Potential Evapotranspiration (PET) is a crucial component of the water cycle and energy balance. Typically, empirical models such as Thornthwaite (TH) and Penman-Monteith (PM) are employed to calculate PET. However, the TH model only considers temperature as a climatic input, leading to lower accuracy, whereas the PM model utilizes a wide range of meteorological parameters, providing much higher accuracy. Nonetheless, the PET derived from these methods generally has low spatial resolution, as it is calculated for a single point. Therefore, in this study, to achieve highly accurate PET estimates with fewer meteorological parameters and enhanced spatial resolution, an optimized PET estimator model (KPET) was developed, which utilizes precipitable water vapor (PWV) derived from the Global Navigation Satellite Systems (GNSS). Initially, the difference between PET calculated by the PM and TH models (DPET) was determined. Then, the relationship between DPET and PWV, temperature, pressure, and spatiotemporal parameters was analyzed, leading to the selection of a quadratic polynomial regression model. Finally, the new model was developed by adding the fitted DPET to the initial PET derived from the TH model. Los Angeles was chosen as the study area, and the statistical results demonstrated the satisfactory performance of the proposed KPET model in the region. The statistical metrics of RMSE, MAE, and the correlation coefficient were used to evaluate the KPET and TH models relative to the PM model, and the evaluation results were compared. The statistical results, averaged across all control stations, showed that the KPET model achieved an RMSE of 0.42 mm, an MAE of 0.33 mm, and a correlation coefficient of 0.98, outperforming the TH model, which had an RMSE of 1.84 mm, an MAE of 1.71 mm, and a correlation coefficient of 0.92. Additionally, the improvement in RMSE across the control stations was found to be 72.25%, 82.6%, 79.7%, 78.95%, and 62.72%, respectively. These findings indicate that the KPET model developed in this study can accurately and reliably estimate PET using limited meteorological data in the region. |
