4.1. Observations at site 1
Assessment of landslide and shoreline abrasion intensity at Site 1 is based on the volume of material displaced per unit area. According to the laser scanning results, changes were evaluated in the upper part of the slope, where a block type landslide was developing, and in the lower part, where an abrasion scarp was formed as a result of wave wash (Figure 6). The evolution of the situation at study site 1 was analyzed by comparing the results of laser scanning (Table 5) and those obtained during the 2019 UAV survey (Table 6).
Ground displacements in the abrasion scarp area vary independently of the seasons, which is explained by the water level fluctuations of the reservoir. First, there is an accumulation of soil in the lower part of the slope due to landslide processes. In the case of high-water levels, the soil is eroded by wave abrasion. During water drawdown the foreshore width does not allow the waves to erode the shore, and material accumulates again at the foot of the slope. This explains why denudation near the abrasion scarp prevails in 2012-2013, and accumulation processes prevailed in 2013-2014.
In general, for study site 1 denudation processes are dominant during the observation period (Table 8). Volume of landslide changes per unit area showed that intra-season and inter-annual dynamics of the process are irregular. The high intensity of the denudation process in the autumn-winter period of 2013 is noted because of a smooth increase of solar radiation and, as a consequence, low evaporation passing into snowmelt runoff. The comparable rates of the denudation process in the summer-autumn period of 2013 and autumn-winter period of 2014 are explained by a significant erosive (>10 mm/day) rainfall and, in contrast, by a less erosive winter precipitation (Figure 7). However, despite the fundamentally different patterns of intra-annual variability of denudation processes, the inter-annual net material loss at site 1 remains, rather constant (ca. -0.03 m3/m2).
The analysis of the dynamics of transverse profiles (Figure 8 Profile 1) showed that in the zone of block landslide development, the mass movement in the summer-autumn period is smaller than in the autumn-summer period. In the summer-autumn period of 2013, the landslide body did not change its position. Displacements of masses occur mainly during the snowmelt period. So, in the autumn-summer period of 2013, the landslide mass moved downslope by 4 m, and in 2014 by 3 m in plan.
The graphs show that in November 2012, there was an accumulation of soil masses on the abrasion part of the bank slope. From November 2012 to July 2013, the accumulated material was intensively eroded. The reduction of the process intensity characterizes the summer-autumn period of 2013. Significant re-activation was observed in the period from November to June 2014.
A significant difference is observed between the 2012-2014 and the 2019 slope profiles. The landslide body is not seen on the 2019 profile which is almost straight. Over the last five years, the landslide body has been completely transformed, the profile of the slope has reached a dynamic equilibrium and has become linear. The average thickness of the transformed soil layer on the flattened sections was 1.5 m or 0.5 m/year during 2012-2014 and 2 m or 0.4 m/year during 2014-2019 (Figure 9).
Based on the 2003-2006 survey, the authors recommended to install landslide-control measures, directed to protect the pier under construction. The essence of the measures consisted in the modelling of the landslide slope with subsequent flattening at 2007, installing engineering structures at the foot of the slope to protect it from wave action and to prevent the development of abrasion processes. The implemented measures turned out to be very effective. At present (2019), no landslide reactivation processes are observed on the treated part of the slope, which can be seen in profile 2 (Figure 8).