A note on determining the scale of turbulence

Recent observations of wind inclination near the ground in stationary conditions, are shown to possess properties consistent with an effectively complete description of the turbulence on the low-frequency side. The results are therefore particularly suitable for determining the conventional time-scale of turbulence, and the values obtained are found to be appreciably greater than many previous estimates. The results are then used to examine the accuracy of estimating the scale from data which are less complete on the low-frequency side (shorter sampling duration), by making assumptions about the shape of the whole spectrum, and it is seen that assumptions which have hitherto seemed reasonable for this purpose can in fact lead to a two-fold or greater underestimation of the scale. It is inferred that for adequate low-frequency specification (hence reliable estimation of scale) the sampling duration over which measurements should be made and repeated many times must be at least 40 times the time-scale.