Samples were routinely analysed for wet density (WD), percentage dry weight (%DW) at 105 °C, and loss-on-ignition analysis (LOI) at 550 °C, and data used to calculate diatom accumulation rates (DAR) in the sediment core. LOI is an approximation of organic content in sediment cores (Bengtsson and Enell, 1986). Each sample was prepared for diatom analysis using the procedure outlined in Mackay et al. (1998), minimizing additional dissolution of valves by omitting chemical pretreatments during preparation. Divinylbenzene microspheres (mean diameter 6.4 μm) were added to enable diatom concentrations and accumulation rates to be calculated (Battarbee and Kneen, 1982). Suspensions were settled on glass coverslips and permanent slides made using Naphrax slide mountant (refractive index 1.73). Diatoms were counted using oil immersion phase contrast light microscopy at ×1000 magnification.
Fig. 3 is the diatom stratigraphy of dominant phytoplankton taxa for BAIK38 expressed as relative percentages, plotted against the age scale. Zone 1 (c. 880 AD–c. 1180 AD) is dominated throughout by the autumnal blooming species Cyclotella minuta, while Aulacoseira baicalensis and Synedra acus (both of which bloom in spring) are also present in lower but similar proportions (c. 15%). Zone 2 (c. 1180 AD–1840 AD) is characterised by an increase in C. minuta values in excess of 80% relative abundance, which are sustained virtually throughout the zone. During this time, other taxa are present at only very low abundances, while some (e.g., Stephanodiscus meyerii and S. acus) are frequently absent. Zone 3 (c. 1840 AD–1994 AD) is characterised firstly by a decline in relative abundance of C. minuta to its lowest levels in the profile, up until c. 1950 AD. This decline is accompanied by concomitant increases in A. baicalensis and, to a lesser extent, Aulacoseira skvortzowii and S. meyerii.
Diatom-inferred snow depth reconstructions for BAIK38 using uncorrected taxa (Fig. 5a–c) show similar trends throughout the study period, with all or only five taxa in the model; snow depth levels are marginally higher in zone 2 in comparison to zones 1 and 3. However, error values are large in comparison to the changes observed. The snow depth reconstruction using corrected diatom abundances (Fig. 5d) shows a somewhat different response. Low values characterise the period coincident with the MWP (between c. 880 AD and c. 1180 AD), which increase into the LIA, reaching peak depths between 1500 and 1775 AD. After then, snow depth values decline to their lowest values in this study by c. 1900 AD. In recent decades, snow depth values appear to increase slightly again up to the top of the core dated at 1994.
Preservation differences can be used as correction factors to recalculate the relative abundances of each of the five dominant plankton taxa in BAIK38 and are depicted in Fig. 4. The resulting profile shows that Synedra acus is now the dominant taxa in zone 1 of the core, with other taxa being present at abundances generally less than 10%. At the zone 1/2 boundary, S. acus declines and is replaced by Cyclotella minuta and, to a lesser extent, Aulacoseira skvortzowii and Aulacoseira baicalensis. This profile is different from the relative abundance profile in Fig. 3, as S. acus values decline to very low values by c. 1400 AD, and C. minuta increases to peak values between c. 1525 and 1650 AD. Furthermore, the profile indicates that A. baicalensis remains common throughout this zone. Towards the zone 2/3 boundary, taxa more characteristic of warmer waters increase earlier than previously suggested at c. 1750 AD.