地震造成生命以及人民財產的損失相當大，所以關於如何預測地震成為一項重要的課題。本研究主要是想探討地震與潮汐行為之間的互動，更加瞭解地震觸發的機制，希望可以進一步地提供地震預測的線索。潮汐效應對地震活動影響的研究已經有100多年的歷史，但它長期以來仍然倍受爭議。在本研究中，我們考慮了台灣、夏威夷和南加州分別從1973年初記錄到2008年末的地震目錄。我們使用經驗模態分解法(Empirical Mode Decomposition, EMD)分解非線性和非穩態的地震資料成數個分量，而這些分量可能會包含物理意義，這樣的特性方便我們找出可能的觸發來源。因此我們對拆解出來的分量做傅立葉轉換和希爾伯特轉換得到頻譜，從這些頻譜去詳細研究時間序例的特徵以求發現隱藏在數據背後的潮汐效應。之後為了提高訊號雜音比，我們還根據理論地球固體潮的相同位特性來疊加 每日地震數目的時間序列。希望在疊加後的頻譜裡，原本較弱的潮汐效應訊號會變得更加明顯。接著，我們比較未去餘震和2種已去餘震的地震目錄的結果差異，發現餘震效應並不會對本研究提出的時間序列分析方法有太大的影響。為了確定所得到的結果並非隨機，所以我們加入顯著性分析，將地震目錄的時間隨機打亂，再重新分析，結果得知這3個研究區域的信賴水準都在80%以上。綜合以上的研究結果，我們的研究顯示潮汐觸發效應會隱藏在地震發生時間的資料裡而且可以透過經驗模態分解法解析出來。也就是說，我們的結果指出月潮汐力可能是觸發地震的因素之一。The tidal triggering effect on earthquake occurrence has been controversial for a long time. In this work, three catalogues of earthquakes occurring from 1973 to 2008 in Taiwan, Hawaii and Southern California regions are considered. Because earthquake data are nonlinear and/or non-stationary, we propose an empirical mode decomposition (EMD)-based approach to analyze the temporal variations in earthquakes for investigating tidal triggering effect. The original signal can be extracted into many components through EMD-based approach, which may contain different physical meanings. We then utilize spectral analysis on these components to study the tidal triggering effect. In order to increase signal-to-noise ratio, we apply stacking method to earthquake count time series based on inphase phenomena of theoretical Earth tides. We also compare non-declustered catalogue with two kinds of declustered catalogues. Finally, statistically significant tests are taken into account. In summary, our studies show that the tidal effect, particularly, the lunar tidal effect, is hidden in the raw earthquake occurrence data and can be extracted through the proposed approach. Namely, our results indicate that the lunar tidal force is likely to be one of earthquake triggering mechanisms.
