Research Of Reduction Calculating Technology And Method Of Runoff

China is a country with a relatively poverty water resource, the per capita water resource is just 2185 cubic meters at 2004; it is a quarter of world per capita water resource. It estimates that the population of china will be one thousand and six hundred million at 2030, and then the per capita water resource will be low to 1760 cubic meters. The poverty water resource is a severe test which china has to face. How to distribute and use the water resource reasonablely, and then make it play an active role in the economic development is very important. The accurate appraisals of water resource are the precondition to exploitation water resource and keep it sustainable development. One of the important content of water resource appraisal is the quantity appraisal. As the rapid development of the national economy, human build a mount of hydraulic engineering, which change the nature condition of the runoff and the actual measurement of the runoff couldn't reflect the nature station. If we want to get the accurate amount of the water resource, we have to do the reduction calculation for the runoff, the accuracy of which direct impact that whether the hydraulic engineering could be planning reasonablely and whether the water resource could be used reasonablely.At present, most of the reduction calculating methods of runoff are based on the traditional methods; the principle of such method is simple and easy to understand, so some experts are interested of these methods. Whereas, as the development of the economy and urban, the influence factor of human is increase, which made the conditions of the catchments more complex. The traditional methods couldn't adapt the request. How to find out a new method to make it go up with the date is a new topic in the water resource appraisal.The research is based on the Pingshang reservoir in Shanxi. Through the research of the data, we build the traditional model first: Mass balance modeling and Rainfall-Runoff model. The Rainfall-Runoff model couldn't simulate the nonlinear relationship perfectly between the annual runoff and its factors of influence. So we establish the new model: the Genetic BP Neural Networks. In addition, the forecast model based on the wavelet decomposition is introduced in this paper, which gained more exact result of the reduction calculation of runoff based on existing data.The following results are acquired though the research: contrast with the Rainfall-Runoff model, the Genetic BP Neural Networks is nonlinear, which is concordance with the hydrology system, so it is more applicability, and the calculation precision is improved; the forecast model based on the wavelet decomposition can find the nature rule at different dimension, and the calculating precision is higher than the other models. Based on the research of different models, we develop software for the reduction calculation of runoff with high use value and actual significance, which can improve the hydrologist's work efficiency, reduce the labor intensity and shorten the assessment time.The research enriches the technology methods in the reduction calculation of runoff, and it has the reference value of practical use.