Predicting the inflow into the dam reservoir using artificial neural network model based on PERSIANN-CDR and CMC data (case study: ZayandehRoud Dam)

Ramtin Moeini ¹ ( Department of Civil Engineering, Faculty of civil engineering and transportation, University of Isfahan, Isfahan, Iran )
Mohammadali Alijanian ² ( Department of Civil Engineering, Faculty of civil engineering and transportation, University of Isfahan )
Mina moradizadeh ³ ( Surveying and Geomatics Engineering Department , Faculty of civil engineering and transportation, University of Isfahan )

Submited date : 2023-12-17 Accepted date : 2024-02-14

Keywords: Artificial neural network, Rainfall, satellite-based data, water equivalent to snow, Zayandeh Rood Dam,

Abstract :

However, the scale of satellite-based data and the need for their exponential scaling are the uncertainties of these data. In this research, the performance of PERSIANN-CDR and CMC satellite data for rainfall and snow estimation and determining the inflow values into the dam reservoir is investigated. Therefore, by considering different combinations of input data, different models are proposed and the input flow to the dam reservoir is predicted using the artificial neural network (ANN) model. Here, the the ZayandehRoud dam reservoir of the Gavkhoni drainage basin is selected as a case study. The results shows that the best R2 and RMSE values for rainfall (snow) estimation data based on the PERSIANN-CDR satellite (CMC) are 0.49 (0.34) and 60.90 (41.56) mm. In other words, the results show the proper performance of satellite-based data for rainfall and snow estimation. Therefore, these data are used for creating the ANN model to determine the inflow values into the reservoir of ZayandehRoud dam reservoir. The results show that the values of R2, RMSE and NES for training data (validation and testing) of ANN model are equal to 0.72 (0.74), 56.08 (75.178) MCM, and 0.85 (0.86) respectively. In other words, the results show the proper performance of satellite-based data for estimating and determining the inflow into the ZayandehRoud dam reservoir using ANN model.