Structures are designed for wind loads representative of the atmospheric boundary layer. However, wind flow is an extremely random natural phenomenon both in time and space. Hence, turbulence can only be described stochastically using probabilistic techniques. Extreme wind events are even more random and hence full-scale measurement-based studies have always been the backbone of wind engineering of structures. Very few field measurement-based studies have been possible in India till date in spite of the long coastline and multiple cyclonic storms annually. The lack of an extensive database may be attributed to huge expenditure associated with installation, data acquisition and maintenance of continuous high-precision instrumentation. The present study deals with analysis of cyclonic wind speed data obtained for a Northern Indian Ocean cyclone, Vardah (December 2016). Spectral analysis reveals a progressive increase in energy content in the large eddies which has not been documented over time for hurricanes and typhoons. This study makes an attempt to numerically quantify turbulence scales using temporal correlation (Taylor’s frozen turbulence hypothesis), as well as a few alternative approaches. The applicability of Taylor’s frozen turbulence hypothesis to atmospheric phenomena showing high variations may be questionable. The objectives of this research were two-fold and interesting revelations were obtained even though a conclusive answer to many questions will need further research and a more extensive database.