To achieve the basic idea in the weather services producing system, some key issues have been solved respectively as follows.
Firstly, some imperative and crucial tools related weather mapping are prior developed by using ArcGIS Geoprocessing interface, since contemporary geoprocessing procedure only focuses on data transformation and has no capacity of implementing the mapping process. We develop an “Export Picture” tool and an “Update Legend” tool to extend the ability of geoprocessing greatly. The “Export picture” tool allows user to produce a number of weather maps with input parameters, for instance, GIF, BMP, and SVG format as well as output resolutions of picture, while the “Update Legend” tool ensures the picture with an updated legend showing the display style, which is very important for public service products.
Another issue must be tackled is that we should make the MICAPS (a very popular forecast system software widely used for decades in china meteorology division) format data file be shared with GIS community to be further processed, thus we propose a proper method to convert meteorological data to GIS format complying with GIS concept data model. The most important is to make sure that meteorological data are less loss information while converting to GIS format. Thus a data conversion tool is also developed for accessing meteorological data source. Meanwhile, some other important tools are also well designed and developed to build complex models for making services products successfully, such as accessing Oracle database tool, ftp transmission tool and layer dynamical rendering tool.
When users are capable of using Model Builder application (a tool of ArcGIS Desktop) to design the complex models to output services products, chaining with these extension tools together with basic available geoprocessing tools, thus execution of these models conveniently, timely and automatically become another most important issue. We proposed a distributed architecture comprised of server and client application. Tasks can be scheduled by a light level client interface program, including setting of parameters of model and running time, and so on. A task can be added, modified, and deleted for multi-user management if here has installed several client terminals, but the tasks setting information is stored in central relationship database of the server machine. The server is responsible for models running centrally which requires the GIS environment. The application of server uses a timer to find whether a task should be executed from the central database. When it is the time for a task's turn, the task can be invoked timely by the execution engine deployed on Sever, and the running status of the task is also recorded in the database for further monitoring. The design of parallel running pattern is induced to allow multi-users to run models at the same time, therefore it improves the application recurrence and efficiency. It also allows users to execute a task by client program, when this situation occurs, the message is sent to the server immediately, and the server engine is invoked to repose the remote calling, then return the execution information to the client as a result. The distributed architecture has greatly enhanced the sharing and scalability.
With the end of this paper, some running instances and examples are also given with pictures and relevant model diagrams. At present, the operational system using GIS Geoprocessing works well and stably in National Meteorological Center, China, and products with high quality of visualization and rich representation have been available over the Internet websites (for instance: http://www.weather.com.cn) and decision-making reports as well.
Keywords: Weather Services Products; Mapping Tool; GIS; Geoprocessing; Distributed Architecture.
Supplementary URL: