不同数字滤波方法在风暴轴研究中的性能比较

Performance of Different Digital Filters in Storm Track Study

比较了Butterworth带通滤波器和不同截断方式下的对称滤波器在风暴轴研究中对天气尺度瞬变波动的滤波性能。理论分析表明:当n≥9时,在天气尺度为2.5～6.0d的定义下,对称滤波器对天气尺度瞬变波动的滤波性能优于Butterworth带通滤波器。针对逐日观测资料,对称滤波器在研究天气尺度瞬变波动问题时取截断参数n=15相对最优。在此截断参数条件下,正弦截断对称滤波器的滤波性能优于梯形截断对称滤波器,且两者均优于Butterworth带通滤波器,这种优势在风暴轴偏强的厄尔尼诺年体现得更为明显。在样本序列较短的情况下,为了得到天气尺度瞬变波动,利用Butterworth带通滤波器并选取带通滤波参数为2.5～8.0d的做法是可取的。

The performance of Butterworth filter （BF） and symmetrical filter with different truncation modes in filtering the synoptic-scale（2.5-6 d） transient wave is compared in storm track study. Result reveals that, when the cutoff parameter n is greater than and equal to 9, the symmetrical filter can better reserve the synoptic-scale signal and suppress the non-synoptic scale signal than BF does. In the study of storm track with day-to-day observations, the symmetrical filter is comparative well-performed when the cutoff parameter equals to 15. The performance of symmetrical filter using sinusoidal cutoff is better than that using trapeziform cutoff in the condition presented above; meanwhile, their performance is both superior to that of BF, especially in El Nino years with stronger storm tracks. If the size of samples is not large enough,it is worthy to use BF to get the transient wave with the band-pass of 2.5-8 d.