Hyperspectral imaging and high-resolution biogenic sedimentary proxies of the mid-elevation Lake Khorlakel (Western Caucasus)

Hyperspectral imaging of lake sediment cores offers a promising analytical method being both high-resolution and non-destructive. In this study, we investigate the potential of an industry-standard hyperspectral (VNIR) camera combined with a custom developed software package to provide indirect data on past changes in sediment accumulation conditions of a mid-elevation lake in the Caucasus Mountains. The software allows for automated core fragment stitching, crack removal and spectral analysis. We introduce the technique of wavelength correlation mapping as a tool to trace within-core variations of spectral properties (i.e. absorption features). The absorption feature around 670 nm wavelength marks a characteristic distinction between organic-rich and organic-poor sections of the studied sediment core. This part of the spectrum corresponds to absorption of electromagnetic radiation by chlorophyll a, a pigment ubiquitous in photosynthetic organisms. In this study, we use the techniques of relative absorption band depth (RABD) and area (RABA) with various spectral spans to calculate and compare the 670 nm derived spectral indices. The selected RABD_660;670 index serves as a high-resolution (229 μm/px) proxy for downcore variations of autochthonous organic matter and provides the reconstruction of lake productivity for the period 3000–1000 BP with subannual resolution. The results are supported with independent biogenic proxies – total organic carbon, loss-on-ignition and concentration of bromine in the sediment.