1. What factors affect temporal coherence in snow?
Temporal coherence in snow is affected by melting and wind. A high temporal coherence is observed at L-band for dry snow with a month temporal baseline. Temporal coherence decreases with increasing frequency. A median temporal coherence of about 0.5 is observed at 10.2 GHz and 16.8 GHz even after 60 days. Vegetation cover decreases the temporal coherence significantly at high frequencies. Temperature is the most critical variable affecting temporal coherence among other variables. SWE accumulation profile retrieval is successful for short temporal baselines and low frequencies in non-vegetated areas. The retrieval is poor using the 12-day repeat-pass data at C-band. 6-day repeat pass C-band data showed good performance for small SWE changes but poor performance for large SWE changes between the interferometric pairs due to phase ambiguity caused by large SWE change. More studies are needed to better model temporal coherence at different frequencies, different land types, and different environmental conditions. L-and C-band are desirable frequencies for differential interferometry due to low temporal coherence and low penetration depth at frequencies higher than 10 GHz.
read more
2. How is interferometric phase related to SWE change?
Leinss et al. (2015) established an approximation linking interferometric phase directly to SWE change. The approximation depends on incidence angles and snow density. Matzler's model (1987) is used for calculating in equation 1. The unwrapped phase is converted to SW E using equation 4. Temporal decorrelation is high at C-band, but a 6-day repeat time improves temporal coherence significantly over snow compared to the normal 12-day Sentinel-1 repeat time. Pixels with less than 0.35 temporal coherence are not considered reliable. Temperature is also a factor, with near surface air temperature above zero indicating wet snow. The reference point for calibrating unwrapped phase or SW E is either stable targets or the average of in situ SW E. Using the average of all in situ stations improves estimation results.
read more
3. How can tropospheric noise be mitigated in Interferometric Synthetic Aperture Radar (InSAR) measurements?
Tropospheric noise can be mitigated in InSAR measurements by using a global atmospheric weather model to predict the radar phase delay due to variations in atmospheric pressure and water vapor content. The European Center for Medium-Range Weather Forecasts (ECMWF) ERA5 model of atmospheric variables provides hourly estimates on a 30 km global grid based on assimilation of surface and satellite meteorological data. The Python-based Atmospheric Phase Screen (PyAPS) software is used to interpolate the grid and convert the variables into a radar phase delay. This processed data is then integrated into the Miami InSAR Time-series software in Python (MintPy) to crop the atmospheric delays to match the spatial extent of the interferograms and project the delays into radar line-of-sight (LOS). The retrieved SWE using Sentinel-1 interferometric phase is compared with in situ stations and LIDAR results, and any retrieved value with temporal coherence less than 0.35 and temperature higher than 0°C is discarded. The correlation and RMSE between the retrieved and in situ SW E are analyzed, and the results show a high correlation (0.82) with an RMSE of 0.76cm SWE. The correlation and RMSE between the entire time series of retrieved and in situ SW E for each station are also evaluated, with most stations showing a correlation greater than 0.6 and RMSE less than 2cm. The correlation and RMSE between the in situ stations and retrieved SW E for each Sentinel-1 acquisition date are also analyzed, with most dates showing a correlation greater than 0.4 and RMSE less than 1cm. The SW E ambiguity of Sentinel-1 is relatively small, making the unwrapping challenging for snow storms. However, among all 31 stations in the Sentinel-1 frame, 24 of them have a mean temporal coherence of more than 0.35, indicating that their Sentinel-1 data can be used for SWE estimation. The correlation between the retrieved and in situ SW E data sets is 0.62, demonstrating the effectiveness of using InSAR technique for SWE retrieval.
read more
4. How is SWE retrieval using InSAR spaceborne data promising for future missions?
SWE retrieval using InSAR spaceborne data is promising for future missions due to its high correlation with in situ values and LIDAR images, with less than 2cm RSME compared to in situ values in 16 stations. The method's temporal coherence, phase unwrapping, and phase ambiguity are the main constraints. However, the NASA-ISRO SAR mission (NISAR) launch can potentially improve the method's effectiveness by using lower frequencies like L-band, which enhances temporal coherence and reduces SWE ambiguity. This makes SWE retrieval using InSAR spaceborne data a viable candidate for future SWE missions.
read more