基于聚合经验模态分解方法的OJ 287射电流量变化周期分析

Analysis on periodic variations of the radio flux of OJ 287 with ensemble empirical mode decomposition

BL Lac天体OJ 287是一个重要的Blazar天体,光变具有准周期特点.本文从密歇根大学射电天文台数据库收集了OJ 287射电波段4.8,8.0和14.5GHz超过30年时间的观测数据.由于天文观测资料的复杂性,一些传统寻找周期的方法不是太理想.近几年迅速发展起来的聚合经验模态分解(EEMD)特别适合于具有非线性和非平稳动态变化特性的Blazar天体光变规律的研究.运用EEMD方法对OJ 287的3个射电波段流量分别进行多时间尺度分解,各获得6个代表不同时间尺度局部特征信息的本征模态函数分量和一个趋势项.

BL Lacertae object OJ 287 is one of the Blazars, which shows convincing evidence of periodic variations. We have collected the 4.8, 8.0 and 14.5 GHz data for OJ 287 from the University of Michigan Radio Observatory database over three decades. Due to the complexity of the variability data, some algorithms for period searching are not ideal yet. An alternative to traditional periodicity analysis is the ensemble empirical mode decomposition （EEMD）. This method can analyze the cyclic components of complicated nonlinear and non-stationary processes. Using the EEMD analysis, the light curve of OJ 287 at radio frequency can be resolved into six independent intrinsic mode functions that have different average periods and trends. We find possible periods of 18.9, I 1.9, 5.7 and 2.4 years in 4.8 GHz, 12.2,5.2 and 2.4 years in 8.0 GHz and 21.8, 12.0, 4.3 and 2.4 years in 14.5 GHz. The most common periods are 12.0 years and 2.4 years, which imply that the emissions in these bands may originate from the same radiative process. The results also confirm that a radio variability period of OJ 287 is in agreement with the optical variability period of about 12.0 years. The result shows that the components of different variability periods can be separated properly with the EEMD that is a nonlinear and non-stationary signal processing method.