Interferometric synthetic aperture radar data from 2021 for landslides at Barry Arm Fjord, Alaska

Subaerial landslides at the head of the Barry Arm fjord remain a tsunami threat for the Prince William Sound region in southern Alaska. Tasked RADARSAT-2 synthetic aperture radar (SAR) data from two ultrafine beam modes (2 m), U19 and U15, were used to measure landslide movement of slopes near the toe of the Barry Glacier between 21 May 2021 and 5 November 2021. Data were acquired every 24 days, with U19 beginning on 21 May 2021 and U15 beginning on 28 May 2021. For a few planned acquisition dates, scenes were not captured because of technical issues. Interferometric synthetic aperture radar (InSAR) deformation maps (interferograms) are provided in wrapped phase (line-of-sight (LOS) phase in radians between 0 and 2π). For landslide areas of interest (Landslides A and B, Figure 1) with more than 25% pixels with coherence values > 0.25, unwrapped phase deformation maps are also provided with displacement in centimeters. These products were created following the same methodology as described in Schaefer et al. (2020), which provides InSAR data for the same landslides in 2020. See methodology file in Schaefer et al. (2020) for processing details. All scenes were filtered by averaging values over 4 lines in both the range and azimuth direction, resulting in 8 m pixel interferograms. For removing topographic phase contributions, this data release uses a different digital terrain model (DTM) than the one that was used in Schaefer (2020). Herein, we use a spliced DTM which combines the 5 m-horizontal resolution IFSAR DTM acquired in 2010, and a 1-m horizontal resolution lidar-derived DTM acquired by the Alaska Division of Geological and Geophysical Surveys (DGGS) on 26 June 2020 (Daanen et al., 2020). In contrast, Schaefer et al. (2020) used the 5 m IFSAR DTM. Non-negligible phase contributions from topography likely remain in areas outside of the lidar-derived DTM (Figure 1).

For each time frame between SAR acquisitions, we describe interferogram results for Landslide A and Landslide B on the western slope of the Barry Arm fjord near the terminus of the Barry Glacier (Figure 1). The kinematic elements of both landslides are adopted from mapping by Coe et al. (2020); Landslide A is composed of four kinematic elements (Kite, Prow, Core, Tail) and Landslide B is composed of one kinematic element (Wedge), which are referred to in the scene descriptions below. All displacement values provided in the descriptions of individual interferograms are given in the LOS direction. 
 
Data are provided in Geostationary Earth Orbit Tagged Image File Format (GeoTIFF) and Keyhole Markup Language (KMZ) formats. For each time frame, a Tagged Image Format (TIF) image is provided that shows an aerial image acquired by ©Planet 2021 on 28 August 2021 overlain with the landslide kinematic elements, the wrapped and unwrapped (when available) interferograms, scale, and satellite look direction information. All coordinates provided are in the World Geodetic System 1984 (WGS84) coordinate system.

Summary of movement during the 2021 monitoring period: We present InSAR detected deformation for two landslides on the southeast-facing slope at the Barry Arm fjord between 21 May and 5 November, 2021. Early (21 May – 21 June) and late (12 October - 5 November) scenes contained substantial noise, likely because of differences in snow cover or adverse atmospheric conditions. These environmental conditions resulted in very few coherent pixels and thus insufficient data to detect landslide displacement during these times.
 
During time periods with moderate-to-extensive coherent pixels (28 May – 18 September), four areas had apparent LOS movement: (1) Areas in the upper part of the Core ranging between 0.005 - 0.05 km2 (near -148.159°, 61.153°) moved downslope a total of ~2 cm between 28 May and 21 June, increased to ~5 cm between 15 July and 8 August, and then back to ~ 2 cm between 25 August and 18 September. (2) A 0.3 km2 area in the upper part of the Kite (centered on -148.170°, 61.140°) moved consistently throughout the summer and early fall, totaling ~ 8 cm of cumulative movement between 21 June and 1 September. This area was incoherent outside of this time frame. (3) Small amounts of movement (<2 cm per month) may have occurred in a 0.1 km2 area near the contact of the Core and Tail (centered on -148.142°, 61.152°) between 28 May and 8 August. Phase gradients in late summer and incoherence in spring and fall prevented accurate interpretation of this area outside of this time frame. (4) Small amounts (~2 cm) of movement may have occurred in a 0.05 km2 area of the northeast toe of Landslide B (centered on -148.1200°, 61.1680°) between 15 July and 8 August, but a lack of coherent pixels during this time interval and throughout the year makes this interpretation tenuous. All cumulative movement mentioned in this paragraph was determined by summing movement from individual interferograms, and using only sequential (not overlapping) time frames. Overall, during 2021, we did not detect landslide-wide downslope movement, such as what occurred in October 2020 (Schaefer et al., 2020).
 
Other advantageous and adverse conditions that impacted our results for 2021 included the following: Compared to our 2020 data release (Schaefer et al., 2020), the use of a high resolution lidar-derived DTM in this data release resulted in improved removal of topographic phase effects. Phase contributions from topography likely remain in areas outside of the lidar-derived DTM (Figure 1). Between 8 August and 18 September, scenes suffered from a phase gradient effect likely caused by hydrostatic effects of local gradients in the air pressure and temperature, making absolute displacement measurements challenging. The upper part of Landslide B suffers from layover, which occurs where steep topography causes points farther from the satellite to return first and distorts the image, preventing measurements.

21 May 2021 – 14 June 2021: Deformation interpretations for this time frame are not possible because of the very low signal-to-noise ratio, which is partially a result of differences in snow cover. Noise appears as speckle in this interferogram and in all the subsequent interferograms.

28 May 2021 – 21 June 2021: As with the previous time frame, substantial noise limits the ability to calculate deformation during this time interval. However, coherent pixels in Landslide A suggest that a small amount (~ 2 cm or less) of deformation occurred in a 0.05 km2 area of the upper Core (centered on -148.156°, 61.152° and in a 0.1 km2 area near the contact between the Core and Tail (centered on -148.142°, 61.152°). Measurements over Landslide B are not possible due to the very low signal-to-noise ratio.  

21 June 2021 – 15 July 2021: This time frame has more coherent pixels than the previous time frame. In Landslide A, ~ 2 cm of movement may have occurred in a 0.3 km2 area in the upper part of the Kite (centered on -148.170°, 61.140°), and < 2 cm of movement may have occurred in the 0.1 km2 area near the contact between the Core and Tail, as in the previous time frame (centered on -148.142°, 61.152°). The upper part of Landslide B suffers from layover, which occurs when steep topography causes points farther from the satellite to return before points closer to the satellite, which subsequently distorts the image and prevents measurements. Coherent pixels in the lower part of Landslide B indicate either that the landslide did not move during this time interval or that it moved at the millimeter scale. 

15 July 2021 – 8 August 2021: As in the previous time frame, coherent pixels in Landslide A  indicate that ~ 2 cm of deformation may have occurred in a 0.3 km2 area in the upper part of the Kite (centered on -148.170°, 61.140°) and < 2 cm of deformation may have occurred in the 0.1 km2 area near the contact of the Core and Tail (centered on -148.142°, 61.152°). Additionally, a 0.005 km2 area in the upper part of the Core appears to have moved ~5 cm (centered on -148.159°, 61.153°). Small amounts (~ 2 cm) of movement may have occurred in a 0.05 km2 area of the northeast toe of Landslide B (centered on -148.120°, 61.1680°), but a lack of coherent pixels makes this interpretation difficult.   

8 August 2021 – 1 September 2021: This image suffers from a phase gradient effect that extends along the entire length of the southeast-facing slope. This effect was likely caused by local gradients in air pressure, moisture, and temperature, which resulted in a false displacement signal of ~1.5 cm across the southeast-facing slope in the unwrapped interferogram. Though the phase gradient effect makes absolute measurements difficult, the 0.3 km2 area in the upper part of the Kite (centered on -148.1700°, 61.1400°) deviates from the overall trend of the phase gradient and ~ 2 cm of displacement appears to have occurred during this time interval.

25 August 2021 – 18 September 2021: As with the previous time frame, this image suffers from a phase gradient that extends along the entire length of the southeast-facing slope. The phase gradient and areas of incoherence make interpretations of deformation challenging. However, ~ 2 cm of movement appears to have occurred in the 0.3 km2 area in the upper part of the Kite (centered on -148.170°, 61.140°) and in the 0.005 km2 area near the upper part of the Core (centered on -148.159°, 61.153°). 

1 September 2021- 25 September 2021: This image suffers from substantial noise compared to the previous image. Coherent pixels in Landslide A indicate that the landslide either did not move during this time interval, or it moved at the millimeter scale. Measurements over Landslide B are not possible due to the very low signal-to-noise ratio.

12 October 2021 – 5 November 2021: Snow at higher elevations during this time interval results in substantial incoherence near the top of Landslide A and Landslide B. Coherent pixels in Landslide A do not indicate any movement. Measurements over Landslide B are not possible due to the very low signal-to-noise ratio.

References cited:
Coe, J.A., Wolken, G.J., Daanen, R.P., and Schmitt, R.G., 2021, Map of landslide structures and kinematic elements at Barry Arm, Alaska in the summer of 2020: U.S. Geological Survey data release, https://doi.org/10.5066/P9EUCGJQ.

Daanen, R.P., Wolken, G.J., Wikstrom Jones, Katreen, and Herbst, A.M., 2021, Lidar-derived elevation data for upper Barry Arm, Southcentral Alaska, June 26, 2020: Alaska Division of Geological & Geophysical Surveys Raw Data File 2021-1, 9 p. https://doi.org/10.14509/30589

Schaefer, L.N., Coe, J.A., Godt, J.W., and Wolken, G.J., 2020, Interferometric synthetic aperture radar data from 2020 for landslides at Barry Arm Fjord, Alaska (ver. 1.4, November 2020): U.S. Geological Survey data release, https://doi.org/10.5066/P9Z04LNK.

Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.