The 13 source types: how each one is computed

Every emission enters AERMOD as one of thirteen source types, and the choice is not cosmetic: it selects which physics the model applies — plume rise or none, a single release point or a numerical integration over a shape, downwash eligibility, even which output types are possible. The type is declared on each source's first input record, which "identifies the source type and dictates which parameters are needed and/or accepted" (AERMOD User's Guide, EPA-454/B-26-001, §3.3.1).

The point family: POINT, POINTCAP, POINTHOR

What they represent: "releases from stacks and isolated vents, as well as other kinds of sources" (§3.3.2.1) — the classic industrial release.

How they're computed: the full plume-physics treatment from the earlier pages — momentum and buoyancy plume rise from the stack parameters, the convective three-plume split, stable-hour rise limits, and (for sources near buildings) the PRIME downwash algorithms with direction-specific building dimensions, which apply only to this family (§3.3.2.1).

The inputs: emission rate, release height, exit temperature, exit velocity, and inside stack diameter. Each has a distinct physical job: height sets where the plume starts; temperature (versus ambient) drives buoyancy; velocity and diameter drive momentum and stack-tip downwash. Three input conventions are worth knowing:

• Exit temperature 0.0 means "always ambient" — the plume neither rises from buoyancy nor sinks (§3.3.2.1).
• A negative exit temperature means "this many degrees above ambient, every hour" — useful when a process tracks ambient conditions.
• Below-ambient releases are outside AERMOD's physics: "Such releases should be modeled with a dense gas model" (§3.3.2.1).

Capped and horizontal variants. A rain cap kills vertical momentum; a horizontal stack redirects it. POINTCAP and POINTHOR encode EPA's Model Clearinghouse procedure: the vertical exit velocity is set effectively to zero (0.001 m/s) while "an effective stack diameter that maintains the actual flow rate" preserves the plume's buoyancy (§3.2.2.4) — the heat still lifts the plume; only the jet is gone. Under PRIME downwash the treatment adapts: a capped stack's initial plume radius becomes twice the stack diameter (the plume mushrooming under the cap), and a horizontal stack's full exit velocity becomes lateral, aimed downwind (§3.2.2.4). Practical consequence: modeling a capped vent as a plain POINT credits it with vertical momentum it does not have and can understate ground-level impacts near the source.

VOLUME: the initially-dilute release

What it represents: "building roof monitors, multiple vents, and conveyor belts" (§3.3.2.2) — emissions that are already spread out when they enter the air.

How it's computed: instead of plume rise, a VOLUME source starts the plume with initial lateral and vertical dimensions — the modeler tells the model how big the "cloud" already is at release. From there dispersion proceeds normally. The inputs are the emission rate, release height, and those two initial dimensions; larger initial dimensions mean a more dilute plume from the first meter, which lowers close-in concentrations. (Since version 23132 the type also carries aircraft- emission parameters through the hourly-emissions file.)

The area family: AREA, AREAPOLY, AREACIRC

What they represent: "low level or ground level releases with no plume rise (e.g., storage piles, slag dumps, and lagoons)" (§3.3.2.3).

How they're computed: no plume rise at all. The model integrates the contribution of the emitting surface numerically — "in the regulatory default mode, the Romberg numerical integration is utilized for all receptors" (§3.3.2.3) — which is why the model "will estimate concentrations (and/or deposition) at receptors located within the dimensions of the source" (§3.3.2.8) — a receptor standing on the pile is a legitimate receptor. One subtlety: AREA and LINE sources omit the lateral-meander enhancement that POINT and VOLUME sources get, unless a non-regulatory ALPHA option adds it (§3.3.2.8). The three variants differ only in geometry: a rectangle (with optional rotation), an arbitrary polygon traced vertex by vertex, or a circle the model converts to an equal-area polygon (§3.3.2.4–.6). Inputs are the emission rate per square meter, the release height, the shape, and an optional initial vertical dimension — a mechanically disturbed pile can justify a nonzero one.

The rate-per-area convention is the classic pitfall: total emissions must be divided by the emitting area, and a wrong area silently rescales every concentration downwind.

OPENPIT: the excavation

What it represents: emissions from within a mine pit or quarry below grade.

How it's computed: the pit is entered as a rectangle with its top elevation and its length, width, and volume; the model computes an "effective depth… based on the length, width, and volume of the pit," and the release height (measured from the pit base, 0.0 meaning the pit floor) "cannot exceed the effective depth" (§3.3.1, §3.3.2.7). From that geometry the algorithm "generates an effective area for modeling emissions from the pit," whose "size, shape and location … is a function of wind direction" (§3.3.2.7) — the escaping fraction leaves from the downwind portion of the pit, which is the physics a flat AREA source at grade would miss. Two consequences: a pit cannot be subdivided into smaller sources, and "receptors should not be located within the boundaries of the pit" — the model sets them to zero (§3.3.2.7).

Lines and roads: LINE, RLINE, RLINEXT

LINE is the simple elongated release — a conveyor gallery, a long low vent — entered as two endpoints, a width, a rate, and a release height (§3.3.2.8).

RLINE is the roadway type: "near-surface releases from mobile sources… a travelled roadway with either single or multiple lanes" (§3.3.2.9). Computationally it is "Romberg numerical integration of point sources" along the line. Two version facts matter: RLINE became a regulatory option usable with DFAULT in version 24142, and it can account for terrain elevations since 23132 (§3.3.2.9) — for transportation conformity work, EPA's Office of Transportation and Air Quality guidance governs the details.

RLINEXT is the extended research variant (per-endpoint release heights, additional geometry) and remains "a non-regulatory ALPHA option" (§3.3.2.10).

BUOYLINE: the smelter special

Aluminum smelter potlines emit hot gases along long parallel buildings — too buoyant for LINE, too elongated for points. BUOYLINE implements the Buoyant Line and Point (BLP) model for exactly this geometry (Model Formulation, EPA-454/B-26-003, §2): each line carries its own emission rate and release height, and the group shares buoyancy parameters. If you are not modeling something shaped like a potline building, this is not your type.

SWPOINT: the research frontier

SWPOINT exists "as a research tool" for sidewash — the lateral shift of a building's wake cavity when wind strikes an elongated building obliquely, deflecting the plume sideways (§3.3.2.12). It requires the non-regulatory ALPHA flag; in a permit demonstration you will encounter it only by deliberate agreement with the reviewing authority.

Cross-cutting rules worth pinning

• Downwash belongs to the point family — direction-specific building dimensions apply to POINT, POINTHOR, and POINTCAP only (§3.3.2.1).
• Deposition is not available for RLINE, RLINEXT, BUOYLINE, or SWPOINT — concentration output only (§3.2.2.6; see Deposition & depletion).
• Variable emissions apply across types — temporally varying factors and hourly emission files (the EMISFACT and HOUREMIS inputs, §3.3.11–.12) attach to sources of any type, which is how a limited-schedule source is modeled at its permitted hours (see Emission inputs).

Choosing the type: represent the release, not the paperwork

The test is always physical: what does the emission look like in its first few seconds? A jet from a stack → POINT. The same jet under a rain cap → POINTCAP, because the momentum is gone but the heat is not. A dusty pile disturbed by loaders → AREA (perhaps with an initial vertical dimension), because there is no jet at all. Roof monitors venting a building's air → VOLUME, already dilute at release. A haul road → RLINE. Choosing by what's convenient to enter, rather than by the release physics, is one of the quiet ways demonstrations go wrong in either direction.

In PlumeSmart

All thirteen types are available in the source editor with their type-specific parameter forms, the ALPHA-only types labeled as non-regulatory, and per-source variable-emission schedules — the model input written for each run records exactly the type and parameters used.