Page - 354 - in Book of Full Papers - Symposium Hydro Engineering

Given from fig 2 and fig3, the regressive results agree well with the observed data. The effects of the formula can be accepted. Furthermore, a detailed analysis based on physical meanings should be done: (1) The full regression model and selecting regression model of Y1: the main influencing factors on Y1 are outflow discharge (the exponent number is 0.417), sediment concentration flowing into the reservoir (the exponent number is 0.415), sediment concentration flowing out of the reservoir (the exponent number is -0.070), local terrain factor in the tributaries (the exponent number is -0.031), local terrain factor in the main channel (the exponent number is 0.030) and the location of the front of the delta (the exponent number is 0.951). Considering the importance of those factors, the location of the front of the delta, outflow discharge and sediment concentration flowing into the reservoir take the first three places and agree well with the physical meanings. Additionally, although the exponent number of the local terrain factor in the tributaries is small enough, the correlation coefficient of that is significant. So the local terrain factor should also be chosen. Then the final form of the regressive model for Y1 is shown as follows: ?̂? 1 = 0.534?̂? 2 ∗0.4?̂? 3 0.4?̂? 4 −0.1X̂6 (R 2=0.638) [8] The exponent number and independent variables are all simplified and the results are shown in fig 4. (2) The full regression model and selecting regression model of Y2: the main influencing factors on Y2 are outflow discharge (the exponent number is 1.373), sediment concentration flowing into the reservoir (the exponent number is - 0.430), sediment concentration flowing out of the reservoir (the exponent number is 0.248), local terrain factor in the tributaries (the exponent number is 0.317), local terrain factor in the main channel (the exponent number is 0.320) and the location of the front of the delta (the exponent number is 0.719). Considering the importance of those factors, outflow discharge, the location of the front of the delta, and sediment concentration flowing into the reservoir take the first three places. Based on the physical meanings, these three factors should all have positive correlations with the depositing volume in the tributaries, while the sediment concentration flowing into the reservoir showed an opposite relation in this model. The physical interpretation is: the sediment flowing into the reservoir area is not equal or positive related to the depositing volume in the reservoir area, because there is a drastic process of the sediment exchange between the channel bed and water flow. Even though clear water flows into the reservoir area, it also may cause a sediment transporting process and a deposition in the river mouth of the tributaries. As a result, the sediment concentration flowing into the reservoir is deleted from the model. Similarly, we can get the final form of the regressive model for Y2 : ?̂? 2 = 85.57?̂? 2 ∗1.4?̂? 4 0.3?̂? 5 0.3?̂? 6 0.7 (R2=0.820) [9] 354