Review on “Parameterization of downward longwave radiation based on long-term baseline surface radiation measurements in China” by Junli Yang, Jianglin Hu, Qiying Chen and Weijun Quan

General comments:

This study investigates different empirical parameterizations of the surface downward longwave radiation regarding the adequacy for their use in China. In addition, the authors develop a new empirical parameterization and perform a comprehensive evaluation using data from 7 stations from the Chinese Baseline Surface Radiation Network. The authors conclude that the parameterizations and associated coefficients they derive are suitable for the determination of downward longwave radiation over China. The paper is well written, fairly straightforward and clearly structured. The applied methods are sound.

My main reservation with this study is the relatively limited applicability of its results,  being only of use for the determination of downward longwave radiation in China. To make the paper more attractive for readers outside China and to increase its impact in the community, information on the applicability of these parameterizations outside China would be valuable. For example, the Baseline Surface Radiation Network (BSRN, www.bsrn.awi.de) with numerous worldwide distributed high quality radiation stations would provide a framework to test and calibrate these parameterizations under more diverse geographical and climatological conditions. This would then allow to investigate the more general applicability of the parameterizations and make it more interesting for the worldwide readership of ACP. While such a broader analysis might be challenging to achieve within the limited time of a revision phase, I would find it at least useful if the authors could add a discussion of the potential and limitations of these parameterizations for their use outside of China, in order to provide guidance to readers interested to apply them for the determination of downward longwave radiation in other parts of the world.

While I think the English is overall adequate, there are still numerous minor issues as indicated in the technical comments. As I certainly not have caught all of them, I encourage the authors to doublecheck the manuscript in this respect, ideally with the help of a native English speaker.

I recommend the acceptance of the manuscript after revisions as outlined above and below.

Specific comments:

L69ff In addition to the climate zones, it would also be interesting to know if the stations are located in an urban, industrial or rural setting. This can give some indications to what extent the measurements could be influenced by local anthropogenic pollution sources.

Table1: it would be worthwhile to include in this table also the measurement period for each site.

L103: are there also collocated upper air soundings (radiosondes) available at some of the stations? BSRN recommends the high quality radiation stations to include upper air soundings for the interpretation of the measured fluxes and testing of models.

L130: I think the structure and formulation of the Brunt model and the Weng model should be explicitly described here or in the method section.

L131: It would be good to describe precisely how the clear sky hourly data were identified as being clear sky.

L142ff: ok here come the formulations of the different parameterizations which I expected earlier on (comment L130). Maybe this part could be described in the method section in a paragraph describing the different parameterizations used in this study together with their formulas.

L150: by eye it is hard to recognize much difference between the 2 models (red and black curve) for the dry conditions (e ≤17.5 hPa ), thus hard to fully appreciate the improved performance of the Wang model for dry conditions.

L185ff: it is not clear to me how the structure of the parameterizations has been established. Why do they have precisely this form and not e.g. another one?

L194ff: I understand the independent clear-sky dataset is independent in the temporal sense, i.e. the data stem from another period (from 2018 onward rather than before  2018), however still from the same stations. Two questions here: 1.) why not all 7 stations have been used, but only 4? 2.) Is there a chance to do a validation also at independent stations (not only independent times)? Basically one could use the entire worldwide BSRN dataset for this (see general comments). This would have the advantage that one could also get an idea on the performance of these parameterizations in other parts of the world under different climate regimes.

L202ff: similar comment as above, why only 3 stations are used here for a validation and not all seven?  Again also an evaluation with (spatially) independent stations would be interesting, ideally even outside China.

L207: Why should more samples necessarily help to reduce the MBEs?

L239ff/Figure 5: I assume this validation uses hourly values? And uses measurements from all 7 stations? This should be mentioned in the text or the figure caption.

Technical comments:

L29: add  “e.g., “ in front of the references, as there are many other and also earlier papers dealing with DLR.

L31: same comment, add  “e.g., “ in front of the references, as there are many other application papers of DLR. There are several other places in the manuscript where an “e.g., “  in front of the references would be appropriate, as other papers could equally well be cited. The authors may check on this throughout the manuscript

L40: hereinafter refer to > hereafter referred to as

L44: the presence of cloudS

L59: in terms of > based on

L63: to estimation of  >  for the estimation of

L90: influencing > influences

L106: strictly quality controlled > strict quality controls

L140: hereinafter refer to > hereafter referred to as

L155: to have basis in physics > to be based on physics

L166: Circles represent data pairs > Circles represent hourly data pairs

L170: are in consistent >  are consistent (inconsistent has the opposite meaning!)

L217: can overestimate > overestimates

L241 & L250:  and that > and the one

L252 & L253 & L254 & L255: could underestimate > tend to underestimate

L269: three parameterization models> all three parameterization models

L278: improve accuracy > improve the accuracy

L279: this sentence sounds awkward and needs reformulation

L280: to establish > to be established

L286: whereas > however

L289: station > stations

Title: Parameterization of downward longwave radiation based on long-term baseline surface radiation measurements in China

Author(s): Junli Yang et al.

MS No.: acp-2022-794

MS type: Research article

Scientific aspect:

The search for empirical expressions for longwave downwelling irradiance for China is a welcome contribution. The measurement of longwave down welling irradiance has been rare worldwide, and especially in China. In this sense, this reviewer does not demand the expansion of the scope to larger regions. To build an empirical relationship between the longwave down welling irradiance with widely available climatic elements requires accurate irradiance measurement. The required accuracy is made not only of an instrumental accuracy, but also of the traceability to the international standard. This latter point is very important for any long-term observations, and is the basis for the Baseline Surface Radiation Network. This development comes from a bitter experience to realize serious differences among the longwave calibration methods practiced by many countries. It has become necessary to establish the global standard in longwave calibration, which is materialized as the World Standard Group of pyrgeometers at the World Radiation Centre in Davos. Within the BSRN, there was only one Chinese station, Xianghe, which joined the BSRN, more than 10 years later than other sites, and ceased to operate already in 2015. The continued functioning of this site was an international wish. It is not a constructive direction for each country to establish own baseline radiation network. If this is done, however, like Chinese Baseline Radiation Network (CBSRN), its traceability to the World Standard must be established. This point is missing in the presented paper, reducing the trustworthiness of the accuracy of the proposed equations. The current status of the BSRN is summarized in Driemel et al. 2018, Earth Syst. Sci. Data,10, 1491-1501. Li et al., 2013 may present the information on the CBSRN, but this literature is not accessible for most readers. Its main content can be introduced in the paper.

            The empirical relationships under the cloudless sky are quite straightforward, as the depth of the atmospheric emission effectively reaching the surface is quite thin as the authors pointed out. The mathematical shapes adopted in this sort of calculation are usually grey body emission. There are, however, at least two original proposals, which are independent of the graybody preoccupation. These original proposals are made by Swinbank’s 1963 QJRMS and Ruckstuhl’s 2007 JGR papers. Swinbank is quoted in the manuscript but his originality is not appreciated.

            The attempt to expand the empirical relationships to all sky conditions causes a problem, mainly owing to the diversity of the clouds and the limit in our observational documentation. Nevertheless, one of the earliest proposals was attempted by H.M. Bolz (1949) in Zeitschr. Meteorol. This is a systematic introduction of the effect of clouds. Relating to this matter, the description from Line 182 to Line 192 must be reformulated. It is necessary to present how the authors consider the shapes of Equations (5), (6) and (7) are justified, and how each independent variables offer the targeted results. There seem to be a slight confusion in expressing Greek variables also.

Formalities:

• Generally, quoting earlier works for substantiating the point in the papers must be done carefully. Just quoting many papers does not support the point authors wish to make. As an example, let’s take the first two sentences in the introduction, Line 28 to 32. The importance of the longwave downwelling radiation was not realized only in 1994 or 2020. These papers are rather recent papers in this subject. This reviewer suggests the authors to quote the first and most original paper on the subject and then several recent and best papers. Not all papers quoted in these lines do not necessarily represent the best knowledge of the present time.
• Introduction can be shorter. Numerical details can be summarily presented in Conclusions.
• Line 44 to 49: in this discussion, Bolz’s and Ruckstuhl’s works can make a constructive contribution.
• Line 180, Table 2: Under Column Network/Site, Line Swinbank (1963), H.M.A.S. Diamantina in the Indian Ocean should be added to Aspendale and Kerang. The observation on board Diamantina over the Indian Ocean provided the measurement in very high humidity, and played an important role in generalizing Swinbank equation.
• Line 185-191: There are confusions in Greek letters.
• The analyses in Line 193 Section 4.3, and Line 236 Section4.4. are well done and very useful.
• Line 273-275: The seven CBSRN sites are all confined in the continental interior regions and do not represent the climate of the maritime regions. This bias should be considered.
• Line 373-379: The order of references, Liu, M.Q., et al. and Li, M.Y., et al. should be reversed. Likewise, Line 381-384: Niemelä et al., and Monteith can be reversed.