Over land, stable conditions are usually characteristic of the nocturnal boundary layer, but can also persist for several months in polar regions during winter. Episodic development of turbulence (also known as ‘bursting') and coherent structures (patchy turbulence) are ubiquitous features of stably stratified flows. These seemingly random yet highly organized features pose great challenges in signal processing. Commonly used techniques (e.g., variable interval time averaging - VITA) are not always sufficient to characterize details in these flow patterns. In this study, we develop a rigorous statistical technique to detect the existence of intermittent high turbulence events which are not captured by standard stationary second-order models. The proposed methodology is based on the concept of an incoherent signal and does not make any assumption about the process spectrum. Observational data from the Cooperative Atmosphere-Surface Exchange Study – 1999 (CASES-99), in Kansas, and the Investigation of Sulfur Chemistry in the Antarctic Troposphere – 2000 (ISCAT00), in Antarctica, field campaigns are utilized to gain a better understanding of the differences between turbulent bursting episodes in polar and mid-latitude regions.