Supporting information for "Global and regional Pleistocene benthic δ18O stacks on age models with and without astronomical tuning"

Constructing accurate age models for Pleistocene marine sediments is crucial for our understanding of glacial-interglacial cycles and other climatic processes. Benthic foraminiferal δ18O stacks, a proxy for ice sheet and climate evolution, are often used for stratigraphic alignment and chronology development in deep-sea sedimentary records, in combination with biostratigraphy, paleomagnetism, and radioisotopic constraints. Selection of an appropriate benthic δ18O alignment target influences the derived chronology at a given site, and divergent regional trends in benthic δ18O highlight the need for ocean-specific benthic δ18O stacks. The specific scientific question to be addressed by a study may also influence whether the alignment target should include astronomical tuning. Here, we introduce three benthic δ18O stacks – Atlantic, Pacific, and global – with three distinct chronologies for the global stack that incorporate astronomical forcing constraints to various degrees. The new global stack utilizes data from 224 cores and includes 50% more data than the previous “ProbStack” (Ahn et al., 2017). Hand-tuned regional and global stacks, intended as updates to the “LR04” stack (Lisiecki and Raymo, 2005), incorporate chronologies transferred from absolutely dated archives during 0-654 thousand years ago (ka) and an astronomically forced ice sheet model during 654-2700 ka. Due to the heterogeneous nature of the age constraints used for these stacks, we call them BIGSTACKmixed, BIGSTACKmixedA, and BIGSTACKmixedP. For applications where astronomical tuning should be minimized, we present a global stack primarily constrained by geomagnetic reversal age estimates, BIGSTACKmagrev. We also develop a third age model, BIGSTACKauto, which uses an automated optimization algorithm to “minimally tune” the stack to the pervasive ~41 kyr obliquity cycle, while avoiding assumptions about astronomical phase relationships. This suite of stacks offers flexibility in choosing δ18O stratigraphic alignment targets, to allow a wide range of applications in paleoceanographic hypothesis testing.构建精准的更新世（Pleistocene）海相沉积物年代模型，对于理解冰期-间冰期旋回及其他气候过程至关重要。底栖有孔虫δ¹⁸O堆叠序列（benthic foraminiferal δ¹⁸O stacks）作为冰盖与气候演化的代用指标，常与生物地层学、古地磁学以及放射性同位素约束方法结合，用于深海沉积记录的地层对比与年代学构建。选取合适的底栖δ¹⁸O对比靶标会影响特定站位的衍生年代序列，而底栖δ¹⁸O存在显著的区域分异趋势，这凸显了构建专属大洋的底栖δ¹⁸O堆叠序列的必要性。研究拟解答的具体科学问题，同样会影响地层对比靶标是否需要纳入天文调谐环节。本文介绍三套底栖δ¹⁸O堆叠序列——大西洋型、太平洋型与全球型；其中全球型序列包含三套差异显著的年代模型，三者对天文强迫约束的纳入程度各不相同。这套全新的全球型序列整合了224个岩芯的数据，相较于此前的"ProbStack"（Ahn等，2017），数据量提升了50%。作为"LR04"堆叠序列（Lisiecki与Raymo，2005）的更新版本，手动调谐得到的区域与全球型序列，在0~654千年前（ka）阶段采用了来自绝对定年档案的传递年代标尺，而在654千~2700千年前阶段则引入了受天文强迫的冰盖模型年代框架。鉴于这些序列所采用的年代约束具有非均一性，我们将其分别命名为BIGSTACKmixed、BIGSTACKmixedA与BIGSTACKmixedP。针对需要尽量减少天文调谐的应用场景，本文还提供了一套主要基于地磁反转年龄估算约束的全球型序列BIGSTACKmagrev。此外，我们还构建了第三套年代模型BIGSTACKauto，该模型通过自动化优化算法将序列"最小限度调谐"至普遍存在的~41千年（kyr）倾角周期，同时规避了关于天文相位关系的假设。这套堆叠序列为δ¹⁸O地层对比靶标的选取提供了灵活性，可广泛应用于古海洋学的假说验证研究中。