Rainfall Threshold For Initiation Of Debris Flows In A Small Catchment And Siphon Drainage Prevention Method

Excessive water in a channel plays an important role in triggering debris flows,especially for small catchments,due to the features of a small area,short channels.In small catchments,intense rainfall can trigger a flood within a short period of time.In the southeastern coastal areas of China,many typhoon rainstorm will happen in summer which trigger debris flows in lots of small catchments.In this context,short-duration rainfall events were analyzed,and field measurements were made for a partially-modified landscape in order to compare the variation in the amount of runoff.Two rainfall threshold models for initiation of debris flow are proposed.We also propose a new approach for the separation of water and sediment and drainage in order to drain water away from channels and reduce the potential for debris flows.The main advances are as follows.(1)we collected real-time rainfall and runoff data in channels.The monitoring results indicate that the runoff and rainfall had the same fluctuation trends.the characteristics of the rainfall determine generation ability of runoff.We choose the rainfall depth(P),average intensity IAVE,maximum 30-min intensity(I30),maximum 10-min intensity(I10),the initial runoff(Qint)and the duration of the rainfall event(D)were selected as the characteristics of the rainfall.The data were analyzed using grey relational analysis to confirm the main characteristics of the rainfall that influenced the runoff.The result shows the first-ranked characteristic was maximum 30-min intensity.(2)There is a non-linear relationship between value of rainfall characteristics and runoff generated ability.the rainfall characteristics with high values of grey relational grade,namely precipitation depth,average intensity,maximum 30-min intensity,maximum 10-min intensity,and initial runoff,were selected as input factors.Then Three non-linear models,i.e.,the Moving Least Squares(MLS)model,the Genetic Support Vector Machine(GASVM),and the Genetic Back-Propagation neural network(GABP)model were considered for forecasting variations in the runoff.The results show the MLS models performed better than the GABP and GASVM model.At same time,the peak discharge was obtained by using the NAM model.The results show that the NAM model can provided accurate results.The NAM model can be used to predict the runoff value.(3)Two rainfall threshold models for initiation of debris flow are proposed accounting for the characteristics of underlying land surface.The two models are critical discharge model and critical depth model,respectively.The rainfall threshold was determined for debris flows associated with a 2-year return period by both the two models.This consequence squares with the facts.Contrast the two predicted models,the critical depth model not only can predict the occurrence of the debris flows,but also can calculate the intensity and influenced area of debris flow.(4)A new prediction method of debris flow average velocity is proposed by using MLS model.The results indicate MLS model possesses the good ability of self-learing and self-organization.The model also has capacity of catching the subtle variation of the influencing factors and presenting different functions when influencing factors are different.All the advantages can result in the higher precision of prediction.(5)we propose a new approach for the separation of water and sediment and drainage in order to drain water away from channels by siphon drainage and reduce the potential for debris flows.A field experiment was also developed and tested in a small catchment in which debris flow was an issue.The monitoring results showed the siphon drainage system can reduce the peak discharge effectively.The distribution of the maximum flow depth calculated by FLO-2D model shows the depth of the surface flow decrease significantly after separation.To a certain extent,the siphon drainage pool can be effective in preventing the occurrence of a debris flow.