Optimization Of SCS-CN Model Parameters About Runoff Prediction In Mountainous Regions

In mountainous regions,because of its unique topography,the process of rainfall production and confluence is more complicated,and the potential water environment problems can not be underestimated.Therefore,it is of great significance to establish a simple and effective surface runoff prediction model to solve the problem of waterlogging,strengthen the rain and flood management and construct sponge city.SCS-CN model proposed by NRCS,it is widely used because of its reasonable basic assumptions,simple calculation method and easy access to the required data.However,the model is not suitable for all regions,so the parameters of model need to be optimized to improve the prediction accuracy.Taking the typical mountain city——Chongqing as an example,this paper selects the undeveloped area of Banan district as the research object,sets up three land use methods:bare land,permeable pavement and grassland in combination with the construction measures of sponge city,and uses artificial simulated rainfall experiment to obtain the rainfall-runoff series data under different underlying surface conditions,then analyzes the model parameters and the sensitivity of each influencing factor,and uses the standard SCS-CN model and its improved model to determine and verify the parameter rate,and evaluates and compares to select the most suitable parameter optimization method.The main conclusions are as follows:(1)Analysis of the correlation and sensitivity between the model parameters shows that the conditions of the underlying surface are the same:surface runoff is negatively and positively correlated with?and CN values;when rainfall is constant,the larger the CN value,the less sensitive the predicted runoff and?are to changes in the CN value;the constant CN,changes the rainfall,the predicted amount of runoff decreases with increasing?,and?is more sensitive to rainfall during small rainfall events.It can be seen that the model parameters will affect each other and directly or indirectly affect the predicted runoff value.(2)Discuss the influence of other factors on runoff prediction:when rainfall is consistent,predicted runoff error will increase with the increase of rainfall intensity and rainfall duration;surface runoff will increase when the slope is increasing,and the impact of slope on large rainfall events is much greater than that of small and medium-sized rainfall events;when the soil water content in the early stage is high,the corresponding calculated runoff value will be too large.In addition,the average runoff coefficients of the three land use methods are:bare land>permeable brick>grassland.So that when other conditions are the same,the output and velocity of runoff have the same rules.In summary,these impact factors need to be considered when using models to predict surface runoff.(3)This study discussed a total of 8 models that are the standard SCS-CN model and its improved models.At the stage of model calibration,the results show that most of the calculated values of the models are lower than the measured values,and the conditions of different underlying surfaces will be different.Among them,the prediction accuracy of the W1 model(standard SCS-CN method)is the worst.In the improved models,the fitting degree and model evaluation index of the W7(step size method optimization parameters based on the previous soil moisture content to estimate the initial loss)and W8(use the huang slope formula to modify the CN value based on W7)is better than other models.(4)Comprehensive comparison of the degree of linear fitting and the model evaluation parameters in the period of model calibration and verification,and the parameter optimization methods applicable to the three underlying surface are:bare soil underlying surface:?=0.01(step method),CN takes the value of CN₃from the standard look-up table value and the standard conversion formula;permeable pavement underlay:optimize the parameters?and CN using the step smethod to obtain the optimal value?=0.01,CN₃=94.64;green roof underlay:based on the inverse calculation of the measured rainfall runoff data.The?value(0.03)is the most suitable for the optimal CN₃value(89.6)obtained by the step size method.In addition,on the basis of selecting the optimal parameters,the initial loss of soil is estimated by considering the previous soil water content,and the CN value is corrected by the Huang slope correction formula,it can improve the accuracy of model prediction to a large extent for three kinds of underlay conditions.