| 摘要: |
| 基于时长38天的海表风场实测数据，应用经验模态分解（Empirical Mode Decomposition，EMD）和小波分解（Wavelet Decomposition，WD）这两种数据处理方法首先对涡相关法中的截断时间尺度（Cutoff Time scale，CTS）进行估算，结果显示：基于EMD与WD方法估算出的CTS一般都在40秒左右（EMD的结果略小），远远小于传统涡相关法中CTS的取值（固定为10分钟），且EMD和WD的使用使得每一段数据都能够根据自身的湍流特点而获得合适的CTS；EMD方法和WD方法有效的去除了计算结果中的非湍部分，且对通量传输方向的刻画也更加合理，极大提高了通量的计算精度，所得通量与传统方法计算的通量偏差平均值高达45%；研究还对EMD和WD的优缺点进行了对比分析，结果表明EMD相比于WD有更高的自主性，而WD对信号的分离程度则更高。 |
| 关键词: 海气通量 涡相关法 截断时间尺度 经验模态分解 小波分解 |
| DOI：10.11693/hyhz20180300060 |
| 分类号:P731 |
| 基金项目:国家自然科学基金项目，41576013号。 |
附件 |
|
| ESTIMATE OF CUTOFF TIME SCALE IN THE EDDY COVARIANCE METHOD FOR AIR-SEA FLUX BASED ON EMPIRICAL MODE DECOMPOSITION AND WAVELET DECOMPOSITION |
|
JIANG Hao1, ZHAO Zhong-Kuo2, FAN Wei1, SONG Jin-Bao1
|
|
1.Ocean college, Zhejiang University, Zhoushan 316021, China;2.Guangzhou Institute of Tropical and Marine Meteorology, China Meteorological Administration, Guangzhou 510640, China
|
| Abstract: |
| We conduct a 38-day at-sea wind measurement 6.5km off-coast on a tower from Feb.4 to Mar.12, 2015 near Maoming City, Guangdong, South China. The sea surface measurement data were processed using Empirical Mode Decomposition (EMD) and Wavelet Decomposition (WD) to estimate the cutoff time scale (CTS) in eddy covariance method. The results show that the values of CTSs estimated by the EMD and WD methods are about 40 s (the result of EMD is slightly smaller), which are much smaller than that of CTS in the traditional eddy covariance method (fixed at 10 min), and the use of EMD and WD makes each segment of data able to get a suitable CTS according to its own characteristics. The EMD and WD method could remove the non-turbulence part of the calculation results effectively, and in addition, the characterization of the flux transmission direction is more reasonable, which greatly improves the calculation accuracy of the flux. The mean deviation of the flux calculated by the new method reached as high as 45% from that by the traditional method. The advantages and disadvantages of EMD and WD are also compared, showing that EMD has higher autonomy than WD, while WD has higher separation of signals. |
| Key words: air-sea flux eddy covariance method cutoff time scale empirical mode decomposition wavelet decomposition |
