Prediction of sediments discharge in watershed with two tank models

Abstract

Uncontrolled erosion would cause considerable damages, such as soil fertility decline, water structures damage and reservoirs sedimentation. As the data for the sedimentation rate are limited, several models have been developed to predict the surface erosion and the rate of sedimentation. However, the availability of sufficient, diverse and extensive data is needed for the implementation of the models, both for the model calibration and the verification. The result of the analysis shows that both of the Water Tank Models that represent the erosion-sedimentation rate process, in which Water Tank 1 being the three-tank cascade system and Water Tank 2 being two-tank cascade system, are not optimum. This can be observed from the values of volume error (VE), relative error (RE), root-mean-square error (RMSE) and correlation coefficient (R) that show the effect of 1.5 hours of rain period in the sedimentation rate. The field condition shows considerable sedimentation, on the other hand, the models’ simulations show decreasing sedimentation rates. The optimum model’s parameters for Water Tank 1 and Water Tank 2 are 924.51%-1049.26% for the relative error, 50.81% - 121.42% for the volume error, 0.9 for the correlation coefficient and 6703.59-17,297.85 for the root-mean-square error. The parameters and constant’s values of the models are different relative to the drainage basins’ condition.