Sea ice is of great interest due to its effect on the global climate, the Earth's ecosystem, and human activities. Microwave remote sensing has proven to be an effective way to measure many of the characteristics of sea ice. In particular, several algorithms map the daily sea ice extent using a variety of instruments. Enhanced resolution images generated from the Scatterometer Image Reconstruction (SIR) algorithm can be used to generate a high resolution ice extent map. Previous algorithms using SIR images were developed for scatterometers which are no longer operational. The Advanced Scatterometer (ASCAT) is a newer scatterometer which has different characteristics from the earlier scatterometers. The previous algorithms do not perform as well when applied to ASCAT. This thesis presents a new algorithm for ASCAT developed to discriminate sea ice from the open ocean and create daily maps of the ice extent. It is developed from previous algorithms used on earlier scatterometers. The algorithm uses an iterative Bayes decision rule to classify pixels as sea ice or ocean. Digital image processing techniques are used to reduce misclassifications. The ice maps from the new algorithm are compared with the NASA Team sea ice concentration maps from the Advanced Microwave Scanning Radiometer for the Earth Observing System (AMSR-E). The comparisons include: difference in area, distance between ice edges, number of missed and false detections. The new ice maps are also compared with the Remund-Long algorithm for the QuikSCAT satellite using the same metrics. The ice edge is verified with high resolution Synthetic Aperture Radar (SAR) data. The new ice maps perform similarly to previous ice mapping algorithms for scatterometers.



College and Department

Ira A. Fulton College of Engineering and Technology; Electrical and Computer Engineering



Date Submitted


Document Type





remote sensing, sea ice, scatterometer, pattern recognition, ASCAT