Authors: Mason Dorseth (Florida Institute of Technology), Jean C. Perez (Florida Institute of Technology), Sofiane Bourouaine (Johns Hopkins Applied Physics Laboratory), Juan Carlos Palacios Caicedo (Florida Institute of Technology)

Spectral and correlation analysis are essential tools to investigate the turbulent properties of the solar wind from in-situ spacecraft observations. Most of these statistical techniques are based on estimators of autocorrelation functions (ACF), power spectral densities (PSD) as well as other important statistical averages. The PSD is often computed as the squared amplitude of the fast Fourier transform (FFT), which requires uniformly spaced, contiguous data. However, real data often has missing points that need to be filled in before applying the FFT, commonly with linear interpolation. Data gaps become more of a problem when conditioning is applied to the signal and many gaps are introduced to cover sections of the signal with undesirable properties. The PSD can also be calculated from ACF using the Wiener-Khinchin theorem. The advantage of using ACF is that it is resilient to data gaps and does not require interpolation. In this work, we artificially introduce gaps to numerically simulated, homogeneous plasma turbulence signals and to high resolution Wind magnetic field signals to investigate the consistency of the PSD estimator, i.e. its convergence to the true ensemble-averaged power spectral density and its sensitivity to the total gap percentages (TGP). This technique is then applied to one-year long conditioned Wind data to study spectral and correlation properties of the fast solar wind, which allows us to use a larger statistical sample than conventional FFT methods.