Configuration of user interface for evaluation of runoff in Orontes basin using artificial intelligence

Introduction. Surfacе runoff is of great importance for water resources formation processes in a river basin. The purpose of this study is to develop a user interface for a numerical solution to the problem of evaluating the runoff in the El-Asi (Orontes) River basin in Syria using artificial intelligence models. Materials and methods. The method of artificial neural networks was used to design the user interface. The task was solved in three stages: training, verification and testing. Several types of model algorithms were tested. Efficiency values were compared for different models using correlation coefficients and the mean root square error. Results. The authors have found that feed-forward artificial neural networks and back propagation artificial neural networks show the best result if used to make hydrological forecasts and simulate nonlinear functions; that’s why they were applied to the user interface. The developed software tool saves the user’s time and effort, because it generates a very large group of models pursuant to various parameters and functions. It selects the best model according to the effectiveness criteria (such as correlation coefficient R and mean root square error MRSE), and allows the performance of other operations, such as the graphical representation of output data, the structure of the model used or the onset of evaluation of runoff values. The software is developed in the operational environment of MATLAB. Conclusions. The software tool is simple and user friendly; it complies with the user-focused methodology, which is easily implemented from the moment the software, is launched through the sequence of activated interface commands. It is recommended to expand the use of artificial intelligence models for forecasting and evaluating elements of the hydrological cycle, especially in absence of source data.