Image analysis as a method to quantify sediment components

Varved lake sediments, with their annual to seasonal resolution, have a high potential for inferring past environmental and climatic conditions. To fully utilize the information present in varved records, high-resolution analyses, which often are time-consuming and difficult to perform, are desirable. The investigation reported here aims at (i) developing image analysis as a method for estimating annual accumulation rates of sediment components such as minerogenic matter, organic matter and biogenic silica, and (ii) assessing the relative importance of these components for changes in varve thickness. Image analysis was used to digitize the grey-scale variations and to measure the varve thickness of 540 varves (476-1015 AD) from Lake Kassjön in northern Sweden. From the 35 cm long digitized sediment sequence, 108 consecutive five-year samples were cut out quantitatively, and relationships between grey-scale variations and sediment dry mass and individual sediment components were assessed. There is a strong correlation between corrected grey-scale (i.e. the product of grey-scale and varve thickness) and the dry mass accumulation rate (r = 0.90, p < 0.001). With a stepwise multiple regression a significant model (R² = 0.81) between corrected grey-scale and the accumulation rates of minerogenic matter (r = 0.90, p < 0.001) and biogenic silica (r 0.26, p < 0.012) was obtained. Considering the minor contribution and weak significance of biogenic silica, image analysis can be used as a fast and non-destructive method to infer past annual accumulation rates of dry mass and minerogenic matter in Kassibn. The model of the relationship between changes in varve thickness, and water content and accumulation rates of sediment components has little predictive power (R² = 0.45). The result shows that the varve thickness in Kassjön, at least during the period 476-1015 AD, is not determined by a single sediment component but partly depends on interactions between the major sediment components.