Quantification of Bioturbation in Hemipelagic Sediments via Thin-section Image Analysis

Traditional ichnofabric-descriptive (semiquantitative) schemes rely upon subjective estimates of the degree to which original sedimentary fabrics have been disrupted. This paper presents an objective method of quantifying bioturbation in hemipelagic sediments based on a thin-section image-analysis technique. In order to perform these analyses, optical or backscattered electron micrographs are processed in order to obtain black and white images. In these images black pixels are silt-size, larger grains and the clay-size sedimentary matrix are white pixels. Once the initial processing has been accomplished, an index, H, is defined to quantify the horizontal orientation of the larger grain size fraction of the sediment. H is low for bioturbated fabric and high for laminated sediments. In hemipelagic suspension deposits, H can be used to quantify bioturbation, assuming that no other sedimentary processes have disrupted the original depositional fabric. The quantitative bioturbation scheme described here is consistent with existing semiquantitative bioturbation schemes; in addition, it is objective and provides higher-resolution descriptions of biological processes that disrupted the sediment after deposition. Applying this quantitative scheme should refine future paleoecological reconstruction derived from lacustrine and marine successions. INTRODUCTION, AIM, AND RATIONALE The amount, type, and distribution of bioturbation provide information about depositional conditions. Because these observations are very useful tools that aid the reconstruction of conditions at and shortly after deposition (e.g., Kennett and Ingram 1995; Savrda 1995), numerous qualitative and semiquantitative classifications of ichnofabric have been proposed (e.g., Reineck and Singh 1975; Savrda et al. 1984; Droser and Bottjer 1986; Miller and Smail 1997). These schemes have been widely used in paleoceanographic (e.g., Savrda 1995) and paleoclimatic (e.g., Behl and Kennett 1996) studies and in research on relative sea-level change (e.g., Taylor and Gawthorpe 1993). Unfortunately, they depend mainly upon visual estimates of sedimentary fabric, and this subjectivity inevitably introduces error into paleogeographic reconstruction. In order to minimize this source of error, this paper presents a quantification method, which is independent of subjective analyses, to define the degree of bioturbation. The method described here builds upon recent developments in thin-section image analysis (Francus 1998; Francus and Karabanov 2000) that quickly and objectively quantify how much the coarse sediment fraction deviates from the horizontal. In fine-grained hemipelagic sediments this deviation is assumed to measure the effects of bioturbation. In order to generate a quantitative descriptive scheme of bioturbation, it is necessary to first describe the fabric of undisturbed hemipelagic sediments. With this in place the effects of bioturbation and other mechanisms of sediment disruption of the primary fabric can be assessed. Then, the validity of the H index and robustness of the image-analysis technique can be compared with measures obtained using existing semiquantification * Present address: Department of Geosciences, University of Massachusetts, Amherst, MA, 01003, U.S.A. schemes. Finally, H index analysis is applied to recent hemipelagic sediments from Lake Baikal, Siberia, Russia (Francus and Karabanov 2000) and Lake Vico, Latium, Italy (Leroy et al. 1996) to test the robustness of the technique and provide working examples of the method. BACKGROUND TO HEMIPELAGIC SEDIMENTATION Hemipelagic sediments are believed to accumulate from fine particles suspended in the water column, which settle to the sediment–water interface (e.g., Grimm et al. 1996). In the sediment this process is thought to cause platy clay and coarser-grained particles that have not flocculated to be oriented subparallel to the sediment–water interface, forming an oriented, primary plasmic fabric (e.g., Kuehl et al. 1988). In clay-rich sediments deposited by flocculation, however, the depositional fabric is commonly more random and less well oriented (e.g., O’Brien and Pietraszek-Mattner 1998). Surprisingly, to our best knowledge, no laboratory experiments have been done to reproduce the primary settling fabric of hemipelagic sediments to corroborate this assessment. Extensive microscopic observation and review of published micrographs of hemipelagic sediments (e.g., Krinsley et al. 1998), however, reveals that elongated particles are commonly oriented mostly horizontally. Where elongated particles are not horizontally oriented, they are usually contained within aggregates or pellets that sedimented either directly from the epilimnion or from redeposition (Chang and Grimm 1999). These aggregates and pellets are easy to identify by thin-sectioning the sediment and then using either conventional optical or backscattered electron (BSE) methods to image the sediment fabric (e.g., Brodie and Kemp 1995). The laminae that are characteristic of hemipelagic sediments are produced by variations in sediment input. In hemipelagic settings (in both marine and lacustrine settings) sediment is rarely supplied to the sediment– water interface as a slow continuous rain of debris. Instead, sediment-trap records unequivocally demonstrate that it is supplied episodically and/or seasonally and that these variations are linked to either variations in primary production or sediment flux (Grimm et al. 1996). In some settings the individual laminae are subtly graded. Where grading is present it is usually attributed to particle sorting in the water column (e.g., Stow and Bowen 1980; Lamoureux 1999). Subtly graded bedding, however, is not common, because clay minerals have settling velocities much lower than that of fine-grained quartz fraction (e.g. Doyle et al. 1983). In hemipelagic sediments, bioturbation is the dominant source of syndepositional and postdepositional disturbance. It disrupts the primary fabric, causing it to exhibit a range of textures from ‘‘slightly disturbed laminae’’ to ‘‘totally homogenized bedding of massive appearance’’ (Droser and Bottjer 1986). Segall and Kuehl (1994) have produced a multi-scale description of some bioturbated sediment and shown that ‘‘bioturbated sediments do not display grain orientation . . .’’. On the basis of 234Th mixing coefficients, Kuehl et al. (1988) reported that bioturbation-induced disruption of plasmic fabric suppresses the original preferred orientation. Finally, microscopic observations of bioturbated sediments (e.g., Brodie and Kemp 1995, fig. 8) reveal that the particles within them exhibit a random orientation. These observations suggest that it should be possible to quantify bioturbation by estimating the disruption of the original horizontal fabric, assuming that other mechanisms that disrupt primary fabric can be de-