Sea Ice and Ocean Environmental Applications of Spaceborn SAR
Doctoral thesis, 2000
The worldwide focus on our changing environment has led to an increased need to observe and characterise a range of environmental phenomena. The Arctic Ocean is particularly sensitive to climate variations. With satellite borne synthetic aperture radar (SAR) satellites we can identify, observe and characterise Arctic sea ice parameters in a climate context with better accuracy than conventional systems. Algorithms for automatic identification and characterisation of sea ice distribution, circulation, and meso-scale features have been developed and validated, when possible, with in-situ ground-truth measurements collected during the Arctic Ocean Expedition 1996. Results have been used to validate and guide developments of Arctic meso- and large-scale numerical models and low spatial resolution passive microwave satellite ice concentration algorithms.
The application of large swath SAR data to map sea ice distribution and drift shows that there occasionally exist large areas of comparatively low ice concentration (75-95%) in the high Arctic. Large swath SAR (i.e. ScanSAR) images can today cover several hundred numerical model grid points with ice retrieval accuracies significantly better than for the passive microwave sensors. Large discrepancies exist between different Arctic numerical models, and between SAR ice concentration retrievals and models. Large discrepancies also exist between the different conventional SSM/I ice concentration algorithm retrievals, and between SAR and SSM/I retrievals (up to 35 %). For mesoscale features that have a significant impact on the Arctic climate, SAR retrievals show similar correlation to a mesoscale model as to an SSM/I sub pixel model. The culmination of these investigations shows the value and the need for careful consideration of the observation retrieval accuracy. Arctic numerical model (and passive microwave ice concentration algorithm) validation and assimilation with high-resolution SAR observations is highly necessary, and must be done more extensively.
New exciting scientific satellite prospects are being development with multi polarisation (linear and quad), multi resolution (3-1000m), and multi frequency SAR systems. Studies of new missions, selection of imaging modes of operation, and new image products show that the spatial and temporal resolution, and the accuracy of Arctic ice parameter retrievals, will be significantly improved in the near future.
sea ice applications