Analysis of the surface hydrology of the Greenland Ice Sheet using Sentinel-1 Synthetic Aperture Radar
Surface meltwater on the Greenland Ice Sheet (GrIS) has recently received much attention due to its ability to pond to form features such as firn slush zones and supraglacial lakes. This study uses imagery from the recently launched Sentinel-1 Synthetic Aperture Radar sensors to investigate supragla...
Main Author: | |
---|---|
Format: | Master Thesis |
Language: | English |
Published: |
Scott Polar Research Institute
2016
|
Subjects: | |
Online Access: | https://doi.org/10.17863/CAM.8755 https://www.repository.cam.ac.uk/handle/1810/263413 |
Summary: | Surface meltwater on the Greenland Ice Sheet (GrIS) has recently received much attention due to its ability to pond to form features such as firn slush zones and supraglacial lakes. This study uses imagery from the recently launched Sentinel-1 Synthetic Aperture Radar sensors to investigate supraglacial lakes in southwest Greenland. This novel method can overcome some of the limitations of optical sensors that are more commonly used for investigations of the surface hydrology of the GrIS, including an inability to image through cloud or in darkness, without compromising either the high temporal or spatial resolution of the imagery. A semi-automated algorithm is developed to detect supraglacial lakes from the Sentinel images during the summer of 2015, which is relatively successful when compared to Landsat NDWI classification of supraglacial lakes. The lake detections are then utilised in a fully-automated rapid lake detection algorithm to investigate supraglacial lake dynamics through the summer (June to September). A total of 48 drainages were detected from the Sentinel data over the ~42,000 km2 study site for lakes greater than 0.2 km2. The analysis is then extended into the early winter months (October and November), where the Sentinel imagery allows the accurate determination of the date of lake freeze-over; the mean freeze-over date for 300 lakes is DoY 240.1 ? 3.7 days from the Sentinel imagery, which is beyond the limit of useful Landsat scenes available for 2015. Furthermore, the Sentinel imagery has allowed the detection of buried lake areas, as well as the date of buried lake appearance and disappearance, which has not previously been carried out in the literature. |
---|