Fine structure of the turbulence in EPRs

It can be seen from the above consideration that the theory of EPR flows based on the introduction of the distributed force and sources of substances agrees well with experimental data, at least qualitatively. At the same time, the notable scattering can motivate looking for a refined theory, especially as concerned to the algebraic turbulence model (3.131). Second-order turbulence models that are expressed in terms of the differential equations for vT and the associated quantities should certainly fit better, 

Some paradoxes of the turbulence in canopies, or EPRs, were pointed out by Raupach and Thom in their state-of-art review of 1981, [522], The first phenomenon is the value of the drag coefficient of elements that constitute the EPR. The highly precise measurements in aerodynamic tubes brought values that depend on the obstacle shape, the flow turbulence level, and the mutual disposition of obstacles but vary near cf 0.5 for spheres and cf 1 for cylinders in the working range of the local Reynolds number 103 Re 105. The same coefficient determined from the field measurements in forests turned out to be several times less (in this case, the indirect calculations were performed). A similar paradox takes place for the exchange coefficients. 

Raupach and Thom [522] explained this fact by shading some EPR elements by another ones ( shelter effect ). However, the mechanism of this effect has not been clarified is it a result of the vorticity behind obstacles, or of the extreme turbulence level, or of the basic difference between the natural forest canopy and laboratory models A possible consequence of aeroelasticity of canopy elements was also listed, [522], All these mean that the purposeful experiments should still be carried out as well as the mathematical models should be developed that account for a particular behavior of EPR elements (like the freedom of droplets to be drifted by the flow). 

There is no explanation for the phenomenon mentioned. It is aimed to attract an 

The accuracy of each final curve E = E(f) depends on the parameters of the measurement procedure described. Most important is the sampling frequency fs that determines the time interval between samples, At = l/fs. The block size was taken as N = 28 = 4096 and so the sampling time was t = N At (the power of 2 being dictated 

---