NO₂ modeling: the tier ladder & every method

Nitrogen dioxide is the one criteria pollutant where AERMOD does chemistry — and where the modeler's method choice can move the answer by a factor of two. This page covers why, every available method, what data each needs, and how the choice plays out when a demonstration is failing.

Why NO₂ is different

Combustion sources mostly emit nitric oxide (NO), not NO₂ — yet the National Ambient Air Quality Standard (NAAQS) is written for NO₂. The Guideline states the chemistry problem exactly:

…NO2 sources co-emit NO along with NO2 and any emitted NO may react with ambient ozone to convert to additional NO2 downwind. Thus, comprehensive modeling of NO2 would need to consider the ratio of emitted NO and NO2, the ambient levels of ozone and subsequent reactions between ozone and NO, and the photolysis of NO2 to NO.

Source · 40 CFR 51 App. W §4.2.3.4(a).

In plain terms: how much of the modeled NOx plume is actually NO₂ at the receptor depends on three things — how much NO₂ the stack emits directly (the in-stack ratio), how much ozone the plume can grab to convert NO on the way, and sunlight driving some NO₂ back to NO. Because that's genuinely complicated, the Guideline on Air Quality Models (Appendix W) requires a tiered screening approach (§4.2.3.4(b)), each tier trading conservatism for data.

The three tiers, with an example

Say your modeled 1-hour NOx design value is 210 µg/m³ against the NO₂ standard of about 188 µg/m³ (100 ppb). (All numbers here are illustrative.)

Tier 1 — full conversion. "Assume a total conversion of NO to NO2" (§4.2.3.4(c)): every microgram of NOx counts as NO₂. Your design value is 210 — fails. Zero extra data, maximum conservatism; if Tier 1 passes, you're done and nobody argues.

Tier 2 — the ARM2 ratio. Multiply the Tier 1 result by the Ambient Ratio Method 2 curve — equilibrium NO₂/NOx ratios "derived from national data from the EPA's Air Quality System (AQS)," bounded by "a minimum ambient NO2/NOX ratio of 0.5 and a maximum ambient ratio of 0.9" (§4.2.3.4(d)) — the ratio applied to each concentration comes from an empirical formula fitted to national monitoring data. In our example the peak draws the ×0.90 cap and lands at 189 — still fails, by one microgram. Still no site data needed; ARM2 is a regulatory option usable with the DFAULT switch.

Tier 3 — the detailed screens. Now the analysis earns its refinement with real data. The three methods usable "for most sources" (§4.2.3.4(e)):

With a nearby ozone monitor's hourly record and a documented in-stack ratio, the PVMRM run comes in at 148 — passes — because on the hours that produced the 210 peak, there simply wasn't enough ozone in the air to convert most of that NO before it reached the receptor.

Tier 3 has a procedural price: use "shall occur in consultation with the EPA Regional office in addition to the appropriate reviewing authority" (§4.2.3.4(e)) — which is why the modeling protocol conversation happens before the analysis, not after.

Two more methods exist inside AERMOD but outside the regulatory set: TTRM/TTRM2 (the Travel Time Reaction Method, alone or paired with OLM/PVMRM/ARM2) are non-regulatory ALPHA options that cannot be used with DFAULT (AERMOD User's Guide §3.2.2, re-verified in the 26135 edition).

The inputs, and how each one moves the answer

The refinement power of Tier 3 lives in three inputs — each with a direction of influence worth knowing before you gather data:

• In-stack NO₂/NOx ratio (ISR). The fraction of NOx leaving the stack already as NO₂. Higher ISR → higher modeled NO₂ — that primary NO₂ needs no ozone at all. The Guideline accepts "manufacturer test data, State or local agency guidance, peer-reviewed literature, and/or the EPA's NO2/NOX ratio database" as justification (§4.2.3.4(d)); a measured ISR of 0.05 for a low-NOx burner versus a default assumption can be the whole demonstration.
• Ambient ozone. The conversion fuel. More ozone → more conversion → higher NO₂. Supplied as a single annual value, a temporally varying table, or an hourly file (the OZONEVAL / O3VALUES / OZONEFIL inputs — AERMOD User's Guide §3.2.5.1). Hourly data is the refinement: the worst NOx hours are often stable nights when ozone is depleted — a single annual ozone value applies summer-afternoon ozone to a winter night and overstates conversion. Note the model's nighttime floor: by default AERMOD applies an ozone value based on the minimum of 40 ppb and the preceding 24-hour maximum for stable nighttime hours (the NOMINO3 option removes the floor; UG §3.2.2).
• The ambient equilibrium ratio. OLM, PVMRM, and TTRM cap the final NO₂/NOx at a default ambient equilibrium of 0.90, adjustable with justification (UG §3.2.5).

Choosing a method when the run is failing

The practical decision tree, assuming Tier 1 and 2 have failed:

• Isolated buoyant stack, peak well downwind → PVMRM or GRSM: the distance-dependent conversion means a plume that travels before touching ground has converted realistically, not instantly.
• Peak close to the source, or many low-level sources → OLM's simpler ozone-limit is often adequate — and easier to defend.
• Photolysis or travel-time effects matter, and an hourly ambient NOx record exists → GRSM, the newest regulatory option (promulgated with version 24142).
• Whichever you choose, the same run should fix the cheap things too: hourly ozone instead of an annual value, and a documented ISR instead of a conservative default. Those two inputs are usually worth more than the method switch itself.

In PlumeSmart

The pollutant-treatment step offers every method — ARM2, OLM, PVMRM, GRSM, and the ALPHA travel-time options with their regulatory status labeled — with the ozone inputs (single value, varying table, or hourly file) and in-stack ratio fields alongside, so a tier escalation is a settings change on the same project rather than a rebuild.