David I. Havelock
Inst. for Microstructural Sciences, Natl. Res. Council, Ottawa, ON K1A 0R6, Canada
Fully developed turbulence follows the well-known Kolmogorov spectrum and, within the inertial subrange, is governed by a single parameter (epsilon) called the viscous dissipation rate. Under stationary conditions the intensity of sound scattered from turbulence follows an exponential distribution with mean intensity I[inf 0] determined by (epsilon). In a more realistic turbulence model, the viscous dissipation rate for a given scattering volume fluctuates with a log-normal distribution. The corresponding fluctuations in I[inf 0] cause the intensity distribution to deviate from the exponential distribution. In particular, the tail of the distribution is raised, providing more frequent occurrences of higher intensity levels. This effect impacts on target detection probability in acoustic remote sensing applications. It is shown that the deviations from an exponential distribution are clearly observable in direct measurements of sound intensities within a refractive shadow near the ground. The variance (sigma) of the fluctuations in the dissipation rate is estimated by comparing the measured and theoretical distributions. It is also shown that (sigma) cannot be obtained directly from short-term estimates of I[inf 0].