VERTICAL COMPONENT OF TURBULENCE
IN CONVECTIVE CONDITIONS
S. J. CAUGHEY AND C. J. READINGS
Metetrro/ogicul Research Unit, R.A.F. Cardington. Bedford. U . K .
1. INTROIXJCTION
The magnitude of the variance of the vertical wind component 0: is of
great importance in the study of the diffusion of contaminants as well as for
its bearing on the energy balance in the atmospheric boundary layer.
However, although its behavior is quite well understood in the constantstress layer (e.g.. Panofsky and Mazzola, 1971), very little is known higher up
in the atmosphere. Panofsky and McCormick (1960) argued that its value is
determined by the height above the ground z and by the local rate at which
turbulent energy is added to the system by the processes represented by the
convective and mechanical production terms in the turbulent energy balance
equations. Applying dimensional analysis to these assumptions gives
( 1 )
6, = A(z(c, + 8&2))”3
where A and 6 are empirical constants; F., is the rate at which mechanical
energy is being added and E~ is the rate at which energy is added by convection. d is introduced to “allow for the possibility that convection is more
efficient than wind shear in producing a vertical component of turbulence.”
Although there is no a priori reason for S being independent of height and
stability, there are no observations that contradict this hypothesis. Of
course, higher up in the atmosphere z may not be the appropriate length
scale, and at any height a, may be affected by other processes such as
advection. This approach of defining 6, in terms of local parameters has
recently been developed further by Yokoyama (1971) and Fiedler and
Panofsky (1972).
The present paper reconsiders this approach using hitherto unpublished
data that have been obtained at heights of up to 1 km, during the past few
years at two different sites---one in England (Cardington, Bedfordshire) and
the other in the United States (Fort Eglin. Florida). A11 the measurements
were made with the Cardington turbulence probe (see Readings and Butler,
125
IN CONVECTIVE CONDITIONS
S. J. CAUGHEY AND C. J. READINGS
Metetrro/ogicul Research Unit, R.A.F. Cardington. Bedford. U . K .
1. INTROIXJCTION
The magnitude of the variance of the vertical wind component 0: is of
great importance in the study of the diffusion of contaminants as well as for
its bearing on the energy balance in the atmospheric boundary layer.
However, although its behavior is quite well understood in the constantstress layer (e.g.. Panofsky and Mazzola, 1971), very little is known higher up
in the atmosphere. Panofsky and McCormick (1960) argued that its value is
determined by the height above the ground z and by the local rate at which
turbulent energy is added to the system by the processes represented by the
convective and mechanical production terms in the turbulent energy balance
equations. Applying dimensional analysis to these assumptions gives
( 1 )
6, = A(z(c, + 8&2))”3
where A and 6 are empirical constants; F., is the rate at which mechanical
energy is being added and E~ is the rate at which energy is added by convection. d is introduced to “allow for the possibility that convection is more
efficient than wind shear in producing a vertical component of turbulence.”
Although there is no a priori reason for S being independent of height and
stability, there are no observations that contradict this hypothesis. Of
course, higher up in the atmosphere z may not be the appropriate length
scale, and at any height a, may be affected by other processes such as
advection. This approach of defining 6, in terms of local parameters has
recently been developed further by Yokoyama (1971) and Fiedler and
Panofsky (1972).
The present paper reconsiders this approach using hitherto unpublished
data that have been obtained at heights of up to 1 km, during the past few
years at two different sites---one in England (Cardington, Bedfordshire) and
the other in the United States (Fort Eglin. Florida). A11 the measurements
were made with the Cardington turbulence probe (see Readings and Butler,
125