**4. Discussion**

The calculation of *ET*<sup>0</sup> is affected by a variety of climatic factors. Ma et al. (2010) [49] studied the influence of main climatic factors on *ET*<sup>0</sup> in mountainous plateau areas and found that the change of wind speed had the most significant impact on *ET*<sup>0</sup> at each site. Luo et al. (2010) [50] conducted a sensitivity analysis on *ET*<sup>0</sup> and main meteorological factors in the main agricultural areas of Tibet, and the results showed that *ET*<sup>0</sup> in the whole region had a declining trend over the past 50 years. The meteorological element that had the most significant impact on *ET*<sup>0</sup> was *Rs*. Similar results were obtained in our study, where the accuracy of *ET*<sup>0</sup> was affected by the error of *Rs*. Xie et al. (2017) [51] analyzed the impact of changes in meteorological factors on *ET*<sup>0</sup> in China's main grain-producing areas from 1961 to 2013, in which *ET*<sup>0</sup> showed a saw-tooth decline. The changing characteristics of main meteorological factors in the study area and the response of *ET*<sup>0</sup> are similar to the results of our study, showing regional and seasonal variations. Overall, our study suggests that the errors of meteorological factors in the Qinghai-Tibet Plateau region and the boundary region of the climate zones are more significant than in other regions, with the highest errors observed in summer.

Due to the incomplete understanding of the physical mechanism of weather changes and limited observational data, there is still a specific error in the reanalysis data [52], and the magnitude of this error tends to vary with different climatic factors. Temperature is a meteorological variable with minor errors, usually less than 10% [53,54]. Similar results were found in our study, in which the *R*<sup>2</sup> of *T*max and *T*min are generally greater than 0.9 in the seven climatic zones. Due to the influence of topography, the errors of wind speed and relative humidity are usually large [26], and similar results were obtained in our study.

It is worth mentioning that the weather stations in our study are affiliated with the China Meteorological Administration. The ground of the weather station is usually covered with short grass under adequate irrigation conditions. However, areas in the grid system do not necessarily have lush vegetation. Therefore, there might be some differences in the environmental factors between the two types of systems, especially for the radiant energy (i.e., *R*<sup>2</sup> < 0.65 in the seven climate zones for the R*s*CLD estimation in our study). This may lead to the problem of overestimating or underestimating the reanalysis data, which indirectly explains why the estimated *ET*<sup>0</sup> values in some areas fluctuate severely in our study. In addition, the variation of wind speed is greatly affected by the terrain and the type of underlying surface, and it is not easy to obtain the average wind speed in a specific area. Similar results were obtained in our study, where the overall UCLA accuracy is not satisfactory.

Finally, this study can provide an idea for economically underdeveloped countries and contribute to improving the reanalysis data set and the accuracy of *ET*<sup>0</sup> estimation. Therefore, when other developing countries establish regional climate models, they can consider their own terrain and climate characteristics and establish a more local model.

#### **5. Conclusions**

*ET*<sup>0</sup> data set based on reanalysis products can make up for the time discontinuity and spatial insufficiency of surface meteorological platform data, which is significant for water resources planning and irrigation system formulation. However, a rigorous evaluation of reanalysis products must be carried out to see if they have value in application. This study evaluated the ability of the CLDAS dataset officially published by the Chinese meteorological system for *ET*<sup>0</sup> estimation. Results indicate that the temperature data of CLDAS have high accuracy in all regions except the Qinghai Tibet Plateau (QTP) region. In contrast, the accuracy of the total radiation data is average, and the quality of relative humidity and wind speed data is poor. The overall accuracy of *ET*<sup>0</sup> is acceptable except for QTP, but there are many stations with large errors. Among seasons, RMSE is the largest in summer and smallest in winter. There are also inter-annual differences in the *ET*<sup>0</sup> of this data set. Overall, the CLDAS dataset is expected to have good applicability in the Inner Mongolia Grassland area, Northeast Taiwan, the Semi-Northern Temperate zone, the Humid and Semi Humid warm Temperate zone, and the subtropical region. However, there are certain risks in other regions. In addition, of all seasons, summer and spring have the slightest bias, followed by autumn and winter. From 2017 to 2020, bias in 2019 and 2020 are the smallest, and the areas with large deviation are in the south of climate zone 3, the coastal area of climate zone 6, and the boundary area of climate zone 7.

**Author Contributions:** Conceptualization, L.-F.W. and L.Q.; methodology, L.-F.W., G.-M.H. and L.Q.; data curation, L.Q. and X.-G.L.; writing—original draft preparation, L.-F.W., Y.-C.W., G.-M.H. and L.Q.; writing—review and editing, L.-F.W., H.B., X.-G.L. and L.Q.; project administration, L.-F.W. and S.-F.W. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Science and Technology Project of Jiangxi Provincial Department of Education (GJJ180925), National Natural Science Foundation of China (51979133 and 51769010) and Natural Science Foundation of Jiangxi province in China (20181BBG78078 and 20212BDH80016).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analysis, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.
