An example of atmospheric effect in differential interferograms in the Piedmont Po Plain area (Vercelli-Trino, Northern Italy)?

With differential interferometry from Interferometric Synthetic Aperture Radar (InSAR) it is possible to monitor subsidence or uplift along line-of-sight. One popular satellite constellation for interferometry is Sentinel-1, operated by ESA. It is composed by two satellites, Sentinel-1A and Sentinel-1B, carrying a C band SAR sensor.

Sentinel data are freely available from the Copernicus Open Access Hub (https://scihub.copernicus.eu/).

I downloaded a few Sentinel-1 images of the Piedmont Po Plain (Vercelli-Trino Vercellese and surroundings) to try deriving some interferometric results to interpret.

Fig. 1: map of the studied are in the Piedmont Po Plain (Northern Italy). Google satellite basemap. Created with QGIS.

To see the temporal evolution of the results, a set of four images for the following 2020 periods were used:

• 05/01/2020: S1B_IW_SLC__1SDV_20200501T171426_20200501T171453_021391_0289B0_577E
  
• 05/13/2020: S1B_IW_SLC__1SDV_20200513T171427_20200513T171454_021566_028F1C_D01D
  
• 05/25/2020: S1B_IW_SLC__1SDV_20200525T171428_20200525T171455_021741_029439_E72E
  
• 06/06/2020: S1B_IW_SLC__1SDV_20200606T171428_20200606T171455_021916_02997C_907D
  

Three differential interferometric results were obtained, using SNAP software by ESA and following the indications presented in tutorials and videos published by ESA. The calculated LOS displacements were masked to hide areas were results have low coherence.

The displacements are presented in the following Fig. 2A-C, where Fig. 2A refers to the 05/01->13 time interval, Fig. 2B to the 05/13 -> 25 interval and  Fig. 2C refers to the final 05/25->06/06 time interval.

 (A)

(B)

(C)

Fig. 2: masked displacements along Line-Of-Sight for the three time intervals.  A) 05/01->13 time interval; B) 05/13->25 time interval; C) 05/25->06/06 time interval. The central E-W line is the profile trace used to derive Fig. 3. Google satellite basemap. Created with QGIS.


We can see that in the first time interval (05/01->13) the regional displacements are mainly negative, in the second (05/13->25) positive with a clear E-W gradient, while the situation is more nuanced in the third one (05/13->06/06), without a clear prevalence of positive or negative values.

A profile of the three displacement rasters, along a E-W direction (red line in Fig. 2), was created using the 'gst' Python module (https://gitlab.com/alberese/gst), as described in the 'Trino' Jupyter notebook (https://gitlab.com/alberese/gst/-/blob/master/docs/interferometry/Trino.ipynb).

Fig. 3: Profiles of the masked displacement rasters along the trace in Figs 2A-C. The first period (I) is in red, the second (II) in green and the third (III) in blue. Created with Python modules matplotlib and gst.

The profiles depict the same situation as visible in map, and evidence, near to rightmost profile end, in correspondence of Vigevano town, a local maximum for the first period while the second period presents a local minimum and in the third period there is a flat (Fig. 3).  

It is difficult to consider these variations, both regional and local (Figs. 2 and 3) as related to topographic surface movements, since no constant trend is observed in the three periods, but the displacements during the second period somewhat reverse those of the first periods. The observed results are therefore interpreted as the results of atmospheric effects, particularly strong in the 05/13 image, that impacted with opposite results the displacements of the first (05/01->13) and second period (05/13->25). The results of the third period are more similar to those expected for a no-movement differential interferogram and possibly reflect only local atmospheric effects for both the 05/25/2020 and 06/06/2020 images.