The Research And Implementation Of The Integrated Processing Platform For Numeric Marine Forecast Data

The movement of the atmosphere and the marine is restricted to some physical laws including hydrodynamics and the thermodynamic etc. With given boundary limits, and solving these equations by numeric methods, analysis the future situation from the current, this process is called numeric simulation. The numeric marine forecast is to predict the wind, wave or current by numeric method. Now the numeric simulation is widely applied in atmosphere and marine forecast. Some common models include WRF, WW3, SWAN and FVCOM, etc. The products from these models are stored in NetCDF files or can be converted to NetCDF files by easy way.Due to the flexibility of NetCDF format, the storage of these data shows unique character from each other. For example, use different dimension name, variable name or logic data structure. The reading and writing method will even be different when dealing with files which all stored in NetCDF formats. This personality resulted in great flaw while implementing these data into GIS platform:the system dealing with one kind of NetCDF data cannot handle another, it has to reconstruct the program to recognize a product from another model. For example, the East China Sea Forecast Center has developed a bunch software holding similar function but different data source. This style of GIS software developing causes huge waste in labor and computing resource.In order to enhance the efficiency of using the simulation products, this dissertation explores the convention in storing numeric marine forecast data in NetCDF files. Unify variety of data arrangement methods in a single structure. Based on this convention, an integrated data processing platform for numeric marine forecast data has been developed. The products of numeric simulation can be displayed with the help of this platform. A parallel computing platform on VMware virtualization platform has also been applied in order to accelerate the processing speed.The main achievements in this dissertation include: (1) The design of convention for the storing of numeric marine forecast data.On summarizing common data structure in storing numeric forecast data in NetCDF files, this dissertation present a convention for the standardization of storing numeric marine forecast data from different models in a single structure.(2) Provide a method to convert numeric forecast data to GIS compatible data.Data stored in NetCDF files cannot always be recognized by GIS software. This dissertation explores the method of convert standardized NetCDF files in to GIS compatible data (including raster and vector, for example, line and point), and so to make them can be displayed in GIS systems.(3) Accelerate the processing speed with parallel computing and virtualization.The processing and generating GIS data from NetCDF files is time costly due to some method of ArcGIS cannot be paralleled in a single computer. In this dissertation, the virtualization technology and an algorithm of parallel computing on virtual machines is applied to shorten the time needed for data conversion.(4) Develop the "Integrated Processing Platform for Numeric Marine Forecast Data"The Integrated Processing Platform for Numeric Marine Forecast Data is developed based on the method mentioned above. And this system provides data services for other GIS systems such as the precise forecast system of Oufeitan tidal flat.