Some problems of landslide monitoring using satellite radar imagery with different wavelengths: Case study of two landslides in the region of Greater Sochi

The problems of processing and interpreting the data provided by radar satellite interferometry for the conditions of landslides covered by vegetation are analyzed in two case studies of landslides in the Northern Caucasus in the region of Kepsha and Mamaika villages in the vicinity of the railway tunnels. The estimates of the displacement fields are obtained by the method of persistent scatterers using the StaMPS program package. The five-year experience of landslide monitoring shows that in the unfavorable conditions of satellite radar interferometry, proper selection of the strategy of satellite image processing is vital. In the present paper, we discuss, in particular, the crop selection, the selection of the master image, reference area, and digital elevation model. For the landslide located in the sparsely populated region near Kepsha village, we used the data from the ascending and descending tracks of the long-wavelength ALOS and shorter-wavelength ENVISAT satellites. For the landslide in the region of Mamaika village with a large number of different buildings serving as good scatterers for radar signals, we used the images from the ENVISAT and from TerraSAR satellite, which transmits even shorter waves. The average line-of-sight (LOS) displacement velocities V_LOS for the landslide near Kepsha village measure at most 10 mm per annum, which means that this landslide has remained stable at least since 2004. The landslide in Mamaika village is significantly more active. The average LOS displacement velocities in the active part of this landslide attain 60 mm per annum. The artificial corner reflector installed on the segment of the landslide where natural scatterers of radar signal are absent made it possible to estimate the LOS displacement velocity on this segment of the slope at 49 mm per annum.