Dispersion Modeling of Commodity and Structural Fumigation - PowerPoint PPT Presentation

1 Dispersion Modeling of Commodity and Structural Fumigation Applications Rick Reiss, Exponent Presented at Kansas State University May 12 th , 2010

2 Overview of Presentation  Risk assessment process for bystander exposure to fumigants  Use of dispersion modeling to estimate downwind concentrations  Example studies with methyl bromide used to characterize emissions

3 Background  Fumigants are generally highly volatile  Emissions after treatment can lead to downwind exposures to bystanders  Regulators are interested in minimizing exposures  The solution was to establish buffer zones around applications, which restrict entry for a period of time after the application  EPA recently established national buffer zones for most fumigants  The principal tool used by EPA was PERFUM

4 The Risk Assessment Process Hazard Identification Hazard Identification Can the substance cause Can the substance cause illness or disease? illness or disease? Dose Dose-Response Response Risk Characterization Risk Characterization Assessment Assessment What is the risk of disease? What is the risk of disease? What dose is necessary? What dose is necessary? Exposure Assessment Exposure Assessment What levels are people What levels are people exposed? exposed? Source: Modified from NAS 1983 (pg. 21) 4

5 Basis for Buffer Zone Estimation  Use of air dispersion modeling to estimate downwind concentrations over range of meteorological conditions  Comparison of concentration estimates with toxicity reference concentrations to estimate risk

6 Possible Buffer Zone Definitions Whole field buffer zone MOE > 100 5% MOE < 100 Maximum concentration buffer zone

7 Factors Influencing Downwind Concentrations  Source-specific  Application rate  Treatment and aeration length  Air exchange rates  Volatility of fumigant  Sealing  Location-specific  Meteorological conditions  Wind speed, wind direction, atmospheric stability  Terrain  Nearby buildings (downwash)

8 Dispersion Modeling Theory

9 Building Downwash

10 Building Downwash – Another Look

11 Steps in Modeling Analysis  Estimating emission rates using real-world field data  Characterization of source of interest  Application rate  Length of treatment and aeration  Air exchange rates  Estimating range of downwind concentrations using historical meteorological datasets

12 General Methodological Options for Emissions  Option #1: Assume a percent release during treatment and aeration  Option #2: Assume an air exchange rate during treatment and aeration and use it to calculate the hourly emissions.

13 Schematic of Building Air Flow – Ventilation Model V=volume Q=air flow Q C(t)=concentration Air Exchange Rate (R) = Q/V

14 Dose Response  Traditional Approach with Animal Studies  Exposure animals for a range of doses and measure chemical-related effects  Determine the No Observed Effect Level (NOEL)  Apply a 100X uncertainty factor to NOEL to determine the “safe” dose for risk assessment  Includes 10X uncertainty factor to account for uncertainty in extrapolating between animals and humans  Includes a 10X uncertainty factor to account for variation in the human population 1

15 Dose-response for methyl bromide  NOAEL of 40 ppm based on a rabbit study exposures at Days 6-16 of gestation  Agenesis (failure to develop) of the gall bladder  Fused sternebrae  Uncertainty factor of 30  Reference concentration = 1.3 ppm (1300 ppb) over 4 hours

16 Using Field Studies to Estimate Emissions – Dispersion Modeling in Reverse

17 Review of Methyl Bromide Historical Studies (early 1990s) Location Location Volume Volume Treatment Treatment Aeration Aeration Period (hr) Period (hr) Period (hr) Period (hr) (ft 3 ) (ft Watsonville Watsonville 1065 1065 0.2 0.2-0.5 0.5 0.2 0.2-0.5 0.5 Bakersfield Bakersfield 18,290 18,290 2 2 Madera Madera 320,000 320,000 90 90 2 Stanislaus Stanislaus 1,450,000 1,450,000 24 24 24 24 Sutter Sutter 3,100,000 & 3,100,000 & 24 24 24 24 6,800,000 6,800,000

18 Loss Rates During Treatment in Mid-Sized Warehouse in Madera 10000 Building Concentration (ppm) 1000 100 C(t) = 7710 ppm exp(-0.05 h -1 t) R 2 = 0.99 10 1 0 20 40 60 80 100 Time (hr)

19 Loss Rates During Aeration in Large Processing Plant in Stanislaus 10000 Exhaust Stack Concentration (ppm) y = 3979 ppm exp(-1.7 h -1 t) 2 = 0.90 R 1000 0 10 20 30 40 Time (mins)

20 Air Exchange Rates from Historical Studies Study Study Scenario Scenario Duration Duration ACH (hr) ACH (hr) (hrs) (hrs) Watsonville Watsonville A 0.17 0.17 0.01 0.01 A 0.28 0.28 0.04 0.04 Bakersfield Bakersfield A 2 0.01 0.01 Madera Madera T 90 90 14 14 A 2 0.9 0.9 Stanislaus Stanislaus T 24 24 23 23 A 24 24 0.4 0.4 Sutter County Sutter County T 24 24 41 41 A 24 24 0.3 0.3-0.7 0.7

21 Characteristics of New Methyl Bromide Studies Site A Site B Site C Application Rate 1.0 1.0 1.0 (lbs/1000 ft 3 ) (nominal) 24 g/m 3 30 g/m 3 18 g/m 3 Initial Concentration 6000 ppm 8000 ppm 5000 ppm Change in Concentration -60% -50% -65% during Treatment Duration to aerate 50% of 0.5 h (passive) 3.0 h (active 1.0 h (active) fumigants and passive)

22

23

24 Emission Rates for New Studies 10000 Building air- Methyl Bromide Concentration (ppm) exchange rate is defined by the rate of change in fumigant y = 5550 e -0.05x R 2 = 0.98 concentration. 1000 0 5 10 15 20 25 30 Time from Treatment (h)

25 Emission Rate Estimates for New Studies Site A Site B Site C Air-Exchange Rate 14-35 hr 17 hr 17 hr Treatment Fumigant mass 44 to 70% over 24 55% over 20 hours 62% over 24 loss hours hours Air-Exchange Rate 0.5-3.5 hr 0.3-2.8 hr(both 0.7 hr (active active and only) (passive) Aeration passive) Fumigant mass 18 to 78% in 1 22 to 86% in 1 hour 63% in 1 hour loss hours

26 Summary of Measured Air Exchange Rates  Treatment  Range of 14-41 hr, consistently across studies  Length between 10 minutes and 90 hours  Often lose >50% of mass, by 3-5% per hour  Aeration  Range of 0.01-3 hr  Larger rates for smaller buildings  Typically, similar, but somewhat less, than the rated fan capacity  Length between 10 minutes and 3 hours

27 Application to Mid-Sized Warehouse in Madera 6 5 4 7 3 15 16 8 2 1 9 13 14 11 10 12

28 Modeled versus Measured Fits for Madera Treatment Period Are Quite Good 1400 Measurements AERMOD 1200 ISCST3 1000 Concentration (ppb) 800 600 400 200 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Sampling Location

29 Point Comparisons for New Studies Treatment - 3 Sites Summary Aeration - 3 Sites Summary 600 300 y = 0.26x + 56.76 R 2 = 0.07 y = 0.86x - 0.45 500 250 R 2 = 0.43 Measurements (ppb) Measurements (ppb) 400 200 300 150 200 100 100 50 0 0 0 100 200 300 400 500 600 0 50 100 150 200 250 300 Predictions (ppb) Predictions (ppb) Model captures the range of concentrations well, but local wind pattern around building structures makes it difficult to predict the spatial distribution.

30 Maximum Concentrations for New Studies Comparison of predicted maximum concentration in the model domain and measured maximum concentration. Site A Site B Site C Treatment Measured (ppb) 170 280 120 Predicted (ppb) 317 326 290 Aeration Measured (ppb) 113 40 565 Predicted (ppb) 303 519 970 EPA Level of Concern (4 hrs) = 1300 ppb

31 Conclusions  Risk assessment with dispersion modeling is a practical method to address bystander exposure and establish buffer zones  There is comparatively less data on emissions than for field applications  More information would be helpful to refine the risk assessment