Validation and Comparison of Climate Reanalysis Data in the East Asian Monsoon Region

Round 1

Reviewer 1 Report

This study by Kim and Lee validates and compares three popular climate reanalysis products (ERA5, JRA55, and NCEP2) along with the CRU gridded observational data, with a focus on the annual and seasonal variations of temperature and precipitation in the East Asian monsoon (EAM) region. The authors use observational station data in the EAM region as the target, perform temporal and spatial analyses, and conclude that ERA5 shows overall the best performance.

I think the topic of this manuscript is interesting to a wide range of audience whoever uses these datasets in research, hence of importance. The manuscript is in high quality: the text is well written and organized with a clear structure and a publication-ready format; the analyses are to the point and presented with overall instructive figures. That said, some minor revision and clarification is needed, which I list in the comments below. Once those have been addressed, I recommend the work be accepted for publication.

General concept comments

• The authors mentioned that observational station data may involve errors from the observational instruments and methods in 2nd paragraph of the introduction section, yet they use the observational station data as the target for the validation of the reanalysis products. If the target for validation may involve errors, how well can we trust the conclusions based on such target? Is it possible to alleviate the impacts of the errors in the target? Some justification and/or explanation is needed here.

Specific comments

• The green edge color in Fig. 4 makes the red face color of the circles difficult to observe and is thus distractive. I suggest remove the edge color for the sites with significant statistics, and use gray edge for sites with insignificant statistics to make the figure clearer.
• Still about Fig. 4, it is not clear to me why a discrete colormap is used when a continuous colormap totally fits. I find the discrete colormap a bit confusing. For example, what is the color for a coefficient value equals to 0.5 in the current setting? I suggest use a continuous colormap as in other figures.
• The metric Bias in Eq. (4) shows that it is the summation of the difference values between the reanalysis value and the station value. Should it be absolute value of the differences? Otherwise, positive and negative values will just cancel out each other.
• The authors might consider another metric in their analysis, the coefficient of efficiency (CE) by Nash and Sutcliffe (1970). It is sensitive to bias and errors in signal amplitude.
• Why is CRU not compared in Fig. 8? I understand that the title of the section is "the best reanalysis data", but in this scenario, CRU seems an alternative option that people can use, and it is analyzed in other figures. Perhaps a complete comparison can be put in the supplementary information.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

the manuscript the correlations and differences of climate reanalyses with weather observations and suggested the best climate reanalysis for the EAM region. The three reanalyses of ERA5, JRA55, and NCEP2 along with the gridded observation (CRU) were evaluated using the correlation coefficients (Pearson, Spearman, and Kendall), difference statistics (RMSE and bias), and Taylor diagrams, comparing their annual and seasonal temperatures and precipitations with those from the total of 537 weather stations across China, North Korea, South Korea, and Japan. I suggest the manuscript should be revised before being accepted.

(1) in 2.3. Methods section, Normalized RMSE  should be used instead of RMSE, NRMSE=RMSE/mean

(2)Pearson ,Spearman and Kendall correlation coefficients should consider   distribution type of data, I suggest authors analyze if the data series is nomal distribution.  

（3）please clarify the innovation of the manuscript.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

I suggest the mauscript can be accepted in the current form.