Progress on a Landsat 8 Image Mosaic of Antarctica and Early Applications

Abstract:

The polar regions, and the large ice sheets in particular, are quickly evolving harbingers of global change. Therefore, it is important that we measure and monitor the ice sheets in a consistent and repeatable manner over time. While kilometer-scale resolution sensors like MODIS and VIIRS can do this on a daily basis, the higher spatial and radiometric resolution of Landsat 8 supports a more quantitative measure of polar change over decameter spatial scales and weekly to seasonal timescales. To date, Landsat 8 has collected over two years of imagery of exceptional radiometric quality and geolocation accuracy, and with unprecedented acquisition rates for the poles. Building upon lessons learned from the Landsat 7 Antarctic mosaic (LIMA; Bindschadler et al., 2008, Remote Sensing of the Environment), this project harnesses the multispectral imaging capabilities of Landsat 8 with Google’s Earth Engine to produce time-series data sets for monitoring the remote, vast, polar ice sheets in a fraction of the time it would take to do so with traditional computational methods (or earlier manual efforts). We focus on building cloud-free, multi-temporal Landsat 8 mosaics (‘data cubes’) by applying a range of strategies (e.g., normalized indices, statistically based thresholds, image cross-correlation). These can then be used to quantify ice sheet surface and snow cover properties over time to study ice sheet change (e.g., ice sheet morphology, ice shelf extent, snow grain size, melt and melt pond extent, surface wind direction from drift orientation).