A comparison of three water discharge forecasting models for monsoon climate region: A case study in cimanuk-jatigede watershed Indonesia

Forecasting the discharge of the water resources model of a river or reservoir is crucial in making decisions on water resource management. Although numerous popular discharge forecasting models have been developed, real-time forecasts remain challenging. This study evaluates discharge forecasts using model of water balance model (FJ Mock and NRECA), autoregressive stochastic (Chain Markov). This study compared the accuracy of the discharge forecast results produced by the F.J. Mock, NRECA, and Markov Chain by five statistical indicators. Based on the simulation results, these models all have an accuracy level of probability in discharge ranging from above 70% which indicates a significant relationship between each model and are suitable alternatives for forecasting discharge and indicating an increase or decrease in discharge of up to 70% is predictable using this model. In comparison to other models, the highest correlation (r) is the Markov Chain model (0.84) with KGE (0.81) and the next sequence is the NRECA, and F.J. Mock. Therefore, the most accurate, precise, and representative water source model alternative for forecasts is the Markov Chain model. The FJ Mock model and the NRECA model are physical-based rainfall-runoff models, while the Markov model is time series generation model. In addition, this model is to be selected as the basis for modeling in forecasting river flow or optimal management of a reservoir, as well as determining the future discharge, especially in monsoon climate regions.