利用连续小波变换识别大气温度斜坡结构

Using Wavelet Transform to Identify Atmospheric Temperature Ramp-Like Structure

将模拟的分形布朗运动信号叠加在理想的温度斜坡结构信号上,运用条件采样技术以及Donoho和Johnstone阈值方法成功地将斜坡(ramp)信号识别和分离.并将这一方法运用到实际大气温度测量数据中.分析表明,实际大气存在温度斜坡结构,周期约为30 s。在不同稳定度条件下其斜坡结构形状有差异.小波变换可有效地识别温度斜坡结构.

Simulating fractional Brownian motion signal is added to the ideal temperature ramp structure. Using the conditional sample technique, and Donoho and Johnostone threshold method, the ramp signal is successfully identified and separated. This method is applied to the actual atmospheric temperature data. The analysis shows that there are the temperature ramp structure in actual atmosphere, and the period is about 30 seconds. Under different stability condition, the form of ramp structure may be different. And the wavelet transform can identify the temperature ramp structure efficiently.