Provenance information from recent and ancient sedimentary archives is obscured by several factors and for disentangling these intermingled signals, analysis by multiple methods is paramount. In sedimentary provenance analysis (SPA), single-grain methods determining mineralogy, chemical composition, or radiometric ages are of key importance but are mostly applied to sand-sized sediments or sedimentary rocks. Finer grained sediments or sedimentary rocks are usually analyzed by whole-rock geochemical means and seldom by single-grain methods. Considering the abundance of fine-grained sedimentary archives, a strong need for single-grain, multi-method analyses of silt-sized sediments is obvious. Thus, we propose a workflow that is optimized for sample throughput and correlative analysis of fine-grained sediments based on machine learning methods. The feasibility of the workflow is demonstrated by differentiating three Central European loess-paleosol-sequences. The increased sample throughput enables access to sedimentary archives at high spatial and/or temporal resolution, which will open up new research pathways in SPA of silt-sized sediments.

中文翻译：

---

从露头到光谱——应用于中欧黄土的粉粒沉积物模态矿物学的自动化方法

来自现代和古代沉积档案的来源信息因多种因素而变得模糊，为了解开这些混合信号，多种方法的分析至关重要。在沉积物来源分析 (SPA) 中，确定矿物学、化学成分或辐射年龄的单颗粒方法至关重要，但主要应用于沙粒大小的沉积物或沉积岩。较细颗粒的沉积物或沉积岩通常采用全岩地球化学方法进行分析，很少采用单颗粒方法。考虑到细粒沉积档案的丰富性，对粉粒沉积物进行单粒、多方法分析的强烈需求是显而易见的。因此，我们提出了一种基于机器学习方法针对样品吞吐量和细粒沉积物相关分析进行优化的工作流程。通过区分三个中欧黄土-古土壤序列证明了该工作流程的可行性。样品通量的增加使得能够以高空间和/或时间分辨率获取沉积档案，这将为淤泥尺寸沉积物的 SPA 开辟新的研究途径。