The scatter of annual net accumulations at a given place

36 stakes spaced 10 m apart was set up in a square pattern at ten different places between 550 m and 3100 m a.s.l. on the Mizuho Plateau. The purpose of this stake-farm method was to obtain reliable representative values at each place, and also to examine the scatter of 36 annual net accumulations at each place. Figure 2 shows that the coefficient of variation S/M, where M is the average of 36 values of annual accumulation in 1981 and S is the standard deviation, was small (less than 1.0) especially in the high accumulation zone. Besides, the value of S/M did not depend on the roughness of snow surface (Satow 1984).

Fig. 2. Roughness of snow surface R; average of thirty-six values of annual accumulation M in 1981 (m of snow depth); standard deviation S of the thirty-six values; and coefficient of variation S/M, plotted against elevation of 36-stake farms (after Satow 1984).

Average of annual net accumulation during the period from 1968/72 to 1981

From data of annual net accumulations measured by single stake method in 1968/72-1981, the averages of annual accumulations were obtained at 2 km intervals along the traverse routes in the area between 550 m and 2230 m a.s.1. Figure 3 shows that annual net accumulation is higher (0.4-0.8 m per snow depth/year) near the coastal region (as far as about 160 km from the coast), which is strongly influenced by low pressure activities, and decreases inland.

Fig. 3. Average annual net accumulations 1968/1972 to 1981 against elevation, along the traverse route (dashed line in Fig.1) on Mizuho Plateau. Dashed vertical lines denote standard deviations for the same period.

Analysis of a 30 m core in a high accumulation zone

Glaciological studies of a 30 m firn core were carried out at Station S18 in the coast area near Syowa Station, where annual net accumulation was assumed to be about 0.5 m snow depth a year. The features of profiles of density, depth hoar level (Watanabe 1978b), oxygen isotope composition δ¹⁸O and stratigraphy of the core indicate that there are no great difficulties in determining annual layers, and thus in obtaining a year-by-year dating of the core. A part of the core analysis is shown in Figure 4. In this way, the annual variation of snow accumulations at the station was obtained from 1914 to 1981 (Figure 5). it was found that annual net accumulations were higher during the period between 1947 and 1960.

Power spectrum analysis was made of layer-to-layer variations of annual accumulations of the core. Then the power spectrum was computed by applying the mximum entropy method (MEM) developed by J.P. Burg (Hino 1977). In the power spectrum as shown in Figure 6, two most prominent peaks indicate periodicities of 45 years and 18 years. At present the relation is unclear between these periodicities of accumulation and variations of other meteorological factors such as pressure, temperature and wind speed.

δ¹⁸O profile against depth

The δ¹⁸O profile (Figure 4) shows typical seasonal variations, providing a useful method of determining annual snow layers. But further inland the profile becomes more complicated and does not ususally show much seasonal variation (Watanabe and others 1981). This is thought to be connected with the snow accumulation depth, the season of surface formation by deposition-erosion process and snow metamorphosis, such as depth hoar formation. Inland the difference of mean monthly temperatures between summer and winter reaches near 40 °C (Satow 1978) and there frequently exists a strong temperature gradient near the surface. Actually the authors once measured the gradient to be 500°C/m in snow, just near the surface, in winter. Here the effect of temperature gradient is considered as one explanation of δ¹⁸O profile-metamorphism of snow in the inland region.

Fig. 4. Vertical profiles of oxygen isotope composition (δ¹⁸O), depth hoar level (DH) and density (D) in part of a 30 m core at S18, together with vertical profile (S) of stratigraphie elements.

Fig. 5. Variations of annual net accumulation from 1914 to 1981 (dashed line), based on the 30 m core analysis at S18. The solid line shows 5 year moving averages.

Fig. 6. Power spectrum of annual net accumulations from 1914 to 1981 at S18, using MEM method. Peaks A, B, C and D indicate periodicities of 45 years, 18 years, 8.2 years and 6.7 years respectively.

Some models of δ¹⁸O- metamorphism of snow with temperature gradient

Experiments were made to examine characteristics of δ¹⁸O profile changes of snow with temperature gradient. Each different temperature gradient was given vertically in a block sample, which was surrounded by thermal insulating materials. Figure 7 shows a schematic diagram of the experiment apparatus. The result of one such experiment is shown in Figure 8, where the snow sample dimensions were 0.15(W) x 0.15(D) x 0.18(H) m. After these experiments. developments of vertical snow texture were observed in most cases.

Based on experimental results under different conditions, the authors propose several models of metamorphism of δ¹⁸O profile with temperature gradient (Figure 9a and 9b). These models can be divided into two groups: one with a cap, such as an ice crust or ice lense, and the other without. Figure 9a and 9b shows the speculated models under the condition that the upper side is colder and the underside is warmer. Which case of these models should be adopted to explain the actual δ¹⁸O profile-metamorphism depends on the circumstances of accumulation depth, strength of temperature gradient of snow, features of snow structures (density or porosity), etc.

Fig. 7. Schematic diagram of experimental apparatus. 1: snow sample, 2: thermal insulating material, 3: thermometer, 4: thermal detector, 5: heater.

Fig. 8. Result of one experiment concerning δ¹⁸O metamorphism of snow and temperature gradient for 37 days. T: snow temperature, S: stratigraphy, O: oxygen isotope composition (°/°°) and D: density of the snow sample. Solid lines in diagrams of δ¹⁸O and D show values before the experiment, dashed lines those after the experiment.

Fig. 2. Metamorphism process model of δ¹⁸O profile of snow with temperature gradient under a cap. The upper side is colder and the under side warmer. Solid lines of δ¹⁸O profiles indicate the first profiles at initial stage 1, dashed lines profiles at the next stage 2, of the metamorphism. ‘A’, ‘B’ and ‘C’ represent three different types of first δ¹⁸O profiles, (b) As for Fig.9(a), in uncapped snow.

As mentioned above, the explanation of the δ¹⁸O profile of snow should be considered very carefully in regard to the δ¹⁸O -metamorphism of snow with temperature gradient, especially of snow in the irregular accumulation zone and the low but uniform accumulation zone, with small accumulation, strong temperature gradient of snow and developments of hoar at depth.