High frequency (HF) radar field practitioners usually discard the phase content of acquired signals due to the assumption that it is random and uniformly distributed. Traditionally, the phase component is removed in the process of computing the spectrum via Fourier analysis. This statistically means that the HF radar data is linear and stationary. However, events occurring in nature are inherently nonlinear and nonstationary and exist in noisy environments. These uncertainties are not easily tractable because they break the assumptions upon which traditional signal processing tools are built. This research consequently investigates the validity of these assumptions by statistically analyzing HF surface wave radar (HF-SWR) field data. The results from this research show that the HF-SWR data is nonlinear, second-order improper or noncircular and nonstationary. This intimates that the phase is informative and it should be considered in order to achieve a robust analysis of HF radar data.
