Hydrologic model calibration has evolved to incorporate as much information as possible from spatiotemporal data in the calibration process to assure the consistency of the solution. Hydrologically consistent models are expected to replicate the signatures calculated based on measured data. However, the literature shows that incorporating many signatures as objective functions degrades the effectiveness of multi-objective optimization algorithms. A novel methodology is developed based on the Principal Component Analysis (PCA) to optimize only three calibration objectives while benefiting from most of the information content in many model calibration metrics. The proposed method is compared against two conventional calibration methods for calibrating three different case studies that represent different model structures and different landscapes. Results indicate that the PCA model calibration achieved 10% more consistent solutions compared to conventional calibration approaches. Furthermore, it is found that the model performance metrics are aggregated based on their underlying correlation that is affected by their physical interpretation. Although no single metric could guarantee finding a consistent solution, it was found that, the number of metrics that met an acceptability threshold set by the user would be the most reliable indicator for the consistency of a solution. This indicator has a strong correlation of 0.8 with the actual consistency index of a solution which can be calculated based on solution performance in the validation period.