The generation of gridded ocean climatologies from irregularly distributed measurements

Current developments in range-dependent acoustic propagation and forecasting ocean-atmosphere models will necessitate the provision of high-resolution gridded climatological data. The generation of representative, fully gridded ocean acoustic climatologies from disparate sets of measurements poses a number of significant problems, not least because of the diversity and volume of the data involved. This paper describes the development of set of methods to support such a process. The approach adopted uses a novel two-stage parametrization process from individual profiles and interpolation/extrapolation based on the parameters rather than the raw observation data. Due consideration is also given to statistical measures of the data and preprocessing to remove suspect values