增進海底電磁儀資料與分析方法並建構西太平洋上部地函電導率構造

Abstract

地層中各岩體不同的導電率（electrical conductivity）可以反映不同物性，如溫度越高或含水量越多的物質組成（例如：部分熔融、熱液）其導電率就越高，進而可了解地層下方的電性構造。而海域的電性構造可透過佈放在海底的海底電磁觀測儀（Ocean Bottom Electro-Magnetometer，簡稱OBEM），觀測儀器下方與周遭的綜合電磁場訊號，並以大地電磁法（magnetotelluric，簡稱MT）逆推獲得地底下導電率隨深度的變化情形。本研究對西太平洋海底電磁場數據進行分析，資料來自27個站點共計46台OBEM，主要目的是增進對西菲律賓海盆（West Philippine Basin，WPB）和四國-帕里西維海盆（Shikoku-Parece Vela Basin，SPVB）底下導電率結構的理解。本研究通過結合Chave and Thomson（2004）提出的standard 和generalized兩種遠程參考法（remote reference method，RR），成功降低了大地電磁響應（MT response）的誤差並提高觀測與預測電場之間的相干性（coherence），其中generalized RR在較短週期內表現出色，得到誤差更小且相干性更高的大地電磁響應。而部分站點觀測紀錄長達三年，更使得大地電磁響應誤差減少約40%，凸顯了在海洋環境中進行長期數據採集的重要性。所記錄的電磁響應包含三維地形影響，在利用電磁響應逆推一維地層導電率前須考量三維地形效應進行地形修正，本研究針對每個測站的初次電磁響應逆推其一維地下電性構造做為初始模型，結合三維地形順推計算該測站理論電磁響應與觀測電磁響應比較後重複迭代，最後收斂求得代表一維地下構造導電率。經過考慮地形效應後，我們首次觀察到使用最終一維地下電性構造結合三維地形順推求得的磁場感應向量（induction vectors）與觀測資料之磁場感應向量非常吻合。由於構造逆推過程中並沒有納入磁場感應向量擬合，表明此研究中在電性構造與電磁響應之逆推結果一致。顯示三維地形修正的重要性外，也代表各測站並無明顯二維電性構造的特徵，所求得的一維地下構造導電率可代表各測站下方電性構造進行討論。本研究透過分析各測站的最終一維電性構造，發現在WPB的T01、T02和T04站點下方30至125公里深處存在顯著的高導電率異常。我們將電性構造轉換為溫度，並與地函橄欖岩固相線對比，我們發現如果地函含水量介於0至0.02 wt%之間，上述區域會發生部分熔融，此結果與前人地震波速慢異常觀測結果相吻合。根據每測站所推得的熱構造，本研究所有測站在150至300公里深處區間，沒有發現部分熔融的可能性。此外，我們從一維電性構造變化情形估算低導電層（lower conductivity layer，LCL）的厚度，在WPB和SPVB站點下方的LCL厚度相似，約為50至70公里，而在太平洋板塊西緣下方，LCL則明顯更厚，約為210公里。我們注意到WPB 和 SPVB 的 LCL厚度都比1,300°C 等溫線深度淺，而該等溫線通常定義為岩石圈-軟流圈熱邊界層。這種差異的一種可能解釋是部分熔融的增加會改變導電率曲線。隨著岩石圈底部和軟流圈頂部導電率的增加，LCL可能會變得更薄。本研究重新分析海底電磁儀資料，經過地形修正得到各測站下方一維電性構造，進一步了解岩石圈與軟流圈的物理特性，推估西太平洋存在部分熔融的區域與深度。Characterizing the physical properties of the lithosphere-asthenosphere system (LAS) beneath the oceans, such as seismic velocity and electrical conductivity, is critical for understanding its complex nature. In this thesis, electromagnetic observations from 46 Ocean Bottom Electro-Magnetometers (OBEMs) across 27 sites in the Philippine Sea and the western Pacific Ocean were reanalyzed, some of which offer up to 3 years of recovery data. This expanded on previous one-year data analyses by utilizing all available data and combining both standard and generalized remote reference techniques (Chave& Thomson, 2004) with land geomagnetic stations as reference sites. Our improvements to seafloor magnetotelluric (MT) responses yielded significantly reduced error bars across the entire period band and increased coherence between observed and predicted electric fields. For some available sites, three years of continuous data allow us to further reduce response errors by approximately 40%, highlighting the value of long-term data acquisition. This study investigates the influence of three-dimensional (3D) topography on MT responses and subsurface conductivity structure. We conducted 1D conductivity inversions for each site, incorporating the 3D topographic heterogeneity overlying the 1D mantle model. To minimize the root-mean-square misfit between observed and predicted MT responses, we iteratively refined the 1D inversions on topographically corrected responses. After accounting for topographic effects in MT responses, we noticed for the first time that the induction vector calculated using the final 1D model and 3D topography closely matched the induction vector calculated by observed magnetic data. Since we didn’t incorporate the induction vectors in inversion analysis, this suggests that non-1D features of MT and induction vectors are roughly explained by considering topography effect. This underscores the importance of topographic corrections in marine EM studies and the resulting 1D subsurface structure effectively represents the true electrical properties beneath each measurement station.Our 1D conductivity models revealed a high conductivity anomaly at depths of 30 to 125 km beneath sites T01, T02, and T04 in the western West Philippine Basin (WPB). By comparing the thermal structure derived from these conductivity profiles with the mantle peridotite solidus, we inferred the need for partial melting at these depths, assuming a water content range of 0 to 0.02 wt. This finding is consistent with previous seismic observations of slow-velocity anomalies. Additionally, our 1D conductivity models allowed us to estimate the thickness of the lower conductivity layer (LCL). The LCL beneath WPB and SPVB is approximately 50 to 70 km thick, while beneath the Pacific Plate, it reaches around 210 km. Interestingly, the LCL thickness in WPB and SPVB is shallower than the 1,300°C isotherm depth, typically associated with the thermal lithosphere-asthenosphere boundary. Increased partial melting in the top of asthenosphere may have altered the conductivity curve, resulting in a thinner LCL.Overall, by reanalyzing seafloor electromagnetic data and applying topographic corrections, we derived 1D electrical conductivity profiles beneath each station. This allowed us to gain insights into the physical properties of the LAS and estimate the extent and depth of partial melting in the western Pacific.