Suppose that the initial distribution of  is dependent only on
 is dependent only on  and
 and
 and that
 and that  ,
,  and
 and  are constant. Then Eq. (14) reduces to
 are constant. Then Eq. (14) reduces to
 
 :
:
 
 
 is constant. Then the solution to Eq. (16) is
 is constant. Then the solution to Eq. (16) is
 
 
 kgm
 kgm of
pollutant at
 of
pollutant at  in the plane
 in the plane  . Eq. (17) is the Gaussian ``puff''
solution because it describes the release of a small ``puff'' of pollutant.
. Eq. (17) is the Gaussian ``puff''
solution because it describes the release of a small ``puff'' of pollutant.
An important quantity is the mean square distance  which the
pollutant has spread from
 which the
pollutant has spread from  . This is given by
. This is given by
 
 
 .
.
In this case we assume that  so that the equation for
 so that the equation for
 is
 is
 
 and also
let
 and also
let  . Then
. Then
 
![\begin{displaymath}
C(X,Y,t)={Q\over 4\pi t\sqrt{\varepsilon _x\varepsilon _y}}{...
...ver \varepsilon _x}+{Y^2\over \varepsilon _y}\right )\right ].
\end{displaymath}](img75.png) 
 and
 and 
 . Hence
. Hence
![\begin{displaymath}
C(X,Y,t)={Q\over 2\pi\sigma_x\sigma_y}{\rm exp}\left [-{1\ov...
... \sigma_x^2}+{Y^2\over \sigma_y^2}\right )\right ].\eqno{(22)}
\end{displaymath}](img78.png) 
This is the most general and practical case where  .
The equation for
.
The equation for  is
 is
 
 in line with the cases of plane and line sources but
it is difficult to think of real cases where there is a mean vertical velocity
 in line with the cases of plane and line sources but
it is difficult to think of real cases where there is a mean vertical velocity
 so we keep
 so we keep  . The transsformed equation for
. The transsformed equation for  is then
 is then
 
![\begin{displaymath}
C(X,Y,z,t)={Q\over 8(\pi t)^{3/2}\sqrt{\varepsilon _x\vareps...
...er \varepsilon _y}+
{z^2\over \varepsilon _z}\right )\right ].
\end{displaymath}](img85.png) 
 ,
, 
 and
 and
 so that
 so that
![\begin{displaymath}
C(X,Y,z,t)={Q\over (2\pi)^{3/2}\sigma_x\sigma_y\sigma_z}
{\r...
...\sigma_y^2}+{z^2\over \sigma_z^2}\right )
\right ].\eqno{(24)}
\end{displaymath}](img87.png) 
 
Fig. 2. Concentration distributions at various times after pollutant release for the diffusion of an instantaneous plane source (Eq. (20)).
 
 and
 and  .
.
 
Fig. 3. Shear flow  .
.
If we neglect diffusion, then the equation for  is
 is
 
 Eq. (26) becomes
 Eq. (26) becomes
 
 , i.e. the concentration is independent of time moving with
the local flow
, i.e. the concentration is independent of time moving with
the local flow  . Consider as in the previous section the case of an
instantaneous plane source at
. Consider as in the previous section the case of an
instantaneous plane source at  ,
,  . Then at
. Then at  
 
 .
Therefore
.
Therefore
 
 
Fig. 4. A line (or plane in 3-D) of pollutant
initially lying on  is stretched and rotated by a shear flow
 is stretched and rotated by a shear flow  .
At time
.
At time  the
 the  extent of the line of pollutant is denoted by
 extent of the line of pollutant is denoted by  .
.
Consider the horizontal extent  of the pollutant. At time
 of the pollutant. At time  it is
clearly
 it is
clearly
 
 . This
contrasts with the rate proportional to
. This
contrasts with the rate proportional to  for diffusion, so it would
appear that advection is always more important than diffusion, except perhaps
for a short time after release. However, we shall see in the next section
that this is not the case.
 for diffusion, so it would
appear that advection is always more important than diffusion, except perhaps
for a short time after release. However, we shall see in the next section
that this is not the case.
 
Fig. 5. An initial slab (a) of pollutant is rotated and stretched by shear (b), leading to enhanced diffusion in the transverse direction (c). The spreading of the line of maximum concentration is clearly reduced by the diffusion.
The processes described here are known as Taylor's mechanism.