Study On Data Mining Technology's Application In Runoff Characteristics Analysis

Water resource was one of the most basic elements for supporting the all lives on the earth and realizing sustainable development the change of runoff factor played a leading role.And it had a profound effect on resource environment and regional economy development.The study of runoff evolution law was the prerequisite and foundation of rational exploitation and effective utilization of water resource.so it is important theoretical and practial sense for mining and identifying change law of river runoff.Traditional runoff evolution law of commonly used statistics and hydrological model simulation method to carry on the research.In recent years,the impact of global climate change and human activities,the runoff of the spatio-temporal evolution showed a lot of uncertainty,the traditional analysis method has some limitations,has been unable to meet the demand.River basin management agencies collect each department has accumulated massive business and monitoring data,the data with large quantity,wide,heterogeneous and distributed features and how to use these historical data,analyzes the evolution law of runoff,the runoff in the long-term prediction becomes a problem need to be solved immediately.As a new technology,data mining has the advantages of timely and effective treatment of large data,and does not depend on the complexity of the mechanism.It is an effective means to realize the conversion of historical data to useful knowledge.In this paper,the author firstly studied the application of data mining method in the analysis of the evolution law of water cycle,and summarized the method system.Secondly,he treated the Yalong River DaLuo to Ertan interval as the study area,adopting remote sensing image,runoff,precipitation and evapotranspiration data as experiment data,using neural network,decision tree classification,regression analysis,correlation analysis and other data mining methods to analyze the various influence factors on the effects of watershed interval runoff.The results show that since the Ertan Reservoir vegetation in reservoir area as a whole gets better and better every year,precipitation in the study area as a whole is decreasing year by year,DaLuo discharged water reduction is direct cause of reduced runoff of Ertan,and precipitation is the main reason for the interval runoff reduction and interval production flow can be weakened.The main reason is that the underlying surface vegetation changed for the better and vegetation interception increases.In this paper,the main results are as follows: the author proposed a thick cloud shadow detection algorithm based on TM image by comparing the two image's spectral characteristics of clouds and cloud free images.The algorithm is constructed to detect thick cloud from TM image;in order to analyze the variation of the vegetation around the Ertan Reservoir,this paper used multi-source remote sensing image as experimental data and laid basis on the methods of decision tree classification and regression,and the author analyzed the spatial and temporal variation of the vegetation around the reservoir.