DIFFUSION FROM A CONTINUOUS SOURCE AT SEA
32 1
Since this value of m could be fitted to the data without appreciable loss of
significance, the value of €3 was determined in each case. For the twelve
experiments in 1970-1971. referred to above, the values of B ranged from 0.4
to 1.6 cm/sec with a median of 1.0 cm/sec.
Table I1 summarizes the values of B for the various groups of experiments.
It is seen that the Holy Loch values are appreciably lower than those for the
Irish Sea except for the area off the Isle of Man where the currents were
weak. Otherwise no clear distinction can be made between the different
areas of the Irish Sea.
5. VERTICAL DIFFUSION
Determinations of the vertical variance 0,' and the corresponding eddy
diffusion coefficient K, were depkndent on measurements of concentration
at only three depths and so were less reliable than those of 0; and K,. No
attempt will be made here to describe the methods of analysis used. It is
sufficient to say that the standard deviation in the vertical was an order of
magnitude less than in the horizontal. The value of a, was between 1 and
3 m for t = 30 min and values of K, were in the range 1.2 to 33 cm2/sec.
Attempts were made to fit a power law relation between of and t, i.e.
(7)
c,' = ht"
analogous to Eq. (2) for a : . These led to very variable values of n and the
method must be considered unreliable, not only in view of only three depths
bcing available but also because the effects of a finite initial variance would
be more severe.
6. PEAK CONCENTRATIONS
If the distribution of concentration were Gaussian in both the lateral and
vertical directions, the peak concentration on the axis of the plume would be
given by
(8)
Cp = Q/2ltVaYa,
where Q is the mass rate of release of dye at the source and V is the current
speed at the time of release. If 0: and 0,' vary according to the power laws in
Eqs. (2) and (7) respectively, then
(9 1
c p = At-4
where y = i ( m + n ) and A = Q / [ 2 n V ( ~ h ) ' / ~ ] .
From the Irish Sea data of 1970-1971, regression lines were fitted to
bilogarithmic plots of C , against t for each release, leading to values of q
32 1
Since this value of m could be fitted to the data without appreciable loss of
significance, the value of €3 was determined in each case. For the twelve
experiments in 1970-1971. referred to above, the values of B ranged from 0.4
to 1.6 cm/sec with a median of 1.0 cm/sec.
Table I1 summarizes the values of B for the various groups of experiments.
It is seen that the Holy Loch values are appreciably lower than those for the
Irish Sea except for the area off the Isle of Man where the currents were
weak. Otherwise no clear distinction can be made between the different
areas of the Irish Sea.
5. VERTICAL DIFFUSION
Determinations of the vertical variance 0,' and the corresponding eddy
diffusion coefficient K, were depkndent on measurements of concentration
at only three depths and so were less reliable than those of 0; and K,. No
attempt will be made here to describe the methods of analysis used. It is
sufficient to say that the standard deviation in the vertical was an order of
magnitude less than in the horizontal. The value of a, was between 1 and
3 m for t = 30 min and values of K, were in the range 1.2 to 33 cm2/sec.
Attempts were made to fit a power law relation between of and t, i.e.
(7)
c,' = ht"
analogous to Eq. (2) for a : . These led to very variable values of n and the
method must be considered unreliable, not only in view of only three depths
bcing available but also because the effects of a finite initial variance would
be more severe.
6. PEAK CONCENTRATIONS
If the distribution of concentration were Gaussian in both the lateral and
vertical directions, the peak concentration on the axis of the plume would be
given by
(8)
Cp = Q/2ltVaYa,
where Q is the mass rate of release of dye at the source and V is the current
speed at the time of release. If 0: and 0,' vary according to the power laws in
Eqs. (2) and (7) respectively, then
(9 1
c p = At-4
where y = i ( m + n ) and A = Q / [ 2 n V ( ~ h ) ' / ~ ] .
From the Irish Sea data of 1970-1971, regression lines were fitted to
bilogarithmic plots of C , against t for each release, leading to values of q