Data compression for 3-D oceanographic fields using empiricalorthogonal functions

摘要

Naval oceanographic and acoustic information systems need to becapable of handling and manipulating large quantities of environmentaldata. This is required to support the use of acoustic models in futuresonar applications, which require a complete 3D description of the oceanenvironment. Increasingly, use is being made of 3D numericaloceanographic forecast models to provide this information (e.g. FOAM, aglobal Forecasting Ocean Atmosphere Model which is being developed foroperational use by DRA and the UK Meteorological Office). However, thequantities of data that such models generate are significantly greaterthan those which can be handled by current naval communications systems.An efficient technique is therefore required for compressing 3Doceanographic data to enable it to be transmitted quickly and withlittle or no loss of accuracy to an at-sea user. The authors describe a3D data compression technique that uses empirical orthogonal functions(EOFs) and variable bit length encoding, to achieve high compression.The use of EOFs for data compression is not new. Clancy (1992) andGarther et al. (1991) describe the application of this technique in theUS Navy's Oceanographic Data Distribution System (NODDS). However, thepresent article gives new results on the accuracy of reconstructedenvironments and the impact of data compression on derived acousticfields