Characterizing Intermittency in the Stable Arctic Atmospheric Boundary Layer

Characterizing Intermittency in the Stable Arctic Atmospheric Boundary Layer

Mohammad Allouche, Elie Bou-Zeid, Jose Fuentes (Penn State), Marcelo Chamecki (UCLA), Sham Thanekar (Penn State), Cedrick Ansorge (U of Cologne)

Polar regions have experienced the most rapid rate of environmental change over the past decade. Dynamic changes in the Arctic sea ice coverage are having a major influence on the transport of scalars and on atmospheric dynamics and chemistry, while over land, the potential for severe permafrost melt looms. To elucidate physically the surface-atmosphere exchange processes that modulate such change in these areas, our understanding of the stable atmospheric boundary layer (SABL) needs to advance further. The SABL remains one of the most challenging topics in boundary layer meteorology in terms of modeling and parameterization. To that end, field observations from Barrow, Alaska are analyzed in this study. Specifically, we focus on the intermittent turbulence regime observed in the strongly stable case. The atmospheric surface layer (ASL) under such regime is characterized by transient and heterogeneous switching between turbulent and laminar state, known as “intermittency”.

Intermittent behavior is marked by rare flux or turbulent events of large magnitude, separated by some periods that are much more quiescent; that is, the flow variability has some “bimodal feature”. Here, different statistical tools are assessed to detect and characterize intermittency; namely, the probability density function of the outliers, the coefficient of variation, and the hourly kurtosis. The analysis is based on how 1-minute average quantities vary over a one-hour period. The first method quantifies the number of sensible heat flux H spikes (statistical outliers). These spikes are defined as the number of burst events that lie outside the range of ± one standard deviation from the mean. The second method relies on the coefficient of variation (CV=standard deviation over the mean), while the third method measures the kurtosis “4th moment” in each period.

The results indicate that these different measures do not identify the same periods as being intermittent. Therefore, the physical differences between the inferences made based on the various metrics are investigated. Subsequently, the failure of eddy diffusion theory during intermittent periods is investigated, and alternative approaches for turbulence closure based on the vertical velocity variance or on stochastic “surface renewals” by advecting turbulent patches are examined.