Study On The Application Of Instantaneous Unit Hydrograph Method To Flood Forecasting Of Fengle River Basin

Flood disaster is one of the most serious natural disasters, and flood forecast is an important non-engineering measure which can prevent flood disaster. Forecasting flood accurately based on the different characteristics of each basin and each rainfall has become a fatal issue which to be solved urgently. Basin flow characteristics in flood forecast model are different because of the difference of temporal and spatial characteristics of each rainfall in fixed basin, that is, the parameters n, K of every flood which used for runoff calculation based on Nash instantaneous unit hydrograph are different. Therefore, taking Fengle River basin as a study area, the daily flow processes of 1984 to 2013 were simulated based on Xin’anjiang model, and with the goal of maximizing the certainty coefficient between the calculated runoff depths and the observed ones, the genetic algorithm was adopted to calibrate the parameters. Then taking the soil water content calculated by daily model as the initial value of hourly model, the hourly simulation model was built based on the runoff calculation and flow simulation of hourly floods, choosing 30 floods from 1984 to 2013, and with the goal of maximizing the average certainty coefficient of flow processes of 20 hourly floods, the genetic algorithm was adopted to calibrate n, K of every flood and SM.The influence factors of n, K, including average rainfall intensity I, the largest rainfall infiltration P24, rainfall gravity center L and index P+Pa were analyzed, and then the relationship between n, K and their influence factors was built based on multiple regression, finally the relationship was applied to the flood forecast in Fengle River basin in this paper.The certainty coefficients of simulated daily runoff process were 0.82 and 0.80 in calibrated period and validated period, respectively. The average relative errors of runoff depth were 10% and 8%, the average flood peak errors were 13% and 12%, the peak time errors were 3.6h and 5.4h, the percents of pass were 100%,80%,85% and 90%,90%,80%, the average certainty coefficients were 0.88 and 0.84 in 20 calibrated floods and 10 validated floods in hourly model, respectively. The relationship between expection of instantaneous unit hydrograph nK, and rainfall gravity center L, index P+Pa was positive correlation, and the relationship between expection of instantaneous unit hydrograph nK and average rainfall intensity I, the largest rainfall infiltration P24 was negative correlation, the average relative error of runoff depth was 13%, the average flood peak error was 9%, the peak time error was 3h, and the corresponding percents of pass were 100%,80%,85% in 4 floods which used for validating regression relationship. The smallest certainty coefficient was 0.80 (NO.840830), the largest one was 0.99 (NO.090628), and the average was 0.89.The results showed that the daily and hourly flood forecast models built based on instantaneous unit hydrograph was reasonable, the calculation accuracy was high. The order of certainty coefficient sensitivity was depth of runoff, flood peak time and peak discharge. The influence factors of n, K, including average rainfall intensity I, the largest rainfall infiltration P24, rainfall gravity center L and index P+Pa were analyzed, and the relationship between instantaneous unit hydrograph parameters and their influence factors was reliable.