Separation of Climate Signals

Kamath, C ; Fodor, I

Lawrence Livermore National Laboratory

关键词: Volcanoes;    Ambient Temperature;    Southern Oscillation;    Climates;    99 General And Miscellaneous//Mathematics, Computing, And Information Science;   

DOI  :  10.2172/15002083
RP-ID  :  UCRL-ID-150775
RP-ID  :  W-7405-ENG-48
RP-ID  :  15002083

美国|英语

来源: UNT Digital Library

【 摘 要 】

Understanding changes in global climate is a challenging scientific problem. Simulated and observed data include signals from many sources, and untangling their respective effects is difficult. In order to make meaningful comparisons between different models, and to understand human effects on global climate, we need to isolate the effects of different sources. Recent eruptions of the El Chichon and Mt. Pinatubo volcanoes coincided with large El Nino and Southern Oscillation (ENSO) events, which complicates the separation of their contributions on global temperatures. Current approaches for separating volcano and ENSO signals in global mean data involve parametric models and iterative techniques [3]. We investigate alternative methods based on principal component analysis (PCA) [2] and independent component analysis (ICA) [1]. Our goal is to determine if such techniques can automatically identify the signals corresponding to the different sources, without relying on parametric models.