Streamflow Simulation Data from Differentiable HBV and LSTM Models Using CAMELS Datasets

Dataset Description This dataset contains streamflow simulations and observations used in the following study: Li, P., Song, Y., Pan, M., Lawson, K., and Shen, C. (2025). Ensembling Differentiable Process-based and Data-driven Models with Diverse Meteorological Forcing Datasets to Advance Streamflow Simulation. EGUsphere, 1–74. https://doi.org/10.5194/egusphere-2025-483 Data Structure The data are stored in a 4-dimensional array with the following dimensions: axis = 0 — LSTM simulations under Daymet, NLDAS, and Maurer forcings, followed by δHBV simulations. axis = 1 — Different random seeds used for ensemble runs. axis = 2 — 531 basins, with corresponding basin IDs provided in the accompanying ids file. axis = 3 — Daily time steps starting from 1980-10-01 (inclusive). Variable Definitions ALL, PUB, and PUR indicate the temporal, PUB, and PUR simulation experiments, respectively. OBS represents observed streamflow, with units of mm/day.