5 Conclusion
This study is the first to document information on the comparison of 13 reference crop evapotranspiration models against lysimeter measurements in a humid alpine meadow, northeastern Qinghai-Tibetan Plateau, we found that the radiation-based models performed better than combination models and temperature-based models. In addition, the combination models and most radiation-based models tended to underestimate measured ET, whereas temperature-based models tended to overestimate measured ET during the whole study period. Specifically, all models tended to underestimate ET during the growing season and overestimate ET during the non-growing season, suggesting that these models should be calibrated or modified by local lysimeter data when extrapolated to other regions. Furthermore, the 1963 Penman and FAO-24 Penman models demonstrated better performances than the recommended FAO-56Penman-Monteith (PM), suggesting that older Penman equations may be superior to the standard FAO-56 Penman-Monteith model, especially considering the good performance of the 1963 Penman model in this study. Given the outstanding performance of the DeBruin-Keijman and Priestley-Taylor models, which require few meteorological inputs, we thus recommend that these two models can be used in other alpine meadows that have similar climates to that of the study region, to improve ET estimation.