Study On Dynamic Warning Method Of Mountain Torrent Disaster In Ungauged Small Basin

Mountain torrents disasters have the characteristics of strong sudden and great destructive power,which often cause casualties and destroy farmland and houses.Dynamic warning method of mountain torrents is an important non engineering measure to reduce the loss of mountain flood disaster.As an important indicator of mountain flood early warning,dynamic critical rainfall is mainly used in basins with long-term and complete hydrological data.It fully considers the impact of early soil moisture content,dynamic changes of cumulative rainfall and rainfall intensity on early warning indicators.However,the randomness of rainfall process and the correlation between rainfall and rainfall duration are often ignored in the determination of dynamic critical rainfall,which resulting in the uncertainty of critical rainfall.In addition,the non-uniformity of rainfall time history distribution has a great impact on the dynamic critical rainfall.However,the uniformity of rainfall time history distribution is less considered in the existing research of mountain torrent dynamic early warning.Most small basins in hilly areas lack measured rainfall and flood data,which can not provide the measured rainfall process for critical rainfall calculation,and can not calibrate the parameters through the rainfall flood process.Therefore,in this paper,the measured rainfall process in the Xiagushan basin is used as the rainfall event sample to estimate the rainfall characteristics of Zhongtang basin,and the random rainfall process is generated according to the principle of probability transfer.Then,taking the Xiagushan basin and Lixin basin as the reference basins,taking Zhongtang basin as the target basin,this paper studies the parameter regionalization.On this basis,calculating dynamic critical rainfall,carrying out the research on the dynamic early warning method of mountain flood disaster in small basin lacking data.The main research contents and achievements of this paper are as follows:(1)Applicability research of HEC-HMS model and its parameter regionalization.In this paper,the spatial data such as Digital Elevation Mode,land use and soil type were processed to obtain the basin digital model.The Soil Conservation Service Curve Number loss method was selected for runof volume computation,the Soil Conservation Service Curve Unit Hydrograph method was used for transforming rainfall excess to runoff,and the Muskingum model was used for channel fow.Then calibrating the model parameters of Xiagushan basin and Lixin basin,conduct parameter regionalization research,and taking Zhongtang basin as an example to verify.The results show that HEC-HMS has strong applicability in small basins in Henan Province.The regionalization scheme obtained by the Xiagushan basin which is close to Zhongtang basin and has similar underlying surface characteristics,is more suitable for Zhongtang basin.(2)Stochastic simulation of rainfall process in ungauged areas.In this paper,it is assumed that Zhongtang basin is a ungauged basin.Because Xiagushan basin is adjacent to Zhongtang basin and has the same hydrological and climatic characteristics,the historical rainfall process of the Xiagushan basin were taken as the rainfall event sample to estimate the rainfall characteristics of Zhongtang basin,and random rainfall events are generated according to the principle of probability transmission.This paper used Copula function to establish the joint distribution of rainfall depth and rainfall duration,randomly generated the combination of rainfall depth and rainfall duration according to its joint distribution function,then verified whether the distribution is consistent with the original rainfall samples and the characteristics of historical rainfall events in Zhongtang basin.The results show that it can be considered that the depth and duration of randomly generated rainfall are from the same population as the characteristics of original rainfall samples,and it can be considered that the rainfall depth and duration of rainfall generated by rainfall events of Xiagushan basin are consistent with the distribution of rainfall characteristics of Zhongtang basin.Next,this paper dealt with the rainfall events dimensionless,established the rainfall information matrix I,determined the distribution function of rainfall proportion in each dimensionless period,and generated random rainfall patterns.Finally,synthesized the rainfall process with rainfall depth and rainfall duration,and constructed the rainfall data set,which provides the rainfall data basis for the calculation of dynamic critical rainfall.(3)Study on dynamic early warning method of mountain torrents.In this paper,the Gini coefficient was used to characterize the uneven degree of random rainfall time history distribution,the spatial distribution characteristics of Zhongtang basin were statistically analyzed,the rainfall distribution proportion of each station was set according to its distribution characteristics,and the random rainfall process of each station was determined.The random rainfall process of each station was substituted into the model for trial calculation,the data combination of critical rainfall and influencing factors(moisture content,cumulative rainfall,rainfall intensity and Gini coefficient)was obtained.It is found that the greater the soil moisture content,cumulative rainfall and rainfall intensity,the more uneven the rainfall time distribution and the smaller the critical rainfall.The multiple regression curve is obtained by fitting data combination of critical rainfall and influencing factors.The dynamic early warning method of mountain torrent was constructed and applied by combining the actual rainfall and predicted rainfall,comprehensively considering the factors such as soil moisture content and accumulated rainfall in the early stage.Finally,the method is applied to an example.The results show that the dynamic early warning method considering the soil water content,cumulative rainfall,rainfall intensity and rainfall time history distribution uniformity has strong applicability in Zhongtang basin,and can provide technical reference for mountain flood early warning.