3:45 PM - 4:00 PM
[MTT37-07] InSAR tropospheric correction using dense GNSS tropospheric products
Tropospheric delay noise is one of the most significant sources of error in synthetic aperture radar interferometry (InSAR). In particular, it is often the major obstacle when one tries to detect mm-level small displacements of subsurface fault locking/slip and magma movement because of the similarity in the spatial wavelengths of the signals. In unfavorable atmospheric conditions, tropospheric delay noise can reach 10 centimeters or even more.
If a dense GNSS network is in operation in the target area, using GNSS tropospheric products is the best way to correct for the tropospheric noise in InSAR (Kinoshita, 2022). In the current situation in Japan where the data of the SoftBank GNSS network are provided to the member organizations of the Japanese CSESS consortium in addition to the publicly-available GEONET GNSS data of the Geospatial Authority of Japan, the GNSS-based noise correction method has much larger advantages over other methods such as the ones based on numerical weather models.
In this presentation, we report our initial results on the effectiveness of the SoftBank GNSS data for InSAR tropospheric correction. The test area is in eastern Shikoku around the Median Tectonic Line. Similar to the iterative tropospheric decomposition model (Yu et al., 2018), we modeled the tropospheric delay to be the sum of the functional and random components, where the former is composed of a bilinear ramp and topography-correlated terms and the latter is the spatial interpolation of the residual using a Gaussian kernel.
We applied the correction method to ionosphere-corrected InSAR displacement data processed from the ALOS-2 satellite imagery. We use images acquired over a short interval, so that we can assume the real crustal deformation signal is negligible compared to the tropospheric noise. In our example case, the standard deviation decreased from 10.5mm to 8.1mm (23% noise reduction) when we used both the tropospheric delay products of the GEONET F5 solution and that obtained from the SoftBank GNSS data processed by GipsyX.
(Acknowledgements) The SoftBank's GNSS observation data used in this study was provided by SoftBank Corp. and ALES Corp. through the framework of the "Consortium to utilize the SoftBank original reference sites for Earth and Space Science". The GEONET F5 tropospheric delay product was provided by Geospatial Spatial Authority of Japan. InSAR processing was conducted using the RINC software (Ozawa et al., 2016).
(Figure caption) (a) InSAR displacements obtained from the ALOS-2 data (descending orbit, Path 21, acquisition dates 2023.11.24 and 2024.1.19). (b,c,d) Tropospheric correction terms obtained from the GEONET F5 solution, from the SoftBank-GipsyX solution, and from the both, respectively. (f,g,h) InSAR displacements corrected with (b,c,d). (e) Elevation of the analyzed area. Red lines correspond to active faults. Units are in meters except for (e).
If a dense GNSS network is in operation in the target area, using GNSS tropospheric products is the best way to correct for the tropospheric noise in InSAR (Kinoshita, 2022). In the current situation in Japan where the data of the SoftBank GNSS network are provided to the member organizations of the Japanese CSESS consortium in addition to the publicly-available GEONET GNSS data of the Geospatial Authority of Japan, the GNSS-based noise correction method has much larger advantages over other methods such as the ones based on numerical weather models.
In this presentation, we report our initial results on the effectiveness of the SoftBank GNSS data for InSAR tropospheric correction. The test area is in eastern Shikoku around the Median Tectonic Line. Similar to the iterative tropospheric decomposition model (Yu et al., 2018), we modeled the tropospheric delay to be the sum of the functional and random components, where the former is composed of a bilinear ramp and topography-correlated terms and the latter is the spatial interpolation of the residual using a Gaussian kernel.
We applied the correction method to ionosphere-corrected InSAR displacement data processed from the ALOS-2 satellite imagery. We use images acquired over a short interval, so that we can assume the real crustal deformation signal is negligible compared to the tropospheric noise. In our example case, the standard deviation decreased from 10.5mm to 8.1mm (23% noise reduction) when we used both the tropospheric delay products of the GEONET F5 solution and that obtained from the SoftBank GNSS data processed by GipsyX.
(Acknowledgements) The SoftBank's GNSS observation data used in this study was provided by SoftBank Corp. and ALES Corp. through the framework of the "Consortium to utilize the SoftBank original reference sites for Earth and Space Science". The GEONET F5 tropospheric delay product was provided by Geospatial Spatial Authority of Japan. InSAR processing was conducted using the RINC software (Ozawa et al., 2016).
(Figure caption) (a) InSAR displacements obtained from the ALOS-2 data (descending orbit, Path 21, acquisition dates 2023.11.24 and 2024.1.19). (b,c,d) Tropospheric correction terms obtained from the GEONET F5 solution, from the SoftBank-GipsyX solution, and from the both, respectively. (f,g,h) InSAR displacements corrected with (b,c,d). (e) Elevation of the analyzed area. Red lines correspond to active faults. Units are in meters except for (e).