Sensitivity Analysis For Parameters On Two Common Calculation Methods Of Small Basin Flood

Detailed information is always unavailable in small basins, due to the low density of hydrologic station. The small basin flood which result has many empirical parameters is applied widely, as lack of the basic rainfall data in some small basins, most of calculation formula for small basin flood is established by a series of assumptions and model generalization. Then, the calculation parameters is reduced greatly, but the result error is not large. As the parameters has reduced, the method is simpler, but the less parameters has greater influence on the result.The paper introduces two kinds of calculation methods for small watershed flood, namely IWHR reasoning formula method and linpingyi method. Meanwhile, this paper calculated three practical examples for each of the two methods by EHP software. We could screen out the more uncertain parameters through the main parameter uncertainty analysis of two methods, and then the sensitivity coefficient will be calculated by+0.5%、±1%、2%、±5%、±10%、±15%and±20%in seven different ranges of unrepeated disturbances for screened parameters. Comparing with the sensitivity coefficient, the influence of each parameter error for the result precision of peak discharge can be known.This paper not only separately analyses the sensitivity of respective parameters for reasoning formula and linpingyi method, moreover, the common parameters involved in two methods are compared and analyzed, so as to draw the general trend. The main results are as follows:1. In reasoning formula method, the sensitivity coefficient absolute value of loss parameter μ,confluence parameter m, longitudinal gradient J and maximum precipitation in24h H24is|QH24h|>|Qm|>|QJ|>|Qμ|. Among the four parameters,|QJ|and|Qμ|is relatively small and less than0.5, the sensitivity coefficients of P and J that is less showed they has less effects on the result of peak flow. Secondly, m which has more effects on the result of peak flow, mainly distributed in0.8～1.2. The maximum|QH24h|is more than1,showed H24is the most influential factor among the above parameters. In addition, the sensitivity coefficient of parameter has larger fluctuation with the parameter range in the5%. 2. In linpingyi method,|Qμ|and|QJc|is relatively small and less than0.5, the sensitivity coefficients of μ and Jc that is less showed they has less effects on the result of peak flow.|QNc|is commonly less than1,but close to1,showed Nc has more effects on the result of peak flow. The maximum|QH24h|is more than1,showed H24is the most influential factor among the above parameters. In addition, the sensitivity coefficient of parameter has larger fluctuation with the parameter range in the5%.3. The common parameters of reasoning formula and linpingyi motheod contain and H24h-The sensitivity coefficient of J(c) is between0and1,showed there was a positive correlation between J (c) and QJ(c) which is the sensitivity coefficient of J(c), on the other hand, J(c) play a role in narrowing the error of peak flow. With the parameters of J(c) increased, QJ(c) generally showed a trend of decrease. By comparison, QJ(c) of reasoning formula method is larger than that in linpingyi method. QH24which is the sensitivity coefficient of H24h is more than1,which showed there was a positive correlation between H24h and QH24, on the other hand, the variation of H24h expand the error of flood peak. With the parameters of H24h increased, QH24generally showed a trend of increase. There was no obvious correlation when QH24in reasoning formula method or linpingyi method.