Figures & data
Table 1. Model hyperparameter setting.
Table 2. The characteristics of acceptable results for RPE, RMSE, MAE, and NSE.
Table 3. Summary of the training dataset (count = 460).
Table 4. Summary of the testing dataset (count = 92).
Table 5. Optimal PSO-ML hyperparameters for different lead times in rainfall-runoff simulation.
Table 6. Optimal PSO-ML hyperparameters for different lead times in EFI-runoff simulation.
Figure 11. Rainfall, EFI, and runoff time series on the held-out testing samples with different lead times.
![Figure 11. Rainfall, EFI, and runoff time series on the held-out testing samples with different lead times.](/cms/asset/ff607a36-2da9-4b7c-bad6-44fe4d18c821/tgnh_a_2353144_f0011_c.jpg)
Figure 12. Multiple model prediction results on the held-out testing samples obtained by two inputs with different lead times.
![Figure 12. Multiple model prediction results on the held-out testing samples obtained by two inputs with different lead times.](/cms/asset/110f1d11-2bf6-4016-bea2-2c3839a4e078/tgnh_a_2353144_f0012_c.jpg)
Table 7. Evaluation indices comparison of two inputs.
Data availability statement
Data will be made available on request.