Scaling, Anisotropy, and Complexity in Near‐Surface Atmospheric Turbulence

Abstract

The development of a unified similarity scaling has so far failed over complex surfaces, as scaling studies show large deviations from the empirical formulations developed over flat and horizontally homogeneous terrain as well as large deviations between the different complex terrain data sets. However, a recent study of turbulence anisotropy for flat and horizontally homogeneous terrain has shown that separating the data according to the limiting states of anisotropy (isotropic, two‐component axisymmetric and one‐component turbulence) improves near‐surface scaling. In this paper we explore whether this finding can be extended to turbulence over inclined and horizontally heterogeneous surfaces by examining near‐surface scaling for 12 different data sets obtained over terrain ranging from flat to mountainous. Although these data sets show large deviations in scaling when all anisotropy types are examined together, the separation according to the limiting states of anisotropy significantly improves the collapse of data onto common scaling relations, indicating the possibility of a unified framework for turbulence scaling. A measure of turbulence complexity is developed, and the causes for the breakdown of scaling and the physical mechanisms behind the turbulence complexity encountered over complex terrain are identified and shown to be related to the distance to the isotropic state, prevalence of directional shear with height in mountainous terrain, and the deviations from isotropy in the inertial subrange.

Motions on Second-Order Moment Budgets in the Stable Atmospheric Boundary Layer

“…Although from the physical point of view length scales are more appealing, time-scales are usually preferred for point measurements, because the time-space transformation is hampered by the inapplicability of Taylor's hypothesis to submeso motions [10]. Thus, in agreement with other authors [11,31,19], we use this fixed-time-scale separation recognizing, however, that the spectral gap is not always observed and that a residual submeso contribution may affect the range T < 100 s, especially for large stability. Figure 1 shows the stability dependence of the submeso parameters for the second-order moments involved in the uw budget, Eq.…”

“…In this sense, T = 100 s is the time scale that separates small-scale turbulence (T < 100 s) and submeso motions (T > 100 s). In the SBL, an averaging time ∼ 100 s has been used by several authors [e.g., 10,17,18,4,19]. To quantify the strength of the submeso effect (Sect.…”

Scale interactions and anisotropy in stable boundary layers

“…Interestingly, this route does not involve purely two‐component axisymmetric turbulence, but is more of an axisymmetric oblate type. However, Stiperski et al () show that the lack of small‐scale two‐component axisymmetric turbulence is often observed in other datasets. An interesting future analysis would be to investigate the scale‐wise return to isotropy in parallel with our results on the persistence and dimension of the dynamics.…”

Decreasing Wind Speed Extrapolation Error via Domain-Specific Feature Extraction and Selection

“…Model uncertainty is a vexing problem in the wind energy industry. It has been shown that the so-called 'universal' model parameters are outdated and can no longer serve their purpose in efficiently predicting and extrapolating meteorological properties (Sfyri et al, 2018;Stiperski et al, 2019). This problem can be mitigated by the use of machine learning tools which have made great strides in the past few decades.…”