Skip to content

Houston Refinery Modeling Assignments

Advanced tutorials for the PyAERMOD GUI: from meteorological preprocessing through terrain processing to a complete oil refinery dispersion model.

These three tutorials build on Tutorials 1--5 and assume you are comfortable with the PyAERMOD GUI, point/area sources, receptor grids, and basic AERMET concepts.

Table of Contents

  1. Tutorial 6 — AERMET for Houston, Texas
  2. Tutorial 7 — Terrain Processing with AERMAP
  3. Tutorial 8 — Modeling a Simplified Oil Refinery
  4. Assignment Deliverables
  5. Supplemental Reference Tables

Tutorial 6 — AERMET for Houston, Texas

Goal: Generate a complete set of AERMOD-ready meteorological files (.sfc and .pfl) for the Houston Ship Channel area using real station data and site-appropriate surface parameters.

Time: 40--50 minutes

What you'll learn: - How to identify surface and upper air stations for a specific location - How Gulf Coast meteorology differs from inland sites - How to select monthly surface parameters for an industrial/coastal setting - How to run all three AERMET stages and verify the output

Background: Why Houston?

The Houston Ship Channel and surrounding area contains one of the largest concentrations of petroleum refining and petrochemical manufacturing in the world. The meteorology of this region is distinctive:

  • Prevailing winds from the south-southeast (Gulf of Mexico) during summer, rotating to northerly during winter cold fronts.
  • High humidity year-round reduces the Bowen ratio compared to arid inland sites.
  • Sea-breeze circulation can reverse wind direction along the coast during afternoon hours, recirculating pollutants back over populated areas.
  • Atmospheric stability is frequently unstable during summer days (strong solar heating) but can become very stable during clear winter nights.

These characteristics make Houston a challenging and instructive location for air quality modeling.

Step 1: Identify the Meteorological Stations

For a refinery near the Houston Ship Channel (approximately 29.75 N, 95.10 W), we need two weather stations:

Surface Station — Houston Hobby Airport (KHOU)

Parameter Value Notes
Station ID KHOU ICAO code; WBAN 12918
Station Name Houston Hobby Airport
Latitude 29.6454 Decimal degrees
Longitude -95.2789 Negative = west
Time Zone (UTC offset) -6 Central Standard Time
Elevation (m) 14.0 Near sea level — typical for coastal Houston
Anemometer Height (m) 10.0 Standard
Data Format ISHD NOAA Integrated Surface Hourly Data

Why Hobby instead of Intercontinental? Houston Hobby Airport (KHOU) is closer to the Ship Channel and more representative of the coastal/industrial area. George Bush Intercontinental (KIAH) is 35 km north-northwest and sits in a more suburban/forested environment with different roughness and moisture characteristics.

Upper Air Station — Lake Charles, Louisiana

Parameter Value Notes
Station ID 72240 WMO station number
Station Name Lake Charles, LA
Latitude 30.1200
Longitude -93.2200

Why Lake Charles? It is the closest radiosonde station to Houston (~200 km east). The only alternative would be Corpus Christi (~330 km southwest). Lake Charles is more representative of the Houston area's upper-atmosphere conditions because it shares the Gulf Coast environment.

Data Files and Date Range

For this tutorial, process one full calendar year of data. Your instructor will provide the raw data files, or you can download them from NOAA's NCEI archive.

Parameter Value
Surface Data File khou_2023.dat
Upper Air Data File 72240_2023.dat
Start Date 2023/01/01
End Date 2023/12/31

Step 2: Configure Stage 1 in the GUI

  1. Launch pyaermod-app.
  2. On Project Setup, set:
  3. UTM Zone: 15
  4. Hemisphere: N
  5. Datum: NAD83
  6. Map Center Latitude: 29.75
  7. Map Center Longitude: -95.10
  8. Go to Meteorology and select Configure AERMET.
  9. On the Stage 1 tab, enter all the surface and upper air station parameters from the tables above.
  10. Enter the data file paths and date range.
  11. Click Save Stage 1 Configuration.
  12. Expand Preview Stage 1 Input and verify the keywords.

Check yourself: The preview should contain: - A SURFACE block with DATA khou_2023.dat ISHD - An UPPERAIR block with DATA 72240_2023.dat FSL - An XDATES line showing 2023/01/01 TO 2023/12/31

  1. Click Download Stage 1 Input File — save as aermet_houston_s1.inp.

Step 3: Configure Stage 2

  1. Click the Stage 2 tab.
  2. Enter the filenames that Stage 1 will produce:
Parameter Value
Surface Extract File khou_extract.sfc
Upper Air Extract File lch_extract.ua
Start Date 2023/01/01
End Date 2023/12/31
Merge Output File houston_merged.mrg
  1. Save and download as aermet_houston_s2.inp.

Step 4: Configure Stage 3 — Houston Surface Parameters

This is where your knowledge of the Houston area matters. The land around the Ship Channel is a mix of industrial facilities, port infrastructure, open water, and coastal grassland. This is very different from the suburban defaults in the GUI.

  1. Click the Stage 3 tab.
  2. Enter file paths:
Parameter Value
Merge File houston_merged.mrg
Start Date 2023/01/01
End Date 2023/12/31
Surface Output houston_2023.sfc
Profile Output houston_2023.pfl
  1. Edit the monthly surface parameters table using the Houston-specific values below:

Houston Ship Channel Monthly Surface Parameters

Month Albedo Bowen Ratio Roughness (m) Rationale
Jan 0.18 0.8 0.40 Bare deciduous trees, dormant grass, dry winter
Feb 0.18 0.7 0.40 Pre-spring, some rain events
Mar 0.16 0.5 0.50 Vegetation greening, increasing moisture
Apr 0.14 0.4 0.60 Full leaf-out, higher transpiration
May 0.14 0.3 0.60 Lush vegetation, humid air, frequent rain
Jun 0.14 0.3 0.60 Peak humidity, tropical moisture
Jul 0.14 0.3 0.60 Peak summer — hot and humid
Aug 0.14 0.3 0.60 Same as July
Sep 0.15 0.4 0.60 Slight drying, still warm
Oct 0.16 0.5 0.50 Transitional, vegetation still green
Nov 0.17 0.7 0.45 Leaf drop begins, drier air
Dec 0.18 0.8 0.40 Dormant season, some cold fronts

Key differences from the suburban defaults (Tutorial 5):

  • Lower albedo (0.14--0.18 vs 0.15--0.35): Houston's dark industrial surfaces, asphalt, and green vegetation absorb more sunlight than light- colored suburban roofs and snow cover.
  • Lower Bowen ratio (0.3--0.8 vs 0.5--1.5): The Gulf Coast moisture keeps latent heat flux high. Moist surfaces partition more energy into evaporation than sensible heating.
  • Higher roughness (0.40--0.60 vs 0.30--0.50): The industrial Ship Channel area has large structures (tanks, stacks, buildings) that create more aerodynamic drag than low-rise suburban neighborhoods.
  1. Leave the Use Stage 1 surface station checkbox checked.
  2. Save and download as aermet_houston_s3.inp.

Step 5: Run AERMET

In a terminal, run the three stages in order:

aermet < aermet_houston_s1.inp
aermet < aermet_houston_s2.inp
aermet < aermet_houston_s3.inp

When Stage 3 completes, you should have houston_2023.sfc and houston_2023.pfl in your working directory.

Step 6: Verify the Output

Return to the GUI:

  1. Go to Meteorology and switch to Use Existing Files.
  2. Enter the paths to houston_2023.sfc and houston_2023.pfl.

Sanity checks:

  • The .sfc file should have ~8,760 lines (one per hour for 365 days, plus a header). Missing hours are normal — a few percent is acceptable.
  • Open the .sfc file in a text editor. The wind direction column should show a mix of directions with a preference for south-southeast winds (150--200 degrees) in summer months.
  • Mixing heights should range from ~100 m (stable nighttime) to ~2,000--3,000 m (convective daytime in summer).

Discussion Questions

  1. Why does the Bowen ratio matter? If you used a desert Bowen ratio (e.g., 5.0) for Houston, how would that affect the boundary layer parameters? (Hint: think about how sensible heat drives convective mixing.)

  2. What if you used the wrong surface station? If you mistakenly used a station 100 km inland (less humid, more rural), which surface parameters would be most affected?

  3. Sea breeze effects. Houston experiences afternoon sea breezes that can reverse the wind direction. Would you expect to see this in the .sfc file? How might it affect a refinery's downwind impacts?

Checkpoint

  • [x] Surface and upper air station selection for a specific area
  • [x] How Gulf Coast climate affects surface parameter choices
  • [x] The role of albedo, Bowen ratio, and roughness in boundary layer calculations
  • [x] Running AERMET and verifying the output files
  • [x] Why station representativeness matters for regulatory modeling

Tutorial 7 — Terrain Processing with AERMAP

Goal: Use the PyAERMOD GUI and AERMAP to assign terrain elevations to all receptors and sources for a model domain near the Houston Ship Channel.

Time: 30--40 minutes

What you'll learn: - Why terrain elevation matters even in "flat" areas like Houston - How to download digital elevation model (DEM) data - How to run AERMAP to compute receptor and source elevations - How to import AERMAP results back into the GUI - The difference between FLAT and ELEVATED terrain modes in AERMOD

Conceptual Background

In Tutorials 1--5, we used FLAT terrain — meaning every receptor and source sat at elevation zero. This is a simplification. In reality, even modest elevation differences can affect concentrations:

  • Elevated receptors (on a hill or ridge) can intercept the plume at a height closer to the effective plume centerline, increasing concentrations.
  • Depressed receptors (in a valley) may be partially shielded, but channeling effects can concentrate pollutants.
  • Source base elevation determines where the plume originates relative to the surrounding terrain.

AERMAP is the terrain preprocessor that extracts elevation data from digital elevation models (DEMs) and assigns:

  1. Receptor terrain elevation (z_elev) — ground height at each receptor.
  2. Receptor hill height (z_hill) — the controlling hill height that AERMOD uses for terrain-aware plume calculations.
  3. Source base elevation — ground height at each emission source.
Digital Elevation Model (DEM)
         |
    AERMAP preprocessor
         |
    +-----------------+
    | z_elev, z_hill  |  --> Receptor elevations
    | for each        |  --> Source base elevations
    | receptor/source |
    +-----------------+
         |
    Import into AERMOD project

Why Bother Near Houston?

Houston is coastal and mostly flat, but "flat" is relative:

  • The Ship Channel itself sits near sea level (0--5 m elevation).
  • Surrounding neighborhoods are 10--25 m above sea level.
  • The terrain rises gradually to 30--50 m northwest of the city.
  • Even a 15 m elevation difference can influence concentrations when the plume effective height is 50--100 m.

For regulatory work, EPA requires terrain processing for virtually all analyses. Learning the workflow here — where terrain effects are small — prepares you for mountainous or hilly sites where terrain is critical.

Step 1: Define the Model Domain

Before running AERMAP, you need to know the geographic extent of your receptor grid. We will use the domain for the refinery model in Tutorial 8.

Refinery location (approximate center): - Latitude: 29.7350 N - Longitude: -95.1050 W - UTM Zone 15N: approximately X = 279,200 m E, Y = 3,291,700 m N

Receptor grid extent (5 km x 5 km centered on the refinery):

Parameter Value (UTM meters)
X Min 276,700
X Max 281,700
Y Min 3,289,200
Y Max 3,294,200

Step 2: Obtain DEM Data

AERMAP requires elevation data in GeoTIFF or USGS DEM format. The easiest approach is to use the USGS 3D Elevation Program (3DEP) data at 1/3 arc-second resolution (~10 m).

Option A — PyAERMOD's built-in downloader (if available):

PyAERMOD includes a DEMDownloader class that can query the USGS TNM API and download the needed tiles automatically. Your instructor may provide a script or notebook for this. The downloader needs:

  • Bounding box in latitude/longitude
  • South: 29.700
  • North: 29.770
  • West: -95.140
  • East: -95.060
  • A local cache directory for downloaded tiles

Option B — Manual download:

  1. Go to the USGS National Map Viewer (https://apps.nationalmap.gov/downloader/).
  2. Search for your area (Houston Ship Channel).
  3. Select Elevation Products (3DEP) > 1/3 arc-second DEM.
  4. Download the GeoTIFF tile(s) covering your domain.
  5. Save to your working directory.

Tip: For the Houston Ship Channel area, you typically need only one or two 1-degree tiles. The files are approximately 50--100 MB each.

Step 3: Set Up the Project in the GUI

  1. Launch pyaermod-app.
  2. On Project Setup:
  3. Title: Tutorial 7 - Terrain Processing
  4. UTM Zone: 15, Hemisphere: N, Datum: NAD83
  5. Map Center: Lat 29.735, Lon -95.105
  6. Pollutant: SO2 (we will use this in Tutorial 8)
  7. Terrain Type: ELEVATED (this is the key change from previous tutorials)

  8. Averaging Periods: 1-HR, 24-HR, ANNUAL

  9. Go to Source Editor and add a placeholder point source at the refinery center. We will replace this with the full source inventory in Tutorial 8, but AERMAP needs at least one source to process.

Parameter Value
Source ID FCCSTK
X Coordinate 279200
Y Coordinate 3291700
Base Elevation 0
Stack Height 60
Exit Temperature 470
Exit Velocity 20
Stack Diameter 3.0
Emission Rate 5.0
  1. Go to Receptor Editor and add the Cartesian grid:
Parameter Value
X Min 276700
X Max 281700
Y Min 3289200
Y Max 3294200
Spacing 100

This gives a 51 x 51 = 2,601 receptor grid at 100 m resolution.

Step 4: Run AERMAP

AERMAP is a separate executable from AERMOD (download from EPA SCRAM). The process differs slightly depending on how your instructor has set things up:

Option A — Using PyAERMOD's terrain module (recommended):

If the AERMAP executable is installed:

  1. PyAERMOD can generate the AERMAP input file from your project configuration.
  2. The AERMAP input references your DEM files and the receptor/source coordinates from the GUI.
  3. Run AERMAP, which produces two output files:
  4. Receptor output — elevation and hill height for each grid point
  5. Source output — base elevation for each source

Option B — Command line:

Your instructor may provide the AERMAP input file. Run it as:

aermap < aermap_houston.inp

AERMAP typically runs in under a minute for a domain this size.

Step 5: Import AERMAP Results into the GUI

  1. Go to Receptor Editor.
  2. Click the AERMAP Elevation Import tab (the fifth tab).
  3. Upload the receptor output file from AERMAP.
  4. The GUI will match each receptor by (x, y) coordinates (within 0.5 m tolerance).
  5. Each receptor will be updated with its terrain elevation (z_elev) and controlling hill height (z_hill).
  6. Upload the source output file from AERMAP.
  7. Each source's base elevation will be updated automatically.

What you should see:

  • Most receptor elevations will be between 0 and 25 meters (the Ship Channel area is low-lying).
  • Hill heights may be slightly higher, reflecting terrain features within a radius of influence around each receptor.
  • The source base elevation will be close to sea level (likely 3--8 m).

Step 6: Verify the Terrain Data

After importing, go to the Run AERMOD page and expand the Generated Input File. Look for:

  1. In the RE pathway, GRIDCART lines should now include ELEV and HILL rows with varying elevation values (not all zeros).

  2. In the SO pathway, the LOCATION line for your source should have a non-zero base elevation.

  3. The CO pathway should contain MODELOPT DFAULT ELEV (not FLAT).

Compare FLAT vs ELEVATED: If you save this project and create a FLAT-terrain copy (change Terrain Type on Project Setup), you can compare the two input files. The FLAT version omits all elevation data. When you run both through AERMOD with the same meteorology, the ELEVATED version will typically produce slightly higher maximum concentrations because nearby terrain "lifts" some receptors closer to the plume centerline.

Discussion Questions

  1. Terrain resolution trade-off. We used 100 m receptor spacing. What happens if you use 50 m spacing? (More receptors = more computation time, but potentially catches terrain features between the 100 m points.)

  2. Hill height vs terrain elevation. AERMAP reports both z_elev (actual ground height at the receptor) and z_hill (the highest terrain feature influencing the receptor). Why does AERMOD need both? (Hint: the "dividing streamline" concept in AERMOD's terrain algorithm.)

  3. Flat areas still matter. Even though Houston's terrain varies by only ~25 m across our domain, that variation could be significant for a 50 m stack. Calculate the ratio: (terrain variation) / (effective stack height). If this ratio exceeds ~0.1, terrain effects are non- negligible.

Checkpoint

  • [x] Why even modest terrain matters for regulatory modeling
  • [x] How to obtain DEM data for a project domain
  • [x] How AERMAP processes DEMs to produce receptor/source elevations
  • [x] How to import AERMAP output into the GUI
  • [x] The difference between FLAT and ELEVATED terrain modes in AERMOD

Tutorial 8 — Modeling a Simplified Oil Refinery

Goal: Build a complete AERMOD model for a simplified petroleum refinery near the Houston Ship Channel, including multiple source types, realistic emission parameters, terrain, meteorology, and regulatory compliance analysis.

Time: 60--90 minutes

What you'll learn: - How to characterize refinery emission sources for AERMOD - How to combine point, area, and volume sources in a complex facility - How to use source groups to separate contributions - How to set up a multi-scale receptor network - How to perform a regulatory compliance analysis for SO2 and PM2.5 - How to interpret results from a multi-source industrial facility

Facility Overview

You are modeling a simplified petroleum refinery located along the Houston Ship Channel at approximately:

  • Latitude: 29.735 N
  • Longitude: -95.105 W
  • UTM Zone 15N: X = 279,200 m E, Y = 3,291,700 m N
  • Base elevation: ~5 m above sea level

The refinery processes 150,000 barrels per day of crude oil into gasoline, diesel, jet fuel, and petrochemical feedstocks. Its main emission sources include:

Unit Source Type Primary Pollutant Description
FCC regenerator stack Point SO2, PM2.5 Fluid Catalytic Cracking unit — burns off catalyst coke
Process heater stack #1 Point SO2, NOx Crude unit charge heater
Process heater stack #2 Point SO2, NOx Vacuum unit charge heater
Boiler stack Point SO2, PM2.5 Steam generation for process heat
Flare (ground-level equiv.) Point SO2 Emergency/routine flaring, modeled as a short stack
Crude tank farm Area VOC Floating roof tanks with seal losses
Product loading rack Area VOC Truck and railcar loading vapors
Wastewater treatment Area VOC, H2S API separators and aeration basins
Cooling towers Volume PM2.5 Drift from cooling water (mineral salts)
Equipment leaks (fugitives) Volume VOC Valves, pumps, flanges throughout process area

For this assignment, we will model SO2 as the primary pollutant. SO2 is a classic refinery pollutant with a clear 1-hour NAAQS (196 ug/m3, the equivalent of 75 ppb). The principles apply equally to PM2.5, NO2, or any other pollutant — only the emission rates and standards change.

Real-world note: An actual refinery permit application would include hundreds of emission sources and multiple pollutants. We simplify to ~10 sources to keep the assignment manageable while demonstrating all the key concepts.

Step 1: Project Setup

  1. Launch pyaermod-app.
  2. On Project Setup:
Parameter Value
Title Line 1 Houston Refinery SO2 Assessment
Title Line 2 Simplified 150 kbpd refinery - Tutorial 8
UTM Zone 15
Hemisphere N
Datum NAD83
Map Center Lat 29.735
Map Center Lon -95.105
Pollutant SO2
Terrain Type ELEVATED
Averaging Periods 1-HR, 3-HR, 24-HR, ANNUAL

Why these averaging periods? The SO2 NAAQS is a 1-hour standard (75 ppb = 196 ug/m3). We also include 3-hour, 24-hour, and annual for secondary NAAQS comparison and general analysis. Regulatory practice requires demonstrating compliance with all applicable standards.

Step 2: Define Emission Sources

Now add all the refinery sources. Go to Source Editor and add them one at a time in the order below. The refinery is oriented roughly north-south along the Ship Channel, with process units in the center and tank farms to the south.

Source 1 — FCC Regenerator Stack (FCCSTK)

The fluid catalytic cracking (FCC) unit is the largest single SO2 source at most refineries. The regenerator burns coke off the catalyst, releasing SO2, CO, and particulate matter through a tall stack.

Source type: Point

Parameter Value Rationale
Source ID FCCSTK FCC regenerator stack
X Coordinate 279200 Center of the process area
Y Coordinate 3291750
Base Elevation 5.0 Near sea level
Stack Height 60.0 Tall stack — regulatory requirement for FCC units
Exit Temperature 470.0 Hot regenerator flue gas (~197 C)
Exit Velocity 20.0 High velocity through large stack
Stack Diameter 3.0 Large diameter for high flow volume
Emission Rate 5.0 5 g/s SO2 — major source

Intuition: The FCC stack is the "big gun" of the refinery. Its tall stack and hot exhaust give it substantial plume rise (the effective release height could be 120--180 m), so peak ground-level impacts will be 1--3 km downwind despite the high emission rate.

Source 2 — Process Heater #1 (HTR1)

Crude unit charge heater — burns refinery fuel gas to heat crude oil.

Source type: Point

Parameter Value Rationale
Source ID HTR1 Crude unit heater
X Coordinate 279100 West of FCC, within process area
Y Coordinate 3291800
Base Elevation 5.0
Stack Height 40.0 Shorter than FCC — typical for heaters
Exit Temperature 420.0 Hot flue gas (~147 C)
Exit Velocity 12.0 Moderate velocity
Stack Diameter 1.8 Smaller than FCC
Emission Rate 2.0 2 g/s SO2

Source 3 — Process Heater #2 (HTR2)

Vacuum unit charge heater — similar to Heater #1 but slightly smaller.

Source type: Point

Parameter Value Rationale
Source ID HTR2 Vacuum unit heater
X Coordinate 279300 East of FCC
Y Coordinate 3291850
Base Elevation 5.0
Stack Height 35.0 Slightly shorter
Exit Temperature 410.0
Exit Velocity 10.0
Stack Diameter 1.5
Emission Rate 1.5 1.5 g/s SO2

Source 4 — Boiler Stack (BOILER)

Package boiler generating steam for process heat and utilities.

Source type: Point

Parameter Value Rationale
Source ID BOILER Utility boiler
X Coordinate 279050 Utility area, west side of refinery
Y Coordinate 3291600 South of process area
Base Elevation 5.0
Stack Height 45.0 Mid-height industrial stack
Exit Temperature 430.0
Exit Velocity 14.0
Stack Diameter 2.0
Emission Rate 3.0 3 g/s SO2 — second-largest source

Source 5 — Flare (FLARE)

The refinery's ground-level flare (modeled as a short elevated point source for simplicity). During normal operations, the flare burns small amounts of waste gas. During upsets, emissions can be much higher.

Source type: Point

Parameter Value Rationale
Source ID FLARE Ground flare
X Coordinate 279400 East side of refinery, away from process units
Y Coordinate 3291500
Base Elevation 5.0
Stack Height 15.0 Low effective height for ground flare
Exit Temperature 1,273.0 Very hot (~1000 C) — combustion temperature
Exit Velocity 20.0 High exit velocity
Stack Diameter 1.0
Emission Rate 1.0 1 g/s SO2 — normal operations

Note on flare modeling: In real regulatory work, flares are modeled using EPA's flare modeling guidance, which accounts for flame length, radiative heat loss, and effective release parameters. Our simplified point source is an approximation suitable for teaching.

Source 6 — Crude Tank Farm (TANKS)

A rectangular area of large floating-roof crude oil storage tanks. The emission rate represents fugitive losses through the floating roof seals. For this SO2 model, we use a small emission rate representing H2S/SO2 emissions from sour crude storage.

Source type: Area (Rectangular)

Parameter Value Rationale
Source ID TANKS Crude tank farm
X Coordinate 279150 Center of the tank farm
Y Coordinate 3291300 South of the process area
Base Elevation 5.0
Release Height 15.0 Tank top height (~15 m for large tanks)
Half-Width Y 75.0 150 m total north-south
Half-Width X 100.0 200 m total east-west
Rotation Angle 0 Aligned with grid
Emission Rate 0.000010 0.00001 g/s/m2 — small fugitive rate

Emission rate calculation: Area = 150 m x 200 m = 30,000 m2. Total emission = 0.00001 x 30,000 = 0.3 g/s from the entire tank farm. This is small compared to the stacks — fugitive SO2 from tanks is minor.

Source 7 — Loading Rack (LOADRK)

The truck/railcar loading area where refined products are loaded for distribution. Vapor displacement during loading produces emissions.

Source type: Area (Rectangular)

Parameter Value Rationale
Source ID LOADRK Product loading rack
X Coordinate 279350 East side of the refinery
Y Coordinate 3291400
Base Elevation 5.0
Release Height 4.0 Vehicle height
Half-Width Y 15.0 30 m north-south
Half-Width X 40.0 80 m east-west
Rotation Angle 0
Emission Rate 0.000005 0.000005 g/s/m2 — very low SO2

Source 8 — Wastewater Treatment (WWATER)

The wastewater treatment area includes API oil-water separators and aeration basins. H2S and SO2 can be released from exposed water surfaces.

Source type: Area (Circular)

Parameter Value Rationale
Source ID WWATER Wastewater treatment
X Coordinate 278900 West side, buffered from process area
Y Coordinate 3291400
Base Elevation 5.0
Release Height 1.0 At ground level (open basins)
Radius 50.0 50 m radius circular treatment area
Num Vertices 20
Emission Rate 0.000030 0.00003 g/s/m2

Emission rate calculation: Area = pi x 50^2 = ~7,854 m2. Total emission = 0.00003 x 7,854 = 0.24 g/s.

Source 9 — Cooling Towers (COOL1)

Large mechanical-draft cooling towers. Drift (small water droplets carrying dissolved minerals) escapes from the tower. We model these as volume sources because emissions come from the entire tower structure.

Source type: Volume

Parameter Value Rationale
Source ID COOL1 Cooling tower bank
X Coordinate 279250 Adjacent to process area
Y Coordinate 3291550
Base Elevation 5.0
Release Height 12.0 Tower exit height
Initial Sigma-Y 8.0 ~ tower width / 4.3 (EPA guidance)
Initial Sigma-Z 6.0 ~ tower height / 2.15
Emission Rate 0.2 0.2 g/s total (drift PM, minor SO2 content)

Volume source parameters: EPA guidance recommends: - Initial sigma-Y = building width / 4.3 - Initial sigma-Z = building height / 2.15 These represent the initial lateral and vertical spread of the plume caused by the source's physical dimensions.

Source 10 — Equipment Leak Fugitives (FUGITV)

Distributed fugitive emissions from thousands of valves, pumps, flanges, and connections throughout the refinery process area. Modeled as a single large volume source representing the aggregate.

Source type: Volume

Parameter Value Rationale
Source ID FUGITV Process area fugitives
X Coordinate 279200 Center of process area
Y Coordinate 3291700
Base Elevation 5.0
Release Height 5.0 Average equipment height
Initial Sigma-Y 25.0 Large lateral extent (~100 m process area)
Initial Sigma-Z 4.0 Low release height
Emission Rate 0.5 0.5 g/s total from all leak points

Step 3: Create Source Groups

Source groups let you evaluate the contribution of different parts of the refinery separately. This is required for regulatory work and invaluable for understanding where impacts come from.

  1. In the Source Editor, navigate to the Source Groups section.
  2. Create these groups:
Group Name Sources What It Represents
STACKS FCCSTK, HTR1, HTR2, BOILER, FLARE All elevated point sources
FUGITIV TANKS, LOADRK, WWATER, FUGITV All fugitive/area/volume sources
FCCONLY FCCSTK FCC unit alone (largest single source)
ALL All 10 sources Total facility impact

Why source groups? When the regulator asks "What happens if you add a scrubber to the FCC unit?", you can compare the FCCONLY group results before and after to quantify the benefit — without re-running the entire model.

Step 4: Set Up a Multi-Scale Receptor Network

A good receptor network captures both the near-field peak and the broader impact area. We will use three receptor configurations:

Receptor Set 1 — Fine Grid (Near-Field)

This captures the peak concentration close to the refinery.

Parameter Value
Grid Name FINE
X Min 278200
X Max 280200
Y Min 3290700
Y Max 3292700
Spacing 50

This gives a 41 x 41 = 1,681 receptor grid at 50 m spacing, covering a 2 km x 2 km area centered on the refinery.

Receptor Set 2 — Coarse Grid (Far-Field)

This captures impacts several kilometers downwind.

Parameter Value
Grid Name COARSE
X Min 276700
X Max 281700
Y Min 3289200
Y Max 3294200
Spacing 200

This gives a 26 x 26 = 676 receptor grid at 200 m spacing, covering a 5 km x 5 km area.

Receptor Set 3 — Discrete Sensitive Receptors

Add discrete receptors at locations of special concern. These represent nearby communities, schools, and monitoring stations.

  1. On the Discrete Receptors tab, add:
Label X Y What It Represents
Community NW 278100 3292500 Residential neighborhood NW of refinery
Community NE 280300 3292400 Neighborhood NE of refinery
Community S 279200 3290200 Neighborhood south of Ship Channel
School 278500 3293000 Elementary school, 1.5 km NW
Hospital 280000 3292800 Medical center, 1.2 km NE
Monitor 279500 3292200 Air quality monitoring station

Total receptor count: 1,681 + 676 + 6 = 2,363 receptors. With hourly meteorology over one year and 10 sources, this is a moderately large model that may take 5--15 minutes to run.

Step 5: Connect Meteorology

  1. Go to Meteorology and select Use Existing Files.
  2. Enter the paths to your Houston .sfc and .pfl files from Tutorial 6:
  3. Surface file: houston_2023.sfc
  4. Profile file: houston_2023.pfl

If you completed Tutorial 7:

  1. Go to Receptor Editor > AERMAP Elevation Import tab.
  2. Upload the AERMAP receptor and source output files.
  3. Verify that elevations have been applied.

If you did not complete Tutorial 7, you can proceed with FLAT terrain by changing the Terrain Type to FLAT on Project Setup. The results will be slightly different but the analysis workflow is the same.

Step 7: Configure Output

  1. Go to Run AERMOD.
  2. Under Output Configuration:
  3. Receptor table: On
  4. Max value table: On
  5. Enable POSTFILE output:
    • Format: PLOT
    • Averaging period: 1 (1-hour — needed for NAAQS comparison)
    • Source group: ALL
  6. Enable separate plot files for each source group:
  7. STACKS — 1-hour averaging
  8. FCCONLY — 1-hour averaging

Step 8: Validate and Run

  1. Click Validate to check for errors.

Common issues at this stage: - Source coordinates outside the receptor grid (verify all sources are within the FINE grid) - Missing meteorological file paths - Source ID longer than 8 characters

  1. Preview the input file. Expand it and verify:
  2. 10 sources in the SO pathway (5 POINT, 2 AREA, 1 AREAPOL, 2 VOLUME)
  3. 4 source groups defined (STACKS, FUGITIV, FCCONLY, ALL)
  4. Receptor grids FINE and COARSE plus 6 discrete receptors

  5. Correct met file paths

  6. Set the Working Directory and click Run AERMOD.

Runtime estimate: With ~2,400 receptors, 10 sources, and 8,760 hours of meteorology, expect approximately 5--15 minutes depending on your computer.

Step 9: Analyze the Results

Once AERMOD completes, go to Results Viewer.

9A. Interactive Map

  1. Select 1-hour averaging period.
  2. Observe the spatial pattern:
  3. The plume should extend primarily to the north-northwest (downwind of the prevailing south-southeast winds in Houston).
  4. The maximum 1-hour concentration will be located 0.5--2 km downwind of the tallest stacks.
  5. Switch to ANNUAL averaging. The pattern should be more symmetric because annual averages include winds from all directions.
  6. Toggle between the street map and satellite basemap. Can you see the Ship Channel and surrounding neighborhoods?

9B. Statistics — Regulatory Compliance

Go to the Statistics tab and answer these questions:

Question 1: Does the refinery comply with the SO2 1-hour NAAQS?

The SO2 1-hour standard is 196 ug/m3 (equivalent to 75 ppb).

  • Enter 196 in the exceedance threshold.
  • How many receptors exceed this standard?
  • What is the maximum 1-hour concentration?
  • Where is the maximum located (coordinates)?

Regulatory context: Under EPA's Guideline on Air Quality Models, a facility demonstrates compliance if the modeled concentration (including background) does not exceed the NAAQS at any receptor. For SO2, the "design value" is the 99th percentile of daily maximum 1-hour values averaged over 3 years.

Question 2: What is the maximum annual average?

There is no annual SO2 NAAQS (it was revoked in 2010), but the annual average is useful for understanding chronic exposure.

  • What is the maximum annual concentration?
  • At which receptor?
  • How does it compare to the 1-hour maximum? (It should be much smaller.)

Question 3: Sensitive receptor analysis.

Check the concentrations at each discrete receptor:

Receptor 1-hr Max 24-hr Max Annual Above NAAQS?
Community NW
Community NE
Community S
School
Hospital
Monitor

Which sensitive receptor has the highest impact? Is this consistent with the wind direction pattern you observed on the map?

9C. Source Group Analysis

If you generated separate plot files for each source group:

Question 4: Source contribution analysis.

Compare the maximum 1-hour concentrations from each source group:

Source Group Max 1-hr (ug/m3) % of Total
ALL (total) 100%
STACKS
FUGITIV
FCCONLY
  • What percentage of the total impact comes from stacks vs fugitive sources?
  • What percentage comes from the FCC unit alone?
  • If you could eliminate the FCC stack emissions entirely, would the remaining sources still comply with the NAAQS?

9D. POSTFILE Analysis (Advanced)

If you enabled POSTFILE output:

  1. Go to the POSTFILE Viewer tab.
  2. Upload the POSTFILE.
  3. Use the timestep slider to find the hour with the highest concentration.
  4. What were the meteorological conditions during this hour? (You can check the .sfc file for the matching date/time.)
  5. Generate an animation GIF showing the hourly concentration evolution over a 24-hour period that includes the peak hour.

Step 10: Sensitivity Analysis (Assignment Extension)

Repeat the model run with one change at a time to understand how each factor affects the results. This is the heart of engineering analysis.

Scenario A: Taller FCC Stack

Change the FCC stack height from 60 m to 90 m. Keep everything else the same.

  • How does the maximum 1-hour concentration change?
  • How does the location of the maximum shift?
  • Is this consistent with what you learned in Tutorial 2?

Scenario B: Reduced Emission Rate

Simulate the effect of installing a flue gas desulfurization (FGD) scrubber on the FCC unit. Reduce the FCC emission rate from 5.0 g/s to 1.0 g/s (80% removal efficiency).

  • What is the new maximum 1-hour concentration?
  • Does the facility now comply at all receptors?

Scenario C: Background Concentration

Real air has background SO2 from regional sources (power plants, shipping, other refineries). Add a background concentration:

  1. In the Source Editor, go to the Background section.
  2. Select Uniform mode.

  3. Enter 10 ug/m3 as the background for all periods.

  4. How does the background change the compliance picture?

  5. At what background level would the facility fail to comply (if it didn't already)?

Scenario D: Flat vs Elevated Terrain

If you have terrain data from Tutorial 7, run the model with both FLAT and ELEVATED terrain:

  • What is the maximum concentration difference between the two terrain modes?
  • Where do the differences occur? (Hint: look at receptors on higher terrain.)

Report Template

Compile your results into a brief technical report (3--5 pages) with these sections:

  1. Introduction — Describe the facility, pollutant, and modeling objectives (1 paragraph).

  2. Model Configuration — Summarize source inventory (table), receptor network, meteorological data, and terrain data used.

  3. Results — Present your answers to Questions 1--4 above, with supporting maps and tables from the Results Viewer.

  4. Sensitivity Analysis — Summarize Scenarios A--D and what they reveal about the most effective control strategies.

  5. Conclusions — Does the simplified refinery comply with the SO2 NAAQS? What would you recommend to the facility operator if it does not?

Assignment Deliverables

Submit the following files and documents:

From Tutorial 6 (AERMET)

  • [ ] Three AERMET input files (aermet_houston_s1.inp, s2.inp, s3.inp)
  • [ ] Written answers to the three Discussion Questions
  • [ ] Brief description of the monthly surface parameter choices and rationale

From Tutorial 7 (Terrain)

  • [ ] Screenshot of the GUI showing imported terrain elevations in the receptor grid
  • [ ] Comparison of FLAT vs ELEVATED generated input files (highlight differences)
  • [ ] Written answers to the three Discussion Questions

From Tutorial 8 (Refinery Model)

  • [ ] Project JSON file (refinery_project.json) — saved from the GUI
  • [ ] Generated AERMOD input file (refinery.inp)
  • [ ] Technical report (3--5 pages) including:
  • Source inventory table
  • Results maps (screenshots from the Interactive Map)
  • Compliance analysis table
  • Source contribution analysis
  • Sensitivity analysis results (Scenarios A--D)
  • Conclusions and recommendations

Grading Rubric

Component Points Criteria
AERMET configuration (Tutorial 6) 15 Correct station selection, appropriate surface parameters, answers to discussion questions
Terrain processing (Tutorial 7) 15 Successful AERMAP run, imported elevations, FLAT vs ELEVATED comparison
Source inventory (Tutorial 8, Steps 1--3) 20 All 10 sources correctly entered with realistic parameters, source groups defined
Receptor network (Tutorial 8, Step 4) 10 Multi-scale grid design, appropriate sensitive receptor placement
Compliance analysis (Tutorial 8, Step 9) 20 Correct NAAQS comparison, receptor-specific analysis, source contribution breakdown
Sensitivity analysis (Tutorial 8, Step 10) 10 At least two scenarios completed with clear interpretation
Report quality 10 Clear writing, professional tables/figures, sound engineering judgment
Total 100

Supplemental Reference Tables

Houston-Area Meteorological Stations

Station ID Type Lat Lon Elev (m) Notes
Houston Hobby KHOU / 12918 Surface 29.6454 -95.2789 14 Recommended for Ship Channel area
Houston IAH KIAH / 12960 Surface 29.9844 -95.3414 29 Better for north Houston
Galveston KGLS / 12923 Surface 29.2653 -94.8603 5 Coastal, high winds
Lake Charles, LA 72240 Upper Air 30.12 -93.22 5 Closest radiosonde to Houston
Corpus Christi 72251 Upper Air 27.77 -97.50 14 Alternative (far south)

Typical Refinery Emission Rates (SO2, g/s)

These are approximate emission rates for a 150,000 bpd refinery burning low-sulfur fuel gas (50--150 ppm H2S). Actual rates depend on sulfur content, control equipment, and operating conditions.

Source Typical SO2 (g/s) Range
FCC regenerator 3--8 With electrostatic precipitator / third-stage separator
Process heater (large) 1--4 Depends on fuel sulfur content
Process heater (small) 0.5--2
Boiler 1--5 Depends on fuel type and controls
Flare (normal) 0.2--2 Much higher during upset conditions
Tank farm (fugitive) 0.05--0.5 Total; primarily H2S oxidized to SO2
Cooling tower 0.01--0.3 Drift PM, minimal SO2

NAAQS Quick Reference

Pollutant Averaging Period Standard (ug/m3) Standard (ppb) Form
SO2 1-hour 196 75 99th percentile of daily max, averaged over 3 years
PM2.5 Annual 9.0 — Annual mean, averaged over 3 years
PM2.5 24-hour 35 — 98th percentile, averaged over 3 years
PM10 24-hour 150 — Not to be exceeded more than once per year
NO2 1-hour 188 100 98th percentile of daily max, averaged over 3 years
NO2 Annual 100 53 Annual mean
CO 1-hour 40,000 35,000 Not to be exceeded more than once per year
CO 8-hour 10,000 9,000 Not to be exceeded more than once per year
O3 8-hour 137 70 Annual 4th-highest, averaged over 3 years

Land Use Surface Parameters for Coastal Texas

Land Cover Season Albedo Bowen Ratio Roughness (m)
Open water All 0.10 0.0 0.001
Coastal marsh Summer 0.14 0.2 0.15
Coastal marsh Winter 0.18 0.5 0.10
Industrial/port All 0.18 0.8--1.5 0.50--1.00
Residential suburban Summer 0.16 0.5 0.50
Residential suburban Winter 0.20 1.0 0.30
Grassland/pasture Summer 0.18 0.3 0.10
Grassland/pasture Winter 0.20 0.8 0.05
Forest (mixed) Summer 0.12 0.3 1.00
Forest (mixed) Winter 0.18 0.8 0.50