Collaborative Research: Bringing the Late Pleistocene into Focus: Better estimates of Ages and Ocean Circulation through Data-Model Comparison

对过去气候变化的测量对于理解自然气候变率的原因、幅度和速率至关重要。海洋沉积物中含有丰富的微化石，是近全球覆盖的过去气候信息的重要档案。为了将海洋沉积物样本置于环境中，研究人员需要两种信息：(1)对该样本所代表的气候的解释；(2)样本的年龄。年龄信息对于重建全球气候变化的速率和顺序至关重要。一些沉积物可以用放射性碳测定年代。然而，许多样品的年龄只能间接估计，导致年龄的不确定性高达4000年。该研究项目旨在通过对现有放射性碳定年样品的统计分析来改进这些间接年龄估计。间接年龄估计的改进将用于提高过去15万年全球气候变化数据集的准确性。更具体地说，这个项目的目的是改进从底栖三角洲18o（全球冰量和深海温度的代表）的地层排列中得出的间接年龄估计。这些年龄的一个主要不确定性来源是，在几个核心地点用非常精确的放射性碳模型观测到的底栖三角洲18o变化时间差异高达4000年。然而，对于底栖delta18O信号在其他位置的时间差异知之甚少。该项目的研究方法结合了海洋环流模型、古气候数据（来自全球分布的~100个沉积物岩心的底栖三角洲18o和14C）和统计推断。该项目的第一阶段是利用来自约100个沉积物岩心的放射性碳数据，结合海洋环流模型的结果，在全球范围内表征底栖三角洲18o年龄差异（三角洲18o滞后）。第二阶段是应用第一阶段的结果来开发改进的delta18O对准方法。第三阶段是生成一个改进的delta18O年龄模型的新数据库。因此，该项目的最终产品将是(1)描述底栖delta18O区域和全球变化模式的概率叠加（平均值），(2)多代理岩心比对和生成delta18O滞后贝叶斯推断的概率算法，(3)使用底栖delta18O数据的约300个岩心年龄模型数据库，以及(4)基于比较底栖del18O数据和不同环流模式情景的海洋环流变化估计。该项目对气候变化研究和教育具有更广泛的影响。准确的古气候年龄估算对于创建用于测试气候模式有效性和提高对其模拟未来潜在气候变化能力的信心的数据汇编至关重要。通过纳入具有较大不确定性的间接年龄估算数据，往往可以增加古气候数据汇编的空间覆盖范围。通过改进底栖三角洲18o序列的年龄估计和量化序列年龄的不确定性，该项目将允许更明智地选择哪些数据包括在汇编中，并总体上提高准确性。作为该项目的一部分进行的数据模型比较也将更好地限制海洋环流的变化，这有助于描述过去气候变化事件的因果链，并可能确定冰川时期海洋中隔离的碳库。该项目还将为两名研究生提供地球科学和数学方面的跨学科培训。通过开发和传播交互式计算机实验室活动，将研究方法和结果纳入本科和研究生课程。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Measurements of past climate changes are crucial for understanding the causes, amplitudes, and rates of natural climate variability. Ocean sediments contain abundant microfossils that are an important archive of past climate information with near-global coverage. To place an ocean sediment sample into context, researchers need two kinds of information: (1) an interpretation of the climate represented by that sample and (2) the age of the sample. Age information is crucial for reconstructing the rate and sequence of climate changes globally. Some sediments are well-dated using radiocarbon. However, the ages of many samples can only be estimated indirectly, resulting in age uncertainties as large as 4,000 years. This research project aims to improve these indirect age estimates using statistical analysis of the available radiocarbon-dated samples. Improvements in indirect age estimates will be used to increase the accuracy of global datasets of climate changes for the past 150,000 years.More specifically, the aim of this project is to improve indirect age estimates derived from the stratigraphic alignment of benthic delta18O (a proxy for global ice volume and deep ocean temperature). A major source of uncertainty in these ages derives from differences of up to 4,000 years in the timing of benthic delta18O change observed at a few core sites with very accurate radiocarbon models. However, little is known about how the timing of benthic delta18O signals differs at other locations. The research approach for this project incorporates ocean circulation models, paleoclimate data (benthic delta18O and 14C from ~100 globally distributed sediment cores), and statistical inference. Phase one of the project is to characterize benthic delta18O age differences (delta18O lags) globally using radiocarbon data from ~100 sediment cores in conjunction with results from ocean circulation models. Phase two is to apply the results from phase one to develop improved methods for delta18O alignment. Phase three is to generate a new database of improved delta18O age models. Thus, the final products of the project will be (1) probabilistic stacks (averages) describing regional and global patterns of benthic delta18O variability, (2) probabilistic algorithms for multiproxy core alignments and for generating Bayesian inferences of delta18O lags, (3) a database of age models for ~300 cores with benthic delta18O data, and (4) estimates of ocean circulation changes based on comparing benthic del18O data with different circulation model scenarios.This project has broader impacts for climate change research and education. Accurate paleoclimate age estimates are crucial for creating data compilations that are used to test the effectiveness of climate models and improve confidence in their ability to simulate potential future climate changes. The spatial coverage of paleoclimate data compilations is often increased by including data with indirect age estimates that have large uncertainties. By improving age estimates for benthic delta18O alignments and quantifying alignment age uncertainties, this project will allow for more informed selection of which data to include in compilations and overall better accuracy. Data-model comparison performed as part of this project will also better constrain ocean circulation changes, which could help describe the causal chain of events for past climate changes and may identify isolated reservoirs of carbon sequestered in the ocean during glaciations. This project will also provide interdisciplinary training in earth science and mathematics to two graduate students. Research methods and findings will be incorporated into undergraduate and graduate classes through the development and dissemination of interactive computer lab activities.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.