The rapid development of deep learning (DL) is the most significant contemporary evolution of hydrological science, yet its limitations remain underexplored. This study aims to evaluate the effectiveness of DL-based, traditional, and hybrid hydrological models in streamflow simulation especially in data-deficient regions. Daily discharge flow of the four sub-catchments in Samar, Philippines were modeled using HEC-HMS, Univariate Long-Short Term Memory (LSTM) Network, Classical GR4J and DL-assisted, parametrically- optimized GR4J. Results show that the classical GR4J underestimated the discharge, while the LSTM overestimated peaks. The DL-assisted, parametrically-optimized GR4J achieved the highest consistency and balance between realistic peak and low flow estimation, with NSE = 0.63–0.84, IA = 0.85–0.93, LMI = 0.76–0.82, and low MAPE (≤ 0.03) and RSR (≤ 0.02). Although the Univariate LSTM captured general trends well, it underestimated some peaks despite reaching high values in performance metrics. Many studies highlight DL’s hydrological modeling power but ignore its limits and hybrid model benefits. The results highlight that neither DL nor classical models alone are sufficient as DL-assisted approaches yield more reliable and realistic hydrological simulations, particularly in climate-sensitive, data-deficient regions like Samar.

Hydrological time series have been forecasted using various models. A study employed nonlinear dynamical methods, namely the correlation dimension and dynamic vs. stochastic (DVS) algorithm, to analyze daily flow data characteristics. Analysis using the correlation dimension and DVS algorithm revealed that the daily streamflow observed from the St. Johns River near Cocoa, Florida, USA, exhibited chaotic characteristics, whereas the daily inflow to the Soyanggang Dam reservoir showed stochastic properties. However, the nonlinearity of the flows has not been investigated, and the stochastic models have not been fit for flow modeling and forecasting. Therefore, the present study tests the nonlinearity and fits stochastic models of the flow data. In addition, the forecasting results obtained from the DVS algorithm and neural networks were compared with those from the fitted stochastic models. The forecasting results derived from the DVS algorithm and neural networks demonstrated higher accuracy for daily streamflow than for daily inflow. Furthermore, when applying the AR (1) model to daily flow and ARMA (3, 1) model to daily inflow, the results showed that the chaotic nature of the daily streamflow yielded more accurate forecasts. The findings suggest that the dynamic structure inherent in hydrological time series may influence forecasting performance. Notably, the two flows exhibited nonlinearity based on BDS statistics, indicating that nonlinear time-series models may be more appropriate for analysis.