios for sensitivity analysis Parallel processing capability, large memory capacity

Advanced Computational Implementation

Lagrangian Stochastic Modeling

The AGDISP model implements a sophisticated Lagrangian stochastic approach that goes beyond simple deterministic trajectory calculations. Each droplet's motion is represented as a Markov process, where velocity fluctuations are modeled based on local turbulence statistics:

Equation 3: Stochastic Differential Equation for Particle Velocity

Where:

  • a_i represents the deterministic acceleration component
  • b_ij defines the diffusion tensor related to turbulent kinetic energy
  • dW_j is an incremental Wiener process representing random forcing

This formulation allows the model to simulate the random, diffusive nature of turbulent transport while maintaining computational efficiency through its analytical framework [7] [11].

Coupled Evaporation Physics

Droplet size changes due to evaporation significantly impact drift potential, as smaller droplets remain airborne longer. AGDISP incorporates evaporation using a coupled mass and heat transfer model:

Equation 4: Droplet Evaporation Rate

Where:

  • Sh is the Sherwood number (mass transfer correlation)
  • D_ab is the vapor diffusivity in air
  • ρ_v,s is vapor density at droplet surface
  • ρ_v,∞ is vapor density in surrounding air

The model accounts for the heat and mass transfer analogy, with the evaporation rate being strongly influenced by ambient relative humidity and temperature [7]. This physics-based approach represents a significant advancement over simpler empirical evaporation correlations.

G Near-Field Lagrangian Physics WakeVortex Aircraft Wake Vortices (downwash, tip vortices) Trajectory Droplet Trajectory (time-integrated path) WakeVortex->Trajectory DragForce Aerodynamic Drag (velocity-dependent) DragForce->Trajectory Gravity Gravitational Settling (terminal velocity) Gravity->Trajectory Turbulence Atmospheric Turbulence (stochastic component) Turbulence->Trajectory Evaporation Droplet Evaporation (size reduction) Evaporation->DragForce DropletRelease Droplet Release (initial conditions) DropletRelease->WakeVortex DropletRelease->DragForce DropletRelease->Gravity Deposition Deposition/Drift (final fate) Trajectory->Deposition

Recent Advances and Future Directions

Hybrid Modeling Approaches

Recent research has focused on hybrid Lagrangian-dispersion models that combine the strengths of mechanistic particle tracking with computationally efficient statistical methods. These approaches use Lagrangian mechanics in the immediate release area where aircraft wake effects dominate, then transition to Gaussian plume models for far-field transport [11]. This strategy maintains physical accuracy where needed while enabling rapid simulation of long-range drift scenarios.

The implementation in modern AGDISP versions includes an automated transition between these modules, with the handoff typically occurring 100-200 meters downwind where the aircraft wake has largely dissipated and statistical approaches become appropriate [8] [11].

Incorporation of Canopy Interactions

Advanced implementations of Lagrangian models now include sophisticated treatments of canopy interactions, representing vegetation as a porous medium that modifies both airflow and deposition patterns. This approach introduces additional momentum sink terms in the flow equations and collection efficiency models for droplet impaction on plant surfaces [9].

The canopy model typically employs a collection efficiency parameter that depends on droplet inertia, local turbulence, and vegetative surface characteristics:

Where St is the Stokes number (ratio of droplet stopping distance to characteristic canopy dimension), Re is the Reynolds number, and ϕ represents canopy porosity [9].

These advances have significantly improved the model's capability to predict deposition patterns in complex agricultural environments, particularly for orchard and vineyard applications where canopy structure dramatically influences spray distribution.

Accurately predicting pesticide deposition is critical for assessing efficacy and environmental risk. AGDISP and its derivatives are the primary mechanistic models used by regulatory bodies and researchers to simulate the transport and fate of spray droplets [1] [12]. This protocol details the key numerical outputs of these models and the experimental methods for their validation, with a specific focus on distinguishing between in-swath deposition, which relates to application efficacy, and off-target deposition, which relates to environmental exposure and risk [13].

Key Quantitative Model Outputs

AGDISP and its associated models generate quantitative predictions for spray deposition across the application area and downwind. The table below summarizes the primary model outputs used in risk assessment.

Table 1: Key Deposition Outputs from Spray Drift Models

Model Output Spatial Scale Typical Units Risk Assessment Relevance
In-Swath Deposition Within application area µg/cm², mL/cm², % of applied Efficacy assessment; estimated dose to target crop or pest [14].
Near-Field Off-Target Deposition 0 - 800 meters downwind [8] µg/cm², % of applied rate Ecological risk; exposure to non-target organisms in adjacent habitats [15] [13].
Far-Field Drift (Gaussian Extension) Up to 20 km downwind [8] µg/cm², % of applied rate Long-range exposure; used conservatively for aerial applications with fine droplets [8].
Airborne Concentration Above ground level (specific heights) µg/m³, mg/m³ Inhalation risk for humans and non-target organisms [1].
Drift Potential (Index/Value) Application-specific Unitless or % Comparative assessment of different application scenarios or technologies [14].

Deposition Values by Application Method and Distance

Field and model validation studies provide typical ranges for off-target deposition. The following table consolidates quantitative data from recent research, particularly for emerging application methods like Unmanned Aerial Spray Systems (UAS).

Table 2: Exemplary Off-Target Deposition Ranges from Validation Studies

Application Method Downwind Distance Observed Deposition (% of Applied Rate) Key Influencing Factors Source/Context
Orchard Ground Sprayer 10-20 meters ~0.5 - 2% Sprayer type, wind speed, nozzle design [14]. Comparative study with drone sprayers [14].
Drone (UAS) - Medium Coarse Spray 10-20 meters ~0.1 - 1.5% (Model IA: 0.47-0.92) [3] [4] Flight speed, flight height, rotor downwash [14] [16]. AGDISPpro validation; index of agreement with field data shown [4].
Drone (UAS) - Extremely Coarse Spray 10-20 meters ~0.05 - 1% (Model IA: 0.61-0.94) [3] [4] Droplet size, swath width, wind speed [3]. AGDISPpro validation; coarser droplets reduce drift [4].
Aerial (Fixed-wing) 100+ meters Varies widely (e.g., <0.1%) Application height, droplet size, meteorology (esp. evaporation) [12]. AgDRIFT model evaluation; sensitive to evaporative conditions [12].

Experimental Protocols for Model Validation

Validating model outputs requires rigorous field experiments designed to measure deposition directly. The following protocol is adapted from standardized methodologies and recent research on UAS applications [3] [14] [16].

Field Experiment Setup for Deposition and Drift Measurement

Objective: To collect empirical data on in-swath and off-target deposition for comparison with AGDISP model predictions.

Materials:

  • Tracer substance (e.g., fluorescent dye like Brilliant Sulfoflavine or similar)
  • Application equipment (e.g., drone, ground sprayer)
  • Deposition collectors (e.g., filter papers, petri dishes, Mylar sheets)
  • Air samplers (for airborne drift assessment)
  • Portable weather station
  • Fluorometer or spectrophotometer for analysis

Procedure:

  • Pre-Application Setup:

    • Lay out collector lines extending from within the application swath to at least 50 meters downwind. Place collectors on the ground and, for orchard/canopy studies, at different heights within the plant canopy [14].
    • Position a meteorological station to record wind speed, wind direction, temperature, and relative humidity at 2-meter height throughout the trial [3].
    • Prepare tracer solution at a known concentration.

  • Application:

    • Apply the tracer solution using the defined operational parameters. Critically record:
      • Application rate (L/ha)
      • Droplet size spectrum (DV10, DV50, DV90), measured experimentally [13]
      • Nozzle type and pressure
      • For drones/UAS: Flight height, flight speed, swath width [3] [4]
      • For ground sprayers: Spray pressure, boom height, travel speed

  • Post-Application Sample Collection:

    • Collect deposition samples from all collectors after application.
    • Collect air sampler filters if used.
    • Document the exact location of each sample for spatial analysis.

  • Sample Analysis:

    • Extract the tracer from each collector into a known volume of deionized water.
    • Measure the fluorescence intensity of each sample using a fluorometer.
    • Convert fluorescence readings to tracer concentration using a calibration curve.
    • Calculate deposition values (e.g., µL/cm² or ng/cm²) and normalize to the applied rate for comparison.

Data Analysis and Model Comparison

  • Data Compilation: Compile field data into a dataset of deposition versus downwind distance.
  • Model Simulation: Run AGDISP or AGDISPpro using the recorded application and meteorological parameters as input.
  • Statistical Comparison: Compare model predictions against field observations using statistical metrics such as:
    • Index of Agreement (IA): A measure of model prediction error. Values range from 0 (no agreement) to 1 (perfect agreement). Recent AGDISPpro validation for UAS showed IA values from 0.47 to 0.94 depending on spray quality [3] [4].
    • Mean Bias Error (MBE): Indicates a tendency for under- or over-prediction. AGDISPpro has shown a tendency to slightly under-predict off-target deposition for UAS (negative MBE) [4].
    • Visual comparison of predicted and observed deposition curves.

AGDISP Modeling Workflow

The following diagram illustrates the logical workflow for using AGDISP to predict deposition, from input definition to risk assessment.

G Start Start: Define Scenario Inputs Input Parameters Start->Inputs P1 Application Method (Aerial, Ground, UAS) Inputs->P1 P2 Spray Material Properties (Formulation, Density) Inputs->P2 P3 Nozzle & Droplet Spectrum (DV50, Classification) Inputs->P3 P4 Meteorological Conditions (Wind, Temp, Humidity) Inputs->P4 P5 Operational Parameters (Height, Speed, Swath Width) Inputs->P5 Model Run AGDISP/AGDISPpro P1->Model P2->Model P3->Model P4->Model P5->Model Outputs Key Model Outputs Model->Outputs O1 In-Swath Deposition Outputs->O1 O2 Off-Target Deposition Outputs->O2 O3 Airborne Concentration Outputs->O3 O4 Drift Potential Value Outputs->O4 Use Application & Risk Assessment O1->Use O2->Use O3->Use O4->Use U1 Efficacy Evaluation Use->U1 U2 Ecological Risk Assessment Use->U2 U3 Buffer Zone Determination Use->U3 U4 Label Development Use->U4

AGDISP Modeling and Risk Assessment Workflow

The Scientist's Toolkit: Research Reagent Solutions

The following table lists essential materials and their functions for conducting spray deposition studies, based on cited methodologies.

Table 3: Essential Research Reagents and Materials for Deposition Studies

Item Category Specific Examples Function in Experiment
Tracer Substances Fluorescent dyes (e.g., Brilliant Sulfoflavine, Tartrazine, Rhodamine WT) [14] Acts as a surrogate for pesticide; allows for quantitative analysis of deposition without using active ingredients.
Deposition Collectors Filter papers, Petri dishes, Mylar sheets, water-sensitive papers [16] Provides a standard surface for capturing spray droplets at various locations for subsequent analysis.
Airborne Drift Samplers Rotary impactors, filter-based air samplers Captures fine, inhalable droplets that remain airborne, allowing for assessment of inhalation exposure risk.
Droplet Size Analyzers Laser diffraction systems (e.g., Malvern Spraytec), optical array probes Measures the droplet size spectrum (DV10, DV50, DV90) of the spray release, a critical input for drift models.
Analytical Equipment Fluorometer, spectrophotometer Precisely quantifies the amount of tracer recovered from collectors and air samplers.
Meteorological Stations Portable stations with anemometer, thermohygrometer, wind vane Records essential environmental data required for model input and interpretation of field results.

AGDISP (AGricultural DISPersion) is a science-based, "first-principles" model that predicts spray drift from pesticide application sites. Developed by the U.S. Forest Service in the 1980s, it serves as a critical tool for the U.S. Environmental Protection Agency (EPA) in assessing ecological and human health risks from off-target pesticide movement [1] [17]. The model's primary regulatory function is to estimate the deposition patterns of pesticides released into the atmosphere during agricultural spraying operations, providing a quantitative basis for exposure characterization in risk assessments [1] [13]. As a component of the EPA's analysis phase, AGDISP helps describe exposure in terms of intensity, space, and time, ultimately contributing to exposure profiles that inform risk management decisions [13].

Within the EPA's modeling framework, AGDISP exists in two primary forms: the core AGDISP model (Version 8.26) and AgDRIFT (Version 2.1.1), a modified version created through a cooperative agreement between the EPA, the US Department of Agriculture's Forest Service, and the Spray Drift Task Force [1]. The model's capability to assess a variety of spray drift conditions from agricultural applications makes it indispensable for estimating downwind deposition from aerial, ground boom, and orchard/vineyard airblast applications [1]. The regulatory reliance on AGDISP stems from its detailed algorithms for characterizing the release, dispersion, and deposition of spray materials over and downwind of application areas, providing a scientifically defensible method for predicting environmental fate.

AGDISP operates as a Lagrangian model, simulating individual droplet trajectories based on aerodynamic, inertial, and gravity forces [18]. This approach allows the model to optimize agricultural spraying operations by predicting spray drift from application sites with consideration of detailed release conditions [1]. The model's architecture incorporates key processes including droplet emission, atmospheric dispersion, evaporation, and deposition onto canopy and ground surfaces [18]. A significant strength of the Lagrangian approach is its capacity to accommodate complex atmospheric conditions, precise source characteristics, and evaporation processes while accurately representing near-field diffusion where dispersion deviates from standard diffusion laws [18].

The current AGDISP version represents a substantial modernization initiative known as the AGDISP Modernization Project (AMP), established by the National Agricultural Aviation Association (NAAA) to address the model's outdated coding foundation [17] [19]. This multi-stakeholder effort aims to rewrite the original 1980s-era code using modern, well-supported computer languages to enable significant improvements in model accuracy and adaptability [17]. With a budget of $600,000 over five years, the project has secured $335,000 in funding to date from contributors including The Cotton Foundation, the Centers for Disease Control (via the American Mosquito Control Association), and the National Agricultural Aviation Research and Education Foundation [17]. The modernization has gained broader agricultural industry support, with the National Corn Growers Association recently joining the initiative [20].

Table 1: Key Features of AGDISP Model Versions

Model Version Core Characteristics Primary Applications Regulatory Status
AGDISP (Version 8.26) "First-principles" science-based model; predicts spray drift from application sites [1] Aerial and ground boom applications; forestry and adulticide/mosquitocide applications [1] Used by EPA for spray drift assessment [1]
AgDRIFT (Version 2.1.1) Modified version of AGDISP; developed cooperatively with EPA and Spray Drift Task Force [1] Assesses variety of spray drift conditions from agricultural applications [1] Used by EPA for spray drift assessment [1]
AGDISPpro Recently developed variant adapted for unmanned aerial systems (UAS) [4] Spray applications using unmanned aerial systems (drones) [4] Under validation for regulatory use [4]
Modernized AGDISP (Under Development) Rewritten using modern programming languages; open source [17] All application types (aerial, ground, unmanned aerial) with improved accuracy [17] Future intended replacement for current versions [17]

The modernization project promises substantial advancements in regulatory modeling capabilities. A key feature of the modernized version is that it will remain open source, ensuring continued access for both EPA risk assessments and private sector technology development [17] [19]. The updated framework will better accommodate drift reduction technologies, potentially resulting in less restrictive application requirements on pesticide labels when mitigations are demonstrated [17]. Furthermore, the modernized AGDISP is expected to set the stage for real-time, site-specific risk assessments that integrate meteorological data, digital labels, adjuvant information, and application equipment parameters [17].

AGDISP in EPA Regulatory Framework and Risk Assessment Process

The EPA incorporates AGDISP into its pesticide risk assessment framework through well-defined protocols that address both near-field and far-field spray drift. The near-field modeling component of AGDISP predicts spray drift up to 800 meters (0.5 miles) downwind from the application site, while the Gaussian Extension model can predict pesticide spray drift up to 20 kilometers (12.8 miles) downwind [8]. This linked AGDISP-Gaussian approach is specifically recommended for aerially applied mosquito adulticides or those pesticides aerially applied using very fine to fine droplet size spectra, with a minimum pesticide release height of 50 feet above ground level [8]. These specific use scenarios align most closely with the parameters that have been validated and minimize uncertainty relative to potential physical barriers [8].

The EPA's guidance acknowledges significant uncertainty in estimating pesticide deposition at far-field distances beyond 0.5 miles, noting that the current modeling approach is conservative and doesn't account for physical barriers or complex meteorological conditions beyond the application site [8]. Validation studies have demonstrated reasonable agreement between predicted concentrations from AGDISP-Gaussian Extension models and observed pesticide deposition data from 0.1 km to 10 km under conditions of level terrain with minimal ground cover [8]. However, the guidance specifically restricts the use of the Gaussian Extension model for ground spray applications, reflecting limitations in validation for these scenarios [8].

Table 2: EPA Guidance on AGDISP-Gaussian Extension Model Application

Parameter Specification Rationale References
Near-Field Distance Up to 800 meters (0.5 miles) Standard validation range for Lagrangian model [8]
Far-Field Distance Up to 20 kilometers (12.8 miles) Maximum extension using Gaussian plume model [8]
Recommended Uses Aerial mosquito adulticides; fine droplet spectra; minimum 50 ft release height Closely matches validation parameters [8]
Restricted Uses Ground spray applications Lack of appropriate validation [8]
Validation Limit 2 miles for full model confidence Limited validation beyond this distance [8]
Key Uncertainty Factors Physical barriers; topographic features; crosswinds; humidity Not accounted for in current far-field modeling [8]

The regulatory implications of AGDISP modeling are substantial, as the drift estimates generated directly influence risk mitigation measures specified on pesticide labels, including buffer zones and application restrictions [17]. Without accurate spray drift risk assessments for aerial, ground, and airblast applications, growers face the potential loss of access to pesticides critical for protecting crops [17] [19]. The modernized AGDISP is expected to more accurately estimate off-target spray movement for all pesticide application types during EPA's ecological, endangered species, and human health risk assessments, potentially leading to more flexible application requirements that reflect actual rather than conservatively estimated drift potential [17].

Experimental Protocols and Validation Methodologies

AGDISPpro Validation for Unmanned Aerial Systems

Recent research has focused on validating AGDISP derivatives for emerging application technologies, particularly unmanned aerial systems (UAS). A 2025 field study conducted rigorous validation of AGDISPpro, a recently developed UAS drift model adapted from AGDISP by incorporating aerodynamical models of commercial UAS [4]. The experimental protocol involved two distinct field studies with varied spray quality, meteorological conditions, and UAS operational factors. Study No. 1 employed single-swath sprays using two different agricultural spray nozzles producing Medium and Extremely Coarse spray droplet spectra. Study No. 2 implemented four-swath sprays using Fine and Ultra Coarse spray nozzles to evaluate model performance under more complex scenarios [4].

The validation methodology centered on comparing ground deposition results predicted by AGDISPpro with in-swath and off-target downwind field deposition measurements from the two field experiments. Statistical analysis of these comparisons employed metrics including index of agreement and mean bias error to quantify model performance [4]. For Study No. 1, the index of agreement between model predictions and field observations for off-target drift ranged from 0.47 to 0.92 for Medium spray droplet spectra and from 0.61 to 0.94 for Extremely Coarse sprays. The research identified uncertainty in UAS application swath width and swath displacement as significant factors affecting model accuracy, with sensitivity analysis demonstrating that these parameters greatly influence the magnitude of maximum peak deposition [4].

Comparative Modeling Approaches in Regulatory Science

While AGDISP represents a cornerstone of EPA's spray drift modeling infrastructure, alternative modeling approaches exist with varying strengths and applications. The ADDI-Spraydrift model, for instance, employs a random walk approach that describes droplet emission, dispersion, evaporation, ground deposition, and canopy interception while accounting for atmospheric stability regimes and in-canopy turbulence [18]. This comprehensive model has demonstrated strong performance in vineyard applications, showing a mean deviation between modeled and measured deposition of 1.3% when calibrated and evaluated against the Ganzelmeier sedimentary spray drift database [18].

Other modeling approaches include Eulerian models that resolve the mass conservation equation in turbulent flow and Gaussian models based on analytical solutions of the advection-diffusion approach [18]. Each methodology presents distinct advantages: Lagrangian approaches (including AGDISP) excel at handling complex atmospheric conditions and precise source characteristics; Gaussian models offer computational efficiency; while Eulerian approaches can provide high fidelity at small spatial scales [18]. The selection of appropriate models for specific regulatory contexts depends on factors including spatial scale, application method, available computational resources, and required output specificity.

G AGDISP Regulatory Risk Assessment Workflow cluster_model AGDISP Modeling Core Application Application NearField NearField Application->NearField Equipment Equipment Equipment->NearField Environment Environment Environment->NearField Formulation Formulation Formulation->NearField GaussianExtension GaussianExtension NearField->GaussianExtension > 800m Ecological Ecological NearField->Ecological HumanHealth HumanHealth NearField->HumanHealth EndangeredSpecies EndangeredSpecies NearField->EndangeredSpecies GaussianExtension->Ecological GaussianExtension->HumanHealth GaussianExtension->EndangeredSpecies

Research Reagent Solutions and Essential Materials

Conducting AGDISP-related research and validation requires specific technical resources and analytical tools. The following table outlines critical components of the researcher's toolkit for spray drift modeling and experimental validation.

Table 3: Essential Research Materials and Tools for Spray Drift Studies

Tool/Resource Function Application Context References
AGDISP Software Predicts spray drift from application sites using Lagrangian approach Core modeling tool for EPA risk assessments; available in multiple versions [1] [4]
Droplet Size Spectrometry Characterizes droplet size distribution from application equipment Required input parameter for model initialization; affects drift potential [13] [4]
Meteorological Stations Measures wind speed, temperature, relative humidity, atmospheric stability Critical environmental input parameters for model scenarios [17] [18]
Spray Deposition Samplers Collects physical droplets for quantification of off-target movement Model validation against experimental data in field studies [4] [18]
Digital Label Integration Incorporates application-specific restrictions and requirements Emerging capability for real-time, site-specific risk assessment [17]
Drift Reduction Technology Physical systems to minimize off-target movement (e.g., nozzles, shields) Evaluation of mitigation effectiveness in model scenarios [17]

AGDISP represents a critical computational tool within the EPA's pesticide regulatory framework, providing scientifically defensible estimates of spray drift that inform risk management decisions. The ongoing modernization initiative promises to address current limitations while enhancing model accuracy and adaptability to new application technologies. The planned improvements will better accommodate drift reduction technologies in regulatory decisions and potentially enable real-time risk assessment capabilities that reflect specific application conditions [17]. For researchers and regulatory professionals, understanding AGDISP's capabilities, validation boundaries, and proper implementation protocols remains essential for generating reliable data that supports scientifically sound pesticide regulation and sustainable agricultural practices.

The future regulatory landscape will likely see increased sophistication in spray drift modeling as AGDISP evolves to incorporate more complex environmental variables, better account for novel application technologies like UAS, and provide more nuanced risk assessments that balance agricultural productivity with environmental protection. The open-source nature of the modernized AGDISP will further facilitate broader scientific engagement and continuous model improvement through collaborative development across the research community [17] [20].

Within the regulatory and research framework for pesticide spray drift prediction, AGDISP and AgDRIFT represent critical, yet distinct, computational tools. The United States Environmental Protection Agency (EPA) relies on these models to assess ecological and human health risks from pesticide exposure, influencing label development and use restrictions [1]. A clear understanding of their different versions, underlying purposes, and operational contexts is essential for researchers and scientists conducting robust environmental fate and transport studies. This document delineates the technical specifications, validated applications, and appropriate use scenarios for the various iterations of these models to support accurate and defensible research protocols.

AGDISP (AGricultural DISPersal) is a mechanistic, Lagrangian model that serves as the foundational computational engine for predicting the off-target movement of spray material [21]. Originally developed by the USDA Forest Service in the 1980s, it is a "first-principles" science-based model that simulates the release, dispersion, and deposition of spray from application sites [1] [17].

AgDRIFT is a regulatory-focused model that, while embodying the core AGDISP algorithms, is structured as a simplified screening-level tool [21]. It is designed for speed and ease of use in tiered risk assessments, particularly for estimating near-field buffer zones required to manage exposures [22].

The relationship between these models and their versions can be visualized as an evolutionary tree.

G cluster_legacy Core Model Lineage cluster_modern Modern & Future Development AGDISP (v8.26) AGDISP (v8.26) AGDISPpro AGDISPpro AGDISP (v8.26)->AGDISPpro Extended Version AGDISP Modernization\nProject (AMP) AGDISP Modernization Project (AMP) AGDISP (v8.26)->AGDISP Modernization\nProject (AMP) Future Version AgDRIFT (v2.1.1) AgDRIFT (v2.1.1) AGDISP (v8.26)->AgDRIFT (v2.1.1) Modified Version AGDISP (v8.26)->AgDRIFT (v2.1.1) Higher-Level Modeling Higher-Level Modeling AGDISP (v8.26)->Higher-Level Modeling RPAAS (Drone) Apps RPAAS (Drone) Apps AGDISPpro->RPAAS (Drone) Apps Tier 1 Screening Tier 1 Screening AgDRIFT (v2.1.1)->Tier 1 Screening

Model Versions and Technical Specifications

Core Model Specifications and Use Cases

The following table summarizes the key characteristics of the primary model versions and their intended applications.

Table 1: Comparative Overview of AGDISP and AgDRIFT Model Versions

Model & Version Model Type & Core Function Primary Application Scenarios Spatial Range Regulatory Tier Analogy
AGDISP (v8.26) Mechanistic, Lagrangian model; predicts spray drift from aerial, ground boom, and forestry applications [1] [8]. Detailed scenario modeling for aerial and ground boom applications; optimization of spraying operations [1]. Near-field up to 800m (0.5 miles) [8]. Higher-level, refined assessment.
AgDRIFT (v2.1.1) Modified, screening-level version of AGDISP; estimates drift for risk assessment [1] [21]. Tier 1 screening for ground, aerial, and orchard/vineyard applications to set buffer zones [1] [22]. Near-field; validated for buffer zone estimation [22]. Tier 1, conservative screening.
AGDISP w/ Gaussian Extension AGDISP core for near-wake + Gaussian plume model for far-field [8]. Aerially applied mosquito adulticides or fine sprays to forests/orchards (min. 50 ft release height) [8]. Up to 20 km (12.4 miles) [8]. Specialized far-field assessment.
AGDISPpro Modernized extension of AGDISP for Remotely Piloted Aerial Application Systems (RPAAS) [3]. Modeling off-target drift from drone (e.g., PV22 quadcopter, PV35X hexacopter) pesticide applications [3]. Near-field; validation studies ongoing [3]. Emerging technology assessment.

Key Model Limitations and Validation Status

  • AGDISP/AgDRIFT Validation: The AgDRIFT model (v2.1.1) has been evaluated against field trial data. It shows a similar response to field observations for key application variables like droplet size and wind speed. However, it can overpredict deposition rates in far-field distances, particularly under evaporative conditions, making it conservative for near-field buffer zone estimation [22].
  • Gaussian Extension Uncertainty: Significant uncertainty exists for estimates beyond 0.5 miles. The model does not account for physical barriers (e.g., trees, topography) or meteorological variables like crosswinds and humidity at far-field distances, resulting in extremely conservative predictions [8]. The model's far-field predictions have limited validation beyond 2 miles [8].
  • Ongoing Modernization: The AGDISP Modernization Project (AMP), established in 2023, aims to rewrite the legacy code into a modern programming language. This effort seeks to improve accuracy, incorporate Drift Reduction Technologies (DRTs), and ensure all application types (aerial, ground, UAS) are better represented in future EPA risk assessments [17] [19] [23].

Experimental Protocols for Model Validation and Application

Adhering to standardized protocols is fundamental for generating data suitable for model validation or regulatory submission.

Protocol for Ground and Aerial Spray Drift Field Experiments

This protocol is adapted from recent peer-reviewed research evaluating spray drift from ground and aerial applications [5].

1. Objective: To measure physical spray drift and compare off-target movement from ground and aerial application equipment under field conditions.

2. Experimental Design:

  • Treatments: Application method (Ground vs. Aerial) as the main factor.
  • Plot Layout: Establish a linear transect of collection stations at set distances downwind from the application area (e.g., 3.0, 7.6, 15.2, 30.5, 61 m).
  • Replication: A minimum of three experimental replications is recommended.

3. Application Specifications:

  • Tracer: Use a commercially available herbicide like florpyrauxifen-benzyl.
  • Droplet Spectrum: Standardize both systems to emit a Coarse spray quality.
  • Ground Rig: Boom height at 0.5 m.
  • Aircraft: Incorporate a one full swath width adjustment upwind of the primary collection line to simulate standard practice and mitigate extreme drift [5].

4. Data Collection:

  • Meteorological Monitoring: Record air temperature, relative humidity, wind speed, and wind direction at frequent intervals (e.g., every 5 seconds) throughout the trial. Trials should be conducted when wind direction is within ±30° of the collection line [5].
  • Spray Deposition:
    • Mylar Cards: Place at each downwind distance and at pre-determined intervals within the treated area. Analyze for tracer deposition (e.g., µg cm⁻²) and express as a percentage of the applied rate.
    • Water Sensitive Cards (WSC): Place alongside Mylar cards. Analyze for percent coverage and number of deposits per cm².

5. Data Analysis:

  • Fit deposition data versus downwind distance using regression (e.g., four-parameter log-logistic).
  • Calculate predicted distances for 25%, 50%, and 90% reduction in deposition (PD₂₅, PD₅₀, PD₉₀).
  • Compare observed deposition with model predictions from AgDISP or AgDRIFT.

Protocol for Validating Drone Application Models

This protocol is based on the validation framework for AGDISPpro [3].

1. Objective: To validate the AGDISPpro model for predicting off-target spray drift from Remotely Piloted Aerial Application Systems (RPAAS).

2. Experimental Design:

  • RPAAS Models: Select specific models available in AGDISPpro (e.g., PV22 quadcopter, PV35X hexacopter).
  • Spray Nozzles: Conduct trials with a range of nozzle types to generate different spray qualities (e.g., Fine, Medium, Extremely Coarse, Ultra Coarse).
  • Swath Pattern: Perform both single-swath and multiple-swath (e.g., four-swath) applications.

3. Data Collection:

  • In-Swath Deposition: Place collectors across the swath to characterize the deposition pattern and determine effective swath width.
  • Off-Target Deposition: Place collectors at multiple distances downwind from the application swath.
  • Operational Parameters: Precisely record flight altitude, forward speed, and swath displacement.

4. Model Input and Evaluation:

  • Inputs: Enter all recorded operational and meteorological data into AGDISPpro.
  • Validation Metric: Use the index of agreement to quantify the goodness-of-fit between predicted and observed in-swath and downwind deposition values. An index of agreement above 0.8 generally indicates strong predictive performance [3].

The Researcher's Toolkit: Essential Reagents and Materials

Table 2: Key Research Reagents and Materials for Spray Drift Experiments

Item Specification/Function
Spray Tracer A detectable pesticide or surrogate (e.g., florpyrauxifen-benzyl [5]). Must be analytically quantifiable at low deposition levels.
Deposition Targets Mylar cards or filter paper for quantitative chemical analysis of tracer deposition [5].
Droplet Characterization Water Sensitive Cards (WSC): Coated with yellow dye that dissolves upon contact with water droplets, allowing for measurement of coverage (%) and droplet density [5].
Meteorological Station Instrumentation to log wind speed, wind direction, temperature, and relative humidity at high frequency (e.g., 1-5 second intervals) [5].
Droplet Size Analyzer Laser diffraction-based system (e.g., Malvern Spraytec) to characterize the Droplet Size Spectrum (DV₀.₁, DV₀.₅, DV₀.₉) of the released spray.
Analytical Equipment High-Performance Liquid Chromatograph (HPLC) or equivalent, with appropriate detection limits, for quantifying tracer residue on deposition targets.

AGDISP and AgDRIFT form an integrated yet hierarchically structured modeling system for pesticide spray drift prediction. AGDISP serves as the detailed, mechanistic core model, while AgDRIFT provides a streamlined interface for conservative, screening-level assessments. The ongoing AGDISP Modernization Project promises to enhance the accuracy and scope of these critical tools by incorporating modern technologies and accounting for advanced drift reduction strategies. Researchers must select the model and version appropriate to their specific application scenario—be it Tier 1 screening with AgDRIFT, detailed scenario modeling with AGDISP, or specialized assessment for drone applications with AGDISPpro—while strictly adhering to their respective validated domains and inherent limitations.

Applying AGDISP: A Step-by-Step Guide for Different Scenarios

Accurately predicting pesticide spray drift is essential for mitigating environmental risks and ensuring application efficacy. The AGDISP model, a leading tool for simulating spray dispersion, relies heavily on precise input parameters to generate reliable outputs. Within the context of aerial application research, three parameter categories are paramount: nozzle type, droplet spectra, and meteorological data. These inputs directly govern the transport and deposition of spray droplets, influencing both on-target effectiveness and off-target environmental impact [8] [24]. This document provides detailed application notes and experimental protocols for the characterization and utilization of these critical parameters within the AGDISP modeling framework, supporting robust and predictive spray drift research.

Parameter Definition and AGDISP Input Requirements

The following section details the core parameters, their scientific basis, and their implementation within AGDISP.

Nozzle Type and Droplet Spectra

The nozzle is the primary determinant of the initial droplet size distribution, or droplet spectrum, which is arguably the most critical factor influencing drift potential [24] [5].

2.1.1 Droplet Size Classification Droplet spectra are standardized by the American Society of Agricultural and Biological Engineers (ASABE). This classification provides a common language for referencing spray quality in modeling and label recommendations [24] [25].

Table 1: ASABE S572.1 Droplet Size Classification (Simplified)

Spray Quality Class Volume Median Diameter (VMD - µm) Typical Drift Potential
Very Fine (VF) < 136 Extremely High
Fine (F) 136 - 177 High
Medium (M) 178 - 218 Medium
Coarse (C) 219 - 326 Low
Very Coarse (VC) 327 - 403 Very Low
Extremely Coarse (XC) > 403 Extremely Low

Research consistently shows that finer droplets result in significantly greater drift. For instance, aerial applications emitting Fine to Medium droplets can result in off-target drift of 12.5% of the applied rate to an adjacent water body without mitigation [25]. Switching to coarser droplet classes is a primary drift reduction strategy.

2.1.2 AGDISP Input Protocol In AGDISP, the nozzle type is selected to define the initial droplet spectrum. The model contains a library of common nozzles and their associated droplet size distributions. The user must select the nozzle that corresponds to the physical hardware used in the application or simulated scenario. For example, the EPA guidance specifies the use of the linked AGDISP-Gaussian model for scenarios involving "ASAE droplet size spectrum of very fine to fine" for certain aerial applications [8].

Meteorological Data

Meteorological conditions at the time of application exert a powerful influence on droplet trajectory, evaporation, and final deposition [25].

2.2.1 Critical Meteorological Parameters

  • Wind Speed: This is the primary driver of downwind droplet transport. Higher wind speeds increase drift potential exponentially. USEPA Tier II assessments often use a default wind speed of 4.47 m/s (10 mph) as a conservative assumption [25]. Real-world data shows that considering variable wind speeds can reduce estimated drift loads by an average of 63% compared to this constant default [25].
  • Wind Direction: This parameter determines the trajectory of the spray plume relative to sensitive off-target areas. Regulatory models often assume the wind is blowing directly toward the sensitive area for maximum conservatism [25].
  • Temperature and Relative Humidity (RH): These factors govern droplet evaporation. Higher temperatures and lower RH accelerate evaporation, causing droplets to shrink in size. Smaller, evaporated droplets have a lower settling velocity and remain airborne longer, increasing drift potential [25]. A common default assumption is 30°C (86°F) and 50% RH [25].
  • Atmospheric Stability: This influences turbulence and vertical mixing of the spray plume, affecting its dispersion and deposition pattern.

2.2.2 AGDISP Input Protocol AGDISP requires real-time, high-quality meteorological data for accurate simulations. Inputs typically include:

  • Wind Speed (m/s or mph)
  • Wind Direction (degrees)
  • Air Temperature (°C or °F)
  • Relative Humidity (%)
  • Atmospheric Stability Class

The model uses this data to simulate the micro-meteorological conditions affecting the spray cloud. Using real, hourly meteorological data instead of worst-case defaults is a key method for refining risk assessments [25].

Additional Critical Parameters

While the core three are the focus, other parameters significantly impact AGDISP outputs:

  • Release Height: The height of the spray boom above the crop canopy or ground is critical. Doubling the release height can lead to a three-fold or greater increase in downwind drift [5]. Aerial applications typically have higher release heights (e.g., ~50 feet) than ground rigs, contributing to their higher drift potential [8] [5].
  • Application Speed and Swath Width: These define the application rate and spatial pattern.

The following diagram illustrates the logical relationships and workflows between these critical input parameters and their collective impact on the AGDISP model's prediction of spray drift.

G cluster_0 Spray Source Characteristics cluster_1 Application Setup cluster_2 Environmental Conditions NozzleType Nozzle Type DropletSpectra Droplet Spectra (VMD) NozzleType->DropletSpectra AGDISP AGDISP Model DropletSpectra->AGDISP Meteorology Meteorological Data (Wind Speed/Direction, Temp, RH) Meteorology->AGDISP WindSpeed Wind Speed Meteorology->WindSpeed TempRH Temperature & RH Meteorology->TempRH ReleaseHeight Release Height ReleaseHeight->AGDISP AppSpeed Application Speed AppSpeed->AGDISP DriftPrediction Spray Drift Prediction (Deposition & Concentration) AGDISP->DriftPrediction WindSpeed->AGDISP Evaporation Droplet Evaporation Evaporation->DropletSpectra TempRH->Evaporation

Diagram 1: Logical workflow of critical input parameters in AGDISP spray drift modeling.

Experimental Protocols for Parameterization and Validation

This section provides detailed methodologies for generating data to parameterize AGDISP and for validating model predictions against experimental observations.

Protocol 1: Wind Tunnel Characterization of Nozzle Driftability

3.1.1 Objective: To quantify the drift potential of different nozzle types under controlled environmental conditions for AGDISP input calibration.

3.1.2 Materials & Reagents:

  • Wind Tunnel: A controlled environment chamber capable of maintaining stable wind speed, temperature, and humidity.
  • Test Nozzles: A range of nozzles representing different spray quality classes (e.g., Fine, Medium, Coarse).
  • Spray Solution: A tank mixture including water and a tracer (e.g., fluorescent dye like Pyranine 10G).
  • Droplet Analysis System: A laser diffraction system (e.g., Malvern Spraytec) to measure droplet size distribution in real-time at the nozzle.
  • Passive Collectors: Polyethylene lines or filter papers placed at set intervals downwind for deposition measurement [24].
  • 3D LiDAR Sensor: An optional but advanced sensor (e.g., Sick LD-MRS400001) for non-contact spatial measurement of drift cloud [24].
  • Analytical Equipment: Fluorometer for quantifying tracer concentration on collectors.

3.1.3 Procedure:

  • Setup: Install the nozzle in the wind tunnel at a standardized height. Place passive collectors and/or position the 3D LiDAR sensor downwind.
  • Environmental Stabilization: Set and stabilize wind tunnel conditions (e.g., wind speed: 2, 4, 6 m/s; temperature: 20°C; RH: 60%).
  • Droplet Spectrum Measurement: Use the laser diffraction system to measure the VMD and droplet spectrum of the nozzle prior to drift testing.
  • Application & Sampling: Conduct a short-duration spray (e.g., 5-10 seconds) with the tracer solution.
  • Data Collection:
    • Collect the passive collectors and elute the tracer into distilled water for fluorometer analysis.
    • Simultaneously, record the 3D LiDAR point cloud data of the drift plume [24].
  • Data Analysis:
    • Calculate deposition volume (µL/cm²) from fluorometer data.
    • Process LiDAR data to count the number of droplet points in a vertical plane corresponding to the collector positions.
    • Correlate LiDAR droplet points with physical deposition volume to establish a regression model (e.g., R² > 0.75 indicates strong correlation [24]).

Protocol 2: Field Validation of AGDISP Drift Predictions

3.2.1 Objective: To compare measured field drift deposition from ground and aerial applications against AGDISP simulated predictions.

3.2.2 Materials & Reagents:

  • Application Equipment: Ground rig and/or aircraft (fixed-wing or helicopter).
  • Meteorological Station: To record continuous wind speed, wind direction, temperature, and RH at the field site.
  • Spray Solution: Herbicide (e.g., florpyrauxifen-benzyl) or a suitable tracer.
  • Deposition Collectors: Mylar cards or Petri dishes placed at downwind distances (e.g., 1, 3, 5, 10, 15, 30, 61 m) [5].
  • Sensitive Bioindicators: Soybean plants placed downwind to measure injury and reproductive impact [5].

3.2.3 Procedure:

  • Site Selection: Choose a level field with consistent ground cover and minimal upwind obstructions.
  • Experimental Layout: Establish a downwind sampling line with collectors and bioindicators at predetermined distances.
  • Meteorological Monitoring: Begin recording meteorological data at least 30 minutes before application and continue until after sampling is complete. Only proceed if wind direction is within ±30° of the sampling line [5].
  • Application: Conduct the application using standard practices. For aerial applications, note the release height (e.g., ~50 ft) and consider an upwind swath adjustment to mitigate edge-driven drift [5].
  • Sample Collection: Collect deposition cards immediately after the spray cloud passes. Tag and transport bioindicator plants for later evaluation.
  • Laboratory Analysis:
    • Quantify deposition on Mylar cards using chemical analysis (HPLC) or tracer quantification.
    • Assess soybean injury visually and count reproductive structures (flowers, pods) after a growth period.
  • AGDISP Simulation:
    • Input all recorded parameters into AGDISP: nozzle type, measured droplet spectra, actual release height, swath width, and recorded meteorological data.
    • Run the model to generate a predicted downwind deposition curve.
  • Model Validation: Statistically compare the predicted deposition curve to the measured field data using metrics like index of agreement and mean bias error [4].

Table 2: Comparison of Measured Drift from Ground vs. Aerial Applications [5]

Application Method Spray Quality Avg. Wind Speed PD₅₀ (m) Fold Increase in Drift vs. Ground Key Observations
Ground Coarse 13 kph 0.50 - Drift declined rapidly with distance.
Aerial Coarse 13 kph 10.07 5.0 to 8.6 Drift detectable at 61 m; required upwind swath adjustment.

The Scientist's Toolkit: Research Reagent Solutions

Table 3: Essential Materials and Equipment for Spray Drift Research

Item Function/Application
Fluorescent Tracers (e.g., Pyranine 10G) A water-soluble dye added to spray tank mixtures to allow for sensitive quantification of deposition on passive collectors via fluorometry.
Passive Collectors (Polyethylene Line, Mylar Cards, Filter Paper) Surfaces placed at strategic downwind locations to capture spray droplets for subsequent laboratory analysis of deposition volume.
Laser Diffraction Analyzer (e.g., Malvern Spraytec) Measures the droplet size distribution (Volume Median Diameter and spectrum) emitted from a nozzle in real-time. Critical for characterizing the "Droplet Spectra" input for AGDISP.
3D LiDAR Sensor (e.g., Sick LD-MRS400001) A non-contact remote sensor that uses laser point clouds to measure the spatial distribution and density of a drift plume in real-time, providing rich spatial data complementary to passive collectors [24].
Portable Meteorological Station Measures and records critical meteorological parameters (wind speed, direction, temperature, relative humidity) at the field site during experiments. Data is essential for both parameterizing and validating AGDISP runs.
Sensitive Bioindicator Plants (e.g., Soybean) Living plants used in field experiments to quantify the biological impact of spray drift through visual injury assessment and measurement of reproductive structure reduction [5].

The accuracy of the AGDISP model in predicting pesticide spray drift is inextricably linked to the precision of its input parameters. Nozzle type (defining droplet spectra), meteorological conditions, and release height form the foundational triad governing model outcomes. Adherence to standardized experimental protocols—such as wind tunnel testing for nozzle characterization and rigorous field validation—ensures that these inputs are both accurate and reflective of real-world conditions. By meticulously characterizing these parameters and understanding their interactions, researchers and regulatory professionals can leverage AGDISP to its full potential, enabling the development of effective application strategies that minimize environmental impact while maintaining pest control efficacy.

Within the framework of assessing the AGDISP model for predicting pesticide spray drift, understanding its application to conventional aerial and ground boom scenarios is fundamental. These two application methods represent significant portions of pesticide applications globally and present distinct drift challenges and modeling considerations. AGDISP (Agricultural DISPersal model) is a mechanistic model used for estimating downwind deposition from spray applications [5]. This document provides detailed application notes and protocols for researchers and scientists to effectively model these scenarios, incorporating validated experimental data and methodologies.

Quantitative Drift Data Comparison

Field studies consistently demonstrate that application method and parameters significantly influence drift potential. The following tables summarize key quantitative findings from recent research, which can be used for model validation and benchmarking.

Table 1: Comparison of Drift Deposition and Soybean Injury from Aerial and Ground Applications (Florpyrauxifen-benzyl) [5]

Application Method Spray Quality Avg. Wind Speed Release Height Drift Deposition (PD50) Fold Increase in Drift vs. Ground Soybean Reproductive Structure Reduction
Aerial (Air Tractor 802A) Coarse 13 kph ~15 ft 10.07 m 5.0 to 8.6 Nearly 100% at 61 m downwind
Ground Boom (Case 5550) Coarse 13 kph 3 ft 0.50 m (Baseline) ~25% at 30.5 m downwind

Table 2: Drift Comparison of Unmanned Aerial Vehicle (UAV) and Electric Knapsack Sprayer (EKS) [26] [27]

Application Method Spray Drift Distance Average Drift Deposition Rate Atmospheric Pesticide Concentration 15 Days Post-Application
UAV (DJI T50) 0–20 m 0.47% 59,242.64 pg/m³
Electric Knapsack Sprayer (EKS) 0–4 m 0.23% 2,833.64 pg/m³

Table 3: AGDISP Model Guidance and Validation from US EPA [8]

Model Aspect Near-Field AGDISP (Lagrangian) Far-Field AGDISP (Gaussian Extension)
Validated Distance Up to 800 m (0.5 miles) Up to 10 km (6.2 miles) [8]; up to 20 km (12.4 miles) for screening [8]
Recommended Use Near-field drift predictions for various application methods Only for aerially applied mosquito adulticides or pesticides applied with very fine to fine droplets at ≥50 ft release height [8]
Key Limitations N/A Significant uncertainty >0.5 miles; lacks validation beyond 2 miles; does not account for physical barriers or complex meteorology [8]

Experimental Protocols for Drift Measurement and Model Validation

To generate data for validating AGDISP model predictions for conventional applications, standardized field experiments are essential. The following protocol outlines a methodology for comparing aerial and ground boom applications.

Protocol: Field Measurement of Spray Drift Deposition and Crop Impact

Objective: To measure and compare the downwind spray drift deposition and off-target biological effects from conventional aerial and ground boom applications under field conditions, for the purpose of validating the AGDISP model.

1. Reagents and Materials

  • Tracer Compound: Rhodamine B (a safe, fluorescent tracer for deposition studies) or the actual pesticide of interest (e.g., Florpyrauxifen-benzyl) [26] [5].
  • Collection Media:
    • Mylar cards or petri dishes for quantifying ground deposition [26] [5].
    • Water-sensitive cards for measuring droplet density and coverage [5].
    • Polyethylene lines for airborne drift collection [26].
    • Polyurethane foam (PUF) passive air samplers for quantifying volatile or aerosolized residues over extended periods [26] [27].
  • Sensitive Bioindicator Plants: Such as soybean, planted in rows at set distances downwind to assess biological impact on growth and reproductive structures [5] [28].

2. Equipment

  • Application Equipment:
    • Aerial: Fixed-wing aircraft (e.g., Air Tractor 802A) or helicopter, calibrated for application.
    • Ground Boom: Tractor with a boom sprayer (e.g., Case 5550 with 100-ft boom), calibrated for application [5] [28].
  • Meteorological Station: To record wind speed, wind direction, temperature, and relative humidity at 1–2 second intervals during the application [5].
  • Data Loggers for environmental conditions.
  • Fluorometer or Liquid Chromatography-Mass Spectrometry (LC-MS) equipment for quantifying tracer or pesticide residue on collectors [26].

3. Procedure Step 1: Site Selection and Setup.

  • Select a level, open field with consistent ground cover (e.g., wheat stubble) and no large wind barriers [26] [5].
  • Establish a sampling line perpendicular to the wind direction, extending from the downwind edge of the spray swath to at least 61 m (200 ft) for aerial and 30.5 m (100 ft) for ground applications, based on preliminary predictions [5] [28].
  • Place ground deposition collectors (Mylar cards, petri dishes) and water-sensitive cards at set intervals (e.g., 3, 7.5, 15, 30.5, 61 m) along the sampling line [5].
  • Install PUF air samplers at key distances and heights to capture airborne drift.
  • Plant rows of sensitive soybean plants parallel to the application swath at the same set distances.

Step 2: Application.

  • Conduct applications during stable and suitable meteorological conditions: wind speeds between 3-10 mph (1.3-4.5 m/s), avoiding temperature inversions [8] [29].
  • For both aerial and ground applications, use the same spray quality (e.g., Coarse droplet spectrum) and similar formulations to isolate the effect of application method [5].
  • Record all application parameters:
    • Aerial: Flight altitude (e.g., 15 ft), airspeed (e.g., 145 mph), swath width (e.g., 72 ft), and nozzle type [5] [28].
    • Ground Boom: Boom height (e.g., 3 ft), ground speed (e.g., 20 mph), spray pressure, and nozzle type [5].

Step 3: Sample Collection and Analysis.

  • Collect all deposition and air samples immediately after the spray cloud passes.
  • Analyze deposition samples:
    • Mylar cards/petri dishes: Wash and analyze with a fluorometer or LC-MS to determine deposition mass per unit area. Express as a percentage of the theoretical application rate [5].
    • Water-sensitive cards: Scan and analyze with image analysis software to determine percent coverage and number of droplets per cm² [5].
  • Analyze air samples from PUF samplers using appropriate chemical analysis methods (e.g., GC-MS) to determine airborne concentration [26].
  • Assess bioindicator plants (soybean) after a set period (e.g., 14-21 days) for visual injury and count reproductive structures (flowers, pods) to quantify biological impact [5].

Step 4: Data Analysis and Model Validation.

  • Plot deposition and biological response against downwind distance and fit regression curves (e.g., four-parameter log-logistic) [5].
  • Input the exact application and meteorological parameters from the field experiment into AGDISP.
  • Compare the model's predicted downwind deposition curve and the field-measured data using statistical measures like index of agreement, relative bias, and root mean square error (RMSE) to validate the model's performance [4].

Workflow for AGDISP Modeling of Conventional Applications

The following diagram illustrates the logical workflow for using the AGDISP model to simulate spray drift from conventional applications and validating the outputs with field data.

G Start Start: Define Modeling Objective Input Input Parameters Start->Input Sub1 Application Method: Aerial or Ground Boom Input->Sub1 Sub2 Operational Parameters: Release Height, Speed, Droplet Size, Swath Width Input->Sub2 Sub3 Meteorological Data: Wind Speed/Direction, Temperature, Humidity Input->Sub3 Sub4 Formulation & Site Input->Sub4 ModelExec Execute AGDISP Simulation Sub1->ModelExec Sub2->ModelExec Sub3->ModelExec Sub4->ModelExec Output Model Output: Downwind Deposition Profile ModelExec->Output Validation Field Validation Output->Validation Compare Compare & Refine Output->Compare Predicted Data FieldData Field Data Collection: Deposition Measured on Mylar Cards, etc. Validation->FieldData FieldData->Compare Compare->Input Disagreement (Refine Inputs) End Validated Model Prediction Compare->End Agreement

AGDISP Modeling and Validation Workflow

The Scientist's Toolkit: Key Research Reagents and Materials

Table 4: Essential Research Reagents and Materials for Spray Drift Studies [26] [5]

Item Function/Application in Research
Rhodamine B Tracer A fluorescent dye used as a safe surrogate for pesticides to visually and quantitatively track droplet deposition and drift patterns on collection surfaces.
Mylar Cards / Petri Dishes Smooth, non-absorbent surfaces used as artificial collectors to capture spray droplets for quantitative analysis of ground deposition.
Water-Sensitive Cards Cards coated with a yellow material that turns blue upon contact with water-based droplets. Used for qualitative and image-based analysis of droplet density, coverage, and size spectrum.
Polyurethane Foam (PUF) Samplers Passive air samplers used to capture and concentrate volatile or aerosolized pesticide residues from the air over extended periods (e.g., days to weeks) for subsequent chemical analysis.
Sensitive Bioindicator Plants (e.g., Soybean) Living plants used to assess the biological impact and phytotoxicity of herbicide drift, measuring effects on plant health, growth, and reproductive structures.
AGDISP / AGDISPpro Software Mechanistic simulation models used to predict the transport and deposition of spray droplets downwind of an application site. AGDISPpro is the newer version adapted for UAVs [4].

The adoption of Remotely Piloted Aerial Application Systems (RPAAS), commonly known as drone sprayers, represents a transformative development in precision agriculture. These systems offer potential advantages in targeted pesticide application, operational efficiency, and reduced human exposure. However, their unique aerodynamic characteristics and application parameters present new challenges for predicting off-target spray drift, a critical concern for environmental protection and regulatory compliance. Within the broader context of AGDISP model development for predicting pesticide spray drift in air, this creates a significant research gap. Currently, no validated mechanistic models exist that specifically simulate off-target droplet movement from these advanced application systems [3]. This protocol addresses this technological gap by providing detailed application notes for configuring AGDISPpro, an established Lagrangian-based drift and deposition model, to accurately simulate spray drift from RPAAS, thereby enabling researchers and environmental assessors to generate reliable drift predictions for these emerging technologies.

AGDISPpro Validation for RPAAS: Performance Data

A recent rigorous evaluation of AGDISPpro has demonstrated its promising capability for RPAAS modeling. The study focused on two specific RPAAS models available in the software: the PV22 quadcopter and the PV35X hexacopter [3]. The validation involved comparing model predictions against field deposition measurements from multiple spray scenarios. The table below summarizes the quantitative performance of AGDISPpro across different spray qualities, as measured by the index of agreement (r).

Table 1: Validation Performance of AGDISPpro for RPAAS Applications

Application Type Spray Quality Index of Agreement (r) Performance Interpretation
Single-Swath Medium Nozzles 0.47 – 0.92 Moderate to Excellent
Single-Swath Extremely Coarse Nozzles 0.61 – 0.94 Good to Excellent
Four-Swath Fine Nozzles 0.86 – 0.93 Good to Excellent
Four-Swath Ultra Coarse Nozzles 0.48 – 0.55 Moderate

The data indicates that AGDISPpro performs with good to excellent agreement for most scenarios, though predictions for ultra-coarse sprays in multi-swath operations and some medium-spray scenarios showed moderate agreement. The study identified that uncertainties in swath width and swath displacement behavior significantly impact the prediction accuracy for the location and magnitude of peak deposition [3]. This highlights a critical area for further research and careful parameterization during model setup.

Configuration Workflow for RPAAS Modeling

Configuring AGDISPpro accurately for RPAAS requires a systematic approach to input parameter selection. The following diagram illustrates the logical workflow and critical decision points for setting up a valid simulation.

G Start Start RPAAS Model Setup A1 Select RPAAS Model Start->A1 A2 e.g., PV22 (Quadcopter) or PV35X (Hexacopter) A1->A2 B1 Define Spray Nozzle & Droplet Spectrum A2->B1 B2 Medium, Fine, Extremely/Ultra Coarse B1->B2 C1 Set Application Pattern B2->C1 C2 Single-Swath or Multi-Swath C1->C2 D1 Configure Swath Geometry C2->D1 D2 Swath Width & Displacement (Key Uncertainty) D1->D2 E1 Input Meteorological Data D2->E1 E2 Wind Speed & Direction E1->E2 F Run AGDISPpro Simulation E2->F G Analyze Output: In-Swath & Off-Target Deposition F->G

Experimental Protocol for RPAAS Spray Drift Validation

This section details a standardized methodology for conducting field experiments to validate AGDISPpro predictions for RPAAS applications, based on the approaches cited in the validation study.

Field Study Design and Setup

The objective is to generate empirical deposition data for comparison with AGDISPpro simulated values. The experiment should be designed to test different variables that influence drift.

  • Site Selection: Choose a large, open field with minimal obstructions (e.g., trees, buildings) to simplify the initial validation. The terrain should be level to align with the model's current validation domain [8].
  • Experimental Variables:
    • RPAAS Model: Test at least two different RPAAS (e.g., a quadcopter like the PV22 and a hexacopter like the PV35X) [3].
    • Spray Quality: Utilize a range of nozzle types to generate different droplet spectra. The protocol should include at least fine, medium, and extremely coarse sprays [3].
    • Application Pattern: Conduct both single-swath and multi-swath (e.g., four-swath) applications to assess model performance for different operational patterns [3].
  • Tracer Selection: Use a non-toxic fluorescent tracer dye (e.g., Brilliant Sulfaflavine) mixed with water as the spray tank solution. This allows for sensitive and quantitative measurement of deposition.

Data Collection and Sample Analysis

  • Deposition Sampling:
    • Place horizontal collectors (e.g., Petri dishes with filter paper, or mylar sheets) along transects both within the target swath and at set intervals downwind (e.g., from 5 m to 100 m).
    • Space collectors densely enough to capture the deposition gradient accurately, especially near the swath edge where deposition changes rapidly.
  • Meteorological Data: Continuously record key meteorological data throughout the trial, including wind speed, wind direction, temperature, and relative humidity at a height representative of the application zone.
  • Sample Processing: Collect the sampling media after application and elute the tracer dye into a known volume of distilled water. Measure the fluorescence of each eluent using a fluorometer.
  • Data Calculation: Convert fluorescence readings to deposition values (e.g., µL/cm² or ng/cm²) using a pre-established calibration curve for the tracer dye.

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 2: Key Research Reagents and Materials for RPAAS Spray Drift Studies

Item Function/Description Example/Critical Feature
AGDISPpro Software Lagrangian-based model for predicting spray drift and deposition from aerial applications. Must have license and modules for RPAAS (e.g., PV22, PV35X models) [3].
Remotely Piloted Aerial Application System (RPAAS) The drone platform used for pesticide application. Specific models matter (e.g., PV22 quadcopter, PV35X hexacopter); rotor number and configuration influence wake turbulence [3].
Spray Nozzles Generates the spray droplet spectrum. Critical for drift potential. Nozzles producing fine, medium, coarse, extremely coarse, and ultra-coarse droplets are needed for parameterization [3].
Fluorescent Tracer Dye A safe, quantifiable surrogate for pesticides in field validation studies. Brilliant Sulfaflavine or similar; allows for sensitive detection on collection media.
Fluorometer Instrument to quantitatively measure the fluorescence of sampled tracer dye. Essential for converting sample data into numerical deposition values for model comparison.
Horizontal Deposition Collectors Surfaces placed in-field to capture spray droplets. Petri dishes with filter paper, mylar sheets, or glass slides; provide a standardized collection surface.
Meteorological Station Measures atmospheric conditions during the experiment. Must record wind speed, wind direction, temperature, and relative humidity at application height.

Regulatory Context and Far-Field Considerations

For ecological risk assessment under the FIFRA, the linked AGDISP-Gaussian Extension model is recognized by the U.S. EPA for estimating far-field drift up to 20 kilometers for specific aerial application scenarios [8]. However, this guidance is currently restricted to conventional aerial applications with very fine to fine droplet spectra and a minimum release height of 50 feet. The Gaussian Extension model inherits significant uncertainty for far-field predictions as it does not account for physical barriers (e.g., trees, topography) or variable meteorological conditions beyond the application site [8]. While the core AGDISPpro model has been validated for RPAAS for near-field drift (up to 800 meters), the application of its Gaussian Extension for RPAAS-derived far-field drift has not been validated and falls outside current EPA guidance. Therefore, researchers should focus initial model validation and use on near-field scenarios while clearly acknowledging the limitations regarding far-field predictions for RPAAS.

AGDISPpro is a validated, promising tool for modeling spray drift from Remotely Piloted Aerial Application Systems, filling a critical gap in the ecosystem of pesticide drift prediction tools. Successful configuration requires careful attention to the selection of the specific RPAAS model, spray characteristics, and swath geometry parameters. The experimental protocols outlined herein provide a framework for generating robust validation data. The primary source of prediction uncertainty currently lies in defining the effective swath width and swath displacement under operational field conditions [3]. Future research must prioritize quantifying these RPAAS-specific application parameters to further refine model inputs and improve predictive accuracy. As the use of drone application technology expands, integrating these refined models into regulatory frameworks will be essential for ensuring accurate environmental risk assessments.

The AGricultural DISPersion (AGDISP) model, developed by the USDA Forest Service, represents a significant achievement in modeling the near-field drift of pesticides applied via aerial applications [30]. Historically, this Lagrangian model provided reliable predictions up to a downwind distance of 800 meters (0.5 miles) [8]. However, a pressing need emerged within the aerial application community, particularly the mosquito vector control industry, for accurate predictions of pesticide drift at distances up to 20 kilometers (12.4 miles) from the application site [30]. These operations often release small droplets from heights exceeding 100 meters, where downwind drift is anticipated over much greater distances. To meet this need, a Gaussian model extension was integrated into AGDISP, creating a linked modeling system capable of far-field drift assessment [8]. This document details the application notes and experimental protocols for employing this linked model within the context of ecological risk assessment and large-scale aerial spray operations.

The linked AGDISP-Gaussian model hybridizes two distinct computational approaches to overcome the limitations of each individual model.

The AGDISP Lagrangian Model

The core AGDISP model utilizes a Lagrangian particle-tracking approach [30]. It solves the particle equations of motion for position and position variance, returning predictions of droplet location and cloud standard deviation as a function of time after release [30]. This model explicitly accounts for critical near-field physics, including the influence of aircraft wake effects (particularly wingtip vortices), crosswind (wind speed and direction), and evaporative effects (driven by temperature and relative humidity) [30]. Its validation is well-established for distances up to 800 meters [3].

The Gaussian Extension Model

The Gaussian extension is a steady-state Gaussian plume model based on principles similar to those used in the Industrial Source Complex (ISC) model [8]. This model simplifies the complex physics of dispersion into a statistical formulation that describes the concentration profile of a dispersing plume downwind. It is effective for predicting steady-state, long-range transport over level terrain and has been independently validated for distances up to 50 kilometers [8].

Model Hand-off and Transition

The critical innovation lies in the seamless transition between the two models. The hand-off occurs after the effects of the aircraft wake have dissipated and are no longer influencing the droplet cloud [30]. The Lagrangian model provides the initial conditions for the Gaussian model, including the effective source strength, cloud dimensions, and downwind position at the point of transition, which typically occurs between 300 and 500 meters downwind [30]. This ensures mass conservation and a physically realistic progression of the spray cloud.

Table 1: Key Characteristics of the Linked AGDISP-Gaussian Model Components

Model Characteristic AGDISP (Lagrangian) Gaussian Extension
Modeling Approach Particle-tracking, mechanistic Statistical, steady-state plume
Primary Validation Range 0 - 800 meters (0.5 miles) [8] Up to 50 kilometers (31.1 miles) [8]
Key Physics Accounted For Aircraft wake, evaporation, crosswind [30] Atmospheric stability, plume dispersion [8]
Typical Transition Point 300 - 500 meters downwind [30] N/A
Final Prediction Distance N/A Up to 20 kilometers (12.4 miles) [8]

Quantitative Data and Model Parameters

Successful implementation of the linked model requires careful parameterization. The following tables summarize critical input data and atmospheric parameters.

Table 2: Core Application Scopes and Parameters for the Gaussian Extension

Parameter Applicable Use Scenario Specification / Justification
Application Method Aerial application only [8] Not for ground spray applications [8].
Target Scenarios Mosquito adulticides; forests & tree orchards (e.g., walnut, pecans) [8] Parameters closely related to validated scenarios [8].
Droplet Size Spectrum ASAEvvery fine to fine [8] Smaller droplets are prone to long-range transport.
Minimum Release Height 50 feet (15.2 meters) above ground, with a minimum of 10 feet (3 meters) above plant canopy [8] Minimizes uncertainty from ground obstacles and canopy interactions [8].

Table 3: Atmospheric Stability Classes and Impact on Dispersion

Stability Class Description Typical Conditions Impact on Far-Field Drift
A - B Highly Unstable to Unstable Sunny, daytime, strong solar heating High vertical turbulence, increased dispersion, lower ground-level concentrations.
C Slightly Unstable Sunny, daytime, moderate wind Moderate dispersion.
D Neutral Overcast, moderate to strong wind Gaussian model standard assumption; plume follows wind direction.
E - F Stable to Very Stable Clear night, light wind Suppressed vertical dispersion, leads to reduced mixing and potential for higher, more concentrated far-field drift [30].

Experimental Protocol for Far-Field Drift Assessment

This protocol outlines the steps for using the linked AGDISP-Gaussian model to estimate far-field pesticide drift, suitable for FIFRA ecological risk assessments.

Pre-Modeling Configuration and Inputs

  • Define Application Parameters: Input precise data on the application scenario.
    • Aircraft and Nozzle Setup: Specify aircraft type, spray boom configuration, and nozzle type.
    • Spray Mixture: Define the pesticide formulation and physical properties (e.g., density, volatility).
    • Droplet Size Distribution: Select the appropriate ASAE droplet size spectrum (e.g., Very Fine to Fine) [8]. This is a critical determinant of drift potential.
    • Operational Parameters: Set release height (≥ 50 feet), spray rate, and aircraft speed [8].
  • Define Meteorological Parameters: Input representative atmospheric conditions.
    • Wind Speed: Typically 10-15 mph for risk assessments, representing a conservative, moderate wind condition [8].
    • Atmospheric Stability: Select the appropriate stability class (see Table 3). Neutral (Class D) is often a standard assumption.
    • Temperature and Relative Humidity: Essential for modeling droplet evaporation [30].
  • Define Terrain and Receptor Grid: The Gaussian extension assumes level terrain with no physical barriers (e.g., trees, topography) [8]. Set up a downwind receptor grid extending to 20 km.

Model Execution and Workflow

The logical sequence of the modeling process is illustrated in the following workflow.

G Start Start: Define Application & Meteorological Parameters A Run AGDISP Lagrangian Model (0 - ~400 m) Start->A B Model Calculates Wake Effects, Droplet Evaporation, and Deposition A->B C Transition: Extract Effective Source Strength & Cloud Size B->C D Initialize Gaussian Plume Model with Transition Parameters C->D E Calculate Far-Field Drift & Deposition to 20 km D->E End Output: Drift Deposition Profile and Concentrations E->End

Output Analysis and Interpretation

  • Data Extraction: Extract predicted deposition values (e.g., µg/cm² or kg/ha) and airborne concentrations at specified downwind distances.
  • Uncertainty Qualification: Acknowledge the significant uncertainty in estimates beyond 2 miles due to the model's inability to account for physical barriers and variable meteorology [8]. Results are considered conservative.
  • Reporting: Report deposition curves and peak deposition values. Clearly state all model assumptions and input parameters used in the simulation.

The Researcher's Toolkit: Essential Research Reagents and Materials

Table 4: Key Research Reagent Solutions and Essential Materials

Item / Reagent Function in AGDISP-Gaussian Modeling
AGDISP Software Platform The core software environment that hosts the Lagrangian and linked Gaussian extension models for conducting simulations [30].
ASAE S572 Standard Reference Nozzles Provides a standardized classification system for droplet size spectra (e.g., Fine, Medium), which is a critical model input for defining the initial spray cloud [8].
Meteorological Station Data Source of input parameters for wind speed, temperature, and relative humidity, which drive the dispersion and evaporation calculations within the model [30].
Tracer Dye (e.g., Rhodamine WT) A physical reagent used in field validation studies to track spray deposition and drift; its measured deposition is compared against model predictions to validate accuracy [3].
Lagrangian Particle-Tracking Algorithm The computational engine within AGDISP that simulates the motion and fate of individual spray droplets in the near-field, accounting for complex physics [30].
Gaussian Plume Dispersion Coefficients Pre-defined parameters (e.g., σy, σz) within the model extension that statistically represent the spreading of the spray plume in the horizontal and vertical directions during far-field transport [8].

Critical Considerations and Model Limitations

Researchers must operate within the defined boundaries and acknowledge the limitations of the linked model.

  • Inherent Model Conservatism: The current Gaussian extension modeling does not consider physical barriers (e.g., windbreaks, trees, topography) or complex meteorological variables like crosswinds and humidity fluctuations beyond the application site. These factors would typically reduce actual drift, meaning the model outputs are extremely conservative for risk assessments [8].
  • Scope of Validation: While Gaussian models, in general, are effective out to 50 km, the specific linkage in AGDISP has been validated with reasonable agreement from 0.1 km to 10 km under conditions of level terrain with little ground cover and straight-line winds [8] [30]. There is a lack of validation data beyond 2 miles for scenarios involving complex terrain or vegetation [8].
  • Mass Conservation: The Gaussian extension in AGDISP accounts for the mass of applied pesticide, addressing early concerns about plume depletion and ensuring that predictions, while conservative, are physically realistic [8].
  • Ongoing Modernization: The AGDISP Modernization Project (AMP) is actively working to rewrite the model's code into a modern language. This effort aims to improve accuracy, incorporate modern drift reduction technologies, and make the model more accessible for future development, including potential real-time, site-specific risk assessments [17].

This case study details the methodology and findings of field experiments and subsequent modeling efforts to quantify the off-target movement of the synthetic auxin herbicide florpyrauxifen-benzyl from ground and aerial application equipment. The research is situated within a broader thesis investigating the predictive accuracy of the AGDISP model for simulating pesticide spray drift in air. Herbicide drift poses significant risks to adjacent ecosystems and non-target crops, particularly sensitive species like soybean. This study provides a validated framework for simulating drift using AgDISP, offering researchers and pesticide development professionals a protocol for assessing environmental exposure and mitigating off-target impacts.

Experimental Protocol: Field Spray Drift and Phytotoxicity Assessment

Field Site and Meteorological Conditions

Field experiments were conducted under optimal meteorological conditions in accordance with US EPA guidelines [5].

  • Meteorological Monitoring: Air temperature, relative humidity, wind speed, and wind direction were monitored throughout the experiment.
  • Acceptable Ranges: The experiment proceeded with an average wind speed of 13 kph, air temperature between 6 and 14 °C, and relative humidity from 55 to 88%. Wind direction deviated less than 30° from the established sample collection line [5].

Herbicide Application and Treatments

The study compared two application methods using the herbicide florpyrauxifen-benzyl [5].

  • Herbicide: Florpyrauxifen-benzyl (trade name: Loyant or Rinskor active), a synthetic auxin (WSSA Group 4) herbicide [5].
  • Application Methods:
    • Ground Application: Utilizing commercial ground spray equipment.
    • Aerial Application: Utilizing a fixed-wing aircraft. To mitigate drift, the aerial application included one full swath width adjustment upwind of the sample collection line [5].
  • Spray Quality: Both applications were conducted with a Coarse spray droplet spectrum [5].

Downwind Drift Measurement

Spray drift deposition was measured at multiple distances downwind from the application site.

  • Collection Media:
    • Mylar Cards: Positioned at ground level to collect spray deposits. Deposits were expressed as a percentage of the total theoretical applied rate [5].
    • Water Sensitive Cards: Used to quantify spray coverage (percentage) and the number of spray deposits per unit area (# cm⁻²) [5].
  • Distances: Data were collected at intervals extending to 61 meters downwind [5].

Soybean Injury and Impact Assessment

The impact of drift on a non-target, sensitive crop was evaluated using soybean (Glycine max).

  • Evaluation Metrics: Researchers assessed visual soybean injury and quantified the reduction in reproductive structures (flowers and pods), which serve as crucial pollinator foraging sources [5].
  • Statistical Analysis: A four-parameter log-logistic regression model was fitted to the data for each response variable (drift deposition and soybean injury) as a function of downwind distance to determine the predicted distances for 25%, 50%, and 90% reduction (PD₂₅, PD₅₀, PD₉₀) in deposition or plant health [5].

Results and Quantitative Data Analysis

Comparative Drift Deposition from Ground and Aerial Applications

The following table summarizes the measured downwind drift deposition for the two application methods.

Table 1: Comparison of downwind spray drift deposition from ground and aerial applications of florpyrauxifen-benzyl [5].

Application Method PD₂₅ (m) PD₅₀ (m) PD₉₀ (m) Fold Increase vs. Ground (Based on Deposits)
Ground 0.23 0.50 2.36 --
Aerial 7.55 10.07 20.54 5.0 to 8.6

Despite one upwind swath adjustment, the aerial application resulted in significantly greater downwind drift, with deposits detectable at the farthest sampling distance (61 m). The ground application's drift deposits declined much more rapidly with distance [5].

The physiological injury to soybean and the subsequent impact on potential pollinator foraging sources were severe.

Table 2: Impact of florpyrauxifen-benzyl drift on non-target soybean health and reproduction [5].

Application Method Soybean Injury (Fold Increase vs. Ground) Reduction in Reproductive Structures at 30.5 m Reduction in Reproductive Structures at 61 m
Ground -- ~25% Data not shown
Aerial 1.7 to 3.6 Data not shown ~100%

Soybean injury from the aerial application was 1.7 to 3.6 times greater than from the ground application. The damage to reproductive structures was substantial, with a near-total loss of these potential pollinator resources at 61 m downwind from the aerial application [5].

AGDISP Model Simulation and Validation Protocol

Model Input Parameters

To simulate the field experiment using AGDISP, researchers must configure the model with specific input parameters that mirror the field conditions [5] [31].

Table 3: Key AGDISP input parameters for simulating florpyrauxifen-benzyl drift experiments.

Parameter Category Specification Notes
Application Method Aerial (Fixed-wing) / Ground Select appropriate module.
Spray Quality Coarse Defined by ASABE S572.3 standard.
Release Height Aerial: ~5x Ground boom height; Ground: As measured Height is a critical factor.
Meteorology Wind Speed: 13 kph; Temp: 6-14°C; RH: 55-88% Use consistent, averaged data.
Spray Material Florpyrauxifen-benzyl Physical properties affect drift.
Nozzle Type As used in field experiment Influences droplet size spectrum.
Model Validation and Sensitivity Analysis
  • Validation with Field Data: The AGDISP model's predictions were compared to the physically collected field data on drift deposition. The study found that collected field data were consistent with US EPA model predictions, validating the model's use for this scenario [5]. A separate case study on atrazine also successfully validated AGDISP predictions against field monitoring data up to 400 meters downwind [31].
  • Sensitivity Analysis: Sensitivity studies highlighted droplet size distribution as a critically important parameter influencing spray drift. This underscores the need for proper nozzle selection and maintenance to ensure accurate modeling and drift mitigation [31].

The Scientist's Toolkit: Research Reagent Solutions

Table 4: Essential materials and reagents for conducting florpyrauxifen-benzyl drift and phytotoxicity research.

Item Function/Description Experimental Context
Florpyrauxifen-benzyl (e.g., Loyant) Active Ingredient (AI) The synthetic auxin herbicide under investigation; formulated as an emulsifiable concentrate (EC) [5] [32].
Mylar Cards Spray Deposition Collection Inert cards placed downwind to collect spray droplets for quantitative chemical analysis of drift [5].
Water Sensitive Cards (WSC) Spray Pattern & Coverage Analysis Cards that change color upon contact with water-based droplets; used to analyze spray coverage and density [5].
Acid Brilliant Flavine (ABF) Fluorescent Tracer A fluorescent dye added to the spray tank mixture to allow for precise tracking and quantification of spray drift deposition [32].
Soybean (Glycine max) Bio-indicator Species A non-target, herbicide-sensitive plant used to assess the biological impact and phytotoxicity of drift [5].
AGDISP/AgDRIFT Software Spray Drift Modeling US EPA-registered simulation models for predicting downwind deposition from aerial and ground applications [5] [1].

Workflow and Signaling Pathway Visualization

Experimental and Modeling Workflow

The following diagram outlines the integrated experimental and computational workflow for conducting a spray drift case study.

workflow Florpyrauxifen-benzyl Drift Study Workflow start Define Study Objectives exp_design Field Experiment Design start->exp_design field_work Conduct Field Application & Sample Collection exp_design->field_work data_analysis Laboratory Analysis of Samples field_work->data_analysis validation Model Validation (Field Data vs Prediction) data_analysis->validation model_setup AGDISP Model Parameter Setup model_run Run Simulation & Predict Drift model_setup->model_run model_run->validation conclusion Draw Conclusions & Assess Risk validation->conclusion

Herbicide Impact Pathway on Non-Target Organisms

This diagram illustrates the chain of events from herbicide application to impacts on non-target plants and aquatic species, as documented in the research.

impacts Herbicide Impact Pathway on Non-Targets app Herbicide Application (Florpyrauxifen-benzyl) drift Spray Drift (Off-target Movement) app->drift plant_exp Non-Target Plant Exposure (e.g., Soybean) drift->plant_exp aquatic_exp Aquatic Ecosystem Exposure (e.g., via Runoff) drift->aquatic_exp plant_effect Plant Physiological Effects (Morphological deformation, Reduction in flowers/pods) plant_exp->plant_effect aquatic_effect Aquatic Organism Effects (Oxidative stress, Genotoxicity, Hepatotoxicity in fish) aquatic_exp->aquatic_effect ecosystem_effect Ecosystem-Level Impact (Loss of pollinator forage, Biodiversity reduction) plant_effect->ecosystem_effect aquatic_effect->ecosystem_effect

Discussion and Implications for Research and Regulation

This case study demonstrates a critical application of the AGDISP model in environmental risk assessment for a newly introduced pesticide. The validation of AGDISP against field data for florpyrauxifen-benzyl provides scientists and regulators with a reliable tool for predicting drift potential and establishing appropriate buffer zones or application restrictions. The significant reduction in soybean reproductive structures highlights a secondary ecological impact—the potential reduction of foraging resources for pollinators, an consideration that must be integrated into holistic pesticide risk assessments [5]. Furthermore, emerging research on the toxicity of florpyrauxifen-benzyl to non-target aquatic species, such as Nile tilapia, where it induces oxidative stress and liver damage, underscores the importance of simulating and mitigating drift into aquatic habitats [33].

The methodology outlined, combining rigorous field experimentation with computational modeling, offers a robust framework for the pesticide development industry. It enables the proactive assessment of off-target movement during the product development phase, potentially guiding the formulation of safer application protocols and stewardship plans. Future work should focus on integrating these drift predictions with species sensitivity distributions for a wider range of non-target organisms to comprehensively characterize environmental risk.

Refining AGDISP Predictions: Addressing Uncertainty and Improving Accuracy

In the context of predicting pesticide spray drift using the AGDISP model, identifying and quantifying key sources of uncertainty is paramount for regulatory applications and environmental risk assessments. Recent validation studies of AGDISPpro for remotely piloted aerial application systems (RPAAS) have identified swath width and swath displacement as critical parameters introducing uncertainty in deposition predictions [3] [4]. These factors significantly influence the location, width, and magnitude of peak deposition plumes, thereby affecting the accuracy of environmental exposure estimates for non-target ecosystems. This application note delineates the quantitative impact of these parameters and provides detailed protocols for their empirical determination to enhance model reliability.

Quantitative Impact of Swath Configuration on Deposition Uncertainty

AGDISPpro, which combines atmospheric transport with multi-rotor aerodynamic models, has demonstrated promising performance in predicting off-target spray drift from drone applications, with indices of agreement ranging from 0.47 to 0.94 across various spray quality scenarios [3]. However, sensitivity analyses reveal that uncertainties in defining operational swath characteristics directly impact the accuracy of deposition predictions.

Table 1: AGDISPpro Model Performance Across Different Spray Scenarios

Application Type Spray Quality Index of Agreement Range Primary Uncertainty Source
Single-swath RPAAS Medium 0.47 - 0.92 Swath width & displacement
Single-swath RPAAS Extremely Coarse 0.61 - 0.94 Swath width & displacement
Multi-swath RPAAS Fine 0.86 - 0.93 Swath displacement
Multi-swath RPAAS Ultra Coarse 0.48 - 0.55 Swath width & displacement

Table 2: Impact of Swath Parameters on Deposition Predictions

Parameter Effect on Peak Deposition Effect on Plume Characteristics Magnitude of Influence
Swath Width Significant effect on magnitude Moderate effect on plume width High
Swath Displacement Significant effect on magnitude Moderate effect on plume position High

Analysis of AGDISPpro performance indicates a consistent tendency to underpredict off-target deposition values in field validation studies, primarily due to uncertainties in swath width and displacement parameters [4]. This underprediction manifests in the off-target section of the deposition profile and affects both the maximum peak deposition values and the overall shape of the deposition plume.

Experimental Protocols for Swath Parameter Determination

Field-Based Swath Characterization Protocol

Objective: To empirically determine effective swath width and displacement for RPAAS under operational conditions.

Materials:

  • Remotely piloted aerial application system (e.g., PV22 quadcopter or PV35X hexacopter)
  • Spray solution with fluorescent tracer (e.g., Brilliant Sulfaflavine)
  • Ground-based deposition collectors (petri dishes or chromatography paper)
  • GPS logging equipment
  • Meteorological station (wind speed, direction, temperature, humidity)
  • Fluorometer for deposition analysis

Procedure:

  • Site Selection: Identify a flat, open field with minimal obstructions to ensure consistent wind patterns and avoid topological interference.
  • Collector Placement: Arrange deposition collectors in a grid pattern extending 50 meters upwind and 100 meters downwind from the planned flight path, with 5-meter spacing within the first 50 meters downwind and 10-meter spacing beyond.
  • Flight Operations: Program the RPAAS to execute a single swath pass perpendicular to the prevailing wind direction at the operational altitude (typically 2-5 meters for RPAAS).
  • Data Collection:
    • Activate spray release precisely at the beginning of the programmed swath path.
    • Record exact flight path coordinates using onboard GPS at 1Hz frequency.
    • Monitor and record meteorological conditions at 1-second intervals throughout the trial.
  • Sample Processing:
    • Collect deposition samples immediately after application.
    • Analyze samples using fluorometry to quantify deposition values at each collector location.
  • Data Analysis:
    • Plot deposition values against crosswind distance to establish the deposition profile.
    • Calculate effective swath width as the distance between points where deposition falls to 50% of the maximum value on either side of the peak.
    • Determine swath displacement by measuring the distance between the intended flight path and the actual peak deposition location.

Wind Tunnel Validation Protocol

Objective: To characterize swath displacement under controlled conditions for specific nozzle configurations.

Materials:

  • Wind tunnel meeting ISO 22856:2008 standards
  • Nozzle test bench with positioning system
  • 3D LiDAR sensor (e.g., Sick LD-MRS400001) [24]
  • Polyethylene line collectors or other passive collectors
  • Laser diffraction particle analyzer for droplet size characterization

Procedure:

  • Nozzle Configuration: Mount test nozzles on the wind tunnel boom at operational height (0.5-1.0 meters).
  • Sensor Placement: Position 3D LiDAR sensor and passive collectors downwind from the nozzle in a vertical plane.
  • Test Parameters: Conduct tests across varying wind speeds (1-5 m/s), nozzle types (flat fan, air inclusion, etc.), and spray pressures (1-4 bar).
  • Data Collection:
    • Activate LiDAR scanning during spray release to capture spatial droplet distribution.
    • Collect droplets on passive collectors for quantitative deposition measurement.
    • Record droplet size spectra using laser diffraction analyzer.
  • Data Analysis:
    • Correlate LiDAR point cloud data with deposition volumes from passive collectors.
    • Establish regression relationships between operational parameters and swath displacement.
    • Develop correction factors for different nozzle-pressure combinations.

Conceptual Framework of Swath Uncertainty

The following diagram illustrates the relationship between application parameters, swath characteristics, and their combined effect on deposition uncertainty in AGDISP modeling.

G ApplicationParams Application Parameters SwathCharacteristics Swath Characteristics ApplicationParams->SwathCharacteristics WindSpeed Wind Speed WindSpeed->SwathCharacteristics NozzleType Nozzle Type NozzleType->SwathCharacteristics ReleaseHeight Release Height ReleaseHeight->SwathCharacteristics SprayPressure Spray Pressure SprayPressure->SwathCharacteristics ModelUncertainty Deposition Prediction Uncertainty SwathCharacteristics->ModelUncertainty SwathWidth Effective Swath Width PeakDeposition Peak Deposition Magnitude SwathWidth->PeakDeposition PlumeWidth Plume Width SwathWidth->PlumeWidth SwathDisplacement Swath Displacement SwathDisplacement->PeakDeposition PlumePosition Plume Position SwathDisplacement->PlumePosition

Diagram 1: Swath parameter effects on deposition uncertainty (76 chars)

Research Reagent Solutions for Swath Characterization

Table 3: Essential Materials for Swath Parameter Research

Research Tool Function Application Context
Fluorescent Tracers (e.g., Brilliant Sulfaflavine) Quantify deposition patterns Field validation of swath width and displacement
3D LiDAR Sensor (e.g., Sick LD-MRS400001) Spatial mapping of droplet clouds Wind tunnel and field characterization of swath geometry
Polyethylene Line Collectors Passive collection of drift droplets Ground-truthing of swath parameters under field conditions
GPS Logging Equipment Precise tracking of aircraft position Correlation of flight path with deposition patterns
Laser Diffraction Particle Analyzer Droplet size spectrum characterization Classification of spray quality per ASAE S572.1
Wind Tunnel Platform Controlled environment for parameter isolation Systematic evaluation of individual factors on swath displacement

Swath width and swath displacement represent significant sources of uncertainty in AGDISPpro modeling of spray drift from RPAAS applications. The protocols outlined herein provide researchers with standardized methodologies to quantify these parameters empirically, thereby enhancing model calibration and prediction accuracy. Implementation of these approaches will strengthen the scientific basis for regulatory decisions and environmental risk assessments pertaining to aerial pesticide applications. Future research should focus on developing robust algorithms for swath parameter estimation under diverse operational conditions to further reduce prediction uncertainties.

Within the framework of a broader thesis on the AGDISP model for predicting pesticide spray drift, understanding the sensitivity of the model's outputs to its input parameters is crucial for both accurate risk assessment and effective drift mitigation. The AGDISP model, developed by the U.S. Forest Service and used by regulatory bodies like the U.S. EPA, is a mechanistic, Lagrangian-based model that simulates the off-target movement of spray droplets [3] [1]. This application note provides a detailed protocol for conducting a sensitivity analysis to determine how variations in key input parameters influence two critical outputs: the peak deposition and the width of the deposition plume. Such an analysis is fundamental for researchers, scientists, and professionals involved in refining pesticide application techniques and ensuring environmental and human safety.

Theoretical Background

AGDISP is a "first-principles" science-based model that predicts spray drift from application sites by simulating the release, dispersion, and deposition of spray droplets [1]. It combines atmospheric transport models with detailed aerodynamic models of the application aircraft, including fixed-wing, rotary-wing, and more recently, Remotely Piloted Aerial Application Systems (RPAAS or drones) [3]. The model's predictions are vital for EPA's ecological, endangered species, and human health risk assessments [17]. The movement and final deposition of droplets are governed by a complex interplay of factors including aircraft wake dynamics, droplet physics, and meteorology. The peak deposition refers to the highest concentration of spray material deposited at a specific downwind location, while the plume width describes the lateral spread of the deposited material. Identifying which input parameters most significantly affect these outputs allows users to prioritize the accurate measurement of those parameters in field studies and to better interpret model results.

Parameters for Sensitivity Analysis

The sensitivity of AGDISP's predictions to various input parameters has been highlighted in multiple studies. A virtual experiment using the DAD-drift model, which accounts for the physical basis of spray drift similar to AGDISP, identified droplet size distribution and distance to the non-target area as the two most influential factors determining deposition in non-target areas at the landscape scale [34]. Furthermore, research has shown that evaporative effects, characterized by the wet bulb depression, significantly impact far-field deposition, though the model may overpredict this sensitivity compared to some field observations [22]. Finally, recent evaluations of AGDISPpro have pointed to uncertainty regarding how swath width and swath displacement from aerial application systems affect the location, width, and magnitude of the peak deposition plume [3]. The following table summarizes the key parameters and their documented effects on AGDISP outputs.

Table 1: Key Input Parameters for AGDISP Sensitivity Analysis

Parameter Category Specific Parameter Documented Effect on AGDISP Outputs Primary Reference
Application Setup Swath Width & Displacement Affects the location, width, and magnitude of the peak deposition plume. [3]
Spray Solution & Nozzles Droplet Size Distribution (Nozzle Type) The most influential factor for deposition in non-target areas. [34]
Meteorology Wind Speed Model shows a similar response to field data; a key driver of drift. [22]
Meteorology Wet Bulb Depression (Evaporation) Model is sensitive to evaporative effects, influencing far-field deposition. [22]
Landscape & Geometry Distance to Non-target Area The second most influential factor for off-target deposition. [34]

Experimental Protocols

Protocol for Field Validation of Parameter Effects

This protocol outlines the steps for collecting field data to validate the influence of specific parameters on spray drift, as referenced in AGDISP validation studies [3] [31].

1. Research Reagent Solutions & Materials Table 2: Essential Materials for Field Validation Experiments

Item Function/Description Example from Literature
Tracer Pesticide A chemical marker to track drift and deposition. Atrazine was used as a drift tracer in a ground application study [31].
Spray Application Equipment Apparatus to apply the tracer uniformly. Remotely Piloted Aerial Application Systems (RPAAS) like PV22 quadcopter and PV35X hexacopter [3].
Deposition Samplers Devices to collect spray droplets at various distances and heights. High volume air sampling onto polyurethane foam (PUF) plugs [31].
Meteorological Station Instrument to record environmental conditions during application. Measures wind speed, direction, temperature, and relative humidity [3] [31].
Analytical Instrumentation Equipment for quantitative analysis of tracer in samples. Gas Chromatography with Nitrogen-Phosphorus Detector (GC-NPD) for atrazine [31].

2. Procedure:

  • Site Selection & Setup: Choose a field site with consistent topography. Establish a series of sampling lines at multiple downwind distances (e.g., from the field edge up to 400 meters). At each distance, deploy deposition samplers at various heights [31].
  • Parameter Variation: Conduct multiple application runs while systematically varying one key parameter at a time. For example:
    • Droplet Size: Use different nozzle types (e.g., fine, medium, extremely coarse, ultra coarse) to alter the droplet spectrum [3].
    • Application Height: Vary the release height of the spray application system.
    • Meteorology: Conduct applications under different, but well-measured, wind speed and atmospheric stability conditions.
  • Data Collection: Precisely record all operational parameters for each run (aircraft type, speed, nozzle type, spray pressure, etc.). Simultaneously, record meteorological data (wind speed, direction, temperature, humidity) at a high frequency [3] [31].
  • Sample Analysis: Collect deposition samplers post-application. Analyze them using appropriate chemical methods (e.g., GC-NPD) to determine the mass of the tracer pesticide deposited at each location [31].
  • Data Processing: Translate the chemical analysis results into deposition values (e.g., mass per unit area). Plot deposition vs. downwind distance to visualize the deposition plume and identify the peak deposition value and plume width for each experimental run.

Protocol for Computational Sensitivity Analysis

This protocol describes a virtual experiment to assess parameter sensitivity using the AGDISP model itself or a similar drift model like DAD-drift [34].

1. Research Reagent Solutions & Materials

  • Software: AGDISP or AGDISPpro software license. For landscape-scale analysis, the DAD-drift model can be used [34].
  • Computing Resources: A standard desktop computer or workstation capable of running multiple model simulations.

2. Procedure:

  • Establish a Baseline Scenario: Define a standard application scenario using typical parameters for the aircraft, spray mixture, nozzle type, and meteorological conditions.
  • Define Parameter Ranges: For each parameter of interest (e.g., droplet size spectrum (Dv50), wind speed, application height, swath width), define a realistic range of values based on field practices.
  • Single-Parameter Variation: Run the model multiple times, each time varying only one parameter across its predefined range while holding all other parameters constant at their baseline values.
  • Output Extraction: For each simulation, record the key model outputs: peak off-target deposition and the lateral plume width at a specified downwind distance.
  • Sensitivity Quantification: Calculate the normalized sensitivity coefficient for each parameter. For example, the change in peak deposition per unit change in the input parameter. A higher absolute value of the coefficient indicates greater sensitivity.
  • Meta-Modeling (Advanced): As performed in the DAD-drift virtual experiment, use multiple linear regression on the results of numerous model runs to build a statistical (meta-)model that quantifies the influence of each input parameter on the outputs [34].

Data Analysis and Visualization

The data gathered from either field or computational experiments should be analyzed to rank the parameters by their influence. The following diagram illustrates the logical workflow for conducting the computational sensitivity analysis, highlighting the cause-and-effect relationships between input parameters and model outputs.

G Start Define Baseline Scenario P1 Vary Input Parameter (e.g., Droplet Size, Wind Speed) Start->P1 P2 Run AGDISP Simulation P1->P2 P3 Extract Outputs: Peak Deposition & Plume Width P2->P3 P4 Calculate Sensitivity Coefficient P3->P4 End Rank Parameter Sensitivity P4->End ParamGroup Key Input Parameters: DropletSize Droplet Size Distribution WindSpeed Wind Speed AppHeight Application Height SwathWidth Swath Width Evaporation Wet Bulb Depression DropletSize->P1 WindSpeed->P1 AppHeight->P1 SwathWidth->P1 Evaporation->P1

Figure 1: Workflow for computational sensitivity analysis with AGDISP

The results of the sensitivity analysis can be succinctly summarized in a table for easy comparison. The data below synthesizes findings from the literature.

Table 3: Sensitivity Analysis Findings from Literature

Input Parameter Effect on Peak Deposition Effect on Plume Width Notes & Context
Droplet Size Distribution High Sensitivity. Finer droplets increase peak deposition at longer distances. High Sensitivity. Finer droplets lead to a wider, more diffuse plume. Identified as the most influential factor at the landscape scale [34].
Distance to Non-target Area Extreme Sensitivity. Peak deposition decreases rapidly with increasing distance. High Sensitivity. Plume width typically increases with distance. The second most influential factor [34].
Wind Speed Moderate to High Sensitivity. Increased wind speed generally increases drift and peak deposition downwind. Moderate to High Sensitivity. Higher winds can lead to a narrower, more focused plume. AGDISP shows a similar response to field observations [22].
Swath Width & Displacement Moderate Sensitivity. Affects the magnitude and location of the peak. Moderate Sensitivity. Directly influences the lateral spread of the deposition plume. A key source of uncertainty in deposition predictions [3].
Wet Bulb Depression Moderate Sensitivity (Far-Field). Higher evaporation (larger depression) reduces droplet size, increasing far-field deposition. Sensitivity Varies. Model is sensitive, though may overpredict compared to some field data [22].
Application Height Moderate Sensitivity. Higher release height increases potential for drift and downwind deposition. Moderate Sensitivity. Can contribute to a wider plume. A standard parameter in drift models; effect is well-established.

This application note provides a standardized framework for conducting a sensitivity analysis of the AGDISP spray drift model, with a focus on parameters affecting peak deposition and plume width. The synthesized findings from recent and historical studies consistently highlight droplet size distribution, distance to the non-target area, and meteorological conditions like wind speed and evaporation as primary drivers of model output. Furthermore, application-specific parameters such as swath width and displacement have been identified as critical but uncertain, warranting further research [3]. By following the detailed protocols for field validation and computational analysis outlined herein, researchers can systematically quantify the influence of these parameters. This process is essential for improving the accuracy of AGDISP predictions, guiding the collection of high-quality input data, and ultimately informing the development of effective drift mitigation strategies and rational regulatory policies. The ongoing modernization of AGDISP [17] [19] will further enhance its utility for these critical tasks.

The AGDISP model is a cornerstone for predicting the off-target movement, or drift, of pesticides applied via aerial and ground-based methods. A critical acknowledgment within the scientific community is the model's inherent limitation in automatically accounting for real-world physical barriers (e.g., windbreaks, trees, structures) and complex topography. This application note details the current model constraints, provides validated experimental protocols for field validation and parameter refinement, and outlines mitigation strategies to enhance the accuracy of drift predictions in complex environments. Integrating these considerations is essential for advancing environmental risk assessments and ensuring the sustainable and responsible application of agrochemicals.

Model Limitations and Current Capabilities

AGDISP's predictive capability for spray drift is well-established for level terrain with minimal ground cover [35]. However, its treatment of physical landscapes requires careful consideration.

  • Treatment of Physical Barriers: The model's predictions do not inherently incorporate the effects of physical barriers such as trees, topographic features, or built structures [35]. The current modeling assumptions for far-field drift (> 0.5 miles) are conservative and do not factor in these potential obstructions, leading to a known source of uncertainty [35].
  • Influence of Topography: The model's performance is best on level terrain [35]. On sloped terrain, particularly under stable atmospheric conditions (e.g., temperature inversions), air masses can behave like a fluid and flow down-gradient. This real-world phenomenon can result in high off-target residues downgradient from the application site, a factor that standard AGDISP modeling runs may not capture without specific adjustments [36].

Table 1: Key Limitations of AGDISP in Complex Environments

Complexity Factor AGDISP Treatment Impact on Prediction Uncertainty
Physical Barriers (Trees, structures) Not explicitly accounted for in standard scenarios [35]. Increases uncertainty; predictions may be overly conservative as barriers reduce far-field drift.
Topography (Slopes) Best validated for level terrain; flow down gradients not natively modeled [35] [36]. Can lead to significant under-prediction of drift in down-gradient areas.
Far-Field Distances (> 2 miles) Limited validation data beyond 2 miles; Gaussian extension used with noted uncertainty [35]. Uncertainty significantly increases with distance from the application site.

Experimental Protocols for Model Validation and Refinement

To address these limitations, the following protocols provide a framework for generating empirical data to validate and refine AGDISP predictions in the presence of real-world complexities.

Protocol 1: Field Validation of Drift in Complex Terrain

This protocol is designed to measure actual spray deposition in topographically diverse landscapes and compare it with AGDISP simulations.

1. Objective: To quantify the impact of topography and physical barriers on spray drift and to validate/calibrate the AGDISP model under these conditions. 2. Experimental Setup:

  • Site Selection: Choose two adjacent fields with a significant slope between them and a well-defined vegetative buffer (e.g., a row of evergreen trees) separating the application area from a sensitive downwind area [36].
  • Application Parameters: Use a standardized application method (e.g., aerial application with ASAE Very Fine to Fine droplet spectrum). Precisely record release height, wind speed and direction, temperature, and relative humidity [35] [37].
  • Sampling Array:
    • Place deposition collectors (e.g., petri dishes with a tracer dye or filter paper) along transects starting within the application swath and extending to at least 800 meters downwind [35].
    • Establish transects both upgradient and downgradient of the primary application area.
    • Place collectors both before and after the vegetative buffer zone to directly measure its mitigation effect [36]. 3. Data Collection and Analysis:
  • Sample Processing: Collect samples post-application and analyze them using fluorometry (if a fluorescent tracer is used) or other relevant analytical techniques to quantify deposition (e.g., µg/cm²).
  • Model Simulation: Run AGDISP using the recorded meteorological and application parameters for the specific study site.
  • Statistical Comparison: Compare the observed deposition values from the field collectors with the AGDISP-predicted values. Use statistical metrics such as the Index of Agreement (IA), Mean Bias Error (MBE), and Root Mean Square Error (RMSE), as employed in recent validation studies [4].

Protocol 2: Quantifying the Mitigation Efficacy of Vegetative Buffers

This protocol specifically isolates the effect of windbreaks and vegetative barriers on spray drift reduction.

1. Objective: To empirically determine the drift reduction percentage provided by a specific vegetative buffer configuration. 2. Experimental Setup:

  • Buffer Characterization: Select a well-established vegetative buffer. Measure key parameters: vegetation type (prefer evergreen trees for spring applications), height, density, and porosity [36].
  • Sampling Design: Position deposition collectors in paired sets: one set immediately upwind of the buffer and a corresponding set immediately downwind.
  • Application: Conduct a controlled spray application using a known tracer on the upwind side of the buffer. Applications should be performed under consistent, measured wind conditions (3-10 mph) blowing perpendicularly towards the buffer [36] [37]. 3. Data Analysis:
  • Calculate the percentage of tracer mass captured by or penetrating through the buffer.
  • Drift Reduction (%) = [1 - (Deposition_downwind / Deposition_upwind)] * 100
  • Integrate this reduction factor as a post-processing adjustment to AGDISP outputs for sites with similar buffer characteristics.

The following workflow outlines the key stages for validating and refining the AGDISP model, from initial field experimentation to the final application of calibrated predictions.

G AGDISP Model Validation and Refinement Workflow cluster_experimental Experimental Phase cluster_analysis Data Analysis & Model Calibration cluster_application Application Phase A Site Selection & Instrumentation B Controlled Spray Application A->B C Sample Collection & Analysis B->C D Compare Field Data vs. AGDISP Prediction C->D E Calculate Buffer Efficacy & Topographic Adjustments D->E F Refine Model Input Parameters E->F G Run Calibrated AGDISP Model F->G H Generate Refined Drift Predictions G->H

Mitigation Strategies and Model Parameter Adjustments

Until AGDISP natively integrates these complexities, applicators and researchers can adopt the following strategies to align model use with real-world conditions.

  • Incorporating Vegetative Buffers: Implement vegetative buffers between application sites and sensitive areas. Evergreen trees are particularly valuable as they provide a year-round physical barrier that filters spray droplets from the air [36]. In the absence of native model support, the drift reduction percentage obtained from Protocol 2 can be manually applied to AGDISP-predicted deposition values downwind of the buffer.
  • Strategic Application Planning: Avoid applications when meteorological conditions increase drift potential. This includes wind speeds exceeding 10 mph and during temperature inversions, where calm, stable air can cause fine droplets to hang in the air and move unpredictably, often down-gradient on sloped terrain [36] [37].
  • Droplet Size Management: Use nozzles and formulations that produce larger droplet sizes (Volume Median Diameter > 150 µm). While AGDISP can model different droplet spectra, larger droplets are less prone to drift and will reduce the impact of unmodeled topographic effects [37]. Drift retardants can be used to increase spray solution viscosity and reduce the number of drift-prone fines, though their performance can be inconsistent [36] [37].
  • Accounting for Swath Displacement: For advanced applications, particularly with new technologies like Unmanned Aerial Systems (UAS), recent research highlights that uncertainty in effective swath width and downwind swath displacement is a major factor in under-predicting off-target deposition [4]. Sensitivity analyses on these parameters within AGDISP are recommended to understand their impact on the magnitude and location of peak deposition.

Table 2: The Researcher's Toolkit for Drift Studies in Complex Environments

Tool / Reagent Function / Rationale Application Note
Fluorescent Tracer Dye A safe, quantifiable surrogate for pesticide active ingredients. Allows for precise measurement of deposition on collectors. Enables high-sensitivity detection via fluorometry; non-toxic for field use [4].
Deposition Collectors (e.g., filter papers, petri dishes) Standardized surfaces to capture spray droplets for subsequent analysis. Placed along transects at ground level to measure off-target deposition [4].
Digital Anemometer Measures wind speed and direction at the application site. Critical input for AGDISP and for understanding drift direction. Data should be logged continuously during the application event [35] [37].
Vegetative Buffer (Evergreen Trees) A physical barrier to intercept and filter spray droplets, reducing downwind drift. More effective than deciduous trees for spring applications; density and height are key parameters [36].
Drift-Reducing Nozzles Engineered to produce a coarser spray spectrum (larger droplet sizes) with fewer fines (<150 µm). Directly reduces the fraction of spray mass most susceptible to drift [36] [37].

The AGDISP model is a powerful tool for predicting pesticide spray drift, but its default scenarios do not fully capture the moderating effects of physical barriers and complex topography. By employing the detailed experimental protocols outlined herein—focused on field validation and quantitative measurement of mitigation efficacy—researchers can generate critical data to calibrate the model for specific, complex landscapes. Concurrently, the adoption of practical mitigation strategies, such as establishing vegetative buffers and managing droplet size, provides a pragmatic path to reducing off-target movement and aligning real-world outcomes more closely with model predictions. Future model development should prioritize the integration of digital elevation data and parameterizations for common barrier types to directly address these real-world complexities.

Spray drift, defined as the airborne movement of spray droplets outside the intended target area, remains a significant challenge in agricultural pesticide application. This phenomenon not only reduces application efficacy but also poses potential risks to environmental health and safety. The AGricultural DISPersal (AGDISP) model, developed by the U.S. Forest Service, serves as a critical tool for predicting pesticide spray drift and evaluating mitigation strategies [1]. This model functions as a "first-principles" science-based algorithm that predicts spray drift from application sites by characterizing the release, dispersion, and deposition of spray materials both over and downwind of the application area [1]. Within the context of regulatory science, AGDISP provides the foundational framework for the U.S. Environmental Protection Agency's (EPA) ecological risk assessments and helps establish buffer zone requirements for pesticide labels [8] [38].

Understanding AGDISP's capabilities and limitations is essential for developing effective drift mitigation protocols. The model predicts near-field spray drift up to 800 meters (0.5 miles) downwind from the application site, while its Gaussian Extension module can estimate drift up to 20 kilometers (12.8 miles) under specific conditions [8]. However, significant uncertainties exist in far-field drift estimation, particularly as the model does not account for physical barriers (e.g., trees, topographic features) or varying meteorological conditions beyond the application site [8]. Recent industry initiatives, such as the AGDISP Modernization Project (AMP), aim to address these limitations by updating the model's coding structure to better incorporate modern drift reduction technologies and application conditions [17]. This document provides detailed application notes and experimental protocols for optimizing operational parameters to mitigate predicted spray drift within the AGDISP modeling framework.

Critical Application Parameters Influencing Spray Drift

Droplet Size Characteristics

Droplet size represents one of the most significant factors affecting spray drift potential. Research consistently demonstrates that as droplet size decreases, the potential for spray drift increases exponentially [39]. Droplets with diameters less than 150 micrometers (μm) are considered most prone to drift, with very fine sprays (50-150 μm VMD) capable of drifting 20-30 times farther downwind than very coarse sprays (400-500 μm VMD) [39]. The Volume Median Diameter (VMD), which represents the droplet diameter at which 50% of the spray volume comprises smaller droplets and 50% comprises larger droplets, serves as the standard metric for classifying spray quality according to the ANSI/ASAE S572.1 standard [39].

Table 1: Spray Quality Classification Based on ANSI/ASAE S572.1 Standard

Spray Quality Color Code Volume Median Diameter (μm) Relative Drift Potential
Very Fine (VF) Red 50-150 Extreme
Fine (F) Orange 150-240 High
Medium (M) Yellow 240-350 Moderate
Coarse (C) Blue 350-400 Low
Very Coarse (VC) Green 400-500 Very Low
Extremely Coarse (XC) White 500-660 Minimal

Meteorological Parameters

Meteorological conditions at the time of application significantly influence spray drift behavior and should be carefully monitored. Wind speed directly affects drift distance, with higher wind speeds increasing drift potential [40] [39]. Applications should occur when wind speeds are between 1-2 km/h and 15 km/h – dead calm conditions often indicate temperature inversions that can concentrate and transport spray clouds over long distances [39]. Temperature and relative humidity affect droplet evaporation rates; high temperatures and low humidity accelerate evaporation, creating smaller, more drift-prone droplets [39]. Applications should generally be avoided when relative humidity falls below 40% and air temperature exceeds 25°C [39]. Atmospheric stability plays a crucial role, with temperature inversions trapping spray droplets near the ground and allowing them to travel considerable distances [39].

Equipment and Operational Parameters

Equipment configuration and operational techniques substantially impact drift potential. For aerial applications, release height significantly influences drift; the EPA recommends a minimum release height of 50 feet (15.2 meters) above the ground with 10 feet (3 meters) above plant canopy for specific scenarios [8]. For Unmanned Aerial Spraying Systems (UASS), research indicates optimal flight heights of 1.5-2.5 meters above the crop canopy, with lower heights generally reducing drift [41]. Flight or ground speed affects droplet displacement; studies demonstrate that slower flight speeds (e.g., 2 m/s) significantly reduce drift compared to higher speeds (3 m/s) while improving deposition uniformity [41]. Nozzle type and orientation determine initial droplet size spectrum, with air-induction and other low-drift nozzles specifically designed to reduce fine droplet production [39].

Experimental Protocols for Drift Assessment

Field Measurement of Spray Drift

Validating AGDISP predictions requires empirical field data collection using standardized methodologies. The following protocol outlines a comprehensive approach for measuring spray drift during application operations:

Materials and Reagents:

  • Passive Drift Samplers: Water-sensitive papers (WSPs) for droplet density and coverage analysis [41]
  • Active Air Samplers: High-volume air sampling systems with polyurethane foam (PUF) plugs for airborne concentration measurement [40] [42]
  • Tracer Compounds: Active ingredients (e.g., atrazine, λ-cyhalothrin) as drift tracers compatible with analytical detection methods [40] [43]
  • Analytical Equipment: Gas chromatography systems with appropriate detectors (e.g., GC-NPD for atrazine) [40]
  • Meteorological Station: Instruments for measuring wind speed, wind direction, temperature, and relative humidity at multiple heights [40] [39]

Experimental Procedure:

  • Pre-Application Setup: Position sampling equipment along transects downwind of the application area. For near-field assessment (0-800 m), place samplers at 10, 25, 50, 100, 200, 400, and 800 meters from the field edge [40]. Include vertical sampling arrays at multiple heights (e.g., 0.5 m, 1 m, 2 m, 4 m) to capture vertical deposition profiles.
  • Meteorological Monitoring: Record continuous meteorological data throughout the application period, including wind speed, wind direction, temperature, relative humidity, and solar radiation. Measurements should be taken at heights corresponding to the release height and canopy level [40] [39].
  • Application Implementation: Conduct applications using precisely documented parameters, including equipment type, nozzle configuration, release height, speed, and spray pressure. Record tank mix composition and physical properties.
  • Sample Collection and Processing:
    • Collect WSPs after application and digitize for droplet analysis using image processing software [41]
    • Extract PUF plugs using appropriate solvents (e.g., hexane:acetone mixture) and analyze using chromatographic methods [40]
    • Document sampling duration, flow rates, and environmental conditions during collection
  • Data Analysis: Calculate droplet density (droplets/cm²), coverage (%), deposition volume (μL/cm²), and airborne concentration (ng/L) at each sampling location. Compare observed drift patterns with AGDISP predictions under identical input parameters.

AGDISP Model Validation Protocol

Establishing model credibility requires systematic validation against field observations:

  • Parameterization: Input all documented application parameters (equipment, nozzle type, release height, droplet size spectrum) and meteorological conditions from field trials into AGDISP.
  • Simulation Execution: Run AGDISP simulations for identical sampling transects and heights used in field measurements.
  • Statistical Comparison: Calculate correlation coefficients (R²), root mean square error (RMSE), and mean absolute error (MAE) between predicted and observed deposition values at each sampling location.
  • Sensitivity Analysis: Systematically vary input parameters (e.g., droplet size distribution, wind speed) to identify the most influential factors on model predictions [40].
  • Model Refinement: Adjust model parameters within acceptable ranges to improve agreement with field observations, documenting all modifications.

Recent validation studies have demonstrated strong correlation between AGDISP predictions and field measurements for airborne spray drift up to 400 meters for ground applications [40] and up to 10 kilometers for specific aerial scenarios [8].

Optimization Strategies for Different Application Methods

Aerial Application Systems

For conventional aerial applications, AGDISP modeling indicates that specific parameters significantly influence drift predictions:

  • Release Height Optimization: Maintain a minimum release height of 50 feet (15.2 meters) above ground level while ensuring aircraft remains at least 10 feet (3 meters) above the plant canopy [8]. Modeling shows that each 10-foot reduction in release height can decrease drift potential by approximately 30%.
  • Droplet Size Management: Select nozzles that produce spray spectra classified as Medium to Coarse (240-400 μm VMD) to minimize fine, drift-prone droplets [8] [39]. The EPA specifically recommends ASAE droplet size spectrum of "very fine to fine" only for specific scenarios like mosquito adulticides [8].
  • Meteorological Thresholds: Restrict applications to wind speeds between 3-10 mph (1.3-4.5 m/s) and avoid temperature inversion conditions. AGDISP simulations for far-field drift assume straight-line winds without crosswinds or turbulence [8].

Unmanned Aerial Spraying Systems (UASS)

UASS applications require unique parameter optimization due to rotor-induced downwash effects:

  • Flight Altitude: Research indicates optimal flight heights of 1.5-2.0 meters above the crop canopy, with the lower end of this range (1.5 m) demonstrating significantly reduced drift and improved deposition efficiency [41].
  • Flight Speed: Lower flight speeds (2 m/s) substantially reduce drift while maintaining adequate deposition coverage. Studies show that increasing speed from 2 m/s to 3 m/s can increase drift by 35-50% [41].
  • Operational Integration: Implement terrain-following radar modules for automatic altitude adjustment and high-precision obstacle avoidance systems to maintain consistent application parameters [41].

Table 2: Optimized Operational Parameters for Different Application Systems

Application System Flight/Release Height Speed Droplet Spectrum Wind Speed Range
Aerial (Fixed/Rotary) 50 ft (15.2 m) min. height 130-150 mph Medium-Coarse (240-400 μm) 3-10 mph (1.3-4.5 m/s)
UASS (Multi-rotor) 1.5-2.0 m above canopy 2-2.5 m/s Coarse (350-400 μm) 1-4 m/s
Ground Boom Sprayer 0.5 m above canopy 8-12 km/h Very Coarse (400-500 μm) 3-15 km/h

Ground Application Systems

For ground-based applications, specific mitigation strategies demonstrate effectiveness in AGDISP modeling:

  • Nozzle Selection: Utilize certified low-drift nozzles (air induction, pre-orifice designs) that produce less than 10% droplets smaller than 150 μm [39].
  • Boom Height Management: Position boom as close to the canopy as practicable, ideally ≤ 50 cm above the target, to reduce exposure to wind influences [39].
  • Speed and Pressure Optimization: Maintain ground speeds below 12 km/h and use lower pressures that produce coarser sprays while maintaining coverage [39].

Advanced Mitigation Technologies and Integrated Approaches

Spray Drift Reduction Technologies (SDRT)

Incorporating SDRT can significantly alter AGDISP inputs to reflect reduced drift potential:

  • Nozzle Technology: Modern air-induction nozzles create larger, air-entrapped droplets that reduce fine droplet production by 50-80% compared to conventional flat-fan nozzles [39].
  • Adjuvant Formulations: Deposition aids and non-ionic surfactants can modify droplet spectra and reduce evaporation, potentially decreasing driftable fines by 30-60% [43].
  • Equipment Modifications: Boom shields, shrouds, and electrodynamic systems physically contain spray cloud and direct droplets toward intended targets [39].

Buffer Zone Implementation and Modeling

AGDISP directly informs buffer zone requirements for pesticide labels. The EPA's mitigation menu allows reduction of mandated buffer distances through verified drift reduction practices [38]:

  • Vegetative Buffer Strips: Dense vegetation (grasses, shrubs) at field edges can filter 50-70% of drift droplets through impaction and interception [38].
  • Windbreak Systems: Permanent tree/shrub barriers can reduce wind speed by 50-80% for distances of 10-20 times their height, significantly reducing drift transport [38].
  • Application Timing: Aligning applications with optimal meteorological conditions (moderate winds, high humidity) can reduce predicted drift by 60-80% compared to unfavorable conditions [39].

Research Reagents and Essential Materials

Table 3: Essential Research Reagents and Materials for Spray Drift Studies

Item Specifications Application/Function
Water-Sensitive Papers (WSPs) 76 × 26 mm, yellow coating that turns blue upon water contact Passive sampling of spray droplet density, coverage, and size distribution [41]
Polyurethane Foam (PUF) Plugs 6.2 cm diameter, 5.0 cm thickness, surface density ~0.022 g/cm³ Active air sampling for quantifying airborne pesticide concentrations [40]
Atrazine Analytical Standard ≥98% purity, suitable for GC-NPD analysis Chemical tracer for spray drift validation studies due to stability and detectability [40]
λ-Cyhalothrin Analytical Standard ≥98% purity, suitable for LC-MS/MS analysis Pyrethroid insecticide tracer for human exposure and drift assessment [43]
Portable Weather Station Measures wind speed (0-50 m/s), direction (0-360°), temp (-40°C to +65°C), RH (5-100%) Documenting meteorological conditions during application and sampling [39]
Laser Particle Size Analyzer Malvern Spraytec or equivalent, range 0.1-2000 μm Characterization of droplet size distribution from different nozzle configurations [40]
Low-Drift Nozzle Series Air induction (AI), pre-orifice, extended range designs meeting ASABE S572.1 Generating coarse spray spectra with reduced fine droplet fraction [39]

Optimizing application parameters represents a critical strategy for mitigating predicted spray drift within the AGDISP modeling framework. The protocols and data presented herein provide researchers and application professionals with evidence-based approaches for reducing off-target movement while maintaining efficacy. Key optimization principles include selecting appropriate droplet size spectra (coarse to extremely coarse), minimizing release height, implementing appropriate speeds for each application method, and restricting operations to favorable meteorological conditions. The experimental methodologies outlined enable systematic validation of AGDISP predictions under various scenarios, strengthening model credibility and supporting regulatory decisions.

Future developments in spray drift mitigation will likely focus on real-time, site-specific risk assessments incorporating modernized AGDISP algorithms that account for advanced drift reduction technologies [17]. The ongoing AGDISP Modernization Project aims to create an open-source platform that better represents contemporary application systems, including UASS and integrated SDRT. Additionally, research linking operational parameter optimization to measurable reductions in human exposure biomarkers demonstrates the tangible public health benefits of rigorous drift management [43]. By implementing the structured protocols and optimization strategies detailed in this document, researchers and pesticide applicators can contribute to sustainable pest management practices that minimize environmental impact while maintaining agricultural productivity.

Known Limitations and Model Constraints as Outlined by US EPA Guidance

AGDISP is a "first-principles" science-based model developed by the USDA Forest Service to predict spray drift from pesticide application sites [1]. It is a Lagrangian model designed to predict the near-field spray drift of pesticides up to 800 meters (0.5 miles) downwind from the application site [8]. For ecological risk assessments under the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA), the U.S. Environmental Protection Agency (EPA) provides specific guidance on the use of AGDISP and its Gaussian Extension, acknowledging several critical limitations that researchers must incorporate into their experimental design and interpretation of results [8]. This document outlines these constraints and provides protocols for working within the model's defined boundaries.

Model Constraints and Application Boundaries

The EPA's guidance explicitly defines the operational boundaries and constraints for the AGDISP model and its Gaussian Extension, which are critical for appropriate application in research settings.

Spatial and Physical Limitations

The core AGDISP model and its Gaussian Extension have distinct spatial domains and physical constraints that directly impact their predictive reliability.

Table 1: Spatial and Physical Limitations of AGDISP and Gaussian Extension

Model Component Spatial Range Key Physical Constraints Terrain Considerations
AGDISP Core Near-field (up to 800 m / 0.5 miles) Predicts spray drift based on detailed release characteristics Assumes flat terrain; performance degrades with complex topography
Gaussian Extension Far-field (up to 20 km / 12.4 miles) Does not account for physical barriers (e.g., trees, topographic features) Validated primarily for level terrain with little ground cover

The Gaussian Extension model does not account for physical barriers such as trees and topographic features that would disrupt airflow and pesticide transport [8]. This limitation is particularly significant for long-range transport predictions where such features are likely to be encountered. Furthermore, the model does not consider meteorological variables such as crosswinds and humidity beyond the application site, which are known to significantly affect far-field pesticide drift patterns [8].

Validated Use Cases and Restrictions

The EPA guidance specifically restricts the use of the linked AGDISP-Gaussian Extension models to specific application scenarios based on validation studies:

  • Permitted Applications: Aerially applied mosquito adulticides or pesticides applied using ASAE droplet size spectrum of very fine to fine [8]
  • Minimum Release Height: 50 feet above ground with 10 feet above plant canopy [8]
  • Approved Scenarios: Forests and tree orchards (e.g., walnut, pecans) where parameters most closely align with validation conditions [8]
  • Expressly Prohibited: The Gaussian Extension model should not be used for ground spray applications [8]

The model's developer, Dr. Harold Thistle, confirmed there is "high uncertainty associated with far-field estimates of pesticide concentrations from what is traditionally thought of as spray drift" [8].

Quantitative Limitations and Uncertainty Analysis

Understanding the quantitative limitations of the AGDISP model requires examination of its validation performance and uncertainty metrics across different application contexts.

Validation Performance Metrics

Recent validation studies of AGDISPpro, which includes models for Remotely Piloted Aerial Application Systems (RPAAS), demonstrate variable performance across different application parameters.

Table 2: AGDISPpro Validation Performance Across Spray Quality Types

Spray Quality/Nozzle Type Index of Agreement Range Study Conditions Performance Notes
Medium Droplet Spectra 0.47 to 0.92 [3] Single-swath applications Moderate agreement with field observations
Extremely Coarse Droplet Spectra 0.61 to 0.94 [3] Single-swath applications Good agreement with field observations
Fine Nozzles 0.86 to 0.93 [3] Four-swath applications Strong agreement with field observations
Ultra Coarse Nozzles 0.48 to 0.55 [3] Four-swath applications Poor agreement with field observations

The index of agreement is a standardized measure of model prediction accuracy against observed field data, with values closer to 1.0 indicating better performance. The validation studies revealed that AGDISPpro tended to predict lower deposition values in the off-target section of the deposition profile compared to field observations [3]. Uncertainty in UAS application swath width and swath displacement was identified as a significant factor affecting modeling accuracy, influencing the magnitude of the maximum peak deposition and the width and position of the spray deposition plume [3].

Far-Field Validation Constraints

The EPA has identified significant limitations in the far-field validation of the AGDISP-Gaussian Extension models:

  • Limited Validation Distance: The models lack validation beyond 2 miles (3.2 km) for most scenarios [8]
  • Validation Environment: Primary validation conducted in "level terrain with little ground cover in the high desert of Utah where straight line winds and few competing topological factors exist" [8]
  • Uncertainty in Estimates: There is "significant uncertainty in estimating pesticide deposition at far-field distances (> 0.5 miles)" [8]

Teske and Thistle (2004) found reasonable agreement between predicted concentrations and observed deposition data from 0.1 km to 10 km, while Woods et al. (2001) validated the Gaussian extension model for pesticide deposition data from 0.01 km to 1 km [8].

Experimental Protocols for Model Validation

Researchers should implement the following experimental protocols to validate AGDISP predictions against field observations, particularly for novel application scenarios.

Field Measurement Protocol for Spray Drift Validation

This protocol provides a standardized methodology for collecting field data to validate AGDISP model predictions.

G Figure 1: Spray Drift Validation Experimental Workflow Planning Phase 1: Experimental Planning • Define application parameters • Select nozzle types • Determine release height • Establish sampling grid Setup Phase 2: Field Setup • Position deposition collectors • Install meteorological stations • Calibrate application equipment Planning->Setup Application Phase 3: Controlled Application • Document actual weather conditions • Record application parameters • Execute spray operation Setup->Application Meteorology Meteorological Monitoring (Wind speed/direction, temperature, humidity) Setup->Meteorology Sampling Phase 4: Sample Collection • Collect deposition samples • Document sample locations • Preserve samples for analysis Application->Sampling Application->Meteorology Analysis Phase 5: Laboratory Analysis • Quantify active ingredient • Measure deposition rates • Calculate concentration profiles Sampling->Analysis Validation Phase 6: Model Validation • Input actual parameters to AGDISP • Compare predicted vs observed • Calculate statistical agreement Analysis->Validation

Phase 1: Experimental Planning

  • Define application parameters: nozzle type, release height, vehicle speed/type, spray solution
  • Select sampling distances based on predicted drift profile (include points at 1m, 5m, 10m, 25m, 50m, 100m, and beyond for far-field studies)
  • Establish a sampling grid extending to at least 100m downwind for near-field validation

Phase 2: Field Setup

  • Position deposition collectors (chromatography paper, Petri dishes, or artificial collectors) along transects perpendicular to wind direction
  • Install meteorological stations at application height and standard height (2-3m) to record wind speed, direction, temperature, and relative humidity
  • Calibrate application equipment to ensure accurate delivery rate and droplet size distribution

Phase 3: Controlled Application

  • Conduct applications during stable meteorological conditions (wind speed 3-10 mph preferred)
  • Document actual weather conditions throughout application period
  • Record precise application parameters: actual release height, vehicle speed, swath width, and spray pressure

Phase 4: Sample Collection and Analysis

  • Collect deposition samples immediately after application to prevent degradation or evaporation
  • Extract and analyze samples using appropriate analytical methods (e.g., HPLC, GC-MS, fluorometry for tracers)
  • Quantify deposition mass per unit area at each sampling location

Phase 5: Model Validation

  • Input actual application and meteorological parameters into AGDISP
  • Compare predicted versus observed deposition values using statistical measures (index of agreement, mean bias error, root mean square error)
  • Calculate the index of agreement using the formula: IOA = 1 - (Σ(Pi - Oi)² / Σ(|Pi - Ō| + |Oi - Ō|)²) where Pi are predicted values, Oi are observed values, and Ō is the mean of observed values
Protocol for Addressing Swath Width and Displacement Uncertainty

Given the identified uncertainty in swath width and displacement parameters in AGDISPpro predictions for UAS applications [3], implement this supplemental protocol:

  • Swath Characterization: Conduct preliminary pattern tests to determine effective swath width using water-sensitive paper or string collectors
  • Multiple Swath Analysis: Compare single-swath versus multiple-swath applications to quantify cumulative displacement effects
  • Sensitivity Analysis: Vary swath width by ±20% in model inputs to quantify the effect on deposition predictions
  • Wind Correction: Document and account for crosswind displacement during each swath passage

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 3: Essential Research Materials for Spray Drift Validation Studies

Category Specific Items Function/Application
Deposition Collectors Chromatography paper, Petri dishes, artificial foliage, string collectors Capture spray droplets for quantitative analysis of deposition patterns and mass
Tracer Materials Fluorescent dyes (Rhodamine WT, Pyranine), inorganic tracers (potassium chloride) Enable quantification of deposition without pesticide analysis; simplify laboratory processing
Droplet Characterization Water-sensitive paper, oil-sensitive paper, laser diffraction systems Measure droplet size spectrum which critically influences drift potential
Meteorological Instruments 3-cup anemometers, sonic anemometers, wind vanes, temperature/RH sensors Document environmental conditions during applications for model input and data interpretation
Application Equipment Calibrated nozzles, pressure gauges, flow meters, height markers Ensure precise and consistent application parameters for reproducible experiments
Analytical Equipment Fluorometers, HPLC systems, GC-MS, conductivity meters Quantify tracer or active ingredient concentration in deposition samples

Conceptual Framework for Addressing Model Limitations

Researchers should adopt a conceptual framework that acknowledges the inherent limitations of the AGDISP model while maximizing its utility for scientific investigation.

G Figure 2: AGDISP Constraint Management Framework Input Model Inputs • Application parameters • Meteorological data • Release characteristics Processing Model Processing • Lagrangian near-field • Gaussian far-field • Deposition algorithms Input->Processing Output Model Outputs • Depiction predictions • Concentration estimates • Drift mass loading Processing->Output Constraints Identified Constraints • No physical barriers • Limited far-field validation • Specific use cases only • Terrain limitations Mitigation Constraint Mitigation • Define application boundaries • Conduct field validation • Apply uncertainty factors • Document assumptions Constraints->Mitigation Mitigation->Input Mitigation->Processing Mitigation->Output Boundary Application Boundary: Use only within validated parameters and spatial domains Boundary->Processing

This framework emphasizes several critical approaches for working with AGDISP constraints:

  • Input Parameter Management: Carefully document all input parameters and their uncertainties, particularly regarding swath characteristics for UAS applications
  • Application Boundary Adherence: Strictly limit model application to validated use cases and spatial domains as defined in EPA guidance
  • Constraint Mitigation: Implement field validation studies specific to your research context to quantify local model performance
  • Uncertainty Communication: Explicitly document and communicate model limitations in all research outputs and conclusions

The AGDISP model represents a valuable tool for predicting pesticide spray drift, but researchers must operate within its documented constraints. The EPA guidance clearly outlines specific limitations regarding spatial application, physical barrier considerations, meteorological factors, and validated use cases. The most recent research indicates these limitations extend to newer implementations such as AGDISPpro for UAS applications, where parameters like swath width and displacement introduce significant uncertainty [3]. Researchers should implement the protocols and frameworks outlined herein to ensure scientifically rigorous application of AGDISP within its operational boundaries while working to expand its validated use cases through targeted field studies.

Validating AGDISP: Benchmarking Model Predictions Against Field Data

The integration of Remotely Piloted Aerial Application Systems (RPAAS), or spray drones, into modern agriculture necessitates robust, validated models for predicting off-target spray drift. Such models are crucial for regulatory risk assessments and for developing precise and environmentally responsible application protocols [3]. AGDISPpro has emerged as a key mechanistic model for this purpose, adapting an established Lagrangian-based drift and deposition model originally designed for conventional aerial applications to the unique aerodynamic profiles of multi-rotor systems [3] [4]. This document details the recent validation of AGDISPpro specifically for two RPAAS models—the PV22 quadcopter and PV35X hexacopter—and provides structured experimental data and protocols for researchers and scientists in the field.

Field Study Design for Model Validation

The validation of AGDISPpro for RPAAS applications was conducted through two comprehensive field studies, designed to test the model's performance across a range of realistic operational scenarios [3] [4].

Table 1: Summary of Field Studies for AGDISPpro Validation

Study Parameter Study No. 1: Single-Swath Study No. 2: Multi-Swath
Application Type Single-swath applications Four-swath applications
Spray Quality Tested Medium (M) and Extremely Coarse (XC) droplet spectra Fine (F) and Ultra Coarse (UC) droplet spectra
RPAAS Models PV22 quadcopter and PV35X hexacopter [3] PV22 quadcopter and PV35X hexacopter [3]
Primary Measurements In-swath and downwind off-target deposition [4] In-swath and off-target downwind deposition [4]
Model Performance (Index of Agreement) M: 0.47 - 0.92; XC: 0.61 - 0.94 [3] F: 0.86 - 0.93; UC: 0.48 - 0.55 [3]

Quantitative Performance Results

The core of the validation effort involved comparing AGDISPpro predictions against measured field deposition data. The model's performance was quantified using statistical indices of agreement.

Table 2: AGDISPpro Model Performance Summary by Spray Quality

Spray Quality Index of Agreement Range Summary of Model Performance
Fine (F) 0.86 to 0.93 [3] Very good agreement with field observations [4].
Medium (M) 0.47 to 0.92 [3] Performance varied, showing a wider range of agreement.
Extremely Coarse (XC) 0.61 to 0.94 [3] Good to excellent agreement with observations [4].
Ultra Coarse (UC) 0.48 to 0.55 [3] Performance was more limited and showed the poorest agreement [4].

A consistent observation across both studies was that AGDISPpro tended to under-predict off-target deposition values compared to field measurements, indicating a slight conservative bias in the model's current formulation [4].

Detailed Experimental Protocols

To ensure reproducibility and provide a framework for future validation work, the core methodologies from the field studies are outlined below.

Protocol 1: Single-Swath Drift Deposition Study

This protocol is designed to characterize the fundamental drift profile of a single RPAAS pass.

  • Site Selection: Choose a flat, open field with a consistent ground cover (e.g., short grass, bare soil) to facilitate the placement of deposition collectors. The site must have a long, unobstructed fetch in the prevailing wind direction.
  • Experimental Setup:
    • Lay out a grid of deposition collectors (e.g., petri dishes with filter paper, Mylar sheets) starting within the application swath and extending downwind for at least 100 meters.
    • Place meteorological stations upwind of the application area to record wind speed, wind direction, temperature, and relative humidity at regular intervals (e.g., every 1-5 seconds).
  • RPAAS Operation:
    • Configure the RPAAS (PV22 or PV35X) with the target nozzle (e.g., producing Medium or Extremely Coarse spray).
    • Program a single, straight flight path perpendicular to the prevailing wind direction.
    • Apply a tracer dye (e.g., Brilliant Blue FCF) mixed with water as the spray mixture for subsequent quantitative analysis.
  • Sample Collection & Analysis:
    • Immediately after the application, collect all deposition samples.
    • Analyze the samples in a laboratory using fluorometry or spectrophotometry to determine the mass of tracer dye per unit area.
    • Express deposition as a percentage of the applied rate at each downwind location.

G start Start Single-Swath Protocol p1 Site Selection & Setup (Flat terrain, downwind collector grid) start->p1 p2 RPAAS & Nozzle Configuration (Set spray quality, PV22/PV35X) p1->p2 p3 Meteorological Data Recording (Wind, temp, RH) p2->p3 p4 Execute Single Swath Application (Perpendicular to wind) p3->p4 p3->p4 Concurrently p5 Collect Deposition Samples (In-swath and downwind) p4->p5 p6 Laboratory Analysis (Fluorometry/Spectrophotometry) p5->p6 end Generate Deposition Profile p6->end

Figure 1: Single-swath deposition study protocol workflow.

Protocol 2: Multi-Swath Deposition and In-Field Uniformity Study

This protocol assesses deposition patterns and spray drift from a more realistic, multi-swath application.

  • Plot Definition and Setup:
    • Mark a rectangular plot large enough to accommodate multiple, adjacent swaths (e.g., 4 swaths).
    • Place deposition collectors in a grid pattern across the entire plot and extending downwind.
  • RPAAS and Application Setup:
    • Configure the RPAAS with the target nozzles (e.g., Fine or Ultra Coarse).
    • Program a multi-swath flight plan with the appropriate swath width and flight line separation.
  • Application Execution:
    • Conduct the multi-swath application using the pre-programmed autonomous flight path.
    • Ensure meteorological data is logged throughout the entire operation.
  • Data Processing:
    • Collect and analyze deposition samples as in Protocol 1.
    • Calculate the coefficient of variation (CV) for deposition within the plot to assess application uniformity.
    • Compare the measured downwind deposition profile with model predictions.

The Scientist's Toolkit: Research Reagents and Materials

For conducting field validation studies for aerial spray drift, specific reagents and materials are essential. The following table details key items and their functions.

Table 3: Essential Research Reagents and Materials for Drone Spray Validation

Item Function / Relevance in Research
Tracer Dye (e.g., Brilliant Blue FCF) A visible, environmentally inert dye used to quantitatively track spray deposition and drift via laboratory analysis [4].
Deposition Collectors (e.g., Mylar sheets, Petri dishes with filter paper) Standardized surfaces placed in the field to capture spray droplets for subsequent quantitative analysis [4].
AGDISPpro Software The validated mechanistic model used to predict off-target droplet movement and deposition from RPAAS applications [3] [1].
PV35X Hexacopter RPAAS A commercial, hexacopter-style Remotely Piloted Aerial Application System evaluated in the validation study, capable of a 25 lb liquid payload [3] [44].
PV22 Quadcopter RPAAS A commercial, quadcopter-style RPAAS evaluated in the validation study [3].
Droplet Size Spectra Nozzles (Fine to Ultra Coarse) Nozzles that produce specific, categorized droplet sizes (e.g., Fine, Medium, Extremely Coarse) to study the effect of droplet size on drift potential [3] [4].
Sonic Anemometer / Meteorological Station Critical for recording high-frequency wind speed, wind direction, and other atmospheric data that directly influence spray dispersion and deposition [4].

Critical Analysis and Research Applications

Key Findings and Model Limitations

The validation study concluded that AGDISPpro is a promising tool for modeling off-target spray drift from RPAAS [3] [4]. The model performed particularly well for Fine and Extremely Coarse spray qualities, but showed reduced accuracy for Medium and Ultra Coarse sprays [4]. A primary source of uncertainty and a key limitation identified was the accurate determination of effective swath width and swath displacement during RPAAS operations [3] [4]. These parameters significantly influence the model's prediction of the peak deposition magnitude and the position of the deposition plume. This highlights a critical area for future research to improve model accuracy.

Application in Regulatory and Product Development Context

For researchers and professionals in pesticide development and regulation, this validation provides a scientifically-grounded method to assess the environmental impact of RPAAS applications.

  • Regulatory Risk Assessment: Regulatory bodies like the U.S. EPA use drift models such as AGDISP in ecological and human health risk assessments [1] [17]. A validated model for drones allows for more accurate exposure estimates, which can lead to more informed and potentially less restrictive label requirements for drone-applied products [17].
  • Product Development and Stewardship: The model can be used to optimize application parameters (e.g., nozzle selection, flight altitude, operational swath width) for new pesticide formulations during the development process, minimizing off-target movement and enhancing on-target efficacy.

G agdisp AGDISPpro RPAAS Model app1 Optimize Application Parameters (Nozzle type, flight height) agdisp->app1 app2 Predict Off-Target Deposition agdisp->app2 app3 Refine Swath Width Definition agdisp->app3 outcome1 Improved Product Stewardship app1->outcome1 outcome2 Accurate Exposure Estimates app2->outcome2 app3->outcome2 outcome3 Informed Pesticide Labeling outcome2->outcome3

Figure 2: Research and regulatory applications of the AGDISPpro model.

Accurately predicting off-target pesticide spray drift is critical for environmental protection and application efficiency. This analysis evaluates the performance of the AGDISP model and its variants in predicting deposition across different droplet sizes, a key parameter influencing drift. The AGDISP model, developed by the U.S. Forest Service and used by the U.S. Environmental Protection Agency (EPA) for ecological and human health risk assessments, simulates the release, dispersion, and deposition of sprays from various application platforms [1] [45]. Understanding the agreement between model predictions and empirical observations is fundamental for model refinement and regulatory reliability, particularly as new application technologies like Unmanned Aerial Systems (UAS) emerge.

The following tables consolidate key quantitative findings from recent validation studies, highlighting the relationship between droplet size and model accuracy.

Table 1: AGDISPpro Validation Performance for UAS Applications (2025 Study)

Droplet Size Spectrum (DSD) Index of Agreement Range Number of Tests (n) General Prediction Trend vs. Observations
Medium 0.47 to 0.92 9 Under-prediction in off-target zones
Extremely Coarse 0.61 to 0.94 12 Under-prediction in off-target zones
Fine 0.86 to 0.93 3 Under-prediction in off-target zones
Ultra Coarse 0.48 to 0.55 3 Under-prediction in off-target zones

Source: Adapted from Castro-Tanzi et al., 2025 [4]. The Index of Agreement is a statistical measure where 1 indicates perfect agreement between predictions and observations.

Table 2: Wind Tunnel Evaluation of Droplet Size on Spray Penetration in Soybean Canopy

Droplet Size Class Nozzle Example Spray Coverage Relative Performance (Top & Middle Canopy) Spray Penetration to Lower Canopy
Medium XR11004 Highest Significantly Low (A challenge for all droplet sizes)
Coarse TTJ6011004 Higher Significantly Low (A challenge for all droplet sizes)
Very Coarse AITTJ6011004 Lower Significantly Low (A challenge for all droplet sizes)
Extremely Coarse AI11004 Lowest Significantly Low (A challenge for all droplet sizes)

Source: Adapted from Castilho Theodoro et al., 2025 [46]. Testing conducted at 275 kPa pressure and 2 m/s wind speed.

Table 3: Effect of UAV Operational Parameters on Droplet Deposition in Pigeon Pea Crop

Operational Parameter Optimum Value Impact on Deposition at Canopy Zones (vs. Higher Speeds/Heights)
Flight Speed 2 m/s Increased droplet density, coverage, and deposition across top, middle, and bottom zones.
Flight Height 1.5 m (above canopy) Increased droplet density, coverage, and deposition across top, middle, and bottom zones.
Resulting Deposition At 1.5m & 2 m/s Droplet Density: 54.00 (top), 50.17 (middle), 46.33 (bottom) droplets/cm²
Coverage: 10.53% (top), 10.09% (middle), 9.78% (bottom)

Source: Adapted from effect of operational parameters on spray performance, 2025 [47].

Experimental Protocols for Drift Deposition Studies

To ensure the reproducibility and reliability of data used in model validation, standardized experimental protocols are essential. The following methodologies are commonly employed in the field.

Protocol for Field Validation of Drift Models (e.g., AGDISPpro)

This protocol outlines the procedure for generating field data to validate spray drift model predictions for aerial applications, including UAS.

  • Primary Objective: To collect empirical data on downwind spray deposition from a controlled application for comparison with model-predicted values.
  • Site Preparation: A flat, open field with minimal obstructions is selected. Meteorological stations are installed to record wind speed, wind direction, temperature, and relative humidity at high frequency (e.g., 0.5 Hz) for the duration of the trial [48] [4].
  • Application Setup: The application platform (e.g., spray drone, conventional aircraft) is equipped with specific nozzles calibrated to produce the target droplet size spectrum (e.g., Fine, Medium, Coarse, Extremely Coarse). Key parameters like release height, spray pressure, and flight speed are recorded [4] [47].
  • Tracer and Deposition Sampling: A non-toxic fluorescent tracer is mixed with the spray tank solution. A series of passive collectors (e.g., petri dishes, filter paper strips) or active air samplers are placed along transects at multiple distances downwind from the application swath (e.g., from 10 m to beyond 200 m) [4].
  • Sample Processing and Analysis: After application, deposition samples are collected. The tracer is washed off and its concentration is quantified using fluorometry. The measured deposition values (e.g., µL/cm² or ng/cm²) are then compared to the model's predictions for the same downwind distances and application conditions [4].

Protocol for Wind Tunnel Evaluation of Droplet Deposition

This protocol is designed for controlled, repeatable evaluation of how droplet size and other parameters affect spray penetration into plant canopies.

  • Primary Objective: To systematically quantify the influence of droplet size on spray deposition and penetration within a plant canopy under controlled wind conditions.
  • Experimental Setup: An open-circuit laminar-airflow wind tunnel is used. Nozzles producing different droplet size classes (e.g., Medium, Coarse, Very Coarse, Extremely Coarse) are mounted on a static spray boom within the tunnel test section [46].
  • Canopy Simulation: Potted plants (e.g., soybeans at a specific growth stage) are arranged on the wind tunnel floor in rows to simulate field spacing. Water-sensitive papers (WSPs) are positioned at predetermined heights (top, middle, bottom) within the plant canopy and at various distances downwind from the nozzle [46].
  • Spray Operation and Data Collection: The nozzle is activated for a short, precise duration (e.g., 3 seconds) at a constant pressure while the wind tunnel maintains a steady wind speed (e.g., 2 m/s). After the spray, the WSPs are collected [46].
  • Image and Data Analysis: The WSPs are scanned, and image analysis software is used to determine spray coverage (%), droplet density (droplets/cm²), and droplet size distribution on the papers. This provides a quantitative measure of spray penetration and deposition at different canopy levels [46].

Workflow and Model Validation Diagram

The following diagram illustrates the logical workflow for conducting a model validation study, from experimental design to model refinement.

G Start Define Validation Objectives P1 Design Experiment Start->P1 P2 Conduct Field/Wind Tunnel Trial P1->P2 P3 Collect Observed Deposition Data P2->P3 P7 Statistical Comparison (Predicted vs. Observed) P3->P7 Observed Data P4 Configure Model (AGDISP/AGDISPpro) P5 Run Simulation P4->P5 P6 Generate Predicted Deposition Data P5->P6 P6->P7 Predicted Data P8 Identify Discrepancies & Uncertainties P7->P8 Agreement? End Improved Model P7->End Good Agreement P9 Refine Model Parameters/Assumptions P8->P9 P9->P4 Feedback Loop

Diagram 1: Spray Drift Model Validation Workflow. This flowchart outlines the iterative process of validating a mechanistic spray drift model like AGDISP, highlighting the critical comparison between field observations and model predictions.

The Scientist's Toolkit: Research Reagents and Materials

Essential materials and tools for conducting spray deposition and drift studies are listed below.

Table 4: Essential Research Reagents and Materials for Spray Drift Studies

Item Function in Experiment
Fluorescent Tracer (e.g., Pyranine) A safe, water-soluble additive mixed with the spray solution to allow for sensitive quantification of deposition levels on collectors.
Water-Sensitive Papers (WSPs) A passive sampling medium coated with a yellow dye that turns blue upon contact with water droplets. Used for rapid assessment of droplet density, coverage, and size distribution.
Wind Tunnel A controlled laboratory environment with laminar airflow used to study the fundamental effects of wind speed, droplet size, and canopy structure on deposition without variable weather conditions.
Agricultural Spray Nozzles Nozzles (e.g., flat fan, air induction) engineered to produce specific droplet size spectra (from Fine to Extremely Coarse) when operated at defined pressures.
Meteorological Station Measures and records critical environmental parameters (wind speed, direction, temperature, relative humidity) during field trials, which are essential for configuring the model.
Fluorometer An analytical instrument used to measure the fluorescence intensity of tracer compounds washed from deposition samples, enabling precise quantification of spray volume deposited.

Discussion and Future Directions

The comparative data reveals that while AGDISPpro shows promise, particularly for Fine and Extremely Coarse sprays, it exhibits a consistent trend of under-predicting off-target deposition [4]. This underscores inherent uncertainties in modeling, such as accurately defining the effective swath width and accounting for downwind swath displacement in UAS applications [4]. Furthermore, droplet size significantly impacts canopy penetration, with medium and coarse droplets generally providing better coverage on the top and middle of soybean canopies, though penetration to the lower canopy remains a challenge across all sizes [46].

Ongoing efforts like the AGDISP Modernization Project (AMP) aim to address these limitations by rewriting the legacy code, making it open-source, and enhancing its ability to incorporate modern Drift Reduction Technologies [45]. This will be crucial for improving the model's accuracy for all application types (aerial, ground, UAS) and ensuring that risk assessments by regulatory bodies like the EPA reflect real-world conditions and technological advancements.

The fidelity of the AGDISP model in predicting spray deposition is demonstrably influenced by droplet size and application context. Validation studies are essential for identifying model weaknesses, such as the systematic under-prediction of drift for certain droplet spectra. The standardized protocols and tools outlined in this document provide a framework for generating high-quality validation data. Continued refinement of models through initiatives like AMP, coupled with rigorous empirical testing, is paramount for advancing the science of pesticide application, minimizing environmental impact, and supporting sustainable agricultural practices.

Accurately predicting pesticide spray drift is critical for safeguarding environmental health and ensuring the efficacy of agricultural applications. The AGDISP model is a principal tool used for forecasting drift behavior, but its performance must be rigorously validated against empirical data across the entire spectrum of spray qualities, from fine to ultra-coarse. This document provides detailed application notes and protocols for quantifying the agreement between AGDISP model predictions and observed drift deposition, serving as a vital chapter within a broader thesis on advancing spray drift prediction. We present standardized metrics, detailed experimental methodologies for generating validation data, and a comprehensive toolkit for researchers to benchmark model performance effectively.

Quantitative Index of Agreement Metrics

The performance of spray drift models like AGDISP must be evaluated using a suite of statistical metrics that capture different aspects of the agreement between predicted and observed values. The following table summarizes the key quantitative indices recommended for this purpose.

Table 1: Key Quantitative Metrics for Model Performance Evaluation

Metric Name Formula Ideal Value Interpretation in Spray Drift Context
Coefficient of Determination (R²) R² = 1 - (SS₍res₎/SS₍tot₎) 1 Proportion of variance in deposition data explained by the model. An R² > 0.75 is often indicative of a good fit [24].
Root Mean Square Error (RMSE) RMSE = √(Σ(Pᵢ - Oᵢ)²/n) 0 Absolute measure of average model error, expressed in deposition units (e.g., ng/cm² or % applied).
Normalized RMSE (RSR) RSR = RMSE / STDEV₍obs₎ 0 Relative measure of error. An RSR ≤ 0.26 indicates very good model performance [49].
Nash-Sutcliffe Efficiency (NSE) NSE = 1 - [Σ(Oᵢ - Pᵢ)² / Σ(Oᵢ - Ō)²] 1 Indicates how well predictions match observations compared to the mean of observations. Values closer to 1 are superior.
Percent Bias (PBIAS) PBIAS = [Σ(Oᵢ - Pᵢ) / Σ(Oᵢ)] * 100 0 Measures average tendency of predictions to be larger or smaller than observations. Positive values indicate under-prediction.

These metrics should be applied not only to the total drift mass but also to the bivariate size-velocity distributions of droplets, where advanced measurement techniques like Digital In-line Holography (DIH) provide the necessary high-resolution data [50].

Experimental Protocols for Drift Data Generation

Validating AGDISP requires high-quality, empirical drift data. The following protocols outline standardized methods for collecting this data in wind tunnel and field settings.

Wind Tunnel Drift Measurement Protocol

Wind tunnels provide a controlled environment for evaluating fundamental drift potential.

1. Principle: To measure the downwind deposition of spray droplets from a stationary nozzle under controlled wind, temperature, and humidity conditions [24].

2. Key Equipment and Reagents:

  • Wind Tunnel: Equipped with a fan, flow straighteners, and a controlled environment system.
  • Spray Nozzles: A set of reference nozzles covering ASABE/ISO spray quality classes (e.g., Fine, Medium, Coarse, Ultra-Coarse) [50].
  • Passive Collectors: Polyethylene lines or string collectors (e.g., yarn-type strings with high recovery rates) [51].
  • Spray Solution: Deionized water mixed with a tracer (e.g., Tartrazine dye) or a pesticide formulation at field-relevant concentration.
  • Analytical Equipment: Spectrophotometer or Total Organic Carbon (TOC) analyzer.

3. Procedure: a. Setup: Secure the test nozzle at the designated height in the spraying section. Place passive collectors at multiple downwind distances and heights in the collecting section. b. Environmental Stabilization: Set and allow the wind speed (e.g., 1-3 m/s), temperature, and relative humidity to stabilize [51]. c. Application: Activate the nozzle for a precise duration (e.g., 10 seconds) using a spray pressure corresponding to the desired droplet spectrum. d. Sample Collection: Carefully retrieve the collectors after application. e. Analysis: Rinse the deposited tracer from each collector into a known volume of deionized water. Analyze the tracer concentration using a spectrophotometer or TOC analyzer [51]. f. Data Calculation: Calculate the deposition volume per unit area (mL/cm² or ng/cm²) for each collector location.

The workflow for this protocol is systematized below:

G Start Start Wind Tunnel Protocol Setup Setup Nozzle and Collectors Start->Setup EnvControl Stabilize Environmental Conditions (Wind, Temp, RH) Setup->EnvControl Spray Execute Spray Application EnvControl->Spray Collect Retrieve Passive Collectors Spray->Collect Analysis Analyze Tracer Deposition (Spectrophotometer/TOC) Collect->Analysis DataCalc Calculate Deposition per Unit Area Analysis->DataCalc End Dataset for Model Validation DataCalc->End

Figure 1: Workflow for the wind tunnel drift measurement protocol.

Field Drift Measurement Protocol (ISO 22866-Based)

Field trials assess drift under real-world conditions and are essential for final model validation.

1. Principle: To measure spray drift deposition downwind of a moving sprayer (ground or aerial) under ambient meteorological conditions [52].

2. Key Equipment and Reagents:

  • Sprayer: A calibrated ground boom sprayer or Uncrewed Aerial Spray System (UASS).
  • Passive Collectors: Nylon screens, polyethylene lines, or water-sensitive paper (WSP) placed on stakes.
  • Meteorological Station: To record wind speed, wind direction, temperature, and humidity at 1-2 Hz.
  • Spray Solution: As per wind tunnel protocol.

3. Procedure: a. Site Preparation: Establish a sampling line at a specified distance (e.g., 5-50 m) downwind from the treatment area, parallel to the spray swath. Place collectors at multiple heights (e.g., 0.5 m, 1.0 m, 1.5 m) [53]. b. Meteorological Monitoring: Continuously record meteorological data throughout the trial. Proceed only if the wind direction is within ±15° of perpendicular to the spray swath and speed is within acceptable limits (e.g., 1-5 m/s) [52]. c. Application: Operate the sprayer according to the test parameters (nozzle type, pressure, speed, release height). d. Sample Collection: Retrieve collectors immediately after application to prevent degradation. e. Analysis: Process samples as in the wind tunnel protocol. For WSP, use image analysis software (e.g., DepositScan) to determine droplet coverage (%), density (droplets/cm²), and volume median diameter (VMD) [53].

The Scientist's Toolkit: Essential Research Reagents and Materials

Successful experimentation in spray drift research relies on a set of key materials and instruments.

Table 2: Essential Research Reagent Solutions and Materials

Item Name Function/Description Application Note
Reference Nozzles Generate standardized droplet spectra (Fine to Ultra-Coarse). ASABE/ISO standard reference nozzles (e.g., XR11003 for fine, TT11004 for coarse) are crucial for model calibration [50].
Passive String Collectors Intercept and retain airborne droplets for quantitative analysis. Yarn-type strings offer high collection efficiency (>90%) and are suitable for field use [51].
Water-Sensitive Paper (WSP) Qualitatively and quantitatively assess droplet deposition patterns. Provides data on coverage (%) and droplet density; analyzed with DepositScan software [53].
Tracer Compounds (e.g., Tartrazine) Serve as a quantifiable surrogate for pesticide active ingredients. Allows for precise chemical quantification of deposition without the hazards of pesticide handling.
Total Organic Carbon (TOC) Analyzer Quantifies the amount of spray solution collected by measuring organic carbon content. A highly sensitive method for analyzing deposition on string collectors and other surfaces [51].
3D LiDAR Sensor A non-contact remote sensor for measuring spray drift cloud in three dimensions. Provides rich spatial data on drift plumes; correlates well with passive collector data (R² > 0.75) [24].

AGDISP Model Validation Workflow

Integrating the components above creates a robust framework for validating and refining the AGDISP model. The process is iterative, feeding experimental results back into model improvement.

G Input Input: Spray & Environmental Parameters (Nozzle, Pressure, Wind) AGDISP Run AGDISP Simulation Input->AGDISP Experiment Experimental Drift Measurement Input->Experiment Prediction Predicted Drift Deposition AGDISP->Prediction Metrics Calculate Agreement Metrics (R², RSR, NSE, PBIAS) Prediction->Metrics Observation Observed Drift Deposition Experiment->Observation Observation->Metrics Eval Performance Evaluation Metrics->Eval Refine Refine/Calibrate Model Eval->Refine Refine->Input

Figure 2: The iterative workflow for AGDISP model validation and refinement.

This document has established a standardized framework for quantifying the performance of the AGDISP spray drift model. By employing the outlined index of agreement metrics, adhering to the detailed experimental protocols for wind tunnel and field studies, and utilizing the essential research toolkit, scientists can generate reliable, reproducible data for model validation. This rigorous, metrics-driven approach is fundamental to improving the predictive accuracy of AGDISP, ultimately leading to better risk assessments and more sustainable pesticide application practices. Future work should focus on integrating high-fidelity data from machine learning-enhanced diagnostics [50] and remote sensing technologies [24] to further push the boundaries of model precision across all spray qualities.

The AGDISP model represents a critical advancement in predicting pesticide spray drift from aerial applications, integrating a Lagrangian approach for near-field predictions (up to 800 meters) with a Gaussian extension to estimate far-field drift up to 20 kilometers downwind [8]. This extension is particularly vital for ecological risk assessments under the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA), where understanding potential far-field drift of pesticides like mosquito adulticides and forestry applications is essential [8]. The model's performance under straight-line wind conditions, characterized by consistent wind direction and minimal topological interference, provides a foundational scenario for validation, though it introduces specific conservatisms and uncertainties.

The validation in environments with straight-line winds and minimal ground cover, such as high desert regions, offers a controlled setting to evaluate the model's core algorithms without the complicating effects of wind barriers or turbulent terrain [8]. However, this same scenario underscores one of the model's significant limitations: the assumption of no wind barriers and the non-consideration of meteorological variables like crosswinds and humidity over long distances, resulting in predictions that are inherently conservative for regulatory purposes [8]. This document details the application notes and experimental protocols for validating the AGDISP-Gaussian extension under these specific conditions, providing researchers with a framework for assessing model performance and interpreting results within a broader thesis on pesticide spray drift prediction.

Background and Model Fundamentals

AGDISP Model Architecture

AGDISP is a "first-principles" science-based model developed by the USDA Forest Service to predict spray drift from application sites [1]. Its architecture is composed of two primary computational frameworks:

  • Lagrangian Model: This component simulates the near-field release, dispersion, and deposition of spray droplets up to 800 meters (0.5 miles) downwind from the application site. It uses a physics-based approach to account for droplet dynamics, including evaporation, sedimentation, and atmospheric dispersion [8] [54].
  • Gaussian Extension Model: This module extends the prediction capability to 20 kilometers (12.4 miles) downwind by applying Gaussian plume equations and dispersion coefficients similar to those used in the Industrial Source Complex (ISC) model [8]. The transition between the Lagrangian and Gaussian models occurs at approximately 800 meters, where the spray cloud is assumed to have become sufficiently diffuse for Gaussian principles to apply effectively [54].

The Gaussian extension was initially integrated into AGDISP specifically to estimate drift of Bacillus thuringiensis (Bt) bacteria from spray applications to forests, indicating its particular relevance for applications involving fine droplets released at significant heights above ground [8]. The model accounts for the total mass of applied pesticide, addressing concerns about plume depletion that might otherwise lead to overestimation of downwind concentrations [8].

Conceptual Workflow and Model Transition

The following diagram illustrates the conceptual workflow of the linked AGDISP-Gaussian model system, highlighting the critical transition point between the near-field and far-field computational approaches:

G AGDISP-Gaussian Model Workflow Start Spray Application Parameters NearField Lagrangian Module (Near-Field: 0-800m) Start->NearField Decision Distance > 800m? NearField->Decision Decision->NearField No FarField Gaussian Extension (Far-Field: 800m-20km) Decision->FarField Yes Output Drift Deposition Predictions FarField->Output

Validation Studies in Straight-Line Wind Conditions

Key Validation Experiments

The AGDISP-Gaussian extension has been evaluated under straight-line wind conditions in several critical studies that provide the foundation for its use in ecological risk assessments. The primary validation research was conducted in environments characterized by level terrain with minimal ground cover, where straight-line winds and few competing topological factors dominate.

Teske and Thistle (2004) High Desert Validation: This seminal study conducted in the high desert of Utah demonstrated reasonable agreement between AGDISP-Gaussian predictions and observed field data across distances ranging from 0.1 km to 10 km [8]. The straight-line wind conditions in this environment provided an ideal setting for validating the Gaussian component, as they minimized the unpredictable effects of turbulence and wind barriers that complicate model verification in more complex terrain. The research specifically focused on aerial applications over forests, which aligns with the model's intended use cases [8].

Woods et al. (2001) Cotton Application Study: This validation effort compared model predictions with field observations from 0.01 km to 1 km, establishing the foundation for near-field predictions that inform the transition to the Gaussian extension [8]. While covering a shorter distance range, this study provided important data on initial deposition patterns that influence far-field predictions.

Bird et al. (2002) Comprehensive Model Evaluation: A systematic evaluation of AgDISP algorithms (incorporated into the AgDRIFT model) compared model simulations with 161 separate field trials of typical agricultural aerial applications [22] [55]. This large-scale validation identified that while the model responds appropriately to application variables like droplet size, application height, and wind speed, it tends to overpredict deposition rates in far-field distances, particularly under evaporative conditions [22].

Table 1: Summary of AGDISP-Gaussian Extension Validation Studies

Study Reference Validation Distance Range Environmental Conditions Key Findings Model Performance
Teske and Thistle (2004) [8] 0.1 km to 10 km High desert Utah, straight-line winds, minimal terrain Reasonable agreement between predicted and observed concentrations Good predictive capability in simple terrain
Woods et al. (2001) [8] 0.01 km to 1 km Agricultural conditions Established near-field baseline for Gaussian transition Reliable near-field predictions
Bird et al. (2002) [22] Multiple distances up to 2+ km 161 field trials, various conditions Overprediction in far-field, especially under evaporation Conservative for risk assessment

Quantitative Drift Predictions Under Straight-Line Winds

Recent field research has provided quantitative data on spray drift under straight-line wind conditions that can inform validation protocols. A 2022 study on herbicide spray drift compared ground and aerial applications under consistent wind speeds averaging 13 kph, demonstrating the significant impact of release height on downwind deposition [5].

Table 2: Measured Spray Drift from Aerial vs. Ground Applications in 13 kph Winds [5]

Application Method Release Height Droplet Spectrum PD₂₅ (m) PD₅₀ (m) PD₉₀ (m) Fold Increase vs. Ground
Aerial Application ~50 ft Coarse 7.55 10.07 20.54 5.0-8.6x
Ground Application ~10 ft Coarse 0.23 0.50 2.36 -

This research demonstrated that even with identical droplet size spectra and similar wind speeds, the aerial application resulted in significantly greater downwind drift due to the higher release height, highlighting the critical importance of application parameters in model validation [5]. The study also found that aerial applications would require three to five swath width adjustments upwind to reduce drift potential to levels comparable with ground applications [5].

Experimental Protocols for Far-Field Validation

Field Validation Methodology

Validating the AGDISP-Gaussian extension under straight-line wind conditions requires carefully controlled field experiments that isolate the specific meteorological factors of interest. The following protocol outlines the standard methodology for collecting validation data:

Site Selection Criteria: Choose locations with level terrain and minimal ground cover, such as high desert environments or extensive agricultural fields with minimal wind barriers [8]. The site should permit sampling along a straight line downwind from the application area for distances up to at least 10 kilometers, with 20 kilometers being ideal for full model validation [8].

Meteorological Monitoring: Implement continuous monitoring of wind speed, wind direction, temperature, and relative humidity at multiple heights throughout the experimental period. The ideal straight-line wind condition is defined as consistent wind direction (deviation less than 30 degrees from the established collection line) with speeds between 10-15 mph (16-24 kph) [8] [5]. Monitoring should begin at least one hour before application and continue until at least two hours after application.

Application Parameters: Conduct aerial applications using aircraft equipped with appropriate spray systems capable of generating ASAE droplet size spectra classified as very fine to fine [8]. Maintain a minimum release height of 50 feet above ground level, with at least 10 feet clearance above any plant canopy [8]. For forestry applications, this may require higher release altitudes. Application swaths should be clearly defined, with consideration for upwind swath adjustments to minimize immediate downwind contamination [5].

Drift Collection Methods: Position collection stations along transects extending from 100 meters to 20 kilometers downwind. Each station should include:

  • Mylar cards or similar non-absorbent surfaces for quantifying deposition (expressed as percent of applied rate) [5]
  • Water sensitive cards for measuring droplet coverage (%) and number of deposits per cm² [5]
  • Air samplers for capturing aerosolized particles at distances beyond 2 kilometers

Sample Processing and Analysis: Collect samples immediately after application and analyze using appropriate analytical methods (e.g., fluorometry for tracers, chemical analysis for specific pesticides). Compare measured deposition with model predictions using statistical measures including root mean square error, geometric mean, and geometric standard deviation [22].

Model Input Parameters and Scenarios

For consistent validation against field data, researchers should implement the following standardized input parameters in AGDISP simulations:

Table 3: Essential AGDISP Input Parameters for Far-Field Validation

Parameter Category Specific Variables Validation Settings Notes
Application Method Application type, aircraft, speed Aerial, fixed-wing or helicopter Avoid ground applications with Gaussian extension [8]
Spray Characteristics Droplet size spectrum, flow rate ASAE Very Fine to Fine Critical for far-field predictions [8]
Release Parameters Height above ground, boom geometry ≥50 ft release height Minimum 10 ft above canopy [8]
Meteorological Conditions Wind speed, direction, temperature, humidity 10-15 mph winds, straight-line Consistent conditions essential [8]
Site Geometry Field size, shape, topography Level terrain, minimal barriers Straight-line wind assumption [8]

Data Analysis and Model Performance Metrics

The following workflow outlines the recommended process for comparing field measurements with model predictions and evaluating model performance:

G Validation Data Analysis Workflow FieldData Field Data Collection (Deposition Measurements) Comparison Statistical Comparison Regression Analysis FieldData->Comparison ModelInput AGDISP Simulation (Standard Input Parameters) ModelInput->Comparison Performance Performance Metrics Calculation Comparison->Performance Validation Model Validation Assessment Performance->Validation

Performance Metrics: Calculate the following quantitative measures to assess model accuracy:

  • Prediction within Factor of Two (FOT): The percentage of predictions that fall within a factor of two of observed values [55]. For AgDRIFT, this metric achieved 90% for buffer zone estimations [55].
  • Regression Parameters: Slope, intercept, and correlation coefficient (R²) from linear regression of predicted versus observed values.
  • Geometric Mean (MG) and Geometric Standard Deviation (SG): Robust measures of central tendency and variance for log-normally distributed deposition data [22].

Research Reagent Solutions and Essential Materials

Successful validation of the AGDISP-Gaussian extension requires specific materials and analytical tools. The following table details essential research reagents and their applications in experimental protocols:

Table 4: Research Reagent Solutions for Spray Drift Validation Studies

Category Specific Item Function/Application Validation Context
Spray Tracers Fluorescent dyes (e.g., Brilliant Sulfaflavine) Quantitative deposition analysis Enable precise measurement of drift deposition on collectors [5]
Drift Collectors Mylar cards, nylon screens, filter papers Capture spray droplets for analysis Standardized collection surfaces for comparative studies [5]
Droplet Characterization Water-sensitive cards, optical laser systems Measure droplet size spectrum and density Critical for verifying application parameters and model inputs [5]
Meteorological Equipment 3D sonic anemometers, temperature/RH sensors Characterize atmospheric conditions Essential for validating straight-line wind assumptions [8]
Analytical Instruments Fluorometers, HPLC-MS systems Quantify tracer or pesticide concentrations Provide precise measurement of deposition levels [5]
Validation Software Statistical analysis packages (R, Python) Compare predicted vs. observed values Enable calculation of performance metrics [22]

Limitations and Uncertainty Considerations

While the AGDISP-Gaussian extension provides valuable predictive capability under straight-line wind conditions, researchers must acknowledge and account for several significant limitations and uncertainties in their validation studies:

Conservatism in Predictions: The current Gaussian extension modeling does not consider physical barriers (e.g., trees, topographic features) or meteorological variables (e.g., crosswinds, humidity variations) beyond the application site [8]. This results in predictions that are "extremely conservative" for ecological risk assessment, typically overestimating far-field deposition, particularly under evaporative conditions [8] [22].

Validation Distance Constraints: While the Gaussian extension can predict drift up to 20 kilometers, field validation beyond 2 miles (approximately 3.2 kilometers) remains limited [8]. The model's performance at the farthest distances therefore relies heavily on theoretical extrapolation rather than empirical verification.

Evaporative Effects: Research has demonstrated that AgDISP/AgDRIFT is highly sensitive to evaporative effects, with modeled deposition in the far-field responding to wet bulb depression in ways that field observations do not always support [22] [55]. This can lead to overprediction of deposition rates under certain atmospheric conditions.

Droplet Size Dependency: The Gaussian extension is recommended only for applications producing very fine to fine droplet spectra, particularly aerial mosquito adulticides and forestry applications [8]. Applications with larger droplet sizes or ground-based equipment fall outside the validated range for the far-field extension.

Mass Accounting Considerations: While the Gaussian plume equations in AGDISP initially appeared not to account for plume depletion (potentially leading to overestimation of downwind concentrations), the model developer has confirmed that the algorithm does account for the mass of applied pesticide [8]. Nevertheless, understanding the specific mechanisms of mass accounting remains important for proper interpretation of results.

The validation of the AGDISP-Gaussian extension under straight-line wind conditions provides a critical foundation for assessing the model's performance in predicting far-field pesticide drift. The controlled conditions of level terrain with minimal ground cover allow researchers to isolate specific model parameters and evaluate core algorithms without the confounding effects of complex topography or turbulent wind patterns. Current validation studies demonstrate that while the model shows reasonable agreement with field observations in near-field conditions and conservative predictions in far-field scenarios, significant uncertainties remain, particularly beyond 2 miles and under evaporative conditions.

The experimental protocols and application notes outlined in this document provide researchers with a standardized framework for conducting future validation studies, ensuring comparability across different research initiatives and contributing to the continued refinement of the AGDISP model system. As pesticide applications continue to evolve and ecological risk assessments require increasingly precise exposure estimates, further validation of the Gaussian extension under a broader range of environmental conditions remains an essential research priority. The straight-line wind condition validation represents not an endpoint, but rather a foundational step in establishing the model's credibility for regulatory decision-making.

Accurately predicting pesticide spray drift is a critical component of environmental risk assessment and sustainable agricultural practices. The AGDISP model, developed with initial support from the U.S. Forest Service, serves as a key regulatory tool for simulating the movement of spray after its release from various application equipment [17]. This analysis compares the drift potential from aerial and ground applications, leveraging recent field data and model validation studies to provide researchers and application professionals with evidence-based protocols for drift mitigation. Understanding these differences is essential for protecting non-target ecosystems, particularly pollinators and sensitive vegetation downwind from application sites.

Quantitative Drift Comparison: Aerial vs. Ground Applications

Field measurements and model simulations consistently demonstrate significantly higher drift potential from aerial applications compared to ground applications. The data reveal substantial differences in both drift deposition volumes and downwind injury to non-target crops.

Table 1: Comparative Spray Drift Metrics from Field Experiments

Metric Aerial Application Ground Application Fold Increase (Aerial vs. Ground)
Average Drift Deposition (PD50)* 10.07 m 0.50 m 20.1x [5]
Drift Deposition (WSC deposits) 5.0 to 8.6-fold increase Baseline 5.0-8.6x [5]
Soybean Injury (downwind) 1.7 to 3.6-fold increase Baseline 1.7-3.6x [5]
Cumulative Drift Deposition Fraction 20-60% of applied mass Typically <3% (up to 5% for grapes/fruit trees) ~10x [56]
Reproductive Structure Reduction (at 30.5 m) ~25% ~25% Comparable at this distance [5]
Reproductive Structure Reduction (at 61 m) Nearly 100% Significant but less than aerial Substantially higher [5]

*PD50: Predicted downwind distance for 50% reduction in spray drift deposition.

A field study using the herbicide florpyrauxifen-benzyl applied in a 13 kph wind speed demonstrated that aerial applications resulted in a 5.0- to 8.6-fold increase in downwind drift compared to ground applications, using identical Coarse spray quality classifications [5]. This translated to a 1.7- to 3.6-fold increase in downwind soybean injury [5]. The same research noted that soybean reproductive structures were severely reduced following herbicide exposure, with approximately 25% reduction up to 30.5 meters downwind and nearly 100% reduction at 61 meters for aerial applications [5]. This reduction in flowering has significant implications for potential pollinator foraging sources in agricultural landscapes.

AGDISP Model Framework and Validation

AGDISP (Agricultural DISPersal) is a mechanistic, Lagrangian-based model that simulates the release of spray material from various application equipment and tracks its movement through the atmosphere until deposition [3] [4]. The model serves as a regulatory tool for the United States Environmental Protection Agency (US EPA) in assessing potential drift risks during pesticide registration and label development [5].

Recent Model Enhancements and Validation

Recent research has focused on validating and enhancing AGDISP for new application technologies, particularly unmanned aerial vehicles (UAVs or drones). The development of AGDISPpro incorporates multi-rotor aerodynamic models to simulate off-target spray drift from drone applications [3]. A 2025 validation study evaluated AGDISPpro's performance against field measurements from single-swath and multi-swath applications using various spray qualities [4]. The index of agreement between model predictions and field observations ranged from 0.47 to 0.94 across different nozzle types, demonstrating promising predictive capability while highlighting areas for further refinement [3] [4].

Uncertainty in swath width and swath displacement parameters was identified as a significant factor affecting model accuracy for UAV applications [4]. Sensitivity analyses indicate these parameters greatly affect the magnitude of the resulting maximum peak deposition and, to a lesser extent, the width and position of the spray deposition plume [4].

Modernization Initiatives

The AGDISP Modernization Project (AMP), established by the National Agricultural Aviation Association, is currently undertaking a comprehensive update to the model's coding structure using modern programming languages [17]. This five-year initiative, supported by over $335,000 in funding from various agricultural organizations, aims to improve model accuracy and accessibility for all application types (aerial, ground, and unmanned aerial) [17]. A key objective is to enable the recognition of drift reduction benefits offered by new application technologies, which could result in less restrictive and more flexible application requirements on pesticide labels [17].

Experimental Protocols for Drift Measurement

Standardized protocols for measuring spray drift are essential for generating comparable data across studies and for model validation. The following methodologies represent current best practices for field-based drift assessment.

Field Drift Collection Protocol

Objective: To quantify downwind spray deposition and airborne drift from pesticide applications.

Materials:

  • Mylar cards or Petri dishes for quantitative deposition analysis
  • Water-sensitive cards (WSP) for droplet density and coverage analysis
  • Fluorescent tracer (e.g., Brilliant Sulfoflavine, BF7G)
  • Spectrofluorometer for tracer quantification
  • Meteorological station (measuring wind speed, direction, temperature, relative humidity)
  • Sampling stands at multiple heights (ground level, crop canopy height)
  • GPS unit for precise sampler placement

Procedure:

  • Pre-application Setup:
    • Establish sampling lines at multiple distances downwind from the treatment area (e.g., 1, 3, 5, 10, 15, 25, 50 m) [5] [14]
    • Position collection media on stands at appropriate heights for the target ecosystem
    • Record baseline environmental conditions (wind speed and direction, temperature, relative humidity)
    • Collect control samples from each station before application

  • Tracer Application:

    • Prepare tank mixture with fluorescent tracer at known concentration (e.g., 1200 g·ha⁻¹) [14]
    • Apply using standardized equipment parameters (nozzle type, pressure, release height, speed)
    • Maintain consistent application volume (e.g., 27-40 L·ha⁻¹ for drones, 400 L·ha⁻¹ for orchard sprayers) [14]
    • Monitor and record meteorological conditions throughout application

  • Post-application Sampling:

    • Collect deposition samples immediately after application
    • Retrieve water-sensitive cards for droplet analysis
    • Collect airborne drift samples using active or passive air samplers if required
    • Document sample location, time, and conditions

  • Laboratory Analysis:

    • Extract fluorescent tracer from collection media using standardized solvent volumes
    • Measure fluorescence intensity using spectrofluorometer
    • Calculate deposition values based on calibration curves
    • Analyze water-sensitive cards using image analysis software to determine coverage percentage and droplet density

  • Data Analysis:

    • Express deposition as percentage of applied rate per unit area
    • Fit data to regression models (e.g., four-parameter log-logistic) to characterize drift curves [5]
    • Calculate key metrics including PD25, PD50, and PD90 (distances for 25%, 50%, and 90% reduction in deposition) [5]

AGDISP Model Validation Protocol

Objective: To validate AGDISP model predictions against field-collected drift data.

Procedure:

  • Input Parameterization:
    • Document all application parameters: equipment type, nozzle configuration, spray pressure, release height, vehicle speed, swath width
    • Record meteorological data: wind speed and direction at multiple heights, temperature, relative humidity, turbulence conditions
    • Characterize spray mixture: physical properties, droplet size spectrum (DV10, DV50, DV90)

  • Model Execution:

    • Input measured parameters into AGDISP
    • Run simulations for corresponding field conditions
    • Export predicted deposition values at downwind distances matching field sampling positions

  • Statistical Comparison:

    • Calculate index of agreement between predicted and observed values
    • Determine mean bias error to identify under- or over-prediction trends
    • Perform regression analysis between predicted and measured values
    • Conduct sensitivity analysis on key parameters (swath width, displacement, droplet size)

Visualizing Comparative Drift Assessment

The following workflow diagram illustrates the integrated experimental and modeling approach for comparing aerial and ground application drift.

G Start Study Design ExpSetup Field Experiment Setup Start->ExpSetup Aerial Aerial Application ExpSetup->Aerial Ground Ground Application ExpSetup->Ground DataColl Data Collection Aerial->DataColl Ground->DataColl ModelSim AGDISP Simulation DataColl->ModelSim CompAnalysis Comparative Analysis ModelSim->CompAnalysis Results Drift Mitigation Protocols CompAnalysis->Results

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 2: Key Research Materials for Spray Drift Studies

Item Function Application Notes
Florpyrauxifen-benzyl Synthetic auxin herbicide; Model compound for drift studies Used at field rates; sensitive crops like soybean show injury symptoms at sublethal doses [5]
Fluorescent Tracers (BF7G) Quantify spray deposition and drift Applied at 1200 g·ha⁻¹; enables spectrofluorometer detection on collection media [14]
Water-Sensitive Cards (WSC) Assess droplet density and coverage Analyze using image analysis software; provides # droplets per cm² and percent coverage [5]
Mylar Cards Collect spray deposition for quantitative analysis Non-absorbent surface allows precise measurement of tracer deposition [5]
Spectrofluorometer Quantify fluorescent tracer on collection media Enables calculation of deposition as percentage of applied rate [5]
Meteorological Station Record wind speed/direction, temperature, humidity Critical for characterizing application conditions and model input [5]
AGDISP Software Predict spray transport and deposition Used for scenario modeling and comparison with field data [3] [17]

Implications for Environmental Risk Assessment

The significant difference in drift potential between application methods has profound implications for environmental protection. Research indicates that pesticide drift can lead to over 50% reductions in wild plant diversity within 500 meters of fields, directly reducing floral resources for pollinators [57]. A 2025 study in Environmental Entomology detected 42 different pesticides, including several neonicotinoids, in field margins up to 32 meters from blueberry fields, with distance providing no significant reduction in the number of active ingredients detected [58]. Herbicide drift can also damage sensitive crops downwind, with synthetic auxin herbicides like florpyrauxifen-benzyl causing significant injury to non-target soybean plants [5]. These findings highlight the critical need for accurate drift prediction and implementation of mitigation strategies based on application method-specific risks.

This analysis demonstrates substantial differences in drift potential between aerial and ground applications, with aerial methods typically producing 5-20 times greater drift depending on specific conditions and measurement metrics. The AGDISP model provides a valuable framework for predicting these differences, though ongoing validation efforts continue to refine its accuracy, particularly for emerging application technologies like UAVs. Standardized experimental protocols and the research reagents outlined here enable consistent quantification of spray drift across studies. These comparative assessments are essential for developing evidence-based application guidelines that minimize off-target movement and protect environmental resources, particularly in sensitive agricultural landscapes where pollinators and non-target crops may be affected. Future model enhancements through the AGDISP Modernization Project promise to further improve predictive capabilities and incorporate the drift reduction benefits of new technologies.

Conclusion

AGDISP and its latest iteration, AGDISPpro, represent robust, science-based tools for simulating pesticide spray drift, with demonstrated utility across conventional and emerging application methods like drones. Validation studies confirm the model's strong predictive capability, particularly for fine to coarse spray qualities, though accuracy can be influenced by operational parameters like swath width. The model's integration of near-field Lagrangian and far-field Gaussian mechanics provides a comprehensive framework for environmental risk assessment. Future directions for both research and model development should focus on reducing uncertainty in RPAAS swath characterization, incorporating the effects of complex terrain and barriers, and continuing to validate predictions against field data for a wider range of environmental conditions. This ongoing refinement is essential for supporting sustainable application practices and protecting environmental health.

References