Simulating hydrologic pathway contributions in fluvial and karst settings: An evaluation of conceptual, physically-based, and deep learning modeling approaches

•LSTM model outperformed SWAT and LUMP models in simulating total daily discharge.•Improved performance of LSTM was more evident in karst basin than fluvial watershed.•LSTM model with digital filter matched process-based hydrologic pathway estimates.•Process-based models exhibited more realistic time-fractal scaling of hydrologic flow pathways.