The manuscript provides a theoretical derivation of the PT coefficient and evaluate the results by using wet-surface measurements. The manuscript provide some in-depth understanding on the variation of PT coefficient and show that the value is also essential for hydrological simulation and projections. The manuscript is written in a organized structure and the contents is supported by in situ measurements. However, there are still one major issue need to be clarified.

We all know that the Priestley-Taylor model has limitations in its application. The PT model should be applied at appropriate tamporal resolution, however, such issue has been ignored in the present manuscript. In table 3 and Figure 4, the results show significant difference in seasonal data and yearly data. Thus, the obtained values should also be different at different temporal resolutions. Thus, it would be nice to understand whether the dereived relationships also fit at temporal resolution of weeks or ten-days, or at least to indicate that the results are reasonable at ?? temporal resolution. Further, the authors select global flux sites data in the evaluation and the days with soil moistuer lower than 50% of the maximum soil moisture are removed. Then, how to obtain monthly data at these flux sites. Some details should be given to clarify this issue.

In the equations, all the variables have no units in the manuscript and the abbreviations of CSIRO also have no full names. I suggest the author to include a table to include the unites of each variable and to show full names of the abbreviations in the appendix.

This is a a paper that tries to understand the Priestley-Taylor parameter alpha based on theoretical derivations from a coupled mixed-layer model. I find this work very interesting, but at the same time find that it does not link well enough to earlier work on the topic and therefore I recommend the editor to reject the paper until this has been repaired. 

The work in this paper is far from new in my view, and previous papers have presented the matter in more detail than this paper. I would like to ask the authors to make a detailed comparison to the math in the papers mentioned below, and explain where they deviate and where the are similar. Also I would like to ask them which are the actual new insights emerging from their work that the previous papers could not provide.

The papers of Raupach (BLM, 2000, QJ 2001) present an very in-depth analysis of equilibruim evaporation in a partially open system such as an entraining convective boundary layer. These are in my view the best papers written on this topic and should be studied in much more detail by the authors.

Also, the paper of LHomme, 1997, BLM present an alternative derivation for alpha and should be considered in detail as well. 

Then, the paper of van Heerwaarden et al., (QJ, 2009) presents a full derivation of the Priestley-Taylor parameter alpha, from a mixed-layer model perspective and hence does exactly that what the authors of this paper intend to do. However, this paper does not make any detailed comparison. I would like to learn where similarities are found and where differences arise.

Concerning the contents, the authors depart from the earlier papers in calling the non-saturated state non-equilibrium, while the previous mentioned papers show that a non-saturated equilibrium exist for open systems. This might require some extra discussion and maybe some rethinking of the chosen definitions.