Interpreting detrital magnetic signals in speleothems: mechanisms, spatial heterogeneity, and methodological constraints

Abstract

Understanding the variability of Earth’s hydrological cycle is central to predicting the impacts of ongoing and future climate change. In particular, changes in precipitation patterns and the frequency and intensity of extreme hydrological events such as floods and droughts pose significant risks to ecosystems, infrastructure, and human societies worldwide. Instrumental records of rainfall and streamflow, however, are typically limited to the past century or less, constraining our ability to evaluate long-term variability and the full range of natural hydroclimate behavior. As a result, paleoclimate archives are essential for extending hydrological records beyond the instrumental period and for placing recent changes in a broader temporal context. Among terrestrial paleoclimate archives, speleothems (secondary cave carbonates such as stalagmites and stalactites) offer several unique advantages for reconstructing past hydroclimate. Speleothems can grow continuously over timescales ranging from decades to hundreds of thousands of years, often with well-preserved annual or sub-annual layering. Combined with precise geochronological constraints, such as U-Th dating, these characteristics enable the development of high-resolution, accurately dated records of environmental change. Furthermore, cave systems are generally protected from surface erosion and disturbance, allowing speleothems to preserve signals of past climate conditions with relatively high fidelity. A wide range of geochemical and physical proxies preserved in speleothems have been used to infer hydroclimate variability. Stable isotope ratios, trace element concentrations, and growth rate variations have all been interpreted in terms of changes in precipitation amount, moisture source, and recharge dynamics. In addition to these widely used proxies, detrital material incorporated into speleothems has emerged as a promising recorder of hydrological processes. In particular, the presence and properties of detrital particles can reflect processes such as infiltration through the epikarst, soil mobilization, and episodic flooding within cave systems. Despite this potential, the interpretation of detrital-based speleothem proxies remains fundamentally challenging. A central difficulty is that similar signals of detrital enrichment can arise from multiple transport and depositional mechanisms. For example, increases in detrital content may result from high-energy flood events that transport sediment into the cave, from enhanced dripwater flow that carries fine particles through the epikarst, or from aeolian processes that introduce dust into the cave environment. Each of these mechanisms responds differently to climate forcing and may encode distinct aspects of hydroclimate variability. As a result, without independent constraints on the origin of detrital material, interpretations of speleothem records can be ambiguous and, in some cases, contradictory. This challenge is particularly evident in studies of speleothem magnetism, where both positive and negative correlations between magnetic mineral concentration and precipitation have been observed in different cave systems, reflecting differing dominant transport processes. A second, closely related challenge arises from the spatial distribution of detrital material within individual growth layers. Most speleothem proxy measurements are obtained along one-dimensional transects, typically following the growth axis of the sample. These approaches implicitly assume that the composition of each lamina is laterally homogeneous and that measurements taken along a single transect are representative of the entire layer. However, detrital particles are often heterogeneously distributed at micrometer to millimeter scales, forming localized clusters or discontinuous features within individual laminae. In such cases, the measured proxy signal can depend strongly on the specific location of the sampling path. This phenomenon, often referred to as the “nugget effect,” can lead to substantial variability between co-located records and may obscure or distort the true temporal evolution of environmental signals. A third challenge is methodological. Because detrital magnetic signals in speleothems are often weak, spatially complex, and measured at micrometer scales, their interpretation is sensitive to contamination, detection bias, and measurement artifacts. Magnetic microscopy methods can resolve individual grains and localized magnetic features, but this increased resolution also introduces new uncertainties related to sample preparation, source detectability, and the conversion of two-dimensional magnetic field maps into interpretable environmental signals. As a result, improving the reliability of speleothem detrital proxies requires not only identification of signal origin and spatial structure, but also explicit attention to the methodological constraints of high-resolution magnetic measurements. Together these challenges represent a fundamental limitation in the use of speleothem detrital proxies for hydroclimate reconstruction. Addressing these limitations requires approaches that can both identify the origin of detrital material and characterize its spatial distribution within speleothems at high resolution. Traditional bulk measurements and one-dimensional analytical techniques are not well suited to this task, as they integrate signals over relatively large volumes and cannot resolve the grain-scale or two-dimensional structure of detrital phases. Recent advances in magnetic microscopy, particularly the development of the quantum diamond microscope (QDM), provide a powerful new tool for overcoming these limitations. QDM enables the imaging of magnetic fields at micrometer-scale spatial resolution, allowing for the detection and characterization of individual ferromagnetic particles within geological samples. By measuring the magnetic moments of individual grains, QDM provides a means to estimate grain size distributions and to distinguish between different populations of magnetic particles. At the same time, its two-dimensional imaging capability allows for direct visualization of the spatial distribution of detrital material within speleothem laminae. These combined capabilities make QDM uniquely suited to addressing both the mechanistic and spatial challenges associated with detrital proxy interpretation. In this dissertation, I apply quantum diamond microscopy to investigate the origin, distribution, and paleoenvironmental significance of detrital magnetic particles in speleothems. The overarching goal of this work is to improve the reliability of speleothem-based hydroclimate reconstructions by explicitly accounting for both the processes that deliver detrital material to the cave and the spatial structure of that material within growth layers. This work is organized around three primary objectives: (1) to develop methods for distinguishing among different mechanisms of detrital transport using grain-scale magnetic properties, and (2) to evaluate the impact of spatial heterogeneity on the interpretation of speleothem proxy records, and (3) to establish a methodological framework for identifying contamination, detection bias, and measurement artifacts in QDM-based studies of speleothems. Chapter 1 focuses on the identification of flood-derived detrital layers in a speleothem from Malfazido Cave in southern Brazil. In this study, I use QDM to map the magnetic field of the speleothem at micrometer resolution and to extract the magnetic moments of individual grains. By analyzing the distribution of magnetic dipole moments, I demonstrate that discrete, highly magnetic laminae are characterized by a population of coarser ferromagnetic particles that are distinct from the finer-grained background population. Comparison with magnetic measurements of cave mud deposited during modern flood events shows that these coarse-grained populations closely match those found in flood-derived sediments. Combined with optical petrography and correlations with instrumental rainfall and streamflow records, these results provide strong evidence that these laminae represent individual flood events. This work establishes grain-scale magnetic characterization as a robust method for distinguishing flood-derived detrital material from other sources and demonstrates the potential of speleothems to record flood frequency at high temporal resolution. Chapter 2 examines the complementary issue of spatial heterogeneity within speleothem laminae. Using QDM imaging of two co-located stalagmite cores from Sitting Bull Falls Cave in New Mexico, I investigate the two-dimensional distribution of ferromagnetic material and its relationship to previously measured trace element records. The results reveal multiple classes of magnetic enrichment, including laterally continuous features associated with lamina-scale signals and spatially localized, high-intensity sources that cross-cut multiple growth layers. Through resampling of one-dimensional transects from the two-dimensional magnetic maps, I show that peak detection in proxy records is highly sensitive to sampling location. In many cases, prominent features observed in one transect are absent in others, and the variability introduced by this sampling effect is comparable to the differences observed between co-located speleothem records. These findings demonstrate that discrepancies between records need not reflect differences in hydrological forcing or flow paths, but can instead arise from the heterogeneous distribution of detrital material within individual laminae. Chapter 3 addresses the methodological constraints associated with applying the QDM to speleothem magnetic records. It examines how contamination introduced during sample preparation and handling, detection bias toward larger and near-surface grains, and instrumental artifacts in QDM measurements can influence the interpretation of grain-scale magnetic data. This chapter develops a practical framework for distinguishing environmentally meaningful magnetic signals from non-representative material, and for identifying the sample- and instrument-specific conditions under which QDM measurements can be interpreted reliably. In doing so, it places the applications in Chapters 1 and 2 within a broader methodological context and clarifies the limits and requirements of magnetic microscopy-based speleothem proxy development. Taken together, these three chapters provide a framework for improving the interpretation of speleothem detrital proxies by integrating grain-scale characterization with two-dimensional spatial analysis. Chapter 1 demonstrates that magnetic granulometry can be used to identify the mechanisms of detrital transport, enabling more confident attribution of proxy signals to specific hydrological processes such as flooding or dripwater infiltration. Chapter 2 shows that the spatial distribution of detrital material plays a critical role in determining how these signals are recorded in one-dimensional measurements. Chapter 3 demonstrates that reliable interpretation also depends on careful attention to contamination, detection limits, and instrument-specific artifacts. More broadly, this dissertation illustrates the value of moving beyond traditional one-dimensional approaches to speleothem analysis and incorporating multidimensional datasets that capture both the physical properties and spatial structure of detrital signals. By resolving both the origin and spatial structure of detrital material, quantum diamond microscopy provides a new pathway for extracting more reliable and nuanced information about past hydroclimate variability from speleothems. These advances have important implications not only for the study of floods and extreme events, but also for a wide range of speleothem-based proxies that may be influenced by detrital inputs. Ultimately, improving the reliability of paleoclimate reconstructions requires a detailed understanding of the processes that generate and modify proxy signals. By combining high-resolution magnetic imaging with sedimentological and geochemical analysis, this dissertation contributes to a more process-based interpretation of speleothem records and provides new tools for reconstructing hydroclimate variability across a range of temporal and environmental settings.