利用 Globorotaloides hexagonus 重建上新世-更新世東太平洋的最低含氧區

Abstract

先前的研究指出，浮游有孔蟲Globorotaloides hexagonus 喜好生存在缺氧水域中，顯示其作為最低含氧區（Oxygen Minimum Zones, OMZ）研究指標的潛力。然而，由於 G. hexagonus 的運用在古海洋的重建使用相對有限，其與OMZ之間的生地化關係仍須更全面了解。本研究使用取自東赤道太平地區的海洋鑽探計畫 (Ocean Drilling Program, ODP) 1241站位，經由分析樣本中浮游性有孔蟲G. hexagonus的殼體孔隙度變化及其殼體微量元素成分（例如：錳鈣比氧化還原環境的輔助指標），以探討其與OMZ之間的關係，並探討上新世OMZ的演化歷程。研究中鎖定上新世晚期/更新世早期海洋同位素階段 (Marine Isotope Stage, MIS) 96-100（2.55-2.4 百萬年前）和上新世中期MIS M2（3.35-3.25 百萬年前）兩個時期的樣本。首先，我們計算ODP第1241站點 樣本中G. hexagonus 的相對豐度，藉此確定其豐度與OMZ之間是否存在關聯。隨後，我們統計篩選出具有高、低相對豐度的樣本各四個，分析G. hexagonus的、殼體孔隙度（porosity），並運用掃描電子顯微鏡（SEM）拍攝殼體影像，然後使用於前人發表已經訓練的深度學習模型重新訓練，並自動量化孔隙度參數，以克服傳統影像識別耗時及可能破壞樣本的問題。此外，本研究也使用ICP-OES測定殼體錳鈣比及殼層微量元素（錳鈣比）變化，作為氧化還原環境的輔助指標。結果顯示高、低豐度組間的殼體總孔隙度存在顯著差異（p < 0.05），且主要與孔隙的尺寸變化有關，而非孔隙數量，此現象支持此時期的G. hexagonus如同現代樣本可透過調節孔隙度以適應低氧環境的假設。而G. hexagonus相對豐度與孔隙度的資料變化，暗示MIS96-100期間OMZ的厚度與強度發生變化，且此變化與冰期-間冰期循環無關，可能是受而受到與歲差相關的日照變化所影響。另外，研究也發現殼體的錳鈣比值與豐度及孔隙度變化趨勢不一致，顯示其受陸源輸入影響較大，難以直接反映OMZ強度。而MIS M2的樣本資料在豐度及微量元素變化上並不明顯，表明此一暖期時，OMZ呈現相對穩定的狀態。本研究不僅證實了G. hexagonus的豐度與殼體孔隙度可作為重建上新世OMZ變化的有效指標，並揭示其殼體結構對低氧環境的適應機制。未來建議可進一步擴展樣本涵蓋的年代範圍，以連接不同時間區段，來驗證驅動OMZ變化的主要因素。

Exploring the development of Oxygen Minimum Zones (OMZs) during the Plio-Pleistocene (5.3-2.6 Ma) can provide valuable insights into the effects of ongoing climate change. Previous studies suggest that the planktic foraminifer Globorotaloides hexagonus thrives in oxygen-deficient waters, indicating its potential as an indicator for OMZ studies. However, because G. hexagonus has not been used very often in paleo-reconstructions, the biogeochemical relationship between G. hexagonus and OMZs has not yet been fully explored. This study examines three proxies derived from G. hexagonus at Ocean Drilling Program (ODP) Site 1241 in the Eastern Equatorial Pacific, which hosted a significant OMZ during the Pliocene. First focusing on Marine Isotope Stages (MIS) 96–100 (~2.5 Ma) in the Plio-Pleistocene transition, we analyzed G. hexagonus abundance, porosity variations of its test, and test trace metal composition (e.g., redox proxy: Mn/Ca) to explore its relationship with OMZ dynamics. In previous research, modern G. hexagonus also displays porosity variations under different dissolved oxygen levels in the water column. However, detecting pore parameters using image recognition methods is often time-consuming and can be destructive, limiting the efficiency of pore pattern analysis. To implement a more effective and non-destructive approach, we utilized a deep learning image recognition technique to examine G. hexagonus pore patterns. First, we quantified the abundance of G. hexagonus at ODP Site 1241 in the Eastern Equatorial Pacific to determine if a relationship exists between G. hexagonus abundance and oxygen concentrations. We then selected four samples with high or low G. hexagonus abundance and captured their images using scanning electron microscopy (SEM). Building on a previously trained deep learning algorithm for benthic foraminifera, we retrained the model specifically for G. hexagonus SEM images to obtain detailed pore parameter data. The results reveal significant variations in G. hexagonus porosity between high- and low-abundance samples, suggesting that G. hexagonus may adapt to low-oxygen conditions by increasing porosity. Notably, these porosity changes are primarily driven by alterations in pore size rather than the number of pores. The variations in the abundance of G. hexagonus along with the porosity results suggest the OMZ varied during MIS96-100, and could have indicated that the OMZ changed in thickness and intensity independently of glacial-interglacial cycles. Instead, a potential precession-related variability in insolation might dominate. On the other hand, Mn/Ca in the tests, measured with ICP-OES, does not correspond with abundance or porosity data but shows potential influence from the 41-kyr obliquity cycle. We suggest that Mn/Ca in G. hexagonus may reflect terrestrial input in the region rather than the OMZ intensity, and thus may not be a direct indicator for OMZ variation in the paleo-reconstruction. Furthermore, we also look into the interval around MIS M2 (covering 3.35–3.25 Ma, including MIS KM5c). Our data show less variability in abundance and trace metals proxies, suggesting that the OMZ was more stable in the warm Pliocene. Although the current results cover only the MIS96-100 and MIS M2, they provide insight into how each proxy derived from G. hexagonus can be employed for OMZ paleo-studies. Future work will quantify the abundance and trace metals of G. hexagonus across a wider age range throughout the Pliocene (connecting the two studied intervals) to determine the primary drivers of the observed variations.