Kolmogorov's structure functions for the longitudinal and transverse components of locally homogeneous isotropic turbulence are combined vectorially to obtain an expression which permits the evaluation of E (atmospheric dissipation rate) from climatological data. This is used to derive climatological patterns of e in the free atmosphere from Crutcher's upper wind statistics of the Northern Hemisphere. The latter are combined with Kung's boundarylayer values to estimate the distribution of total atmospheric dissipation over the Northern Hemisphere. I. EPSILON AS A FUNCTION OF CLIMATOLOGICAL PARAMETERS In a recent review of the methods of evaluating e (the rate of kinetic energy dissipation in the atmosphere), it appeared that it could be determined from wind variability data. The theory for such an eva1.uation is provided by Kolmogorov's (1941a) second hypothesis of similarity of locally homogeneous isotropic turbulence. In such a field of turbulence and with the z axis along the mean vector wind, this gives for the wind components a t points 1 and 2 a distance x apart U,(X)~= (U~-U~)~=~(U,)~[~-~,(X)]=C(~X)~~~ (1) and U~(X)~= (~~-~~) 2 = 2 (~,)~[ 1-r , (z) J = 4 5 C (~X)~/~, i.e., the space variance or structure function (square of the Eulerian space variability, a,($), which in turn is a function of the standard deviation, u,, and the Eulerian space correlation, ru(z)) is a function only of the separation distance and the intensity of, the turbulence. The latter, by Eolmogorov's hypothesis, is determined by e. Aside from the effect of orientation, equation (1) can be written by dimensional analysis. It can also be derived leading to theoretical as well as empirical values for..&e.constant C. Values assigned Cin the literature include 34 (Kolmogorov, 1941b), 22/a (Obukhov and Iaglom, 1951)) $5 (MacCready, 1953), and approximately 2 (Hinze, 1959; Pond et al., 1963). In this study we use C=2 primarily because this yields the lowest estimates. forE. The x-range of validity of (1) is presumed to be