Step 3: Modeling Protocol

Yes. A combined Modeling Protocol and Risk Assessment Work Plan will be accepted and can be used to reduce redundancy between the documents. Please be sure to clearly define each required submittal in the document (e.g., submit them as separate sections or chapters).

A level 1 risk assessment makes us of pre-calculated dispersion values and simplifying assumption to create a simple table that calculates risk. Because of the many assumptions involved, a level 1 assessment may not be appropriate under the following conditions:

• Complex terrain features, such as elevated terrain, are present as determined by DEQ
• The facility layout is complex to the degree that a simple stack or volume fugitive emission layout may be underestimating risk

Please contact DEQ if you have questions regarding the applicability of a Level 1 for your facility.

A Modeling Protocol for a level 1 and level 2 risk assessment must include identification of exposure locations and reporting of Risk Based Concentrations (RBCs) that will be used in the analysis. Because level 1 and level 2 risk assessments do not require separate risk assessment work plans, the Modeling Protocol must include all exposure scenario assumptions made before submission of the final risk assessment.

A level 3 or level 4 risk assessment requires both a Modeling Protocol and a Risk Assessment Work Plan. Because there is significant overlap between modeling and risk assessment, specifically in the receptor grid and exposure location assignment, a crosswalk of all receptor coordinates and exposure locations should be submitted as an electronic spreadsheet at the modeling protocol stage -- fields should include, at a minimum, the following information for each modeling receptor: identification, UTM Northing, UTM Easting, and exposure classification. All gridded modeling receptors should be included in the crosswalk, including roadways and railways where risk may ultimately not be calculated. Information regarding RBCs and risk calculations are part of the Risk Assessment Work Plan.

​All Toxic Air Contaminants (TACs) in the approved emissions inventory should be accounted for in the modeling protocol. However, only TACs that have a Risk Based Concentration (RBC) in OAR 340-245-8010 Table 2​ need to be modeled. The modeling protocol should include a table of all TACs, by Toxics Emissions Unit (TEU), their corresponding emission rates in [lbs/yr], [lbs/day], modeling input rates in [g/s], and their RBCs. Reported TACs (see OAR 340-245-8020 Table 2) that do not have an RBC should be listed in a separate table of TACs that will not be modeled. These TACs should be included in the uncertainty analysis in the final risk assessment report as a potential source of under-estimated risk.

​The recommended model depends on the level of risk assessment selected:

• A Level 1 risk assessment does not require any modeling by the facility. DEQ conducted AERMOD dispersion modeling to produce dispersion factor tables [OAR 340-245-8050 Table 5. See “Step 4: Risk Assessment Work Plan" FAQs section – “When is a level 1 risk assessment inappropriate?" for more information on when a level 1 can be conducted.
• A Level 2 risk assessment can be modeled using AERMOD-MAKEMET (MM), or AERSCREEN with internally generated MAKEMET.
• A Level 3 or Level 4 risk assessment requires the use of AERMOD with all preprocessors including AERMAP, BPIPPRIM, AERMET, and AERSURFACE. An alternative dispersion model may be used with DEQ approval.

Information on the AERMOD modeling system can be found at https://www.epa.gov/scram/air-quality-dispersion-modeling-preferred-and-recommended-models#aermod.

If the facility is in a location where terrain elevation is a concern, as determined by DEQ, a level 1 risk assessment cannot be used. This is because flat terrain was assumed in the creation of the level 1 dispersion factors. To account for terrain impacts on receptors, AERMAP should be run to generate terrain heights for all sources and receptors prior to modeling in AERMOD. There may be rare situations where AERMAP does not adequately represent the terrain features of a facility and the nearby community, and these features may have a significant impact on the modeling results. In these cases please work with DEQ to establish appropriate terrain assumptions.

​If the facility is conducting modeling for both New Source Review (NSR) and CAO, a single Modeling Protocol may be appropriate. Many of the elements will be the same, and therefore it may reduce redundancy to submit a single document. However, all aspects of both NSR and CAO must be included in the Modeling Protocol and care should be taken to outline each of these requirements. See the Recommended Procedures for Air Quality Dispersion Modeling for a list of these elements. Check in with DEQ before combining the protocol.

​Modeling receptors should begin at the facility boundary.

​Receptors should be placed in a grid with the following spacing (all distances from fenceline):

• 25m spacing along fenceline and out to 200m
• 50m spacing from 200 to 1,000m
• 100m spacing from 1,000 to 2,000m
• 200m spacing from 2,000 to 5,000m
• 500m spacing from 5 000 to 10,000m

Receptors in areas zoned residential with or without homes should be classified as residential exposure locations. Worker exposure should be evaluated in all areas zoned for industrial/commercial use, including parking lots and empty lots, not only buildings. However, a facility may request to reclassify an area to match the current land use if that is different from its zoning classification. See “What if the land near my facility is not currently being used in a manner consistent with zoning map?"

Receptors are usually gridded according to the spacing described in the Modeling Procedures document and are unlikely to align with zoning boundaries or exposure locations such as a house or school. Since the modeled concentration at a receptor will be similar to concentrations at adjacent receptors, the concentration between receptors can be reasonably estimated by taking the value from the nearest receptor, or by averaging the surrounding values. For example, if we want to know the concentration, and subsequently risk, at a house in a commercially zoned area and does not lie under a modeling receptor, we can 1) place a new modeling receptor on the house, label it as a residential exposure location, and re-run the model, or 2) re-label the nearest modeling receptor from a worker exposure location to a residential exposure location, and make the adjustment in the post-processing step without having to re-run the model. The latter option is the easiest and quickest, and will give a similar result.

An electronic spreadsheet crosswalk of receptor coordinates and assigned exposure locations should be submitted with the Modeling Protocol, including, at a minimum, the following information for each modeling receptor: emission release identification, UTM Northing and Easting, and exposure location classification.
​

​It is important to keep in mind that modeling receptors are points in space that describe a concentration surface, which usually changes gradually from one receptor to another. It should be noted that the use of modeling receptors to predict concentrations at exposure locations provides only mathematical approximations of these concentrations, which inherently include uncertainties, and these uncertainties grow larger when the spatial domain resolution is decreased – i.e., the receptor grid spacing increases.

Receptors are usually gridded according to the spacing described in the Modeling Procedures document and are unlikely to align exactly with zoning boundaries or exposure locations such as a house or school. For this reason, modeling receptors should represent the most protective landuse/zoning in the area between the receptor and the next nearest receptor. For example, if a receptor is in the 200m spaced modeling grid and falls on a road inside a residential neighborhood, that receptor should be assigned as a residential exposure location. In some cases it may be necessary to place an additional receptor within the grid spacing to capture concentrations at a specific exposure location (e.g., a home among commercial land use)
​​

​​​If the land use near a facility is currently being used in a manner different than the zoning, you have the option of requesting that the chronic exposure locations be reclassified in the risk analysis to represent the current use. For example, if land surrounding a facility is zoned to allow for a residence or business but there is currently no established home or business at these locations. In this situation, the following is required [OAR 340-245-0210(1)(a)(F)]:

• The approved locations must still be included in the model and acute noncancer risk must be evaluated in these locations.

• The facility must submit annual reports demonstrating that the land use is unchanged.

• If the annual report indicates that land use has changed, the facility must re-evaluate risk consistent with the current land use.

Please note, any occupied residence must be assessed as residential regardless of actual zoning classification – e.g., in the case where homes were permitted to remain in locations where local zoning has been revised to only allow future commercial development.

To request a change to underlying zoning, the facility must complete the AQ521 Form and the AQ522 Table.

​​Yes. DEQ has developed a crosswalk between state-level zoning classifications and recommended exposure locations for use in a CAO Risk Assessment. Facilities may deviate from these recommendations with appropriate documentation from local land use authorities, such as the applicable city or county zoning authority requirements.

​Please include all gridded modeling receptors in the modeling to ensure completeness and accuracy in the modeling results. These locations can be excluded in the risk assessment step if the receptor falls on an area where people do not normally congregate, such as roadways or railways. One exception to this is along the facility boundary. If the boundary is against a roadway or railway, receptors can begin on the other side of the roadway or railway.

​The CAO modeling domain should extend no less than 2 km from the facility boundary. The full extent of the boundary should cover all areas where risk for any exposure scenario is greater than an excess cancer risk or hazard index of 0.5 (i.e., the source permit level) up to a distance of 10km from the facility boundary. The modeling domain may have to be extended after the initial risk analysis to ensure all areas show risk no greater than 0.5.

​​DEQ addresses considerations for estimating and modeling emissions from stationary diesel engines in the “Step 2: Emissions Inventory" FAQs section – please see the question “How should I estimate emissions from stationary diesel engines at my facility?"