APPENDIX 5D CALPUFF DISPERSION MODEL · Appendix 5D: CALPUFF Dispersion Model Socio-economic Impact Assessment December 2006 Page 5D-6 CALPUFF contains algorithms for near-source

Volume 2: Environmental and Sturgeon Upgrader ProjectSocio-economic Impact Assessment Appendix 5D: CALPUFF Dispersion Model

December 2006

Page 5D-1

APPENDIX 5D CALPUFF DISPERSION MODEL

Sturgeon Upgrader Project Volume 2: Environmental andAppendix 5D: CALPUFF Dispersion Model Socio-economic Impact Assessment

December 2006

Page 5D-2


December 2006

Page 5D-3

5D.1 Introduction Ambient air quality models are used to predict air quality changes (i.e., changes to ambient concentrations or deposition) associated with current and future emission scenarios. This section discusses the selection and application of the dispersion model that was used to evaluate the Project.

5D.1.1 Model Types Air quality simulation (or dispersion) models provide a scientific means of relating industrial and community emissions to air quality changes by using mathematical equations to simulate transport, dispersion, transformation, and deposition processes in the atmosphere. Dispersion models can address a wide range of distance scales (100’s of m to 1000’s of km) and time scales (minutes to years). There are two levels of modelling levels of effort: • Screening models estimate maximum short-term (~1 hour) average concentrations for a

wide range of pre-selected meteorological conditions. These models are typically limited to single sources and downwind distances less than 10 km (e.g., the U.S. EPA SCREEN3 model)

• Refined models use sequential hourly meteorological data for a 1 to 5 year period (8760 to 47800 h, respectively). These models can address multiple sources, and predict hourly average concentrations for all source, meteorology, and receptor combinations. The hourly concentrations can be used to predict concentrations for averaging periods that are factors of 24 (i.e., 2, 3, 4, 6, 8 or 12 h), or for longer periods (i.e., seasonal or annual) (e.g., the U.S. EPA ISC-PRIME and AERMOD models). Some refined models can also account chemical transformation, and deposition processes (e.g., the CALPUFF model).

Regulatory agencies have relied on dispersion models as part of the approval process. Numerous models are available for air quality predictions and the appropriate selection depends on project-specific circumstances. In response to the regulatory use of these models, formal guidelines regarding the selection and application of these models have been developed (e.g., Alberta Environment 2003a; Alberta Environment 2003b; U.S. EPA 2005).

5D.1.2 Model Application The application of a dispersion model requires the preparation of input files and the analysis of output files. The input files include the following: • Control/Option information to identify the model run, and to select the available technical

and output features for the selected model. • Source data that identifies the location, the emissions characteristics (e.g., stack height),

and the emission amounts (e.g., SO2 emission rate). • Terrain and receptor data to account for elevated terrain airflow and to provide the

deposition characteristics for the vegetation canopy. • Meteorological data to characterize the airflow and turbulence in the region on an hourly

basis.

The output files include:

• A summary file to identify the model run and provide an overview of the run. • Hourly concentration files for each receptor and meteorological combination. • Hourly deposition files for each receptor and meteorological combination.


December 2006

Page 5D-4

Presentation software is used to re-format the output and to provide concentrations and deposition contour plots that can be superimposed over base maps.

5D.2 Model Selection

5D.2.1 Model Requirements For the impact assessment, the models must have the ability to account for:

• Multiple point and area sources; • Flat and elevated terrain features; • Secondary PM2.5 formation; • SO2 to sulphate (SO4

2-), and NOx to nitrate (NO3-) conversion; and

• Wet, dry, gaseous, and particulate deposition processes.

These features are required to predict ambient concentrations and potential acid input (PAI); and according to Alberta Environment’s (2003a) definition, a refined model is required.

5D.2.2 Candidate Models Table 5D-1 describes the refined dispersion models outlined in the Alberta Air Quality Model Guideline (Alberta Environment 2003a). Of these models, only the CALPUFF model can be used to predict secondary PM2.5 formation and the deposition of acidifying compounds. If deposition was not a requirement, then the ISC3-PRIME or AERMOD models could be used. AERMOD provides a more refined treatment of dispersion relative to ISC3-PRIME. The RTDM and CTDM-PLUS models are primarily single source models designed for very complex terrain applications such as the Alberta Foothills. The CALPUFF model was selected as the preferred model for this assessment.

CALPUFF has two major options with respect to meteorological data:

• The ISC mode assumes a uniform meteorological field over the modelling domain during a given hour. While this is consistent with the ISC-PRIME and AERMOD models, CALPUFF has the advantage of allowing the effluent trajectory to vary from hour-to-hour in a systematic manner as the wind direction varies from hour-to-hour.

• The CALMET mode allows for three-dimensionally varying meteorological field over the modelling domain during a given hour.

For this assessment, the CALPUFF model with the three dimensional CALMET wind field was selected (see Appendix 5C). The CALPUFF model performance was tested by comparing model predictions to selected observations.


December 2006

Page 5D-5

Table 5D-1: Dispersion Models identified in the Alberta Air Model Guideline The SCREEN3 (U.S. EPA 1995) model is a simple Gaussian plume model that uses Pasquill-Gifford dispersion coefficients to characterize atmospheric turbulence and Briggs relationships to determine plume rise. The model calculates the maximum ground-level concentration that occurs below the plume centreline. The model examines a wide range of atmospheric stability class and wind speed combinations (54) to identify the combination that results in the maximum ground-level concentration. Limited mixing conditions are assumed for selected meteorological events. A pre-selected array of 50 distances, ranging from 100 m to 50 km, can be used. An iteration routine is used to determine the maximum concentration and the associated distance to the nearest metre. The Industrial Source Complex ISC3-PRIME (Schulman et al. 1998) and ISC-OLM (Tikvart 1996) models with refined meteorological data are U.S. EPA multi-source Gaussian models capable of predicting both long-term (annual) and short-term (down to 1-h mean) concentrations arising from point, area, and volume sources. Gravitational settling of particles can be accounted for using a dry deposition algorithm; wet deposition and depletion due to rainfall can also be treated. Effects of building wakes can be incorporated. The model has options for both urban and rural dispersion coefficients. AERMOD (U.S. EPA 2004) is a new-generation U.S. air quality modelling system. It contains updated algorithms for convective boundary layers; for computing vertical profiles of wind, turbulence, and temperature; and for the treatment of all types of terrain. It was developed by the U.S EPA, in collaboration with the American Meteorological Society. The Rough Terrain Diffusion Model (RTDM) (Paine and Egan 1987) is a U.S. EPA Gaussian model capable of predicting short-term concentrations arising from point sources in complex terrain. The model cannot address building wake effects. RTDM can be used with routinely available meteorological data relating to wind and stability categories. The Complex Terrain Diffusion Model (CTDMPLUS) (U.S. EPA 1989) is a refined air quality model that is preferred for use in all stability conditions in complex terrain applications. CTDMPLUS is applicable to all receptors on terrain elevations greater than stack top height. However, the model contains no algorithms for simulating building downwash, or the mixing and recirculation found in cavity zones in the lee of a hill. The CALPUFF (Scire et al. 1999) model is a multi-layer, multi-species, non-steady state puff dispersion model that can simulate the effects of time and space-varying meteorological conditions on substance transport, transformation, and removal. CALPUFF can use the three-dimensional meteorological fields developed by the CALMET model or simple, single station winds in a format consistent with the meteorological files used to derive ISCST3 steady-state Gaussian models.

The SCREEN3, ISCST3, ISC3-PRIME, ISC-OLM, AERMOD, RTDM, and CTDM models and corresponding documentation are available from the U.S. Support Centre for Regulatory Air Models (SCRAM) web site (http://www.epa.gov/ttn/scram/). The CALPUFF and CALMET models and documentations are available from the http://www.src.com/calpuff/calpuff1.htm website.

5D.2.3 CALPUFF Model CALPUFF is a multi-layer, multi-species, non-steady-state puff dispersion model, which can simulate the effects of time- and space-varying meteorological conditions on pollutant transport, transformation, and deposition. CALPUFF can use the three-dimensional meteorological fields developed by CALMET model, or simple, single station winds in a format consistent with the meteorological files used to drive the ISCST3 steady-state Gaussian model. However, single-station ISCST3 winds do not allow CALPUFF to take advantage of its capabilities to treat spatially varying meteorological fields.


December 2006

Page 5D-6

CALPUFF contains algorithms for near-source effects such as building downwash, transitional plume rise, partial plume penetration, as well as longer-range effects such as chemical transformation, and pollutant removal (wet scavenging and dry deposition). It can accommodate arbitrarily varying point source and area source emissions. Most of the algorithms contain options to treat the physical processes at differing levels of detail depending on the requirements for the particular model application.

The major features and options include the following.

• Dispersion Coefficients: Several options are provided in CALPUFF for the computation of dispersion coefficients:

• From direct turbulence measurements (σv and σw),

• From similarity theory to estimate σv and σw from surface heat and momentum fluxes provided by CALMET,

• From the Pasquill-Gifford (PG) or McElroy-Pooler (MP) dispersion coefficients, or

• From dispersion equations based on the Complex Terrain Dispersion Model (CDTM).

Options are also provided to apply an averaging time or surface roughness length adjustments to the PG coefficients.

• Chemical Transformation: CALPUFF includes options to parameterize chemical transformation effects using the five species scheme (SO2, SO4

2-, NOx, HNO3 and NO3-)

employed in the MESOPUFF II model, a modified six-species scheme (SO2, SO42-, NO,

NO2, HNO3 and NO3-) adapted from the RIVAD/ARM3 method, or a set of user-specified,

diurnally-varying transformation rates.

• Dry Deposition: A full resistance model is provided to calculate dry deposition rates of gases and particulate matter as a function of geophysical parameters, meteorological conditions, and pollutant species. Options are provided to allow user-specified, diurnally varying deposition velocities to be used for one or more pollutants instead of the resistance model (e.g., for sensitivity testing) or to bypass the dry deposition model completely.

• Wet Deposition: An empirical scavenging coefficient approach is used in CALPUFF to compute the depletion and wet deposition fluxes due to precipitation scavenging. The scavenging coefficients are specified as a function of the pollutant and precipitation type (i.e., frozen versus liquid precipitation).

The following section describes the application of the CALPUFF model specific to the Sturgeon Upgrader air quality assessment.

5D.3 Model Application

5D.3.1 Model Domain The CALPUFF model requires the user to define locations where concentrations are to be calculated, these locations are referred to as “receptors”. The CALPUFF computational domain was selected to represent a 100 by 100 km centered on the Sturgeon Upgrader site; this is the same area evaluated with the CALMET model (Appendix 5C). Deposition predictions are displayed for an 80 x 80 km domain, and concentration predictions are displayed for a smaller 50 x 50 km area, Table 5D-2 also provides the corners of the smaller 50 x 50 km area and the coordinates for the corners of these regions. Two types of receptors within the modelling domain were defined: nested Cartesian grid points, and discrete locations.


December 2006

Page 5D-7

Table 5D-2: CALPUFF Modelling Domain Coordinates (UTM Z12; NAD 83) CALPUFF 80 x 80 km

Area

Easting (m)

Northing (m)

Southwest Corner 318053 5927982 Northwest Corner 318053 6007982 Southeast Corner 398053 5927982 Northeast Corner 398053 6007982

CALPUFF 50 x 50 km Area

Easting

(m) Northing

(m) Southwest Corner 333053 5942982 Northwest Corner 333053 5992982 Southeast Corner 383053 5942982 Northeast Corner 383053 5992982

Figure 5D-1 shows the receptor points used to provide an understanding of the spatial concentration and deposition patterns. The receptors are based on:

• 50 m spacing along the northern plant fence line (452 receptors),

• 50 m spacing along the southern project fence line (225 receptors),

• 50 m spacing for the 2.5 x 2.5 km area centred on the Sturgeon operating area (2624 receptors),

• 800 m spacing for the 10 x 10 km area centred on the Sturgeon operating area (144 receptors), and

• 2000 m spacing for the 80 x 80 km area centred on the Sturgeon operating area (1681 receptors).

The described grid results in 5126 receptor points. The density is greater near the Project operating area to provide a greater resolution and facilitate the determination of the maximum concentrations due to the Project emissions. The indicated discrete receptor grid deviates from the guidance provided by AENV (2003a). The receptor spacing guidance was developed for a wide spectrum of facilities ranging from a single compressor station to facilities as large as the existing and proposed upgraders in the region. The spacing is viewed as being sufficient to provide an indication of maximum values due to Project emissions and the contribution of the Project to high values near other facilities.

In addition, 234 discrete locations corresponding to specific sites of interest were included. Figure 5D-2 shows the locations of these residential (RESI), agricultural (AGRI), public use area (PUA) and commercial (COMM) receptors. These receptors are also identified in Table 5D-3. Agricultural receptors typically include residences. Public use areas can include recreation areas and areas where the public has access. Commercial areas include other industrial facility locations. Most of the discrete receptors are located within a nominal 7.5 km radius of the Sturgeon Upgrader site. These receptors are numbered from 1 to 251 based on an original numbering scheme. During the assessment, a number of receptor locations were removed as they were located within a proposed industrial site.

FIGURE

FIGURE ID

PROJECT SCALE

CHECKED APPROVED

DRAFT DATE

REVISION DATE

PREPARED BY

DRAWNAO MD MD

24/10/2006 OS1589

5D-1AXYS-CAL-078-000Acknowledgements:

Original Drawing by: Jacques Whitford-AXYSConsulting Ltd. Sources: Alberta Government, NRN

3 0 3 6 9

Distance in Kilometres

STURGEON UPGRADER

CALPUFF Dispersion ModelReceptor Grid in the 80 km by

80 km Study AreaProjection:

UTM 12 NAD 83 1:500,000

Jacques Whitford-AXYS

PREPARED FOR

Elk IslandNational Park

RedwaterNatural Area

Fort Saskatchewan

Lamont

Strathcona

Bruderheim

Namao

Bon Accord

Waskatenau

TWP 53

TWP 54

TWP 55

TWP 56

TWP 57

TWP 58

TWP 52

Morinville

Legal

TWP 59Vimy

Gibbons

Redwater

Sherwood Park

RGE 18W4MRGE 19RGE 20RGE 21

RGE 22RGE 23RGE 24RGE 25

Edmonton

North

Saskatc hew

an

R

iver

Redwater RiverSturgeon R

iver

LostpointLake

AstotinLake

Beaverhill Creek

FawayikLake

16

15

21

803

28A

643

28

38 45

825

28

830

829

320000

320000

340000

340000

360000

360000

380000

380000

5940

000

5940

000

5960

000

5960

000

5980

000

5980

000

6000

000

6000

000

CALPUFF ReceptorCommunityPaved AccessUnpaved AccessRailwayProject Fence LineProtected AreaNatural AreaUrban Area

24/11/2006

FIGURE

FIGURE ID

PROJECT SCALE

CHECKED APPROVED

DRAFT DATE

REVISION DATE

PREPARED BY

DRAWNAO MD MD

24/10/2006 OS1589

5D-2AXYS-CAL-083-000Acknowledgements:

Original Drawing by: Jacques Whitford-AXYSConsulting Ltd. Sources: Alberta Government, NRN

2 0 2 4 6

Distance in Kilometres

STURGEON UPGRADER

Discrete Receptors in the50 km by 50 km Study Area Projection:

UTM 12 NAD 83 1:300,000

Jacques Whitford-AXYS

PREPARED FOR

Elk IslandNational Park

RedwaterNatural Area

Fort Saskatchewan

Bruderheim

Namao

Bon Accord

TWP 54

TWP 55

TWP 56

TWP 57

Gibbons

Redwater

RGE 20W4MRGE 21RGE 22RGE 23RGE 24

Edmonton

North BruderheimNatural Area

Northwest of BruderheimNatural AreaAstotin

Natural Area

OpalNatural Area

North

Saskatchewan

River

Redwater River

Sturge

on River

LostpointLake

AstotinLake

Beaverhill Creek

15

21

803

28A

643

28

38 45

825

28

830

829

340000

340000

360000

360000

380000

380000

5960

000

5960

000

5980

000

5980

000

Discrete ReceptorCommunityPaved AccessUnpaved AccessRailway

Project Fence Line7.5 km Radius AreaProtected AreaNatural AreaUrban Area

24/11/2006


December 2006

Page 5D-10

Table 5D-3: Locations of the Discrete Receptors

Air Quality Model ID

Receptor Name (Individual residents are not identified)

Receptor Group

UTM Easting

(km)

UTM Northing

(km) Elevation

(m) 1 FAP – Fort Saskatchewan RESI 353.268 5952.286 629 2 FAP – Ross Creek PUA 354.840 5954.192 626 3 FAP – Station 401 AGRI 357.579 5953.142 623 4 FAP – Range Road 220 PUA 359.917 5957.679 628 5 FAP - Scotford COMM 363.108 5962.286 628 6 FAP - Redwater PUA 361.915 5968.142 716 7 FAP - Lamont RESI 376.063 5958.486 722 8 FAP – Elk Island National Park PUA 376.620 5949.805 621 9 FAP/EC-A RESI 355.844 5958.173 622 10 FAP/EC-B PUA 354.822 5954.185 618 11 FAP/EC-C RESI 359.201 5953.989 626 12 FAP/EC-D COMM 363.112 5962.285 622 13 FAP/EC-E COMM 364.267 5966.735 623 14 FAP/EC-F PUA 376.627 5949.818 716 20 DeGussa Canada Inc. COMM 359.660 5967.34 635 24 . AGRI 357.363 5969.617 639 30 Northwest Upgrading COMM 360.538 5968.123 633 31 Serink D. & L. AGRI 358.852 5970.134 634 32 Smulski J. AGRI 360.455 5968.824 633 33 Derouin D. & P. AGRI 356.910 5969.811 644 35 Derouin L. AGRI 357.169 5970.107 639 36 Serink D. & L. AGRI 359.430 5970.162 634 37 Taranaski lt. AGRI 359.038 5970.443 636 38 Moerman C. & S. AGRI 357.389 5970.54 637 39 Providence Energy COMM 359.505 5965.444 634 41 Groot W. & B. AGRI 355.508 5967.717 643 42 Groot W. AGRI 355.529 5967.416 646 43 Vermeer’s Dairy Ltd. AGRI 358.562 5970.851 636 44 Kasaholf AGRI 356.895 5970.484 641 45 Fermaniuk F. & C. & M. AGRI 355.594 5968.979 642 46 Ouellet AGRI 355.579 5966.874 646 47 Derouin D. & P. AGRI 356.333 5970.162 640 49 Sank R. AGRI 357.310 5965.121 640 50 Nikiforuk Construction Ltd. COMM 355.265 5968.284 642 51 Braun S. AGRI 357.331 5970.870 637 52 Provident Energy Ltd COMM 359.570 5965.070 638 53 Barkshire AGRI 356.922 5970.750 644 54 Lamoureux V. AGRI 355.222 5967.318 648 55 Roe AGRI 356.359 5970.566 639


December 2006

Page 5D-11

Table 5D-3: Locations of the Discrete Receptors (cont’d) Air Quality Model ID

Receptor Name (Individual residents not identified)

Receptor Group

UTM Easting

(km) UTM Northing

(km) Elevation

(m) 56 Klassen B. AGRI 358.944 5971.224 636 57 Lamoureux V. AGRI 355.121 5966.959 648 58 Kamjes T. & C. AGRI 355.509 5966.091 649 59 Derouin AGRI 356.847 5971.006 638 60 Kugler E. & E. AGRI 356.706 5964.867 643 61 Henryk Farm Ltd. AGRI 357.111 5964.533 644 62 Friesen K. & J. Braun S. et al AGRI 357.423 5971.508 634 63 Vermeer’s Dairy Ltd. AGRI 358.904 5971.614 634 64 Shaw J. & S. AGRI 355.600 5970.519 645 65 Kampjes H. & F. Kamjes H. & F. AGRI 355.503 5965.459 650 66 Agrium Products Inc. COMM 361.366 5970.322 639 67 Shaw R. & D. AGRI 355.342 5970.512 648 68 Westralia Farms Ltd. AGRI 355.110 5965.644 655 69 Agrium Products Ltd. COMM 362.330 5968.371 630 70 Shaw S. & K. AGRI 355.661 5971.061 644 71 Shaw C. AGRI 354.812 5970.226 654 72 Hood A. AGRI 354.105 5967.518 654 73 Shaw R. & D. AGRI 354.976 5970.566 652 74 Melnychuk E. & W. AGRI 354.117 5966.992 654 75 Makoweckl D. AGRI 360.558 5971.654 641 76 Maschmeyer R. AGRI 362.051 5965.718 604 77 Hood A. AGRI 353.938 5967.906 653 78 Stewart R. & S. AGRI 355.050 5964.958 652 79 Syvenky R. AGRI 357.025 5963.690 640 80 Kalco Farms Ltd. AGRI 353.954 5968.855 650 81 Shaw C. AGRI 354.488 5970.231 654 82 Comelius A. AGRI 360.044 5972.101 636 83 Kalisvaart M. Jansen K. AGRI 354.069 5966.215 655 84 Gates W. & H. AGRI 355.935 5964.009 647 85 Williams W. & D. AGRI 358.957 5972.529 639 86 Henryk Farm Ltd. AGRI 360.191 5963.702 601 87 Half M. AGRI 353.663 5968.727 651 88 Holmes R. AGRI 356.356 5963.562 641 89 Rawson K.D. AGRI 362.921 5966.497 605 90 Beierbeck C. AGRI 360.044 5972.468 636 91 Kalisvaart M. Jansen K. AGRI 353.535 5967.024 653 92 Shaw W. & A. AGRI 353.528 5968.925 658 93 Moss D. & S. AGRI 355.837 5972.172 648 94 Maschmeyer R. AGRI 362.529 5965.404 624


December 2006

Page 5D-12

Table 5D-3: Locations of the Discrete Receptors (cont’d)



Receptor Group

UTM Easting

(km)

UTM Northing

(km) Elevation

(m) 95 Shaw A. & C. AGRI 355.342 5971.854 643 96 Shaw J. & S. AGRI 354.048 5970.560 652 97 Stross M. & A. AGRI 355.072 5964.030 654 98 Eagle Farms Ltd. AGRI 353.497 5966.221 658 99 Shaw A. & C. AGRI 355.322 5972.152 646 100 Chaba G. & B. AGRI 359.025 5973.081 640 101 Lafond C. & T. AGRI 360.599 5972.771 639 102 Gubersky G. & T. AGRI 355.749 5972.613 644 103 Harriott B. & J. AGRI 357.782 5973.277 649 104 Hood A. AGRI 352.994 5967.351 658 105 Knupp A. & D. AGRI 357.362 5973.285 653 106 Boisjoil G. & T. AGRI 354.522 5964.052 656 107 Whaling AGRI 354.000 5971.326 654 108 Kropp L. F. & K. L. AGRI 363.763 5967.131 617 109 Gaumont C. & J. AGRI 363.613 5966.275 623 110 Steffler B. & G. AGRI 359.392 5962.478 616 111 Cholowski R.F. & V. R. AGRI 363.214 5965.069 624 112 Whaling AGRI 353.960 5971.562 656 113 Shaw W. & A. AGRI 353.302 5970.572 657 114 Jigolyk L. AGRI 356.706 5962.504 637 115 Kopala M. M. Kopala L.C. AGRI 363.628 5965.833 626 116 Kropp L.F. & K.L. AGRI 363.992 5966.967 623 117 Mika D. & J. AGRI 358.527 5973.713 643 118 Johnson B. & W. AGRI 358.189 5973.746 641 119 Motz AGRI 364.134 5968.045 625 120 Battenfelder A. & T. AGRI 357.803 5973.751 643 121 Whaling G. & B. AGRI 353.946 5971.793 655 122 Shaw A. & C. AGRI 355.308 5972.945 650 123 Hampshire D. & S. AGRI 357.410 5973.718 651 124 Roe T. & A. AGRI 352.481 5967.463 662 125 Kupsch A. AGRI 364.199 5968.708 625 126 Kropp L. F. & K. L. RESI 364.191 5967.125 626 127 Whaling G. & B. AGRI 354.021 5972.038 658 128 Lafond C. & T. AGRI 361.070 5973.243 637 129 Steffler B. & G. AGRI 359.333 5962.035 602 130 Boisjoil G. & T. AGRI 353.729 5964.089 654 132 King B. & T. AGRI 353.707 5971.888 651 133 Kroening B.J.Samolf B.J.etal AGRI 360.188 5962.15 630


December 2006

Page 5D-13




Receptor Group

UTM Easting

(km)

UTM Northing

(km) Elevation

(m) 134 Kuglar G. & L. AGRI 352.276 5968.748 659 135 Harris H. & A AGRI 352.217 5967.745 661 136 Shaw W. & A. AGRI 352.416 5969.536 667 137 Formanski J. & G. AGRI 356.542 5973.840 643 138 Reid R. & B. AGRI 352.753 5970.593 657 139 Umbach T. AGRI 354.095 5972.492 661 140 Henkelman P. AGRI 364.527 5967.325 627 141 Sandford D. AGRI 357.403 5974.111 650 142 King B. & T. AGRI 353.336 5971.685 654 143 Arneson D. AGRI 358.971 5974.158 643 144 Kalco Farms Ltd. AGRI 352.309 5969.778 665 145 Marquardt B. & C. AGRI 364.555 5969.016 627 146 Marquardt B. & C. AGRI 364.605 5968.716 624 147 Hoehn R. AGRI 354.99 5973.318 651 148 Doblanko K. & G. AGRI 351.996 5967.977 659 149 Jigolyk S. & P. & A. AGRI 356.944 5961.754 636 150 Moreau D. & M. AGRI 352.175 5966.296 657 152 Whaley G. & S. AGRI 351.953 5968.603 661 153 Henkelman P. AGRI 364.669 5966.981 625 154 Kalco Farms Ltd. AGRI 352.319 5970.215 658 155 Studney S. & J. AGRI 352.023 5969.19 661 156 Westralia Farms Ltd. AGRI 355.013 5962.396 646 157 Moreau D. & M. AGRI 352.256 5965.725 658 158 Shell Canada Ltd COMM 362.12 5962.65 626 159 Serink B. Serink M.T. AGRI 364.898 5968.103 628 160 Fraser L. & D. AGRI 358.527 5974.534 645 161 Groot Farms Ltd. AGRI 351.679 5967.095 659 162 North Bank Potato Farms Ltd. AGRI 354.290 5962.580 644 163 Doblanko K. & G. AGRI 351.607 5967.766 662 164 Lelchner M. & Lelchner W. AGRI 364.828 5969.748 622 165 Reed D. & M. AGRI 360.774 5974.306 638 166 Eagle Foods Ltd. Arneson J. AGRI 359.025 5974.763 641 167 North Bank Potato Farms Ltd. AGRI 353.815 5962.779 640 168 WasyluchaM. & E. AGRI 351.791 5970.167 658 169 Reed D. & M. AGRI 360.599 5974.48 637 170 Hewitt D. & L. AGRI 358.513 5974.911 643 171 Reed D. & M. AGRI 360.558 5974.547 638 172 Hein B. & P. AGRI 364.501 5971.237 623


December 2006

Page 5D-14




Receptor Group

UTM Easting

(km)

UTM Northing

(km) Elevation

(m) 173 Procinsky W. Dumin B. AGRI 353.405 5962.967 638 174 Reed D. & M. AGRI 360.545 5974.614 638 175 R & M Poultry Ltd. AGRI 361.823 5974.104 634 176 Prins R. AGRI 351.900 5964.91 653 177 NanaksarGurdwaraGursikhTemple AGRI 356.410 5961.075 635 178 The Manderley Corporation AGRI 356.825 5960.945 638 179 Chaba D. Moore P. AGRI 356.976 5974.993 644 180 North Bank Potato Farms Ltd. AGRI 353.686 5962.401 631 182 Wolfe AGRI 352.441 5972.200 665 183 Cooreman et al AGRI 351.927 5964.500 641 184 Cooreman et al AGRI 351.766 5964.737 641 185 Procinsku W. Dumin B. AGRI 353.432 5962.521 622 186 Sulka S. & B. AGRI 360.558 5974.924 642 187 AGRI 352.414 5972.382 659 188 NanaksarGurdwaraGursikhTemple AGRI 356.577 5960.756 634 189 MacDougall H. AGRI 352.021 5971.82 672 190 North Bank Potato Farms Ltd. AGRI 353.659 5962.169 619 191 The Manderley Corporation AGRI 356.852 5960.654 634 192 Ford Et al AGRI 352.577 5972.674 660 193 Procinsky W. Dumin B. AGRI 353.33 5962.417 616 194 Anema H. & N. AGRI 353.261 5973.495 655 195 Mc Bride B. et al AGRI 354.082 5974.253 654 196 Ibey J. & F. AGRI 352.041 5972.26 670 197 AGRI 352.414 5972.79 664 198 Steen C. & D. AGRI 357.127 5960.335 626 199 Schultz AGRI 355.369 5975.074 650 200 Watt A. & P. AGRI 352.665 5973.236 665 201 Juniper Hills RESI 360.056 5975.513 644 202 McKinnon J. & C. AGRI 356.566 5960.384 628 203 Kuefler R. AGRI 362.213 5974.723 633 204 Austin AGRI 353.675 5961.587 639 205 Value Creation Inc. (BA Energy) COMM 365.86 5965.61 624 207 Veltman H. & L. AGRI 365.633 5964.87 628 208 Casa Vista RESI 350.881 5965.168 637 209 Armstrong G. AGRI 351.912 5972.694 661 210 Salahub T. & M. AGRI 352.15 5973.026 664 211 Sadoway D. & A. AGRI 364.111 5973.498 628


December 2006

Page 5D-15

Table 5D-3: Locations of the Discrete Receptors (cont’d) Air Quality Model ID

Receptor Name (Individual residents are identified)

Receptor Group

UTM Easting

(km)

UTM Northing

(km) Elevation

(m) 212 Smith W. & R. AGRI 365.523 5971.601 622 213 Sawchuk AGRI 354.673 5960.741 637 214 Salahub T. & M. AGRI 352.129 5973.189 665 215 Waterdale Park RESI 361.659 5975.399 634 216 Sturgeon Valley Estates RESI 350.182 5968.613 654 217 Salahub T. & M. AGRI 352.129 5973.332 665 218 Salahub T. & M. AGRI 352.163 5973.548 661 219 Riverside Park RESI 355.293 5960.158 631 220 Smith W. & R. AGRI 365.644 5972.099 620 221 Schwing B. AGRI 364.77 5973.471 625 222 Hu Haven RESI 354.754 5960.223 639 223 Ag-Oil Alberta Ltd. COMM 352.645 5974.457 653 224 Radke B. & Radke L.D. AGRI 365.157 5962.591 620 225 Ag-Oil Alberta Ltd. COMM 352.421 5974.349 651 226 Schwing B. AGRI 365.12 5973.471 623 227 Bruderheim Natural Area PUA 367.121 5968.75 627 228 Astotin Natural Area PUA 367.069 5965.875 624 229 Kneller E. AGRI 352.502 5974.795 653 230 Libbey H. Lybacki M. et al AGRI 365.698 5973.443 625 231 Hansen J. & C. AGRI 352.529 5975.114 653 232 Madu R. & E. AGRI 352.075 5974.769 655 233 Chichak L.V. AGRI 363.954 5960.564 628 234 lane C. AGRI 365.026 5974.426 625 235 Wolansky W. &K. AGRI 366.215 5972.971 622 236 Cook D. AGRI 366.188 5973.4 623 237 Romariuk D. AGRI 367.062 5971.949 615 238 Lane C. AGRI 365.766 5974.494 623 239 Tancowny T. & S. AGRI 367.048 5973.159 618 240 Sudayko M. & J. AGRI 366.9 5973.414 619 241 Sudayko M. &J. AGRI 367.143 5973.427 617 242 Tancowny T. &S. AGRI 367.385 5973.171 620 243 Gibbons RESI 347.14 5967.01 659 244 Redwater RESI 362.04 5979.69 632 245 Fort Saskatchewan Natural Area PUA 354.582 5955.741 603 246 Bruderheim RESI 372.29 5963.36 625 247 Josephberg RESI 363.65 5953.851 646 248 Redwater Natural Area PUA 371.039 5977.003 631 249 Bon Accord RESI 340.889 5967.709 703 250 Lamont RESI 382.356 5957.181 658 251 Elk Island National Park PUA 375.062 5943.202 724


December 2006

Page 5D-16

5D.3.2 Meteorology The CALMET diagnostic wind field module was used to provide representative wind, temperature and turbulence fields (see Appendix 5C).

5D.3.3 Building Downwash Building downwash was not viewed as being applicable given the stack heights associated with the Project (see Appendix 5A). As the larger process structures tend to be open, they will not result in downwash influences that would be associated with a solid structure of similar size.

5D.3.4 Terrain Coefficients Terrain in the study area was described in the CALMET description (Appendix 5C). While the terrain near the Project is relatively flat (~633 m ASL), there are locations where higher terrain occurs (e.g., Bon Accord area at 720 m ASL, Elk Island National Park at 730 m ASL, and Beaver Hills at 760 m ASL). As a plume/puff passes over complex terrain, it has the potential to move closer to the ground. The plume path coefficient (PPC) method can be used to account for this potential decrease in height above the ground. A PPC of 1.0 assumes that the plume trajectory is parallel to the terrain features.

The default CALPUFF values are 0.5, 0.5, 0.5, 0.5, 0.35, and 0.35 for PG stability categories A, B, C, D, E and F, respectively. The selection of these values is not justified in the user guide. Lott (1984) compared a number of alternate terrain schemes and recommended PPC values of 0.8, 0.7, 0.6, 0.5, 0.4, and 0.3 for Pasquill-Gifford (PG) stability categories A, B, C, D, E and F, respectively. For this assessment, PPC values based on Lott’s evaluation have been adopted.

5D.3.5 Chemical Transformation CALPUFF employs two alternate chemical reaction schemes: the MESOPUFF II and the RIVAD/ARM3 schemes. The RIVAD/ARM3 chemical scheme was selected since the MESOPUFF II scheme is viewed as being outdated (Morris et al. 2003). This chemistry scheme treats the NO and NO2 conversion process in addition to the NO2 to NO3

- and SO2 to SO42-

conversions, with equilibrium between gaseous HNO3 and particulate NH4NO3 (Scire et al. 1999). The selected chemical transformation scheme was applied relative to the prediction of sulphate and nitrate compounds and the associated deposition.

5D.3.6 NO to NO2 Chemistry While the CALPUFF model can predict ambient NO and NO2 concentrations, the calculation has been shown to overestimate ambient NO2 concentrations. For this assessment, the ozone limiting method (OLM) was applied to account for this overestimation. The OLM assumes that the conversion of NO to NO2 in the atmosphere is limited by the ambient ozone (O3) concentration in the atmosphere. The approach assumes that 10 percent (on a volume basis) of the NO is converted to NO2 prior to discharge into the atmosphere. For the remaining NO, the following is adopted:

• If 0.9 (NO) is greater than the ambient O3 concentration then NO2 = 0.1 (NO) + 0.9 (O3). For this case, the conversion is not complete.

• If 0.9 (NO) is less than the ambient O3 concentration then NO2 = 0.1 (NO) + 0.9 (NO) = NO. This is equivalent to the total conversion approach, since there is sufficient ozone to effect the complete conversion.


December 2006

Page 5D-17

In the application of the OLM, the above relationships assume the concentrations are expressed on a ppb basis.

Alberta Environment (2003a) recommends ambient ozone concentrations for 1-h, 24-h and annual averaging periods (i.e., 50, 40 and 35 ppb for rural areas, and 50, 35 and 20 ppb for urban areas). Alternately, hourly ambient ozone data can be used to calculate the NO to NO2 conversion on an hourly basis. For consistency, the hourly ozone data should coincide with the meteorological data used in the modelling. For the application of the OLM approach in this assessment, hourly ozone data from Lamont for 2002 were used to estimate hourly NO2 concentrations. The Lamont ozone data are discussed in Appendix 5B.

5D.3.7 Particulate Formation The CALPUFF model was used to predict secondary PM2.5 formation due to precursor SO2 and NOx emissions. The model predicts particulate nitrate NO3

-, which can exist as an aerosol (i.e., dissolved in a water droplet) or as a particle (e.g., NH4NO3). Similarly, sulphate SO4

2- can also exist as an aerosol (i.e., dissolved in a water droplet) or as a particle (e.g., (NH4)2SO4). As the predicted NO3

- and SO42- concentrations have been assumed to react with ambient ammonia

(NH3) to produce ammonium nitrate and ammonium sulphate, respectively; the predicted sulphate and nitrate are multiplied by the factors indicated in Table 5D-4.

Table 5D-4: PM2.5 Multipliers for SO42- and NO3

- Predicted Parameter SO4

2- NO3-

Molecular Mass 96 62 End Product (NH4)2SO4 NH4NO3 Molecular Mass 132 80 Multiplier 1.375 1.290

NOTE: Multiplier = (Molecular Mass of End Product)/(Molecular Mass of Predicted Parameter)

5D.3.8 Intermittent Sources The CALPUFF model was also used to evaluate air quality changes associated with short-term events (e.g., flaring) that are typically 1 hour or less in duration. When evaluating these sources, each upset flaring event was addressed as an isolated event by adjusting the model input so puffs associated with a given hour did not overlap with puffs from preceding or following hours. Specifically, the event was assumed to occur once every 4 hours, and the model was run four times with staggered emission releases to include all hours in the meteorological data. In this manner, the highest one-hour SO2 concentration due to each intermittent event could be determined. If for some reason, a calm condition was to persist for 4 hours and the predicted puffs were to overlap, the model would overstate the air quality effect by overpredicting the expected concentrations.


December 2006

Page 5D-18

5D.3.9 Deposition The CALPUFF model was used to predict SO2, SO4

2-, NO, NO2, HNO3, and NO3- deposition as

annual averages. The Potential Acid Input (PAI) due to these individual substances was calculated as follows:

[ ] [ ] [ ] [ ] [ ] [ ]−− +++++= 3322

42 NO621HNO

631NO

461NO

301SO

962SO

642PAI

where the PAI is expressed in keq H+/ha/a and the values in the brackets [ ] represent the sum of the predicted wet and dry deposition in kg/ha/a. The multiplication coefficients account for valance and molecular mass differences for the individual species. The neutralizing effect of base cation contribution was addressed by subtracting 0.14 keq H+ /ha/a (see Appendix 5B).

The total nitrogen deposition (N) was calculated as follows:

[ ] [ ] [ ] [ ]−+++= 332 NO6214HNO

6314NO

4614NO

3014N

where the N is expressed in kg N/ha/a and the values in the brackets [ ] represent the sum of the predicted wet and dry deposition in kg/ha/a. The multiplication coefficients account for molecular mass differences for the individual species. No background term was added as the model area is sufficiently large that the contribution from the main sources is accounted for explicitly in the model.

5D.3.10 Interpretation of Predictions Alberta Environment (2003a) recommends discarding the eight highest 1-h predictions at each receptor location during any given year, as these values “are considered outliers and should not be used as the basis for selecting stack height”. This means that the hourly AAAQO values should be compared to the 9th highest prediction, not to the highest prediction. For a one-year period, the 9th highest value corresponds to the 99.9 percentile of predicted concentrations.

For PM2.5, the CWS is applicable to the 98th percentile. The 98th percentile was taken as the 8th highest daily average at each receptor location based on the 365-day simulation period.

5D.4 CALPUFF Performance

5D.4.1 Model Prediction Confidence Uncertainty associated with dispersion model predictions stems from two main areas (U.S. EPA 2005):

• Reducible uncertainty results from uncertainties associated with the input values and with the limitations of the model physics and formulations. Reducible uncertainty can be minimized by better (i.e., more accurate and representative) measurements and improved model physics.

• Inherent uncertainty is associated with the stochastic nature of the atmosphere and its representation. Models predict concentrations that represent an ensemble average of numerous repetitions for the same nominal event. An individual observed value can deviate significantly from the ensemble value. This uncertainty may be responsible for a ±50 percent deviation from the measured values.


December 2006

Page 5D-19

Generally, models are quoted as having a factor-of-two accuracy. Comparison studies have indicated that models can predict the magnitude of highest concentration occurring sometime and somewhere within an area to within ±10 to ±40 percent. Predictions for a specific site and time are often poorly correlated with observed values. This poor correlation can be related to reducible errors in wind direction. For example, an uncertainty of 5º to 10º in the wind direction can produce a concentration error in the 20 to 70 percent range. (U.S. EPA 2005).

The confidence associated with CALPUFF/CALMET dispersion model predictions was determined by comparing the ambient air quality data measured at the FAP sites with model predictions. Any comparison should note:

• The Base Case scenario described in Appendix 5A does not represent existing emissions since the Base Case includes approved facilities that are not yet operating. The model predictions also assume emissions are constant with time, and that no upset or abnormal conditions occur. The comparison provided in this section focuses on existing sources (i.e., the approved but not yet constructed BA Energy Heartland Upgrader is not included) and is referred to as the Existing Case.

• The ambient monitoring data represent contributions from existing facilities in the study area and smaller contributions from sources outside the study area. The model, as applied for this comparison assessment, does not include the contribution from sources outside the study area.

• The existing sources are subject to hour-to-hour and day-to-day variability associated with normal and abnormal emissions. The model, as applied for this comparison assessment, does not explicitly account for abnormal emissions.

The CALPUFF model comparison was undertaken for existing sources and the comparison focuses on SO2, NOx, NO2 and PM2.5 concentration comparisons.

5D.4.2 Sulphur Dioxide Comparisons The comparison of measured and predicted SO2 concentrations provides the best indication of model performance because: the emissions originate from a few, well documented sources; chemical reactions that affect SO2 concentrations are not significant for the associated transport times; are significantly large that they can be measured in the ambient air; and there are a number of locations where ambient measurements are taken. The modelling, however, does not account for upset and abnormal events. Measured and predicted 1-h, 24-h and annual average values were compared.


December 2006

Page 5D-20

5D.4.3 1-h SO2 Comparison Table 5D-5 compares the measured 1-h average values at the FAP stations with the predicted maximum (i.e., the 100th percentile) and 99.9th percentile values at the same locations. The table shows the prediction bias and provides a scatter plot. The comparison indicates that:

• There is an average bias to underpredict by 6 percent, based on comparing maximum values.

• There is an average bias to overpredict by 43 percent, based on comparing 99th percentile values.

• The model is predicting high concentrations where high concentrations are measured, and the model is predicting low concentrations where low concentrations are measured.

• Hourly average SO2 exceedances have only been measured at the Scotford and Redwater stations. The model predictions only indicate a potential for exceedances at these two stations.

This comparison only focuses on high values, which is only one measure of model performance. On average, the model is predicting within a factor of two. On average, there is a reasonable correlation between predicted and measured 1-h average SO2 concentrations.

5D.4.4 24-h SO2 Comparison Table 5D-6 compares the measured 24-h average values at the FAP stations with the predicted maximum (i.e., the 100th percentile) and 99.9th percentile values at the same locations. The table shows the prediction bias and provides a scatter plot. The comparison indicates that:

• There is an average bias to overpredict by 18 percent, based on comparing maximum values.

• There is an average bias to underpredict by 40 percent, based on comparing 99th percentile values.


This comparison only focuses on high values, which is only one measure of model performance. On average, the model is predicting within a factor of two at most locations. The highest underprediction is at the Scotford station. On average, there is a reasonable correlation between predicted and measured 24-h average SO2 concentrations.


December 2006

Page 5D-21

Table 5D-5: Comparison of 1-h SO2 Concentrations with those Predicted for Existing Sources

Maximum Predicted (100th Percentile) 99.9th Percentile Measured Predicted (Existing Case) Measured Predicted (Existing Case)

FAP Station

SO2 Concentration

(µg/m³)

SO2 Concentration

(µg/m³)

Prediction Bias (%)

SO2 Concentration

(µg/m³)

SO2 Concentration

(µg/m³)

Prediction Bias (%)

Fort Saskatchewan 147 101 -31 42 60 +42 Lamont 121 87 -28 50 69 +39 Ross Creek 131 121 -8 79 80 +2 Range Road 220 280 162 -42 53 111 +112 Scotford 463 752 +62 286 245 -14 Redwater 1138 1229 +8 482 848 +76 Average prediction bias -6 +43

NOTES:

A “+” prediction bias indicates a tendency to overpredict.

A “-” prediction bias indicates a tendency to underpredict.

10 100 1000Maximum Measured SO2 (µg/m3)

10

100

1000

Max

imum

Pre

dict

ed S

O2 (µg

/m3 )

LegendFort Saskatchewan 100 & 99.9Lamont 100 & 99.9Scotford 100 & 99.9Ross Creek 100 & 99.9Range Road 220 100 & 99.9Redwater 100 & 99.9


December 2006

Page 5D-22

Table 5D-6: Comparison of 24-h SO2 Concentrations with those Predicted for Existing Sources

Maximum Predicted (100th Percentile) 99.9th Percentile Measured Predicted (Existing Case) Measured Predicted (Existing Case)

FAP Station

SO2 Concentration

(µg/m³)

SO2 Concentration

(µg/m³)

Prediction Bias (%)

SO2 Concentration

(µg/m³)

SO2 Concentration

(µg/m³)

Prediction Bias (%)

Fort Saskatchewan 17.1 39.6 +132 13.5 18.5 +8 Lamont 27.6 23.2 -16 22.9 15.9 -42 Ross Creek 37.5 38.5 +3 30.9 20.6 -45 Range Road 220 33.9 54.6 +61 23.1 20.6 -39 Scotford 129 63.5 -51 116 30.6 -76 Redwater 225 178 -21 185 119 -47 Average prediction bias +18 -40

NOTES:



10 100 1000Maximum Measured SO2 (µg/m3)

10

100

1000

Max

imum

Pre

dict

ed S

O2 (µg

/m3 )

LegendFort Saskatchewan 100 & 99.9Lamont 100 & 99.9Scotford 100 & 99.9Ross Creek 100 & 99.9Range Road 220 100 & 99.9Redwater 100 & 99.9


December 2006

Page 5D-23

5D.4.5 Annual SO2 Comparison Table 5D-7 compares the measured annual average values at the FAP stations with the predicted values at the same locations. The table shows the prediction bias and provides a scatter plot. The comparison indicates that:

• There is an average bias to overpredict by 6 percent.


The model is predicting within a factor of two at most locations. The highest underpredictions are at the Stations 4 and 5. On average, there is a reasonable correlation between predicted and measured 24-h average SO2 concentrations.

5D.4.6 Nitrogen Dioxide Comparisons The predicted NO2 concentrations are based on contributions from industrial and non-industrial sources, and NOx emissions are not as well documented as the SO2 emissions. The non-industrial emissions undergo more daily and seasonal variations than the industrial sources, and these variations were not incorporated in the model. While the SO2 predictions have to account for emission, transport and dispersion processes, the NO2 predictions have to account for emission, transport, dispersion, and chemical transformation processes. The ability for the model to predict NO2 concentrations can therefore be more demanding than that for SO2. Measured and predicted 1-h NOx and NO2 concentrations were compared, and measured and predicted 24-h and annual average NO2 values were compared.

5D.4.7 1-h NOx Comparison Table 5D-8 compares the measured 1-h average values at the FAP stations with the predicted maximum (i.e., the 100th percentile) and 99.9th percentile values at the same locations. The table shows the prediction bias and provides a scatter plot. The comparison indicates that:

• There is an average bias to overpredict by 80 percent, based on comparing maximum values.



The comparison indicates that, on average, the model is predicting within a factor of two for the higher NOx concentrations sites and overpredicting by a factor of about two for the other sites. On average, there is a reasonable correlation between predicted and measured 1-h average NOx concentrations.


December 2006

Page 5D-24

Table 5D-7: Comparison of Annual SO2 Concentrations with those Predicted for Existing Sources

Average Predicted (Existing Case) FAP Station

Average Measured SO2 Concentration

(µg/m³) SO2 Concentration

(µg/m³) Prediction Bias

(%) Fort Saskatchewan 2.9 4.6 +59 Lamont 3.9 3.9 0 Ross Creek 1.7 4.8 +182 Range Road 220 2.6 4.3 +65 Scotford 7.9 5.9 -25 Redwater 8.6 11.9 +38 1 3.4 4.5 +32 2 4.4 3.7 -16 3 2.9 2.2 -24 4 2.3 0.9 -61 5 2.2 0.7 -68 6 3.5 3.2 -9 7 4.4 2.5 -43 8 3.8 2.4 -37 9 2.9 3.0 +3 10 3.4 3.5 +3 Average prediction bias +6

NOTES:



0.1 1 10 100Maximum Measured SO2 (µg/m3)

0.1

1

10

100

Max

imum

Pre

dict

ed S

O2 (µg

/m3 )

Legend12345678910Fort SaskatchewanLamontPoss CreekRange Road 220ScotfordRedwater


December 2006

Page 5D-25

5D.4.8 1-h NO2 Comparison Table 5D-9 compares the measured 1-h average values at the FAP stations with the predicted maximum (i.e., the 100th percentile) and 99.9th percentile values at the same locations. The table shows the prediction bias and provides a scatter plot. The comparison indicates that:




The comparison indicates that, on average, the model is predicting within a factor of two. On average, there is a reasonable correlation between predicted and measured 1-h average NOx concentrations.

5D.4.9 24-h NO2 Comparison Table 5D-10 compares the measured 24-h average values at the FAP stations with the predicted maximum (i.e., the 100th percentile) and 99.9th percentile values at the same locations. The table shows the prediction bias and provides a scatter plot. The comparison indicates that:


• There is an average bias to underpredict by 28 percent, based on comparing 99th percentile values.

• With the exception of Ross Creek, the model is predicting high concentrations where high concentrations are measured, and the model is predicting low concentrations where low concentrations are measured.

This comparison only focuses on the maximum values, which is only one measure of model performance. On average, the model is predicting within a factor of two at most locations. The highest underprediction is at the Ross Creek station. On average, there is a reasonable correlation between predicted and measured 24-h average NO2 concentrations.

5D.4.10 Annual NO2 Comparison Table 5D-11 compares the measured annual average values at the FAP stations with the predicted values at the same locations. The table shows the prediction bias and provides a scatter plot. The comparison indicates that:

• There is an average bias to overpredict by 28 percent.

• The model is generally predicting high concentrations where high concentrations are measured, and the model is predicting low concentrations where low concentrations are measured.

The model is predicting within a factor of two at most locations. The highest underprediction is at the Station 3, and the highest overprediction is at the Lamont station. On average, there is a reasonable correlation between predicted and measured annual average NO2 concentrations.


December 2006

Page 5D-26

Table 5D-8: Comparison of 1-h NOx Concentrations with those Predicted for Existing Sources

Maximum (100th Percentile) 99.9th Percentile Measured Predicted (Existing Case) Measured Predicted (Existing Case)

FAP Station

NOx Concentration

(µg/m³)

NOx Concentration

(µg/m³)

Prediction Bias (%)

NOx Concentration

(µg/m³)

NOx Concentration

(µg/m³)

Prediction Bias (%)

Fort Saskatchewan 875 935 +7 534 641 +20 Lamont 230 678 +195 141 377 +167 Station 401 545 887 +63 302 652 +116 Ross Creek 1693 1497 -12 1195 993 -17 Range Road 220 376 709 +89 263 545 +107 Redwater 412 988 +140 282 614 +118 Average prediction bias +80 +85

NOTES:



100 1000Maximum Measured NOx (µg/m3)

100

1000

Max

imum

Pre

dict

ed N

Ox (µg

/m3 )

LegendFort Saskatchewan 100 & 99.9Lamont 100 & 99.9Station 401100 & 99.9Ross Creek 100 & 99.9Range Road 220 100 & 99.9Redwater 100 & 99.9


December 2006

Page 5D-27

Table 5D-9: Comparison of 1-h NO2 Concentrations with those Predicted for Existing Sources


FAP Station

NO2 Concentration

(µg/m³)

NO2 Concentration

(µg/m³)

Prediction Bias (%)

NO2 Concentration

(µg/m³)

NO2 Concentration

(µg/m³)

Prediction Bias (%)

Fort Saskatchewan 154 139 -10 117 129 +16 Lamont 102 127 +25 75 97 +5 Station 401 113 164 +46 94 128 +14 Ross Creek 282 227 -19 226 171 -39 Range Road 220 245 129 -47 132 115 -53 Redwater 169 173 +2 94 139 +18 Average prediction bias -3 15

NOTES:



10 100 1000Maximum Measured NO2 (µg/m3)

10

100

1000

Max

imum

Pre

dict

ed N

O2 (µg

/m3 )



December 2006

Page 5D-28

Table 5D-10: Comparison of 24-h NO2 Concentrations with those Predicted for Existing Sources

Maximum (100th Percentile) 99.9th Percentile Predicted Measured Predicted (Existing Case) Measured Predicted (Existing Case)

FAP Station

NO2 Concentration

(µg/m³)

NO2 Concentration

(µg/m³)

Prediction Bias (%)

NO2 Concentration

(µg/m³)

NO2 Concentration

(µg/m³)

Prediction Bias (%)

Fort Saskatchewan 104 80 -23 96 64 -34 Lamont 42 59 +40 42 41 -1 Station 401 83 79 -4 78 60 -24 Ross Creek 206 103 -50 186 72 -62 Range Road 220 98 68 -31 88 58 -34 Redwater 94 107 +14 71 62 -13 Average prediction bias -9 -28

NOTES:




10

100

1000

Max

imum

Pre

dict

ed N

O2 (µg

/m3 )



December 2006

Page 5D-29

Table 5D-11: Comparison of Annual NO2 Concentrations with those Predicted for Existing Sources

Average Predicted (Existing Case) FAP Station

Average Measured NO2 Concentration

(µg/m³) NO2 Concentration

(µg/m³) Prediction Bias

(%) Fort Saskatchewan 24.1 24.5 +2 Lamont 5.9 15.1 +156 Station 401 20.0 22.0 +10 Ross Creek 31.6 25.9 -18 Range Road 220 13.5 15.9 +18 Redwater 16.4 16.6 +1 1 23.6 23.3 -1 2 13.8 18.4 +33 3 34.0 9.1 -73 4 9.1 4.9 -46 5 8.9 4.5 -49 6 12.7 14.4 +13 7 15.2 13.4 -12 8 16.2 13.1 -19 9 15.3 14.8 -3 10 15.3 16.0 +5 Average prediction bias +28

NOTES:




1

10

100

Max

imum

Pre

dict

ed N

O2

(µg/

m3 )

Legend12345678910Fort SaskatchewanLamontStation 401Ross CreekRange Road 220Redwater


December 2006

Page 5D-30

5D.4.11 PM2.5 Comparison The measured PM2.5 concentrations include contributions from industrial, traffic and community sources, as well as contributions from other sources. The latter includes forest fires, agricultural operations and windborne dust. These emission sources are not as well documented as the SO2 emissions. The ability for the model to predict PM2.5 concentrations can therefore be more demanding. Measured and predicted 1-h, 24-h and annual average values were compared. The predicted values included primary PM2.5 (i.e., directly emitted) and secondary PM2.5 (i.e., sulphate formed from SO2 emissions and nitrate formed from NOx emissions).

Table 5D-12 provides the comparison for the FAP sites, which indicates that:

• The predicted maximum and 99.9th percentile 1-h predictions are similar to the 99th percentile measured values. The extreme measured maxima values are likely influenced by local or other sources (e.g., forest fires) that were not included in the modelling.

• The predicted maximum and 98th percentile 24-h values are slightly greater than the corresponding measured values for the Fort Saskatchewan and Redwater stations. For the Lamont and Elk Island stations, there is good agreement.

• The annual average predicted and measured values at the Elk Island station are 3.2 and 4.4 µg/m3, respectively. For the other stations, the predictions are in the 6.3 to 8.8 µg/m3

range, and the measured values are in the 5.4 to 7.3 µg/m3 range.

The predicted values show a reasonable agreement with the available measurements.

5D.4.12 CO Comparison Table 5D-13 compares the predicted and measured CO 1-h concentrations for the Fort Saskatchewan monitoring station. The maximum predicted value falls between the maximum and 99.9th percentile measurements. The predicted values show a reasonable agreement with the available measurements.

5D.4.13 H2S Comparison Table 5D-14 compares the predicted and measured H2S 1-h and 24-h concentrations for the Fort Saskatchewan monitoring station. The closest agreement is for the Scotford station (Site D), where the maximum predicted 1-h concentration falls between the maximum and 99.9th percentile measured values, and the maximum predicted 24-h average concentration equals the measured 99.9th percentile. The predicted and measured values at this site are attributable to Scotford Complex emissions. For the other two locations, there is a tendency to underpredict H2S concentrations. This underprediction is likely due to not including H2S from other sources in the modelling (e.g., the refineries in the east Edmonton area or the field oil and gas operations in the Lamont area).

In conclusion, the model is providing a measure of high H2S concentrations near industrial sources where the H2S emissions are known; it is also in these areas where the H2S concentrations are the highest. The model is underpredicting in areas that are distant from industrial sources; this is likely due to not including the smaller emission sources.


December 2006

Page 5D-31

Table 5D-12: Comparison of PM2.5 Concentrations with those Predicted for Existing Sources

Measured PM2.5 Concentration Monitoring

Station

Averaging Period (µg/m3)

Predicted PM2.5 Concentration (Existing Case)

(µg/m3) Comment

1-h maximum 394.1 194.5 1-h 99.9th percentile 65.2 154.8 1-h 99th percentile 28.8 1-h 95th percentile 16.2 1-h 90th percentile 11.9

Predicted is between measured 99.9th percentile and maximum

24-h maximum 64.4 75.8 Measured and predicted are similar 24-h 98th percentile 18.6 39.2 Predicted is larger than measured

Fort Saskatchewan

Annual 5.4 8.8 Predicted is larger than measured 1-h maximum 208.6 116.8 1-h 99.9th percentile 61.6 84.3 1-h 99th percentile 33.0 1-h 95th percentile 20.5 1-h 90th percentile 16.2

Predicted is similar to measured 99.9th percentile

24-h maximum 51.2 56.1 Measured and predicted are similar 24-h 98th percentile 22.1 24.6 Measured and predicted are similar

Lamont

Annual 7.3 6.7 Measured and predicted are similar 1-h maximum 116.0 158.1 1-h 99.9th percentile 56.0 100.5 1-h 99th percentile 31.0 1-h 95th percentile 19.0 1-h 90th percentile 14.0

Predicted is between measured 99th and 99.9th percentiles

24-h maximum 25.1 47.6 Predicted is much larger than measured

24-h 98th percentile 20.1 25.0 Measured and predicted are similar

Redwater

Annual 7.0 6.3 Measured and predicted are similar 1-h maximum 325.6 106.7 1-h 99.9th percentile 51.3 72.4 1-h 99th percentile 24.7 1-h 95th percentile 13.4 1-h 90th percentile 9.8

Predicted is smaller than measured

24-h maximum 37.1 35.1 Measured and predicted are similar 24-h 98th percentile 15.3 13.2 Measured and predicted are similar

Elk Island

Annual 4.4 3.2 Measured and predicted are similar


December 2006

Page 5D-32

5D.4.14 Benzene Comparison Table 5D-16 compares the predicted and measured benzene 24-h and annual concentrations for the EC/FAP VOC monitoring sites. The closest agreement is for the Scotford station (Site D), where the maximum predicted 24-h concentration falls between the 95th and 99th percentile measured values. A perfect agreement would not be expected as the model predictions are based on annual average emissions and fugitive benzene sources can be intermittent events. The predicted and measured values at this site are attributable to Scotford Complex emissions. For the other locations, there is a tendency to underpredict benzene concentrations. This underprediction is likely due to not including benzene from transportation emissions in the modelling.

In conclusion, the model is providing a measure of high benzene concentrations near industrial sources where the benzene emissions are known; it is also in these areas where the benzene concentrations are the highest. The model is underpredicting in areas that are distant from industrial sources; this is likely due to not including the transportation emissions. Lower benzene concentrations are measured in these areas.

Table 5D-13: Comparison of 1-h CO Concentrations (µg/m³) with those Predicted for Existing Sources


FAP Station

CO Concentration

CO Concentration

Prediction Bias (%)

CO Concentration

CO Concentration

Prediction Bias (%)

Fort Saskatchewan 5839 3222 -45 2748 2244 -18

Average prediction bias -45 -18


December 2006

Page 5D-33

Table 5D-14: Comparison of 1-h and 24-h H2S Concentrations (µg/m³) with those Predicted for the Existing Sources

Measured H2S Concentration FAP Sites


Predicted Benzene Concentration (Existing Case)

(µg/m3) 1-h maximum 11.1 0.7 1-h 99.9th percentile 5.6 0.4 1-h 99th percentile 2.8 1-h 95th percentile 1.4 1-h 90th percentile 1.4

24-h maximum 5.2 0.2 24-h 99.9th percentile 3.2 0.1 24-h 99th percentile 2.3 24-h 95th percentile 1.5

Fort Saskatchewan

24-h 90th percentile 1.0



Lamont




Scotford



December 2006

Page 5D-34

Table 5D-15: Comparison of 24-h and Annual Benzene Concentrations with those Predicted for Existing Sources

Measured Benzene Concentration EC/ FAP VOC Sites


Predicted Benzene Concentration (Existing Case)

(µg/m3) 24-h maximum 2.8 0.8 24-h 99.9th percentile 2.7 0.5 24-h 99th percentile 1.9 24-h 95th percentile 1.4 24-h 90th percentile 1.2

A

Annual 0.60 0.07 24-h maximum 4.1 0.7 24-h 99.9th percentile 4.0 0.3 24-h 99th percentile 3.2 24-h 95th percentile 2.3 24-h 90th percentile 1.7

B (FAP Ross Creek monitoring site)


C (FAP Station 401 monitoring site)


D (FAP Scotford monitoring site)


E


F (FAP Elk Island monitoring site)

Annual 0.40 0.01


December 2006

Page 5D-35

5D.5 Summary and Conclusions The CALPUFF dispersion model (Version 6.112, Level 060412) was selected as the primary air quality assessment tool to predict ambient concentrations and deposition. The following were adopted for the application of the model:

• 5377 receptor grid points were selected for an 80 x 80 km CALPUFF receptor area. Receptor density is increased near the FHELP Sturgeon Project. Receptor spacing varies from 50 m nearest the facility to 2 km at the furthest reaches of the modeling domain. This total includes the 231 community and monitoring site receptors, most of which are within a 7.5 km radius of the Sturgeon Project.

• One year of meteorological data for the period January 2002 to December 2002 was selected. The CALMET model (see Appendix 5C) was used to provide the meteorological data for the CALPUFF model.

• Hourly ozone concentrations from Lamont were used. The ozone limiting method was selected to estimate ambient NO2 concentrations from the predicted NOx values.

A comparison between model predictions and measurements indicates that on average, the CALPUFF/CALMET model system is predicting:

• High 1-h average SO2 concentrations that are in the range of -42 to 62 percent of the measurements, with an average underprediction of 6 percent.

• High 24-h average SO2 concentrations that are in the range of -51 to 132 percent of the measurements, with an average overprediction of 18 percent.

• Annual average SO2 concentrations that are in the range of -68 to 182 percent of the measurements, with an average overprediction of 53 percent.

• High 1-h average NO2 concentrations that are in the range of -47 to 46 percent of the measurements, with an average underprediction of 1 percent.

• High 24-h average NO2 concentrations that are in the range of -50 to 40 percent of the measurements, with an average underprediction of 9 percent.

• Annual average NO2 concentrations that are in the range of -49 to 156 percent of the measurements, with an average overprediction of 28 percent.

• PM2.5 and CO concentrations that are reasonable.

• H2S and benzene concentrations that are similar to measurements near identified H2S and benzene sources. Predicted values at more distant locations are underestimated.

These are general conclusions and there are site-to-site variations. Overpredicting annual average SO2 and NO2 concentrations leads to an overestimate of the dry deposition contribution to the PAI.


December 2006

Page 5D-36

5D.6 CALPUFF Model Options For the purposes of organization, the CALPUFF control file defines 18 input groups as identified in Table 5D-16. For many of the options, default values used in the absence of site/project specific data. Tables 5D-17 to 5D-24 identify the input parameters, the default options, and the values used for the Project assessment.

Table 5D-16: Input Groups in the CALPUFF Control File Input Group Description Applicable to Project?

0 Input and output file names Yes 1 General run control parameters Yes 2 Technical options Yes 3 Species list Yes 4 Grid control parameters Yes 5 Output options Yes 6 Sub grid scale complex terrain inputs No 7 Dry deposition parameters for gases Yes 8 Dry deposition parameters for particles Yes 9 Miscellaneous dry deposition for parameters Yes 10 Wet deposition parameters Yes 11 Chemistry parameters Yes 12 Diffusion and computational parameters Yes 13 Point source parameters Yes 14 Area source parameters Yes 15 Line source parameters No 16 Volume source parameters No 17 Discrete receptor information Yes


December 2006

Page 5D-37

Table 5D-17: CALPUFF Model Options Groups 1 and 2

Input Group 1: General Run Control Parameters Parameter Default FHELP Sturgeon Comments

METRUN 0 0 All model periods in met file(s) will be run IBYR - 2002 Starting year IBMO - 1 Starting month IBDY - 1 Starting day IBHR - 0 Starting hour XBTZ 7 Base time zone (7 = MST) NSPEC 5 10 Number of chemical species NSE 3 7 Number of chemical species to be emitted ITEST 2 2 Program is executed after SETUP phase MRESTART 0 2 Write a restart file during run NRESPD 0 24 File updated every 24 periods METFM 1 1 CALMET binary file (CALMET.MET) AVET 60 60 Averaging time in minutes PGTIME 60 60 PG Averaging time in minutes

Input Group 2: Technical Options Parameter Default FHELP Sturgeon Comments

MGAUSS 1 1 Gaussian distribution used in near field MCTADJ 3 3 Partial plume path terrain adjustment MCTSG 0 0 Scale-scale complex terrain not modelled MSLUG 0 0 Near-field puffs not modelled as elongated MTRANS 1 1 Transitional plume rise modelled MTIP 1 1 Stack tip downwash used MBDW 1 2 PRIME Method is used to simulate building

downwash MSHEAR 0 1 (0,1) Vertical wind shear (not modelled,

modelled) MSPLIT 0 0 Puffs are not split MCHEM 1 3 Transformation rates computed internally

using (RIVID/ARM3) scheme MAQCHEM 0 0 Aqueous phase transformation not

modelled MWET 1 1 Wet removal modelled MDRY 1 1 Dry deposition modelled MTILT 0 0 Gravitational settling (plume tilt) is not

modeled


December 2006

Page 5D-38

Table 5D-17: CALPUFF Model Options Groups 1 and 2 (cont’d)

Input Group 2: Technical Options (cont’d) Parameter Default FHELP Sturgeon Comments

MDISP 3 2 Dispersion coefficients from internally calculated sigma v, sigma w using micrometeorological variables (u*, w*, L, etc.)

MTURBVW 3 3 Use both σv and σw from PROFILE.DAT to compute σy and σz (n/a)

MDISP2 3 3 PG dispersion coefficients for rural areas (computed using ISCST3 approximation) and MP coefficients in urban areas when measured turbulence data is missing

MTAULY 0 0 Draxler default 617.284 (s) MTAUADV 0 0 No turbulence advection MCTURB 1 1 Standard CALPUFF subroutines MROUGH 0 1 PG σy and σz is adjusted for roughness MPARTL 1 1 No partial plume penetration of elevated

inversion MTINV 0 0 Strength of temperature inversion computed

from default gradients MPDF 0 0 PDF not used for dispersion under

convective conditions MSGTIBL 0 0 Sub-grid TIBL module not used for

shoreline MBCON 0 0 Boundary concentration conditions not

modelled MSOURCE 0 0 Individual source contributions not saved MFOG 0 0 Do not configure for FOG model output MREG 1 0 Do not test options specified to see if they

conform to regulatory values


December 2006

Page 5D-39

Table 5D-18: CALPUFF Model Options Groups 3 and 4

Input Group 3: Species List-Chemistry Options CSPEC Modelled1 Emitted2 Dry Deposition3 Output Group Number

SO2 1 1 1 0 SO4

2- 1 0 2 0 NO 1 1 1 0 NO2 1 1 1 0 HNO3 1 0 1 0 NO3

- 1 0 2 0 NOx 1 1 0 0 PM 1 1 0 0 VOC 1 1 0 0 CO 1 1 0 0

NOTES: 1 0=no, 1=yes 2 0=no, 1=yes 3 0=none, 1=computed-gas, 2=computed particle, 3=user-specified

Input Group 4: Map Projection and Grid Control Parameters Parameter Default Project Comments PMAP UTM UTM Universal Transverse Mercator FEAST 0 0 False Easting (km) at the projection origin FNORTH 0 0 False Northing (km) at the projection origin IUTMZN - 12 UTM zone UTMHEM N N Northern Hemisphere for UTM projection DATUM WGS-84 NAR-C NAR-C used for output coordinates NX - 100 Number of X grid cells in meteorological grid NY 100 Number of Y grid cells in meteorological grid NZ - 8 Number of vertical layers in meteorological grid DGRIDKM - 1 Grid spacing (km) to match CALMET (see Appendix

5C) ZFACE - 0,20,40,80,160,320,

600,1400,2600 Cell face heights in meteorological grid (m)

XORIGKM - 307 Reference X coordinate for SW corner of grid cell (1,1) of meteorological grid (km)

YORIGKM - 5917 Reference Y coordinate for SW corner of grid cell (1,1) of meteorological grid (km)


December 2006

Page 5D-40

Table 5D-18: CALPUFF Model Options Groups 3 and 4 (cont’d)

Input Group 4: Map Projection and Grid Control Parameters (cont’d) Parameter Default Project Comments IBCOMP - 1 X index of lower left corner of the computational grid JBCOMP - 1 Y index of lower left corner of the computational grids IECOMP - 100 X index of the upper right corner of the computational

grid JECOMP - 100 Y index of the upper right corner of the computational

grid LSAMP T F Sampling grid is not used IBSAMP - 1 X index of lower left corner of the sampling grid JBSAMP - 1 Y index of lower left corner of the sampling grid IESAMP - 100 X index of upper right corner of the sampling grid JESAMP - 100 Y index of upper right corner of the sampling grid MESHDN 1 1 Nesting factor of the sampling grid

Table 5D-19: CALPUFF Model Option Group 5

Input Group 5: Output Option Parameter Default Project Comments ICON - 1 Output file CONC.DAT containing concentrations is created IDRY - 1 Output file DFLX.DAT containing dry fluxes is created IWET - 1 Output file WFLX.DAT containing wet fluxes is created IT2D 0 0 2D Temperature IRHO 0 0 Density IVIS 1 0 Output file containing relative humidity data is not created LCOMPRS T F Do not perform data compression in output file IQAPLOT 1 1 Create a standard series of output files (e.g. locations of

sources, receptors, grids) suitable for plotting IMFLX 0 0 Do not calculate mass fluxes across specific boundaries IMBAL 0 1 Mass balances for each species reported hourly ICPRT 0 0 Do not print concentration fields to the output list file IDPRT 0 0 Do not print dry flux fields to the output list file IWPRT 0 0 Do not print wet flux fields to the output list file ICFRQ 1 24 Concentration fields are printed to output list file every 24 hour IDFRQ 1 24 Dry flux fields are printed to output list file every 24 hour IWFRQ 1 24 Wet flux fields are printed to output list file every 24 hour


December 2006

Page 5D-41


Input Group 5: Output Option Parameter Default Project Comments IPRTU 1 3 Units for line printer output are in g/m3 for concentration and

g/m2/s for deposition IMESG 2 2 Messages tracking the progress of run are written on screen LDEBUG F F Logical value for debug output IPFDEB 1 1 First puff to track NPFDEB 1 1 Number of puffs to track NN1 1 1 Meteorological period to start output NN2 10 10 Meteorological period to end output

Concentrations printed (0=no, 1=yes)

Dry Fluxes printed (0=no, 1=yes)

Wet Fluxes printed (0=no, 1=yes) Mass Flux

Species Printed Saved to

disk PrintedSaved to

disk Printed Saved to

disk Saved to

disk SO2 0 1 0 1 0 1 0 SO4

2- 0 1 0 1 0 1 0 NO 0 1 0 1 0 1 0 NO2 0 1 0 1 0 1 0 HNO3 0 1 0 1 0 1 0 NO3

- 0 1 0 1 0 1 0 NOx 0 1 0 1 0 1 0 PM 0 1 0 1 0 1 0 VOC 0 1 0 1 0 1 0 CO 0 1 0 1 0 1 0


December 2006

Page 5D-42

Table 5D-20: CALPUFF Model Option Groups 6, 7, and 8

Input Group 6: Sub-Grid Scale Complex Terrain Inputs Parameter Default Project Comments NHILL 0 - Number of terrain features NCTREC 0 - Number of special complex terrain receptors MHILL - 2 Hill data created by OPTHILL & input below in Subgroup (6b);

Receptor data in Subgroup (6c) XHILL2M 1 - Conversion factor for changing horizontal dimensions to meters ZHILL2M 1 - Conversion factor for changing vertical dimensions to meters XCTDMKM - - X origin of CTDM system relative to CALPUFF coordinate system

(km) YCTDMKM - - Y origin of CTDM system relative to CALPUFF coordinate system

(km)

Input Group 7: Dry Deposition Parameters for Gases Species Default Project Comments

0.1509 0.13719 Diffusivity from RWDI (2005) 10000.0 1000 Alpha star

8.0 8.0 Reactivity 0.0 0.0 Mesophyll resistance

SO2

0.4 0.033108 Henry’s Law coefficient - 0.22034 Diffusivity from RWDI (2005) - 1.0 Alpha star - 2 Reactivity - 94 Mesophyll resistance

NO

- 18 Henry’s Law coefficient 0.1656 0.15845 Diffusivity from RWDI (2005) 1.0 1.0 Alpha star 8.0 8 Reactivity 5.0 5 Mesophyll resistance

NO2

3.5 3.5 Henry’s Law coefficient for 0.1628 0.1041 Diffusivity from RWDI (2005) 1.0 1.0 Alpha star

18.0 18 Reactivity 0.0 0 Mesophyll resistance

HNO3

0.00000008 0.00000008 Henry’s Law coefficient


December 2006

Page 5D-43

Table 5D-20: CALPUFF Model Option Groups 6, 7, and 8 (cont’d)

Input Group 8: Dry Deposition Parameters for Particles Species Default Project Comments

SO42- 0.48 0.48 Geometric mass mean diameter of SO4

2- [µm] SO4

2- 2.0 2.0 Geometric standard deviation of SO42- [µm]

NO3 - 0.48 0.48 Geometric mass mean diameter of NO3

-[µm] NO3

- 2.0 2.0 Geometric standard deviation of NO3 - [µm]

NOTE: ‘-‘ Symbol indicates that the parameter was not applicable to the Project.


Input Group 9: Miscellaneous Dry Deposition Parameters Parameters Default Project Comments

RCUTR 30 30 Reference cuticle resistance (s/cm) RGR 10 10 Reference ground resistance (s/cm) REACTR 8 8 Reference pollutant reactivity NINT 9 9 Number of particle size intervals for effective particle

deposition velocity IVEG 1 1 Vegetation in non-irrigated areas is active and unstressed

Input Group 10: Wet Deposition Parameters Species Default Project Comments

3.21E-05 3.21E-05 Scavenging coefficient for liquid precipitation [s-1] SO2 0.0 0.0 Scavenging coefficient for frozen precipitation [s-1]

1.0E-04 1.0E-04 Scavenging coefficient for liquid precipitation [s-1] SO42-

3.0E-05 3.0E-05 Scavenging coefficient for frozen precipitation [s-1] 2.847E-05 2.9E-05 Scavenging coefficient for liquid precipitation [s-1] NO

0.0 0.0 Scavenging coefficient for frozen precipitation [s-1] 5.13E-05 5.1E-05 Scavenging coefficient for liquid precipitation [s-1] NO2

0.0 0.0 Scavenging coefficient for frozen precipitation [s-1] 6.0E-05 6.0E-05 Scavenging coefficient for liquid precipitation [s-1] HNO3

0.0 0.0 Scavenging coefficient for frozen precipitation [s-1] 1.0E-04 1.0E-04 Scavenging coefficient for liquid precipitation [s-1] NO3

- 0.00003 3.0E-05 Scavenging coefficient for frozen precipitation [s-1]


December 2006

Page 5D-44


Input Group 11: Chemistry Parameters Parameters Default Project Comments MOZ 1 1 Hourly ozone values from Lamont were used BCKO3 12*80 “__” Background ozone concentration (ppb) BCKNH3 12*10 4.60,4.10,3.40,5.20,5.20,

6.10,6.80, 8.10,5.30, 4.80, 3.50,6.10

Background ammonia concentration (ppb) (Based on FAP measurements, see Appendix 5B)

RNITE1 0.2 0.2 Night-time NO2 loss rate in percent/hour RNITE2 2 2 Night-time NOX loss rate in percent/hour RNITE3 2 2 Night-time HNO3 loss rate in percent/hour MH202 - - H2O2 data input option BCKH202 - - Monthly background H2O2 concentrations

(Aqueous phase transformations not modelled) BCKPMF - - Fine particulate concentration for Secondary

Organic Aerosol Option OFRAC - - Organic fraction of fine particulate for SOA

Option VCNX - - VOC/NOx ratio for SOA Option


Input Group 12: Diffusion/Computational Parameters Parameters Default Project Comments

SYDEP 550 550 Horizontal size of a puff in metres beyond which the time dependant dispersion equation of Heffter (1965) is used

MHFTSZ 0 0 Do not use Heffter formulas for sigma z JSUP 5 5 Stability class used to determine dispersion rates for puffs

above boundary layer CONK1 0.01 0.01 Vertical dispersion constant for stable conditions CONK2 0.1 0.1 Vertical dispersion constant for neutral/stable conditions TBD 0.5 0.5 Use ISC transition point for determining the transition point

between the Schulman-Scire to Huber-Snyder Building Downwash scheme

IURB1 10 10 Lower range of land use categories for which urban dispersion is assumed

IURB2 19 19 Upper range of land use categories for which urban dispersion is assumed

ILANDUIN 20 - Land use category for modelling domain XLAIIN 3.0 - Leaf area index for modelling domain ZOIN 0.25 - Roughness length in metres for modelling domain


December 2006

Page 5D-45

Table 5D-22: CALPUFF Model Option Group 12 (cont’d)

Input Group 12: Diffusion/Computational Parameters (cont’d) Parameters Default Project Comments

ELEVIN 0.0 - Elevation above sea level XLATIN -999 - North latitude of station in degrees XLONIN -999 - South latitude of station in degrees ANEMHT 10 - Anemometer height in metres ISIGMAV 1 1 Sigma-v is read for lateral turbulence data IMIXCTDM 0 - Predicted mixing heights are used XMXLEN 1 1 Maximum length of emitted slug in meteorological grid

units XSAMLEN 1 10 Maximum travel distance of slug or puff in meteorological

grid units during one sampling unit MXNEW 99 60 Maximum number of puffs or slugs released from one

source during one time step MXSAM 99 60 Maximum number of sampling steps during one time step

for a puff or slug NCOUNT 2 2 Number of iterations used when computing the transport

wind for a sampling step that includes transitional plume rise

SYMIN 1 1 Minimum sigma y in metres for a new puff or slug SZMIN 1 1 Minimum sigma z in metres for a new puff or slug

Parameter

SVMIN SWMIN Stability Class

Minimum turbulence

(σv) (m/s) Minimum turbulence

(σw) (m/s) Land Water Land Water A 0.5 0.37 0.2 0.2 B 0.5 0.37 0.12 0.12 C 0.5 0.37 0.08 0.08 D 0.5 0.37 0.06 0.06 E 0.5 0.37 0.03 0.03 F 0.5 0.37 0.016 0.016


December 2006

Page 5D-46

Table 5D-22: CALPUFF Model Option Group 12 (cont’d) Input Group 12: Diffusion/Computational Parameters (cont’d)

Parameters Default Project Comments CDIV 0.0, 0.0 0.0, 0.0 Divergence criteria for dw/dz in met cells WSCALM 0.5 0.5 Minimum wind speed allowed for non-calm conditions (m/s) XMAXZI 3000 3000 Maximum mixing height in metres XMINZI 50 50 Minimum mixing height in metres

- 1.54 wind speed category 1 [m/s] - 3.09 wind speed category 2 [m/s] - 5.14 wind speed category 3 [m/s] - 8.23 wind speed category 4 [m/s]

WSCAT

- 10.80 wind speed category 5 [m/s]

Parameter PLX0 PPC (see text)

Stability Class Wind speed profile exponent Plume path coefficient A 0.07 0.8 B 0.07 0.7 C 0.10 0.6 D 0.15 0.5 E 0.35 0.4 F 0.55 0.3

Parameters Default Project Comments

0.020 0.020 potential temperature gradient for E stability [K/m] PTG0 0.035 0.035 potential temperature gradient for F stability [K/m]

SL2PF 10 10 Slug-to-puff transition criterion factor equal to sigma y/length of slug

NSPLIT 3 3 Number of puffs that result every time a puff is split IRESPLIT 0,0,0,0,0,0,

0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0

0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0

Time(s) of day when split puffs are eligible to be split once again

ZISPLIT 100 100 Minimum allowable last hour’s mixing height for puff splitting

ROLDMAX 0.25 0.25 Maximum allowable ratio of last hour’s mixing height and maximum mixing height experienced by the puff for puff splitting


December 2006

Page 5D-47

Table 5D-22: CALPUFF Model Option Group 12 (cont’d)

Input Group 12: Diffusion/Computational Parameters (cont’d) Parameters Default Project Comments

NSPLITH 5 5 Number of puffs that result every time a puff is horizontally split

SYSPLITH 1 1 Minimum sigma-y of puff before it may be horizontally split

SHSPLITH 2 2 Minimum puff elongation rate due to wind shear before it may be horizontally split

CNSPLITH 1.0e-7 1.0e-7 Minimum concentration of each species in puff before it may be horizontally split

EPSSLUG 1.00E-04 1.00E-04 Fractional convergence criterion for numerical SLUG sampling iteration

EPSAREA 1.00E-06 1.00E-06 Fractional convergence criterion for numerical AREA sampling iteration

DRISE 1.0 1.0 Trajectory step length for numerical rise HTMINBC 500 - Minimum height (m) to which BC puffs are mixed as

they are emitted (MBCON=2 ONLY) RSAMPBC 10 15 Search radius (km) about a receptor for sampling

nearest BCpuff. MDEPBC 1 0 Concentration is NOT adjusted for depletion

NOTE: ‘-‘ symbol indicates that the parameter was not applicable to the Project.


Input Group 13: Point Source Parameters Parameters Default Project Comments

NPT1 - Varies by scenario

Number of point sources with constant stack parameters or variable emission rate scale factors

IPTU 1 1 Units for point source emission rates are g/s NSPT1 0 0 Number of source-species combinations with variable

emissions scaling factors NPT2 - 0 Number of point sources with variable emission

parameters provided in external file

NOTES: Point source parameters are given in Appendix 5A. ‘-‘ symbol indicates that the parameter was not applicable to the Project assessment.


December 2006

Page 5D-48


Input Group 14: Area Source Parameters Parameters Default Project Comments

NAR1 - Varies by scenario

Number of polygon area sources

IARU 1 1 Units for area source emission rates are g/m2/s NSAR1 0 0 Number of source species combinations with variable

emissions scaling factors NAR2 - 0 Number of buoyant polygon area sources with variable

location and emission parameters NOTES: Area source parameters are given in Part A. ‘-‘ symbol indicates that the parameter was not applicable to the Project assessment.

Input Group 15: Line Source Parameters Parameters Default Project Comments NLN2 - - Number of buoyant line sources with variable location and

emission parameters NLINES - - Number of buoyant line sources ILNU 1 - Units for line source emission rates is g/s NSLN1 0 - Number of source-species combinations with variable

emissions scaling factors MXNSEG 7 - Maximum number of segments used to model each line NLRISE 6 - Number of distance at which transitional rise is computed XL - - Average line source length (m) HBL - - Average height of line source height (m) WBL - - Average building width (m) WML - - Average line source width (m) DXL - - Average separation between buildings (m) FPRIMEL - - Average buoyancy parameter (m4/s3)

NOTE: ‘-‘ symbol indicates that the parameter was not applicable to the Project assessment.


December 2006

Page 5D-49

Table 5D-24: CALPUFF Model Option Groups 16 and 17

Input Group 16: Volume Source Parameters Parameter Default Project Comments

NVL1 - Varies by scenario

Number of volume sources

IVLU 1 - Units for volume source emission rates is grams per second NSVL1 0 - Number of source-species combinations with variable

emissions scaling factors IGRDVL - - Gridded volume source data is not used VEFFHT - - Effective height of emissions (m) VSIGYI - - Initial sigma y value (m) VSIGZI - - Initial sigma z value (m)

NOTE: ‘-‘ symbol indicates that the parameter was not applicable to the Project.

Input Group 17: Discrete Receptor Information Parameter Default Project Comments

NREC - 5377 Number of non-gridded receptors

NOTE: Discrete receptors are identified on Figure 5D-2.


December 2006

Page 5D-50

5D.7 References Alberta Environment. 2003a. Air Quality Model Guideline. Publication Number T/689. Available

from www.gov.ab.ca/env/air/airquality/airmodelling.html. pp 36.

Alberta Environment. 2003b. Emergency/Process Upset Flaring Management: Modelling Guidance. Publication Number T/690. Available from www.gov.ab.ca/env/air/airquality/airmodelling.html. pp 36. pp. 6.

Lott, R.A. 1984. Case Study of Plume Dispersion over Elevated Terrain. Atmospheric Environment. Volume 18. pp125-134.

Morris. R.E., C.Tana, and G. Yarwood. 2003. Evaluation of the Sulphate and Nitrate Formation Mechanism in the CALPUFF Modelling System. Presented at AWMA Conference Guideline on Air Quality Models: The Path Forward. October 22 to 24 2003. Mystic, CT.

Paine, R. J. and B. A. Egan, 1987, User's guide to the Rough Terrain Diffusion Model (RTDM)-Rev. 3.20 ERT. Document PD-535-585. ENSR. Acton, MA, 260 pp

RWDI AIR Inc. 2005. NOx Dispersion and Chemistry Assumptions in the CALPUFF Model. CEMA Contract 2003-0034. Prepared for the Cumulative Environmental Management Association.

Schulman, L.L., D.G. Strimaitis, and J.S. Scire. 1998. Development and Evaluation of the PRIME Plume Rise and Building Downwash Model. Submitted to Journal of the Air & Waste Management Association.

Scire, J.S., D.G. Strimaitis and R.J. Yamartino. 1999. A User’s Guide for the CALPUFF Model (Version5.0). Earth Technologies Inc.

Tikvart, J.A. 1996. Application of Ozone Limiting Method. Model Clearinghouse Memorandum #107. U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC, August 15, 1996.

U.S. EPA. 1989. User’s Guide to the Complex Terrain Dispersion Model Plus Algorithms for Unstable Situations (CTDMPLUS): Volume 1. Model Description and User Instructions. EPA/600/8-89/041.

U.S. EPA. 1995. SCREEN3 Model User’s Guide. EPA-454/R-95-004.

U.S. EPA. 2004. AERMOD: Description of Model Formulation. EPA- 454/R-03-004.

U.S. EPA. 2005. Appendix W to Part 51, Revision to the Guideline on Air Quality Models: Adoption of a preferred general purpose (flat and complex terrain) dispersion model and other revisions; Final Rule. (November 9, 2005 Edition).

Documents

APPENDIX 5D CALPUFF DISPERSION MODEL · Appendix 5D: CALPUFF Dispersion Model Socio-economic Impact Assessment December 2006 Page 5D-6 CALPUFF contains algorithms for near-source