Displacement analysis over Baltic landfast ice

The coherence (Figure 22a), interferogram (Figure 22b) and displacement maps (Figure 23) are products of methodology flowchart (Figure 19). The fringes in the interferogram were converted to displacements in the LOS in Figure 23. One fringe corresponds to an approximately 27.5 mm displacement in LOS. The resulting displacements were from –10 cm to 30 cm over landfast ice. The negative sign indicates that the ice has moved away from the satellite by either sinking or moving to the west, and positive indicates movement towards the satellite by either lifting up or moving to the east. (PIII)

The southern part of the polygon was moving away, and the northern part was getting closer to the satellite. These movements created a converging zone between the southern and northern parts and the projected strain was approximately 40 cm across a 20 km distance (Figure 23).

78

The movements are due to horizontal or vertical changes, or a combination of both. As a descending pair was used, it is not possible to separate the horizontal and vertical changes. One reason for a vertical displacement could be sea level tilt. Absolute sea level elevation could not affect interferogram fringes because the landfast ice is afloat, but sea level tilting might affect fringes according to ∆ = ⁄2 , where is the interferometric phase, and

= is the wave number, which was equal to 114.20 m⁻¹ (PI).

(a) (b)

Figure 22. (a) The extracted polygon with high coherence (0.2-0.46) for the image pair of 6 and 18 February 2015 over the Baltic Sea. (b) The interferometric phase.

The Sentinel-2 image was used as background. Water and land are shown in black and green respectively. Figures adapted from PI.

79

Figure 23. Displacement map from the unwrapped 6-18 interferogram. ‘Near’ and

‘Far’ means close to the satellite and far away from the satellite respectively. Figure adapted from PI.

The sea level data showed that in this case, sea level could make only small relative vertical displacements. In the period of study, from 6 to 18 February, the sea level decreased from 41.4 to 36.0 cm in Oulu and from 39.7 to 36.5 cm in Kemi. Thus, the maximum relative change in sea level was 2.2 cm and the distance between the sea level stations was 82 km (PI). The absolute sea level ranged within ±50 cm in the period of study, but the whole water body moved up and down almost coherently leaving tilts across the basin below 5 cm over a 100-km distance. The sea level information was presented in Figure 13.

Another possible reason is ice growth but here changes of freeboard could make vertical displacements only less than one centimeter. Regarding the lateral displacement, that may happen due to thermal expansion. However, one

80

would not expect to see simultaneous thermal expansion and contraction here, as Figure 23 would suggest. (PI)

The only possible reason to have the observed lateral displacements is mechanical forcing by winds or water currents. Water currents were not measured but they tended to be small below the landfast ice. The location of the study area (Figure 23) was within the landfast ice regime during the period.

The landfast ice thicknesses at that time were around 30-50 cm, and in that sheltered part of the archipelago, landfast ice of that thickness remains quite firmly in place (Leppäranta 2013). Outside the landfast ice regime and the study area, there was thin level ice and close ice, based on the operational ice charts (Figure 10b). By overlaying both backscatter intensity images on 6 and 18 February (Figure 11), it was shown that on 6 February, the offshore area in the west of the study area was covered with mainly thin level ice. As could be expected from wind records on 10, 16 and 17 February (Figure 12), with predominantly strong south-westerly winds with high speed, on the order of 15 m/s, the drift ice was compressed towards the landfast ice edge and deformed to ridges and rafted ice. This deformation can be seen as features with bright backscatter intensity on 18 February 2015 (Figure 11b) and as very close or consolidated drift ice with ridge symbols in the operational ice charts.

The southwest wind had a minor influence on the inner part of the landfast ice zone where the fringes were shown. There was a simultaneous dilatation along the boundary, as typically takes place in such forcing conditions (Goldstein et al. 2009).

Landfast ice had heavier deformation with ridging further out near the landfast ice boundary. This area was away from the landfast ice boundary suggesting that the main load of the southwest storm did not reach the study area.

81

There were some small phase jumps in the fringe pattern showing shearing and cracking (Figure 22b). However, these phase jumps were not strong enough to make changes in the fringe pattern. There were also black lines in Figure 22a with low coherence, even 600 meters long. Some of them match the phase jumps in Figure 22b. There are two possibilities to have these phase jumps over the area of study: a) Fractures due to landfast ice displacement, and b) Ice roads. The latter option was evaluated by overlaying one of the SAR backscatter intensity scenes on Google Earth, and it was seen that the lines could be routes between islands used by people for fishing. By looking at Krassinletto island, it was clear that most of the black lines finished there.

There were some tracks that did not seem to reach an island, but they could represent fishing camps on the ice. (PI)

7.2 Relative performance of different SAR features and their