The rates and processes of periglacial mass movement were measured on the vegetation-free debris-mantled slopes in middle reaches of Adventdalen and Reindalen in central Spitsbergen Svalbard (Fig. 1). All slopes are mainly covered with debris of sandstone and shale of Jurassic, Cretaceous and Tertiary age (Fig. 2). A grain size analysis shows that surface layers of these slopes are composed of rubble and frost-susceptible fine materials (Fig. 5). The maximum depth of active layer varies year by year, but the average depth is estimated about 1 meter. Total 17 painted-stone lines were installed nearly horizontally on the slope surface in above 2 areas : 12 lines during the summer, 1988 and -5 during the summer, 1989 in order to detect movement rates of surface materials (Table 1). Eight pieces of flexible glass-fibre tubes (5 mm in diameter) were inserted into the ground vertically in 1988 and 1989. Furthermore measurement of year-round ground temperature and frost heave was achieved at north-facing mountain slope (700 m a.s.l.) of Mt. Skolten, using a data logger at 3-hour intervals from August, 1990 to July, 1991. Sensors were installed at 0, 5, 20, 40, 60 and 100cm in depth.
Annual averages of movement rates measured by painted stone-lines are 0.6 cm to 11.2 cm (Table 1). Movement rates of surface rubble and slope gradients have roughly mutual relation (Fig. 3). Deformation of all painted stone-lines were patterns parallel to the base lines, and this occurred on every type of slope materials independently of the thickness and size of surface rubble layer (Fig. 4). The vertical velocity profiles of the excavated tubes fall into three types : straight profile keeping tilt (Fig. 6 JH-1), concave profile indicating the greatest movement at the surface (5-1, S-2, S-4, R-5) and complex (concave/convex) profile indicating relatively larger movement at depth (AD-4, S-3, R-3). The movement of the tops of all tubes averaging 3.4 cm, which similar to the average rate of movement measured by painted stones. This implies that these three types of deformation were formed by the same processes, namely frost creep and gelifluction.
Only eleven diurnal freeze-thaw cycles were recorded at the ground surface on the northfacing mountain slope (19°) of Mt. Skolten from August, 1990 to July, 1991 (Fig. 7-a). Maximum and average depth of ground freezing occurred during these short-term cycles were 6.5 cm and 3.2 cm respectively. Frost heave recorded as the average 0.47 cm per 1 event, occurred from middle August to middle September, corresponding to the diurnal freeze-thaw cycles (Fig. 8). On the other hand, the heave caused by seasonal frost occurred from 21 to 27 August and the amount was 3.0cm (Fig. 8). The seasonal heave occurred untill the frost table reached less than 30cm in depth. The cumulative amount of diurnal and seasonal frost-heave is 8.7 cm. Accordingly amount of potential frost creep for the year is calculated to 2.7 cm on the 19 degrees slope. The annual average movement rate which was measured by painted stonelines at the same slope is 4.0 cm. This value is lager than the potential frost creep. Since the movement of painted stone-line resulted from a combination of frost creep and gelifluction, the value (1.3 cm) which deduct potential frost creep from movement by painted line is regarded as movement by gelifluction.
An increase of thickness of surface rubble layer without interstitial fine materials is the cause of decrease of rubble movement because ice segregation do not occur in such rubble layer (Fig. 9). Especially the surface rubble attaining more than about 45 cm in thickness suddenly decrease in movement rates by the reason of the layer underlying such thick rubbles maintain the frozen conditions for all the year round.