**5. Conclusions**

This work assessed the actual evapotranspiration estimated by the CR approach in the frozen ground regions of the QTP. The uncertainties associated with observation and modeling of *E* at point scale were analyzed by investigating the parameter determined by different lengths of calibrated periods and parameter deviations on the accuracy of *E* estimation. Finally, two more latest works were compared with the CR-based models in this study at BJ and NAMORS.

Five CR-based models–whether with default or calibrated parameters–performed well in daily *E* estimation: the NSE values were both above 0.7, the accuracy of each CR-based model was close to another, the C2018 and H2018 models performed better among five CR-based models on daily and monthly timescales at four field sites, and the S2017 model performed poorest. The five CR-based models could estimate daily *E* during warm seasons, however, they are hardly applied during cold seasons. Nevertheless, the amount of *E* during cold seasons only account for small proportion of annual evaporation. Therefore, CR-based models with fewer data requirements and parameters provide a feasible approach to estimate daily and monthly *E* without needing detailed information on moisture and vegetation in the frozen ground regions of the QTP.

For calibrated parameter *αe* and *b* in CR-based models, all models are more sensitive to variations of *αe* than *b*. Specifically, within a certain range of deviations of parameter *αe* (for example, deviations from −10 to 30% in this study), the RMSE values will increase about 0.1 mm d−<sup>1</sup> compared with RMSE of *E* estimation with optimized parameter values. Parameter *b* has little impact on the accuracy of *E* estimation. Thus, reasonable parameter *αe* values are of much importance to the performance of CR-based models.

Compared with previous studies about the application of CR-based models on the QTP, this study uses more in situ measurement data and more CR-based models. The present study also demonstrates better capability of CR approach on estimating *E* in high altitude and cold regions over QTP. At present, the CR approach has been developed as a benchmarking tool for large-scale evapotranspiration estimates, however, discrepancies in different spatiotemporal scales with other mainstream evapotranspiration products indicated that with the exception of developing physically CR-based models, hybrid modelling approaches such as combining CR approach with data-driven models are of grea<sup>t</sup> significance to large-scale and high-accuracy ET estimation in the future.

**Author Contributions:** C.S.: Formal analysis, Writing—original draft, Visualization. T.W. (Tonghua Wu): Conceptualization, Supervision, Writing—review & editing. N.M.: Conceptualization, Methodology, Writing—review & editing. J.W.: Conceptualization, Methodology, Writing—review & editing. X.L.: Writing—review & editing. X.Z.: Writing—review & editing. T.W. (Tianye Wang): Writing—review & editing. G.H.: Writing—review & editing. R.L.: Conceptualization. S.Y.: Conceptualization. J.C.: Conceptualization. J.Y.: Conceptualization. C.Y.: Writing—review & editing. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Natural Science Foundations of China (grant number 41771076, 41690142, 42001071), the West Light Foundation of Chinese Academy of Sciences (Tonghua Wu), the "Hundred Talents program" of Chinese Academy of Sciences (Sizhong Yang), the Open Research Fund Program of State Key Laboratory of Cryospheric Science, Northwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences (grant number: SKLCSOP-2020-11). And The APC was funded by Tonghua Wu.

**Data Availability Statement:** The data that support the findings of this study are available from the corresponding author upon request (Tonghua Wu, thuawu@lzb.ac.cn).

**Acknowledgments:** The authors would like to thank Xuelong Chen, Cunbo Han and Ling Yuan for sharing the original and modelling data at BJ and NAMORS.

**Conflicts of Interest:** The authors declare that they have no known competing financial interest or personal relationships that could have appeared to influence the work reported in this paper.
