| The flood forecast in the medium and small basins has always been a difficult problem in front of hydrologists and researchers,which is also of great significance to carry on flood forecasting timely,effectively and accurately for the national livelihood safety and economic development.Studying the physical characteristics on the formation of floods in depth,exploring the relationship between the formation of floods and the three elements of hydrology(rainfall,flow and water level),and taking the advantages of the integration from various subjects such as computer science,physics and geographical information science,will be a key point for the breakthrough of flood forecasting in both theory and method.This study,focusing on the three elements of hydrology,takes a further research in auxiliary analysis of flood control based on the achievements from one of the key social development projects of Major science and technology projects of Zhejiang province,which is a system research and application of auxiliary analysis for flood monitoring,flood control,as well as disaster relief on both the satellite-based observations and ground measurements.The main research contents and related conclusions are as follows:(1)The precipitation data from rain gauges are of high quality,however,it is difficult to characterize the regional rainfall information due to the limited number of sites.While the precipitation information from satellite-based remote sensing observation,with coarse spatial resolution and low accuracy,could provide large area rainfall,synchronously.This study,considering the environment factors with high spatial resolution such as NDVI,DEM and LST,has applied the information from environmental variables with finer resolution to downscale the TRMM data from 25 km to 1 km by using Geographical Weighted Regression algorithm.As a result,the correlation coefficient between average annual TRMM and ground observation was increased to 0.77 from 0.64.(2)The traditional estimation of runoff,mostly,based on empirical functions,which is difficult to explain the mechanisms by the basic principles of physics and hydrology etc.In this paper,the SWAT distributed water model was used to simulate the runoff at Xinsheng basin,on the upper area of the Cao'e River.After the sensitivity analysis and calibration of the model parameters,the value of the ENs and the certainty coefficient R~2 was up to 0.847 and 0.901,respectively,completely meeting the accuracy requirements of SWAT.(3)A fuzzy clustering algorithm was introduced to conduct fuzzy matching water level of the real-time flood from historical flood observation.Firstly,all of historical hydrological data from multiple years over this study area were stored in database after reorganization.Secondly,the representation information of historical floods were excavated and analyzed,and then were clustered by using fuzzy algorithm.Finally,the most similar record,in the history,to the current flood would be selected and provide a reference for forecast analysis on the trend of the current water level.(4)Based on the research on the mechanisms of flooding formation,such as area rainfall estimations by remote sensing observation,stimulation of runoff and data mining on historical hydrological records,we developed a set of decision system with auxiliary analysis for flood control on WebGIS platform by using HTML5,GeoServer,and other mainstream open source technologies,as well as the modular development methods.This system,with a beautiful interface,good utility and simple operation,provides a great convenience for the relevant hydrological workers to carry out support analysis work on flood control. |
PDF Full Text Request