Climate Extremes over the Arabian Peninsula Using RegCM4 for Present Conditions Forced by Several CMIP5 Models
Round 1
Reviewer 1 Report
The article “Climate extremes over the Arabian Peninsula using RegCM4 for present conditions forced by several CMIP5 models” in which the author discussed the findings of temperature and precipitation extremes over the Arabian Peninsula using the data from a regional climate model RegCM4.
Considering the scope of the study in the region and limited published literature in the area I would like to accept this article after the following suggestions and corrections.
Line 39-40, the author wrote “ Observations show that in the recent decades, the precipitation in Saudi Arabia is following a decreasing trend (a drop of 47.8 mm/decade [4]), although the number of heavy precipitation events has increased in the last decade” decrease in precipitation for which time period? In the introduction section, the author referred to studies related to climate models. I would suggest adding an at-least a couple of recent studies on historic observations in the study region. Some of the statements are not clear, i.e. line 125-126, the author wrote “The bias for temperature and precipitation fields is obtained 125 from the model with respect to the Climatic Research Unit (CRU) monthly data”, what does that mean? The bias correction was applied to this data if that which method was used? If not elaborate. The reference list is very odd, if the author selected with a numbered reference type, it should be 1,2… so on; I would suggest changing accordingly. The author defined some of the terminologies like Tmax and Tmin but all variables used like dry days, I would suggest defining, as threshold varies from region to region (all the selected indices must be defined in the methodology part of the paper. Section “Data and methodology” missing methodology part mentioned above. In the results section table 1, the author should remove the column significant level, as p-values provide enough information about significant level. There is no need to write an index like “Ice day”, as such an index is not required for the study region. The author did not mention which trend detection method was applied, linear regression or MK test (applied to both tables 1&2), the method should be mention in the methodology section of the paper Some of the statements i.e. line 242-243 “Warm days (TX90p): Warm days is a percentile index calculated from the daily maximum temperature that varies from region to region over the study area” belongs to the methodology section of the paper. In figure 10 author showed the comparison of different models against the observed data, the data showed a quite good fit until 2010, more particularly for the warm spell duration index it showed a step change afterward; is this data quality issue or extreme heat event; refer to the published literature.
Author Response
Comments and Suggestions for Authors
The article “Climate extremes over the Arabian Peninsula using RegCM4 for present conditions forced by several CMIP5 models” in which the author discussed the findings of temperature and precipitation extremes over the Arabian Peninsula using the data from a regional climate model RegCM4.
Considering the scope of the study in the region and limited published literature in the area I would like to accept this article after the following suggestions and corrections.
Response: Thank you very much for your comments on my article which definitely improved its quality.
Line 39-40, the author wrote “Observations show that in the recent decades, the precipitation in Saudi Arabia is following a decreasing trend (a drop of 47.8 mm/decade [4]), although the number of heavy precipitation events has increased in the last decade” decrease in precipitation for which time period?
Response: The period 1978-2009 is now mentioned in the text.
In the introduction section, the author referred to studies related to climate models. I would suggest adding an at-least a couple of recent studies on historic observations in the study region.
Response: Thanks for this, a few latest references on the observations are added in the revised version. The references are:
Atif, R.M.; Almazroui, M.; Saeed, S.; Abid, M.A.; Islam, M. N.; Ismail, M. Extreme precipitation events over Saudi Arabia during the wet season and their associated teleconnections. Atmospheric Research 2019, 231: 104655.
Saeed, S.; Almazroui, M. Impacts of mid‑latitude circulation on winter precipitation over the Arabian Peninsula. Climate Dynamice, 2019, https://doi.org/10.1007/s00382-019-04862-6.
Almazroui, M.; Saeed, S. Contribution of extreme daily precipitation to total rainfall over the Arabian Peninsula. Atmospheric Research 2020, https://doi.org/10.1016/j.atmosres.2019.104672.
Some of the statements are not clear, i.e. line 125-126, the author wrote “The bias for temperature and precipitation fields is obtained from the model with respect to the Climatic Research Unit (CRU) monthly data”, what does that mean?
Response: Thanks for this, now the definition of bias is mentioned in the revised text.
The bias correction was applied to this data if that which method was used? If not elaborate.
Response: Thanks for this. The simulated data are used as obtained from RegCM4 without bias correction. If bias is corrected with respect to observations, then there will be not much difference between the number of extremes from both simulated and observed datasets.
The reference list is very odd, if the author selected with a numbered reference type, it should be 1,2… so on; I would suggest changing accordingly.
Response: Thanks for this. This is done.
The author defined some of the terminologies like Tmax and Tmin but all variables used like dry days, I would suggest defining, as threshold varies from region to region (all the selected indices must be defined in the methodology part of the paper. Section “Data and methodology” missing methodology part mentioned above.
Response: Thanks for this. The indices are calculated following ETCCDI and using RClimDex where threshold for each station location are used, now it is mentioned clearly.
In the results section table 1, the author should remove the column significant level, as p-values provide enough information about significant level.
Response: Thanks for this. This is done.
There is no need to write an index like “Ice day”, as such an index is not required for the study region.
Response: Thanks for this. This is done.
The author did not mention which trend detection method was applied, linear regression or MK test (applied to both tables 1&2), the method should be mention in the methodology section of the paper
Response: Thanks for this. This is added in the revised version.
Some of the statements i.e. line 242-243 “Warm days (TX90p): Warm days is a percentile index calculated from the daily maximum temperature that varies from region to region over the study area” belongs to the methodology section of the paper.
Response: Thanks for this, followed the suggestion and added under Analysis procedure.
In figure 10 author showed the comparison of different models against the observed data, the data showed a quite good fit until 2010, more particularly for the warm spell duration index it showed a step change afterward; is this data quality issue or extreme heat event; refer to the published literature.
Response: Thank you very much to indicate this point. Yes, simulated indices seem closer to the same from observations till around 2010 and after that large variations are observed. The quality control of data is performed in the same procedure for all years. In our experience, extreme events are occurring more in the recent years as compared to the previous years, in particular extremes in Saudi Arabia are happening more frequently since 2009. This might be a reason and needs more analysis in detail on this issue.
Author Response File: 
Author Response.docx
Reviewer 2 Report
Dear author,
My comments can be found in the attached pdf.
Comments for author File: 
Comments.pdf
Author Response
Comments and Suggestions for Authors
The author presents an interesting analysis on extreme climate indices using several simulations of a regional climate model (RegCM4). The simulations use two land schemes (BATS, CLM) and four different datasets for initial and boundary conditions, ERA-interim reanalysis, GFDL, HadGEM and ECHAM6. Overall the manuscript is well written and the results are well presented. Below I am including some minor comments about the current state of the manuscript. In my opinion a discussion about the two different land schemes should be added before acceptance.
Response: Thanks a lot for your encouraging comments and provides suggestion for the improvement of the manuscript.
Minor Comments
183-183 The minimum temperature estimation of BATS is significantly lower than CLM in all driving model experiments (Fig 2, or more pronounce at Sup 2). It would be very informative for future readers to discuss why this is happening. What is the main difference of BATS and CLM that drives this difference in minimum temperature? 203-204 Same as above, it’s quite important to explain or at least suggest the probable mechanism that drives the precipitation differences between BATS and CLM? There is some discussion about the differences in wind and relative humidity of CLM and BATS simulations (397-398) but still an explanation in the physics/schemes/assumptions behind these two land surface models would be highly beneficial.
Response: Thanks for these. A comparison of BATS and CLM is added in the revised version under experimental setup.
General Comments
86-88 reference is missing
Response: This is done, reference is added.
88-89 The sentence belong to the 2.2 section
Response: This is done, moved in Section 2.2.
92-94 To be precise, most probably the exact horizontal resolution for this ECHAM6 resolution is 1.875x1.875
Response: We followed the published article.
Reference:
Almazroui, M.; Islam, M. N.; Saeed, S.; Alkhalaf, A.K.; Dambul, R. Assessment of Uncertainties in Projected Temperature and Precipitation over the Arabian Peninsula Using Three Categories of Cmip5 Multimodel Ensembles. Earth Syst Environ 2017a, 1(1), https://doi.org/10.1007/s41748-017-0027-5.
107-121 The experimental set up presentation can be improved by using a table.
Response: Yes, we can make a table, however the information provided are almost similar.
125-128 CRU should be introduced in the Data section. The availability of the data that composed the CRU product should also be mentioned. For example the representation of the stations used for CRU over Sahara desert was quite sparse (Tsikerdekis et al., 2019).
Response: This is done, text moved and modified.
225-227 This phrase should be moved or something similar can be indicated at the end of the Introduction.
Response: Thanks for this, the sentence is removed from the revised version.
309 “Importantly” may be replaced with “Notably”?
Response: This is done
309-311 Not clear enough what the author means by “exactly in phase”
Response: The sentence is modified.
After 1996 warm days obs display a sharp increase which all of the experiments are unable to simulate. Is this there something specific about this year?
Response: The experience shows that in the recent years, extremes are more frequent compare to the previous decades, this seems an indication.
435-436 Instead of downscaled I would suggest to use “used as initial and boundary conditions (ICBC)”. Because in truth RegCM uses this GCMs data on the boundaries and the rest is guided by each own physics parameterizations.
Response: Thanks for this. Followed the suggestion.
439-440 A map showing the location of these 27 stations would highly help future readers to get a grasp of this ground station network spatial coverage.
Response: The reference for the station location map along with stations names is referred under Data because already 16 figures are in this manuscript.
There is something odd with the references numbering (they are not in ascending order). I am not sure if this is the author’s fault or may be a product of automated post-processing by the journal.
Response: The reference numberings are now as they appears in the text.
References Tsikerdekis, A., Zanis, P., Georgoulias, A. K., Alexandri, G., Katragkou, E., Karacostas, T. and Solmon, F.: Direct and semi-direct radiative effect of North African dust in present and future regional climate simulations, Clim. Dyn., 53(7–8), 4311–4336, doi:10.1007/s00382-019-04788-z, 2019.
Response: The reference is now cited in the revised version.
Reviewer 3 Report
This study evaluate the simulated climate extremes indices by a regional climate model RegCM4 that downscaling from the CMIP5 database. The conclusions indicate that RegCM4 well-captures the temperature extremes and overestimate the rainfall extremes over Arabian Peninsula for the present climate (1986-2005). However, the analysis results seems not totally agreement with the conclusions, especially in temperature extremes. I suggest the author should modify the paragraph associated with the results of temperature extreme indices. My comments are as follows:
Major
Line 186: In Fig.1, it seems all the experiments do not capture the distribution of maximum temperature. Line 317-319: Fig.10 shows that RegCM4 and observation have large differences in temperature indices after 2000. The RegCM4 do not capture the temperature extremes for entire present climate. Section 3.3: Fig. 16 is missing. Suggest to plot a map indicate the locations of 27 meteorological stations.
Minor
The number of vertical levels in AR5 models are in general over 20 levels. When the author downscaling the horizontal resolution to 25km in RegCM4, is it appropriate to use only 18 levels? Will this low vertical resolution smooth out the convection structure and affect the results of simulating extreme precipitation? Figure 14: lack of the legend of OBS (gray dashed line).
Author Response
Comments and Suggestions for Authors
This study evaluate the simulated climate extremes indices by a regional climate model RegCM4 that downscaling from the CMIP5 database. The conclusions indicate that RegCM4 well-captures the temperature extremes and overestimate the rainfall extremes over Arabian Peninsula for the present climate (1986-2005). However, the analysis results seems not totally agreement with the conclusions, especially in temperature extremes. I suggest the author should modify the paragraph associated with the results of temperature extreme indices. My comments are as follows:
Response: Thanks for the comments on my manuscript. The manuscript is revised based on your suggestion and comments.
Major
Line 186: In Fig.1, it seems all the experiments do not capture the distribution of maximum temperature.
Response: The text is revised and it is now as follows:
In general, the patterns of simulated maximum and minimum temperatures follow the patterns of the CRU data with some variations such as the highest maximum temperature simulated in the southeast Peninsula in all experiments not appearing in the observations. Very close scrutiny indicates that there is a difference between the observation and simulations, which reflects the limitation of climate models in capturing climatic information at local level. Simulated temperatures are higher than the observations in the southeast and lower in the northwest parts of the peninsula.
Line 317-319: Fig.10 shows that RegCM4 and observation have large differences in temperature indices after 2000. The RegCM4 do not capture the temperature extremes for entire present climate.
Response: Thanks for this suggestion. The revised text is now as following:
Fig. 10 shows that RegCM4 and observation have large differences in temperature indices after 2000. Overall, we can conclude that RegCM4 simulations are able to capture few temperature extremes such as warm days, warm spell duration, and cold nights before the 2000 and the RegCM4 simulations do not capture the observed temperature extremes for the entire present climate over the study region.
Section 3.3: Fig. 16 is missing. Suggest to plot a map indicate the locations of 27 meteorological stations.
Response: Thanks for this, the Fig. 16 is added in the revised version. Because already 16 Figures are there, therefore the station location are referred in one of our publication.
Minor
The number of vertical levels in AR5 models are in general over 20 levels. When the author downscaling the horizontal resolution to 25km in RegCM4, is it appropriate to use only 18 levels? Will this low vertical resolution smooth out the convection structure and affect the results of simulating extreme precipitation? Figure 14: lack of the legend of OBS (gray dashed line).
Response: Thanks for this. Yes, the CMIP5 datasets is downscaled to horizontal resolution to 25km using RegCM4 with 18 vertical levels. We admit that for large domain the 23 level is more suitable and most of CMIP5 models data are 20 levels. The gray dash line was hidden and now visible in Fig. 14.
Author Response File: Author Response.docx
Round 2
Reviewer 3 Report
The author has fully replied the questions in first review. The new manuscript has clarified some previous concerns and is revised based on previous comments. I have no more comments and agree to accept in present form.
