小波变换方法具有较好的时频局部特性, 非常适合于分析非平稳的湍流信号。本文对35 m铁塔的超声风速测量数据进行了离散正交小波变换, 计算了各向同性系数(ISO isotropy coefficient)及小波功率谱, 以此对实验场所代表的水陆交际复杂下垫面的近地面层湍流特征进行了研究。ISO系数可以很好地描述实际大气在不同尺度的各向同性特征。根据ISO系数的分布, 我们可以通过设定一定的阈值(ISO=0.7)来得到各向同性小涡的分离尺度(记为ISO_0.7), 即湍流的各向同性尺度范围。研究表明, 下垫面对于边界层湍流各向同性特性具有相当大的影响, 当风从不同下垫面吹过时ISO_0.7尺度的均值具有较明显的差异。同时湍流度和稳定度对湍流各向同性也有一定的影响。对湍流小波功率谱研究表明, 功率谱谱幂率在分离出的小涡所处的频段接近于-5/3。而当风从陆面吹过且湍流度较弱时, 湍流小波功率谱小涡以分离尺度所对应的频率(f_ms)为界具有明显的两段趋势: 高于f_ms的频段, 谱幂率一般接近-5/3; 而低于f_ms的频段的谱幂率接近-3/3。这反映了下垫面特征对湍流功率谱的影响。同时在不同的风向转变过程中下垫面对功率谱的影响具有差异。

It is well known that wavelet transforms have good time-frequency localization, which makes the transforms be especially suitable for analyzing non-stationary signals, like turbulent signals. In this paper, discrete orthogonal wavelet transform is applied to the 35 m tower measured data, and the isotropy coefficient(ISO) and the wavelet power spectrum are computed using the wavelet coefficients, to study the turbulent characteristics of atmospheric boundary layer over complex underlying. ISO could describe the isotropy characteristic of real atmosphere at different scales well. Based on the distribution of ISO, by a threshold of ISO(here, the threshold is 0.7), we obtained the separation scale of isotropy small eddies(names as ISO_0.7). The result shows that the underlying characteristics have a large impact on ISO_0.7, when wind blows from different underlyings, the mean ISO_0.7 scale varies very obvoius. The result also reveals that the turbulent intensity and stability have some impaction too. The study on wavelet power spectrum shows, the high-frequency segment of wavelet power spectrum has a two-segment tendency beside the frequency f_ms which corresponds to the separation scale: The segment with frequency higher than f_ms has a scaling exponent(a_s)close to-5/3, which means, in this smalleddy frequency range, the turbulence is closely isotropic;and the segment with frequency lower than f_ms, has a scaling exponent(a_l)deviating from-5/3, which means, in this frequency range, the turbulence is closely anisotropic. And when wind blows from land and the turbulent intensity is weak, the power spectrum evidently has this two-segment-tendency and the a_l is very close to-1. So it reveals the impaction of underlying. And the different processes of wind direction change have very different impactions.