Atmospheric Structure Simulation: An Autoregressive Model for Smooth GeophysicalPower Spectra with Known Autocorrelation Function

摘要

Within a defined domain, geophysical phenomena often are characterized by smoothcontinuous power spectral densities having a negative power law slope. Frequently, Fourier transform analysis has been employed to generate synthetic scenes from pseudorandom arrays by passing the stochastic data through a Fourier filter having a desired correlation structure and power spectral dependency. This report examines the possibility of producing synthetic structure by invoking autoregression analysis as contrasted with the Fourier method. Since computations that apply multidimensional fast Fourier transforms to large data arrays consume enormous resources and time, the goal of this study is to seek an alternative method to reduce the computational burden. Future editions of the Phillips Laboratory Strategic High Altitude Atmospheric Radiance Code (SHARC) will feature an ability to calculate structured radiance. The methods explored herein provide a process that can complement or in some cases supplement methods presently being used.