COMPUTATIONAL MODELING OF
TURBULENT TRANSPORT'
1. INTROD~JCTION
Inexpensive semiquantitative numerical simulation of turbulent transport
would have many applications. In particular. when applied to pollution
dispersal in an urban atmospheric environment, or in an estuary. it would
permit rational decision making by the government bodies involved. There
are, in addition. problems in oceanography and meteorology (such as thermocline formation) on which a reasonably realistic, though semiempirical.
computational technique could shed light.
Before attempting to construct such a method, we should consider two
basic, related questions. Is the method within our grasp conceptually and
computationally? That is, do we understand turbulence well enough to
model it with acceptable accuracy, and can the model provide simulation at
acceptable cost, with available computational facilities? These are very real
questions for, although we understand a great deal about turbulence, there is
even more that we do not know; also, even techniques that do not exceed the
capacity of the most recent generation of computers can easily exceed the
ability (or willingness) of reasonable men to pay, if computer time must be
accounted for.
A great deal of success has bcen had with direct numerical simulation of
turbulence (Orsmg and Patterson, 1972). This involves no dynamical modeling. However, Fox and Lilly (1972) have shown that such an approach
rapidly recedes beyond economic reach as the Reynolds number increases.
This work supported in part hy the Environmeiital Protection Agency. through its Select
Hcseiirch CiIoup in Meteorology, and in part by the Atmospheric Sciences Section of the
Natioiial Science Foundation under Grant No. GA-35422X. Prepared for presentation at the
Second I[JCX;-I IJTAM Symposium on Atmospheric Diffusion and Environmental Pollution.
('harlottrsville. Virginia, 1973; a preliminary version was presented hy proxy at the IAHRA I H H Intcrniition:il Symposium on Stratified Flows, Novosihirsk. 1972.
I69
TURBULENT TRANSPORT'
1. INTROD~JCTION
Inexpensive semiquantitative numerical simulation of turbulent transport
would have many applications. In particular. when applied to pollution
dispersal in an urban atmospheric environment, or in an estuary. it would
permit rational decision making by the government bodies involved. There
are, in addition. problems in oceanography and meteorology (such as thermocline formation) on which a reasonably realistic, though semiempirical.
computational technique could shed light.
Before attempting to construct such a method, we should consider two
basic, related questions. Is the method within our grasp conceptually and
computationally? That is, do we understand turbulence well enough to
model it with acceptable accuracy, and can the model provide simulation at
acceptable cost, with available computational facilities? These are very real
questions for, although we understand a great deal about turbulence, there is
even more that we do not know; also, even techniques that do not exceed the
capacity of the most recent generation of computers can easily exceed the
ability (or willingness) of reasonable men to pay, if computer time must be
accounted for.
A great deal of success has bcen had with direct numerical simulation of
turbulence (Orsmg and Patterson, 1972). This involves no dynamical modeling. However, Fox and Lilly (1972) have shown that such an approach
rapidly recedes beyond economic reach as the Reynolds number increases.
This work supported in part hy the Environmeiital Protection Agency. through its Select
Hcseiirch CiIoup in Meteorology, and in part by the Atmospheric Sciences Section of the
Natioiial Science Foundation under Grant No. GA-35422X. Prepared for presentation at the
Second I[JCX;-I IJTAM Symposium on Atmospheric Diffusion and Environmental Pollution.
('harlottrsville. Virginia, 1973; a preliminary version was presented hy proxy at the IAHRA I H H Intcrniition:il Symposium on Stratified Flows, Novosihirsk. 1972.
I69