One-dimensional models can enable the assessment of the dynamic behavior of hydropower units during transient operation with minimal computational resources. Since these models do not consider the full three-dimensional flow in hydraulic piping systems, their predictions rely on static performance characteristics obtained typically through reduced-scale measurements. These measurements usually cover only a small portion of the complete operating range of the machine, making it challenging to simulate transient events such as start-up, shutdown, or load rejections with high fidelity. To address this issue, we propose a novel approach for calibrating one-dimensional physics-based models of hydropower units by combining monitoring data with blackbox optimization. Specifically, the performance characteristic curves feeding the one-dimensional model of the power plant are represented by a polynomial whose parameters are optimized by minimizing the difference between simulation results and experimental data. We demonstrate that the proposed method is suitable for a 50 MW Kaplan turbine considering various startup scenarios, including different guide vane and blade opening sequences. The optimization was conducted using different combinations of training and test datasets to assess the validity of the calibrated models. Good agreement between simulation and experimental data was obtained using only a few startup sequences in the training dataset which demonstrate the robustness of the proposed methodology. This may pave the way for the calibration of hydropower unit digital twins for unit monitoring and anomaly detection.