Figure 5. Differences in LOS-change detected by ALOS2 and Sentinel-1 seasonal interferograms (Figures 4a-d and 4f-i). In the last term of seasonal analysis (Figure 4e and 4j), we could not estimate differences due to coherence loss.
Figure 5 shows the differences between ALOS2 and Sentinel-1 InSAR data with nearly identical periods, which may help in cross-validating the measurements and understanding the actual deformation processes. The estimated differences and their 2σ scatter were 0.5±1.2cm (Fig 5a),0.7±2.3cm (Fig 5b),0.3±1.3cm (Fig 5c), and 0.6±0.3cm (Fig 5d), with mean of 0.5±1.5 cm. The differences and their variances were variable over time but apparently indicated some systematic trends. For instance, over the east-facing slopes, the differences were almost always positive (This is discussed more comprehensively in section 5.1).
The Sentinel-1 interferograms for 2017 demonstrate that the progress of deformation was not at a constant rate (Figure 6). The most rapid deformation took place in June (periods 1 and 2) with no substantial deformation in July (period 3) and started to subside again in August (periods 4-6). We found that the subsidence occurred sporadically over time and space and that the burned area did not uniformly subside. For periods 4, 5 and 9 we were unable to perform phase unwrapping at specific locations near the ridge and the boundaries between the burned and unburned areas. These unwrapping errors were responsible for the localized, large differences observed in Figure 5b. We confirmed the presence of low coherence bands along the unwrapping errors, which may suggest large phase jumps due to large displacements during the 12 days; enigmatically, no such line-shaped low coherence was detected in the long-term ALOS2 interferograms. Moreover, Figure 6 demonstrates that the frost heave started in late September, which was missed in the periods (b) and (g) of Figure 4, and that the absence of any deformation signals lasted from early December to May of the following year. We will physically interpret the absence of deformation signals during the coldest season in section 5.3.