Wind profile relationships from the ‘wangara’ experiment

From over 1000 half-hour observations of near-surface wind profiles at Hay, NSW, more than 500 high quality sets of data are selected. In unstable conditions, these closely confirm previous analyses and suggest near equality between z/L and Ri. The data are well described by either the KEYPS relationship (Panofsky) or that of Businger: ϕM = (1–16z/L)−1/4. In stable conditions, a log-linear formulation (ϕM = 1+az/L) is found to give an adequate description of the wind profile up to z/L ≃ 0.5, with some evidence for a slight variation in α between the values 4.0 at neutral and about 6 when z/L ≃ 0.2. The average value of α between these limits is found to be 5.0±0.2. In conditions of very high stablity, a linear profile (du/dz = cu*/(kL)) is suggested above the height z ≃ 10L. In the transition region between the log-linear and linear profile regimes, the log-linear formulation appears to tend towards a purely logarithmic law as stability increases. There is no evidence for any sudden change in behaviour, nor is there any suggestion that a purely logarithmic relationship is ever attained as an average situation. The value of c in the purely linear relationship is found to be between 0.4 and 0.9. The data also indicate that in extremely stable conditions (L ≃ 50cm) the dimensionless gradients of heat and of momentum may differ by about a factor of two, with ϕH being the larger. The roughness length of the site used is found to be 1.2±0.1 mm, considerably less than the values appropriate to earlier experiments performed in the same general area. There is some evidence for an increase in z0 with decreasing wind speed (reaching about 3 mm when the wind at 1 m is about 1 ms−1), in accord with Deacon's hypothesis concerning the form drag of roughness elements. From the point of view of applying a low-level drag coefficient in order to estimate friction velocities, the errors arising from the change in roughness length are sometimes comparable to those resulting from stability effects. Not surprisingly, the data show that in slightly stable conditions dewfall was low, whereas in extremely stable situations dewfall accounted for most of the heat loss from the air. The data used here form part of a much larger body of information obtained during 1967 and known as the ‘Wangara’ experiment.