Day 4

Detailed paper information

Back to list

Paper title Water Level Changes in Small Lakes Using The accumulated Phase Change of Coherent Pixels from DInSAR
  1. Saeid Aminjafari Stockholm University Speaker
  2. Fernando Jaramillo Stockholm University Baltic Sea Centre
Form of presentation Poster
  • A7. Hydrology and Water Cycle
    • A7.01 Inland Water Storage and Runoff: Modeling, In Situ Data and Remote Sensing
Abstract text Monitoring water levels can help hydrological modeling, predict hydrological responses to climatic and anthropogenic changes, and ultimately contribute to environmental protection and restoration. However, measuring lake water levels is easier said than done. The conventional ground-based gauges are now scarce due to limited accessibility, high cost, the labor needed for continuous maintenance, and required security and oversight of equipment. Although satellite altimetry is a standard tool for water level change detection in lakes worldwide, the newest sensors still have limitations regarding coarse temporal and spatial resolution and re-tracking errors from the backscattered signal from non-water surfaces. Changes in water levels can also be retrieved from Differential Interferometric Synthetic Aperture Radar (DInSAR) by measuring the phase change between two Satellite Radar images, but these changes are relative in space and difficult to unwrap.
Here, we develop a new methodology to estimate the absolute water level changes in the only 30 small Northern-latitude lakes gauged in Sweden. Sweden has more than 100,000 lakes covering 9% of the country’s surface area. We aim to evaluate the capability of InSAR in estimating absolute water level changes of lakes in latitudes beyond 55 degrees without the need of unwrapping the phase component, as it is usually done for InSAR studies over water surfaces. With the constraint of a very short temporal baseline (6 days) between pair Sentinel-1 SAR images, we deal with the phase jump in interferograms resulting from sudden changes in water level, and instead of unwrapping each interferogram, we accumulate the phase change of successive pair images across nine months in 2019. We chose only pixels inside the lakes’ surface area that exhibit a steady, coherent behavior across all interferograms and identified the pixels where the DInSAR and gauged estimates of water level change show high linear correlation coefficients (R2 > 0.8). We found lakes with many pixels showing a high correlation, suggesting the capability of DInSAR to determine the direction of water level change in these lakes. The highest correlation between the accumulated phase change and the gauged water level was observed in a pixel on Lake Båven, southeast Sweden (R2 > 0.97), and the lowest correlation was observed in a pixel on Lake Lillglän in the west of the country (R2 > 0.26). The pixels with a high correlation between the accumulated phase change and the gauged water level were located along the lake shorelines, surrounded by forest and wetland land covers. Surprisingly, features on these shores can still enable the double bounce of the radar signal necessary for the interferometric technique, allowing the retrieval of water level change. The high correlation in these pixels shows that the accumulated phase change of the Sentinel-1 twin satellites can help detect the trends of water level change in high latitude lakes surrounded by marsh-dominated wetlands and forests or other shoreline features.