a?
JOOST A. t)t!SLNCit;H AND S. P. S. ARYA
- I
-r
1
FIO. 4. Comparison of the neutral stress profilm from the present model with those of
Wyngaiird cr NI. (this volume).
stress and velocity defect remaining at this level, it should not be con!&ered
ilh the top of the boundary layer.
Altcrnative ways of defining the boundary layer height have been discusscd by Zilitinkevich (1972). The most convenient, especially from a
theorctical point of view, L tkdimensionleoe height t, where the magnitude
of the normalized stress reduces to an arbitrarily specified small fraction. If
we take this fraction to be 1 y ! 4, turns out to be very close or identical to <,
(represented by open circles in Fig. 10). Based on some rough assumptions
about the limiting or critical value of the flux Richardson number and about
the relationship between the boundary layer height and an "effective"eddy
viscosity. Zilitinkevich (1972) predicted 4, - p; 1'2, for the asymptotic case
of large stability. This relationship is also confirmed by the results of our
model: the dotted line in Fig. 10 representing
(30)
nicely fits through the computed points, especially for 42%.