Many current stochastic turbulent simulation methods generate a random vector by prescribing a power spectral density, which specifies the magnitudes of the Fourier components, and then independently sampling the phases from a uniform distribution. It can be shown that this formulation will lead to a Gaussian, stationary process in the time domain, which may not accurately model the non-Gaussian, nonstationary wind of the real world. This presentation will cover the theory of “phase coherence,” a simple modification of the standard stochastic simulation method that can simulate nonstationary processes. An analysis of the prevalence of phase coherence in real data will also be presented. Lastly, a brief demonstration on the effect of phase coherence on the output of a dynamical system will be given to emphasize the importance of including phase coherence in simulations for design purposes.