*3.1.3 Time-lapse camera applications*

The surface kinematics of the Planpincieux Glacier right lobe has been deeply investigated with image analysis of 6-year-long time-lapse monitoring. The data analysis allowed characterising the terminus dynamics and classifying the instability processes that cause break-offs: (i) disaggregation, (ii) slab fracture and (iii) water tunnelling [7]. Disaggregation is the progressive toppling of small ice pieces caused by the movement of the terminus beyond the frontal bedrock cliff. It is the most frequent process and it involves break-offs of limited size, usually lower than 1000 m3 . Slab fracture instability is caused by the aperture of a crevasse orthogonal to the motion direction, located in correspondence to the maximum tensile stress line. When the fracture reaches the bedrock, it triggers a large break-off of an ice lamella that can assume a volume of 104–105 m3 . Water tunnelling refers to the formation of R-channels [52] where a large amount of

#### **Figure 4.**

*RTS measurements of prisms 13, 14, and 2b before the failure of 22/09/2014. Using these data, the break-off was predicted 10 days in advance.*

water can accumulate. The water produces a strong pressure on the frontal cliff that can provoke failure of the terminus. Moreover, the empty tunnels increase the instability and they can collapse themselves.

Besides the visual photographic interpretation, DIC in the Fourier domain was applied to the hourly images, obtaining surface displacement maps at daily resolution. During the monitoring period, the surface displacement pattern was composed of four distinct kinematic domains, which were characterised by different velocity regimes. The presence of kinematics domains indicates the action of high strain rates localised at the domain limits, where large fractures appear (**Figure 5**). The behaviour of the frontal sector is noteworthy, because it reveals the occurrence of a few speed-up periods per year, which culminate with large break-offs (**Figure 6**). These kinematic fluctuations were characterised by well-defined thresholds of initial velocity (*v*<sup>0</sup> ≥ 30 cm day<sup>−</sup><sup>1</sup> ) and acceleration (*a* ≥3 cm day<sup>−</sup><sup>2</sup> ). Moreover, a monotonic relationship (rank correlation coefficient > 0.7, p-value < 0.02) between the velocity peak and the collapsed volume has been observed.

DIC in the spatial domain was applied to the images of the Whymper Serac to measure the displacement in July 2019 (**Figure 7a**, **b**). The available images presented rototranslation that had to be compensated with robust coregistration. Moreover, the smooth texture and low chromatic contrast of the scene lowered the signal-to-noise ratio (i.e. the correlation, see **Figure 7**) and hence many artefacts were present in the displacement maps. Therefore, a robust outlier correction method was applied [53]. The results showed a slight acceleration during July, which was confirmed by RTS measurements.

## *3.1.4 TRI applications*

The Planpincieux is probably the unique glacier where TRI surveys were conducted using four different terrestrial interferometric radar models, namely: GPRI™ (Gamma Remote Sensing, https://www.gamma-rs.ch/rud/microwavehardware/gpri.html), IBIS-L™ (IDS Georadar, https://idsgeoradar.com/products/ interferometric-radar/ibis-fl), FastGBSAR-S™ (MetaSensing, https://www.geomatics.metasensing.com/fastgbsar-s) and GBInSAR LiSALab™ (LiSALab, http:// www.lisalab.com/home/default.asp?sez=6).

#### **Figure 5.**

*(a) Image of the Montitaz Lobe acquired by the monoscopic time-lapse camera. The terminus width is approximately 100 m. (b) Surface deformation map. Different velocity regimes clearly identify the four kinematics domains. (c) Longitudinal conceptual scheme of the glacier lobe (not in scale). The black lines indicate bedrock discontinuities that correspond to the kinematic domain limits. Modified from Giordan et al. [7].*

**133**

and 50 cm day<sup>−</sup><sup>1</sup>

**Figure 6.**

**Figure 7.**

*Close-Range Sensing of Alpine Glaciers DOI: http://dx.doi.org/10.5772/intechopen.92841*

The surface kinematics of the glaciers was surveyed in five TRI campaigns, in 2013, 2014, 2015, 2016 and 2019 (**Table 1**). The first two were conducted using the GPRI™ real-aperture radar (RAR) in Ku band that surveyed the glacier from the valley bottom and the valley ridge opposite to the glaciers. Both campaigns lasted for 2 days and they were able to detect the displacements of the lower portions of the Planpincieux and Grandes Jorasses glaciers, which were approximately 25 cm day<sup>−</sup><sup>1</sup>

*(a) IBIS-L GBSAR surveyed the Planpincieux Glacier area, delimited in black. (b) The cumulative sum of the* 

*interferograms acquired in the period September 4, 2015 to October 14, 2015.*

*Time series of the daily velocity of sectors A and B (see Figure 5) in the years 2014–2019 (from top to bottom). The break-off occurrence is depicted in black dots, while the white circle size is proportional to the volume.*

Ku-band ground-based synthetic aperture radars (GB-SAR). The campaign of autumn 2015 (IBIS-L™) lasted much longer and hence it was possible to recognise the different kinematic domains of the Montitaz Lobe (**Figure 8**). During the campaign, the meteorological conditions were severe and the radar acquisitions were affected by strong APS. To solve the issue, a polynomial APS model that was a function of the topography was developed [9, 42]. In 2016, FastGBSAR-S™ measurements with an acquisition frequency of 10 s were carried out; thereby, the atmospheric disturbance was minimised. Fully polarimetric measurements were experimented, but the very long distance did not allow exploiting the potentiality of such a technology. The last campaign (GBInSAR LiSALab™) began at the end of September 2019 for civil

protection and it is still active during the writing of the present chapter.

respectively. Instead, the following surveys were conducted using

#### **Figure 6.**

*Glaciers and the Polar Environment*

(*a* ≥3 cm day<sup>−</sup><sup>2</sup>

been observed.

*3.1.4 TRI applications*

instability and they can collapse themselves.

was confirmed by RTS measurements.

www.lisalab.com/home/default.asp?sez=6).

well-defined thresholds of initial velocity (*v*<sup>0</sup> ≥ 30 cm day<sup>−</sup><sup>1</sup>

water can accumulate. The water produces a strong pressure on the frontal cliff that can provoke failure of the terminus. Moreover, the empty tunnels increase the

Besides the visual photographic interpretation, DIC in the Fourier domain was applied to the hourly images, obtaining surface displacement maps at daily resolution. During the monitoring period, the surface displacement pattern was composed of four distinct kinematic domains, which were characterised by different velocity regimes. The presence of kinematics domains indicates the action of high strain rates localised at the domain limits, where large fractures appear (**Figure 5**). The behaviour of the frontal sector is noteworthy, because it reveals the occurrence of a few speed-up periods per year, which culminate with large break-offs (**Figure 6**). These kinematic fluctuations were characterised by

). Moreover, a monotonic relationship (rank correlation coeffi-

cient > 0.7, p-value < 0.02) between the velocity peak and the collapsed volume has

DIC in the spatial domain was applied to the images of the Whymper Serac to measure the displacement in July 2019 (**Figure 7a**, **b**). The available images presented rototranslation that had to be compensated with robust coregistration. Moreover, the smooth texture and low chromatic contrast of the scene lowered the signal-to-noise ratio (i.e. the correlation, see **Figure 7**) and hence many artefacts were present in the displacement maps. Therefore, a robust outlier correction method was applied [53]. The results showed a slight acceleration during July, which

The Planpincieux is probably the unique glacier where TRI surveys were conducted using four different terrestrial interferometric radar models, namely: GPRI™ (Gamma Remote Sensing, https://www.gamma-rs.ch/rud/microwavehardware/gpri.html), IBIS-L™ (IDS Georadar, https://idsgeoradar.com/products/ interferometric-radar/ibis-fl), FastGBSAR-S™ (MetaSensing, https://www.geomatics.metasensing.com/fastgbsar-s) and GBInSAR LiSALab™ (LiSALab, http://

*(a) Image of the Montitaz Lobe acquired by the monoscopic time-lapse camera. The terminus width is approximately 100 m. (b) Surface deformation map. Different velocity regimes clearly identify the four kinematics domains. (c) Longitudinal conceptual scheme of the glacier lobe (not in scale). The black lines indicate bedrock discontinuities that correspond to the kinematic domain limits. Modified from Giordan* 

) and acceleration

**132**

**Figure 5.**

*et al. [7].*

*Time series of the daily velocity of sectors A and B (see Figure 5) in the years 2014–2019 (from top to bottom). The break-off occurrence is depicted in black dots, while the white circle size is proportional to the volume.*

#### **Figure 7.**

*(a) IBIS-L GBSAR surveyed the Planpincieux Glacier area, delimited in black. (b) The cumulative sum of the interferograms acquired in the period September 4, 2015 to October 14, 2015.*

The surface kinematics of the glaciers was surveyed in five TRI campaigns, in 2013, 2014, 2015, 2016 and 2019 (**Table 1**). The first two were conducted using the GPRI™ real-aperture radar (RAR) in Ku band that surveyed the glacier from the valley bottom and the valley ridge opposite to the glaciers. Both campaigns lasted for 2 days and they were able to detect the displacements of the lower portions of the Planpincieux and Grandes Jorasses glaciers, which were approximately 25 cm day<sup>−</sup><sup>1</sup> and 50 cm day<sup>−</sup><sup>1</sup> respectively. Instead, the following surveys were conducted using Ku-band ground-based synthetic aperture radars (GB-SAR). The campaign of autumn 2015 (IBIS-L™) lasted much longer and hence it was possible to recognise the different kinematic domains of the Montitaz Lobe (**Figure 8**). During the campaign, the meteorological conditions were severe and the radar acquisitions were affected by strong APS. To solve the issue, a polynomial APS model that was a function of the topography was developed [9, 42]. In 2016, FastGBSAR-S™ measurements with an acquisition frequency of 10 s were carried out; thereby, the atmospheric disturbance was minimised. Fully polarimetric measurements were experimented, but the very long distance did not allow exploiting the potentiality of such a technology. The last campaign (GBInSAR LiSALab™) began at the end of September 2019 for civil protection and it is still active during the writing of the present chapter.

#### **Figure 8.**

*(a-c) Surface displacement maps of the Whymper Serac of the periods July 1, 2019 to July 7, 2019, July 7, 2019 to July 16, 2019 and July 16, 2019–July 24, 2019. (d) Map of the mean correlation coefficient, which displays low values because of the texture smoothness of the snow surfaces. The serac face is approximately 40 m high.*

#### **Figure 9.**

*(a) GPR traces of the Planpincieux (orange line) and Grandes Jorasses (blue line) glaciers. (b-c) GPR profiles of the Whymper Serac and Planpincieux Glacier respectively. The white-red boundary indicates the ice thickness.*

**135**

**Figure 10.**

*from Dematteis et al. [6].*

*Close-Range Sensing of Alpine Glaciers DOI: http://dx.doi.org/10.5772/intechopen.92841*

Glacier and lower than 20 m in the Whymper Serac.

A helicopter-borne 65-MHz GPR survey was conducted in the Planpincieux-Grandes Jorasses glacial complex in April 2014, when 16 GPR traces homogeneously distributed on the glaciers' surface were acquired (**Figure 9**). The noise of the radar data was quite high, because the numerous crevasses caused bounds of the electromagnetic waves and produced echoes and artefacts. Nevertheless, it was possible to estimate the glacier thickness, which was in the range 20–40 m in the Planpincieux

In September 2015, time-lapse photography and terrestrial radar campaigns were conducted simultaneously to measure the Planpincieux Glacier surface kinematics. The actual three-dimensional surface kinematics was obtained by coupling DIC and TRI results. **Figure 10** reports the mean daily velocity map, where the colour represents the velocity module and the arrows indicate direction and versus. The 3D displacement can be obtained only in the areas visible by both the sensors. In the right lobe, the displacement vectors are not uniformly parallel to the surface, because the seracs move downstream as a single body and the ice is subjected to internal deformation. This result is not trivial, as the most common approach to estimate 3D displacement is to project the single movement components along the local slope obtained from the DEM, but this assumption might be misleading in specific cases. The permanent monitoring system of the Whymper Serac is composed of RTS and time-lapse imagery. In July 2019, the data of the two sensors were integrated and represented in an informative bulletin [54, 55], shown in **Figure 11**. Such integration allows evaluating the versus and direction of the principal movement (with

the RTS data) and the distribution of the strain rates (with the DIC results).

*Velocity field of the surface kinematics of the lower Planpincieux Glacier obtained with the integration of DIC and TRI measurements. Colours and arrows represent velocity module and direction respectively. Modified* 

*3.1.5 GPR applications*

*3.1.6 Data integration*
