AirWare User Manual

AERMOD: A Dispersion Model
for Industrial Source Applications

This documentation is extracted from Perry et al. (1998).

Dispersion Estimation

AERMOD represents a fundamental change from the familiar PG approach to one based on more current planetary boundary layer (PBL) parameterizations. The PG curves are based exclusively on short-range dispersion from ground-level sources. However, ISC2 applications are not limited to low-level short-range situations. For example, ISC2 is routinely used to estimate concentrations from large buoyant elevated sources.

The rate of diffusion in the atmosphere varies significantly with height. In the convective boundary layer (CBL), convection dominates dispersion only a short distance above the surface (on the order of the Monin-Obukhov length, L, typically tens of meters). However, near the surface, where the PG dispersion curves are based, vertical dispersion, s z, is dominated by mechanical turbulence characterized by the friction velocity, u*. Both field experiments and numerical simulations have shown that dispersion from elevated sources exhibits a linear dependence with downwind distance, x. However, PG "A" produces a dependency of x3/2 or (x2 beyond 500m) as is expected for surface releases only.

A similar issue exists in the stable boundary layer (SBL). Near the surface, mechanically-generated turbulent velocities (that dominate s z) are of the order of u* (typically about one-tenth of the wind speed). However, above the SBL s z is suppressed by the stable buoyant forces. The ISC2 approach does not consider these vertical variations.

In contrast, the AERMOD dispersion parameters (s y and s z) are determined from known relationships with the turbulence quantities, s v and s w. The parameters s v and s w are either measured or estimated from similarity considerations as functions of the convective scaling velocity (w*), u*, the convective mixed layer height (zi), the mechanical mixed layer height (h), and the height (z) above the surface. These relationships, which have been found to compare well with both theory and observations, provide needed height and downwind distance dependencies in the model.

Gaussian Formulation

Under all atmospheric conditions, the ISC2 model relies on a Gaussian form for both the horizonal and vertical concentration distributions. This assumption is reasonable for neutral and stable conditions and for the lateral distribution under convective conditions. However, it has been known for some time that the vertical distribution under convective conditions is clearly non-Gaussian in the mixed layer0. Pollutant distributions reflect the distribution of vertical velocities that, in a convective boundary layer, are not symmetric in relation to updrafts and downdrafts. Although the area-averaged vertical velocity may be near zero, updrafts tend to be "stronger" while downdrafts tend to cover more horizontal area. This yields a skewed vertical distribution of mass.

For convective conditions, AERMOD relies on a skewed probability density function (pdf) to characterize the vertical distribution. The specific pdf used is bi-Gaussian in form. Therefore, application of AERMOD should help reduce the uncertainty in estimates of the design concentrations when convective conditions are controlling (e.g. simple terrain, tall stack, large buoyant flux).

Boundary Layer Inhomogeneity

ISC2 makes the assumption that the PBL is homogeneous in its treatment of both plume transport and dispersion. However, as a plume grows vertically, variations in both turbulence and mean wind through the plume's depth become significant. These effects are most pronounced near the surface where the vertical gradients of these parameters is greatest. AERMOD accounts for the vertical inhomogeneity of the PBL by "averaging" the parameters of the actual PBL into "effective" parameters of an equivalent homogeneous PBL. With these effective parameters, AERMOD accounts for the inhomogeneity of the PBL, in an averaged sense. Since the effect of the inhomogeneity on transport and dispersion depends on the plume's vertical size, these effective parameters are formulated as functions of downwind distance.

Near the source, where the plume is small, AERMOD uses the actual values of the variables at the height of release. However, at and beyond the point where the plume size is of the order of zi, the effective parameters are calculated as averages over the mixed layer depth. Between these two limits the effective parameters are assumed to be exponentially dependent on downwind distance and the Lagrangian time scale.

Plume Rise and Buoyancy Effects

Three aspects of ISC2's approach to plume rise and buoyancy have been improved in AERMOD. First, for final rise in neutral and unstable conditions, ISC2 uses the Briggs0 model that contains no treatment for rise limited by convective turbulence during the daytime. AERMOD addresses this by superimposing on the distance-dependent plume rise of ISC2 an estimate of the plume's displacement due to random convective velocities.

Second, when plume material reaches the top of the mixing layer, ISC2 completely reflects it back towards the ground. Although this may be appropriate for a plume that is non-buoyant at zi, this approach does not hold for highly buoyant plumes. Such plumes have a tendency to hug the top of the mixed layer, for a time, before they start diffusing back towards the ground. AERMOD has a separate term, the indirect source term, to explicitly account for reflections off the mixed-layer top. This term is similar to the first image source used in ISC2 in that both are mathematical constructs designed to reverse the flux of material that would otherwise leave the PBL. However, where ISC2 immediately reverses the flux, AERMOD delays the reversal based on the plume buoyancy.

Third, ISC2 adopts a simple "all or nothing" approach for estimating plume penetration of an elevated stable layer. If the plume centerline is predicted to be above zi, then ISC2 assumes no mass remains in the mixed layer and thus yields zero ground-level concentrations. Conversely, ISC2 assumes that all mass remains within the mixed layer if the plume centerline is below zi. In contrast, AERMOD computes the fraction0,0 of plume mass that penetrates the layer aloft. Penetration is a function of plume buoyancy, wind speed, zi, and the strength of the stability in the layer above zi. AERMOD employs a third source term in the elevated stable layer to account for penetrated mass that may re-enter the mixed layer and impact at receptor locations. The dispersion of pollutant from this penetrated source is controlled through the use of "effective" turbulent intensities whose values are influenced by turbulence levels in both the stable and mixed layers (see section on PBL Inhomogeneity). This improvement should reduce the excessive sensitivity to zi that is exhibited by ISC2 during conditions of limited mixing.

Next: Terrain handling

© Copyright 1995-2018 by:   ESS   Environmental Software and Services GmbH AUSTRIA | print page