Simulation of Parametric Uncertainty of Hydrological Models using UNEEC-P Framework: Monthly Water Balance Model case Study

Efforts to achieve suitable estimation of parametric or the structural uncertainty of mathematical or conceptual frameworks have been led to develop various probabilistic, possibilistic, and innovative methods. In this paper, uncertain parametric behavior in a monthly water balance model has been studied using the structure of the UNEEC-P method. In the approach proposed by this paper and for the first time, instead of using a variety of regression methods to estimate the upper and lower uncertain bounds, the original conceptual model has been used. The applied conceptual model is a three-parameter monthly water balance model, and the case study of the paper is a small basin with an area of 82 square kilometers in southern France. Also, in order to evaluate the performance of the proposed method, Generalized Regression Neural Network (GRNN) has been used to evaluate the results of the conceptual models. The estimation of parametric uncertainty has been used to simulate the results of GLUE method in a Confidence Level (CL) equal to 90%. In order to measure the accuracy and validity of the new mechanism proposed in this study, in addition to the usual evaluation indicators of similarity and dissimilarity, the AIC index is also used, and different statistical indicators such as Mean Square Error (MSE), Normal Mean Square Error (NMSE), Nash-Sutcliff (NS), Correlation Coefficient (CC), and AIC demonstrate the better performance of proposed method comparing to the GRNN.