Firstly,we describe operational requirement of typhoon forecast business. By the support of historical and real-time typhoon spatial database, the process flow of typhoon forecast product creation is expatiated. The first step is to visualize typhoon track and wind data, the second step called data editing is to smooth typhoon track, edit feature property and validate topology for spatial analysis, the following is data analysis which includes landing area analysis, real-time and future affected area analysis, and similar track search, the last is service product creation by using microsoft office development technology.
Secondly,four key issues of GIS application on production system are discussed in details. GIS spatial analysis aims at extracting and transporting spatial information based on geospatial object location and geometry character. Using Geodatabase(gdb) and Spatial Data Engine(SDE), historical and real-time typhoon database can be built and provide the ability of data record and update. ESRI representation symbolizes data using a flexible and rule-based structure that is stored inside the geodatabase along with data, features can support multiple representations simultaneously; constructing feature filling modes and color representation styles by standard rules can efficiently solve large volume of typhoon data visualization. In order to improve the accuracy and efficiency of typhoon forecast, two similar track searching methods are proposed. One is key area searching which can dynamically customize multiple spatial areas and search the meet historical typhoon tracks. The other is largest curvature method, this method can calculate curvature of the typhoon track and select the largest numbers, around those key points do buffer analysis to search historical data. The paper also proposes two probability ellipse methods of typhoon track forecasting to predict potentially infected districts ahead of schedule. spatial interpolation method firstly densify the typhoon forecast points and calculate each point's weighed error radius, then make the vertical line throughout the densified points, finally acquire and link the control points to construct a polygon and do Bessel smooth for this polygon. Dynamic buffer method based on integral splits the typhoon track by kilometer unit, by the ratio of the length of typhoon track the buffer radius can be calculated, after then do dynamic buffer based on a series of radiuses and union the buffer area.
Finally, we point out that GIS provides a reliable scientific basis for typhoon forecast, and we hope to achieve more applied and efficient typhoon forecast models in the near future.
Key Words: GIS; Typhoon forecast; Representation; Similar track searching; Dynamic buffer; Probability ellipse
Supplementary URL: