Testing assumptions of spheroidal carbonaceous particle (SCP) quantification at very low concentrations: Implications for dating geologic archives

Abstract

Spheroidal carbonaceous particles (SCPs) are preserved in geologic archives such as lake sediments, ice cores, or peatlands, and can provide chronology for cores. SCPs are only produced during the combustion of coal and fuel oil, so they are a reliable global marker of the onset of industrialization and can be used to track deposition from these sources. While the usage of SCPs as chronostratigraphic indicators in recent sediments is common, the enumerative method of quantifying their sedimentary concentrations has remained virtually unchanged, and assumes that one subsample is representative of the entire sediment sample. We test this assumption and explore its implications for the SCP chronological method of dating recent sediments by analyzing multiple subsamples to characterize the precision and accuracy of SCP concentration measurements. Notably, we do not use the conventional SCP quantification method and focus on samples with lower concentrations and larger SCPs than are typically quantified in the literature. However, we base our conclusions and inferences on insights gained from analyses of the effects of subsampling on SCP numerosity (counts), which are translatable insights to all sizes and methods of SCPs quantification which rely on particle counts. We quantified SCPs in sets of 30 subsamples for 14 riverine sites (n = 420). SCP concentrations varied (0 SCPs/gDM – 2141±825 SCPs/gDM), but reflect the typical ranges of SCP concentrations quantified in modern sediments in other environmental settings. For each site, we used a bootstrapping method to approximate the theoretical mean of SCPs at 1–30 subsample sizes, then compared the theoretical mean and relative standard deviation. We found that enumerating 10 subsamples per sample better represents the theoretical mean of SCPs than the enumeration of 1 subsample, especially for lower SCP concentration samples. The greatest chance for falsely reporting the absence of SCPs was when <10 SCPs/gDM were measured in fewer than 10 subsamples, indicating that more replicates could provide greater confidence in SCP-based dating of geologic archives. If aiming to delineate the stratigraphic onset of SCP presence for dating purposes, we recommend enumerating a minimum of 10 subsamples for samples with low SCP concentrations to ensure the reliability of these measurements. We acknowledge that enumerating multiple subsamples is time and resource intensive, but provide some strategies (e.g., limiting subsampling) for minimizing additional cost and argue that the advantages afforded for dating reliability outweigh the costs in paleoenvironmental research.