Key Technologies And Realization Of Early Warning DSS For Snowmelt Flood In Xinjiang

The Northern of Tianshan mountains in Xinjiang Uigur Autonomous Region is a snow-cover-distributed zones with frequent spring snowmelt flood and snow disaster. In spring, the snow melts intensively and causes snowmelt flood easily, which affected the agriculture seriously and livestock production and impacted on the living of the people in this region, meanwhile, brought serious economic losses. The early warning DSS(Decision Support Systems) for snowmelt flood in Xinjiang is an important non-engineering measure on flood control and disaster reduction. This DSS will improve the scientificity of early warning for snowmelt flood and decision making effectively, it also will take full advantage of hydraulic engineering to get the best benefit.This paper focus on research of key technologies and application in the early warning DSS for snowmelt flood, and the author put forward some theories and the approaches of their realization. At the same time, the early warning DSS for snowmelt flood in Xinjiang is established. The primary contents and conclusions are as the following:(1) Based on the structure of B/S(Brower/Server) and J2EE, the general frame of the early warning DSS for snowmelt flood is designed. The development environment and programming way in the DSS is put forward. And the application of J2EE technology in the DSS is realized. It makes the system flexible, expandable, reliable, easy to develop and maintain.(2) After the analysis of the classification and characteristics of the knowledge in the DSS, an object-oriented method which is representation of the knowledge is presented. It can express the case-knowledge and rule-knowledge in the system. The general frame of knowledge-base system of the DSS is designed, and both knowledge-base and knowledge-base management system are built, which solve the problem of sharing and reusing of the knowledge. Thus, the scientific level of the early warning DSS is enhanced dramatically.(3) The parametric elements theory is applied to the research on the models and parameters, and the mature RDBMS(Relational Database Manage System) is integrated with model-base. The relational model-base management system based on parametric elements is presented.(4) A data interface for the open and unified standard parameter in model-base is designed. By means of the object-oriented technique and Java program language, the DSS models and other common models are developed in form of Java class, which realize the reusability, independence, integrity and extendibility of these models. Besides, the problems, the mutual independence between the model and its parameters, the mutual independence between the model-base and database, and the sharing of the parameters, are solved. So the model-base management system can control and manage the selection of the models, combination of the models, and run of the model arbitrarily.(5) By means of the management, which exchanges the data between model-parameter-files and records of model-parameter-relational-table and provides the download and upload of model-parameter, it achieves the aim of user controlling parameter data in real-time. By the help of parameter-dictionary, it achieves the purpose of the user comprehending the model-parameter on time. (6) By the model-factory technique the establishment of the model-object is separated from the application of the model in order to get the remote call of the model in distributed environment.(7) Both DEM(Digital Elevation Model) data format and GeoTIFF(Geographic Tagged Image File Format) file data format is studied, the DSS can fulfill the space analysis of the hydrological character information of a basin independently without supports of GIS software.(8) The distributed snowmelt runoff model in Xinjiang based on DEM is studied. The spatiotemporal dynamic process of the snowmelt flood can be simulated, especially for flow generation and concentration, which can provide the information of decision making for the early warning of snowmelt flood.(9) The application cases of the DSS in the typical research region are offered from the view of practice, which can provide scientific reference for the decision making of early warning about snowmelt flood in Xinjing. At the same time, the key technology is applied and verified in this paper. (10) Different from the traditional DSS, under the structure of B/S it can be managed and run based on Web in the DSS.In summary, research and application of the early warning DSS for snowmelt flood are still blank in China. The research of both model-base management system and knowledge-base management system of DSS is just beginning. This research and successful application of the DSS can fill the blank in this field.