Revision on: Schwitalla et al., “Soil moisture-atmosphere coupling strength over Central Europe in the recent warming climate”

The manuscript provides an analysis of the inter-annual variability in land-atmosphere coupling focusing on the summers from 1991 to 2022 in Europe. The analysis uses the ERA5 reanalysis product and combines different coupling metrics and atmospheric and soil conditions to try to explain the drivers of summer conditions. Although the main idea of the manuscript sounds interesting, the text and explanation of the results are very confusing and in my opinion is not solid enough as it is to be published.

Specific comments:

• The authors decide to use a reanalysis for the study of LA coupling, soil conditions, surface pressure and air temperatures but then they use the E-Obs product for precipitation. Despite the good performance of E-Obs in precipitation, if the aim of the study is to analyze the interconnection between variables and LA metrics in the “reality” of the ERA5 product, I do not agree with the use of E-Obs in precipitation, when all the rest of variables are related to the ERA5 simulated precipitation. The change of E-Obs by ERA5 precipitation may not change the results, since they may agree in the classification of dry and wet summers but in my opinion it is a required step for adding consistency to the analysis.

• The use of ERA5 product for the study of LA coupling is also controversial per se, since land surface models like the one employed in ERA5 have several difficulties in simulating LA interactions (e.g. Dirmeyer et al., 2018) and the authors have not validated the ERA5 simulation of the coupling metrics employed in the analysis. Given that the validation of the ERA5 product simulation of LA coupling would correspond to another article, I recommend adding another reanalysis product or model to the analysis and focus on the agreement between models. If the authors still consider that adding more products is too complicated then I would recommend to add another section to the article discussing the possible effect of ERA5 uncertainties (e.g. in precipitation, LH and SH) reported in the literature on the results (not in general as it was done in line 391).

• The classification of summers in warm dry or wet was done using averages of the whole Europe as domain, although most of the time the authors only comment on the results over Germany and around countries (e.g. section 4.1.1). I think it would make more sense to classify the summer with the averages of an area centered in Germany avoiding the effect of the different regions in this classification and the different patterns that you obtain for the same category (e.g. warm and dry).

• The selection of the domain also complicates the description of the results. The authors focus the analysis over Germany, thus in some part of the text they comment only results over Central Europe (e.g. section 4.1.1) and in other all Europe (e.g. section 3.2), which makes it very confusing because the patterns and results in Europe are specially diverse. I would recommend the authors to focus the analysis in central Europe, avoiding the discussion of north and south areas, or to divide the results section into paragraphs dedicated to the phenomena happening in each region. Right now, the structure of the results includes the explanation of each variable over the whole Europe is very difficult to follow, especially when the authors try to connect results between variables, because it is not clear to which region they are referring. Something similar happens with the selection of warm and dry summers. The chosen criterion leads to very different results for the same category (perhaps because of the spatial variability in the whole Europe used as average for the classification). And then the description of results seems incomplete since the explanations of the authors do not apply for all areas and all summers in the same category. The two results sections need to be revised and re-organize to improve the clarity of results for different regions and years. Perhaps a good idea is to show the results by year (including all maps of the same year in the same figure) and reduce the number of selected years so we can better follow the story that the authors are suggesting.

• The results (sections 3 and 4) are also not clear and not well supported. For example, the analysis is incomplete (e.g. the reader is sent to section 4.3.2 and 3.3 in lines 202 and 203 but they do not exist) or it includes wrong references (e.g. Figure 14 in line 300, line 303 referring to high SH in Fig. 10 when Fig10 shows correlations, and other missing references in the paragraphs like in lines 248 and 249 Fig 6 and Fig8 should be cited). The results in section 3.0 are also supported by references mainly to abstracts in conferences (e.g. line 232) instead of on the results from the ERA5 product. The main justification for using ERA5 data for this analysis is that you have all the environmental variables more or less consistent with each other, so for example if in line 206 you are saying that the dry anomaly in summer is related to dry spring season you should support that with a map based on ERA5 in the supplementary information that supports that claim. The same happens with other explanations of results based on some heavy precipitation or drought events (e.g. lines 169, 238, 353,357-359), are these events really represented in the ERA5 data? Because if not the results that we are seeing are not related to that event.

• Some of the claims based on the results are not easy to follow or see in the maps (e.g. line 200 “By comparing Fig. 6 and 7…”). Perhaps a statistical analysis of spatial correlations between variables could help to reach more robust conclusions. Another example is in the paragraph starting in line 277, there are more coupling hot spots over central Europe for example in 2019 and 2006 but the soil moisture anomalies sometimes are negative and sometimes are positive. The authors should make an effort to explain the results that we are seeing or reduce the number of maps included in the manuscripts explaining the processes leading to warm conditions in particular years and areas. Also the interpretation of land atmosphere coupling should be revised in the manuscript. I am not sure the authors explain clearly the role of soil moisture deficits in the restriction of latent heat flux and the induced increase in temperature. For example, this is the case in the mentioned paragraph (line 277), since both strong and weak coupling are associated with positive soil moisture anomalies. We need more information about what is happening there.

• The conclusion sections could be reorganized, separating the discussion from the conclusions. In this way perhaps it is easier to identify the real conclusions of this study and the new information that we have learned about land-atmosphere coupling, which at the moment is not clear.

Minor comments:

• Line 111: “To complement our analysis, seasonal mean anomalies of 500 hPa geopotential (Lhotka and Kyselý, 2022) and ERA5 volumetric root zone soil moisture were calculated” This is not clear, do you mean geopotential height? Also are both variables calculated from ERA5 or do you use another database for the 500 hPa geopotential?

• There are some minor spelling errors. The text should be revised (e.g. line 64 “suggests”, line 75 “In the preceding…”, line 213 no new paragraph, line 228 “The very…” is a very long sentence and the connectors are not used well, line 169 “caused by the severe'', line 189 “previous” no “precious”, line 199 revise connectors, line 258 remove “in addition”, line 294 “suggests”, line 299 “is more often in...”, line 302 “a considerable increase in the HLCL”)

• Please, correct the order of maps (2003 and 1994) in figure 4.

• Line 236 “preventing a moisture limitation”, please check this sentence. Do you mean here “leading to moisture deficits”?

• Line 254 “show non-significant values” please avoid the “significant” word if you did not apply a significant test. If you did apply a significant test, please provide the details on the text.

• Line 246 “almost weak” Please replace it by another expression e.g. “mostly weak” or just “weak”.

• Line 249 “the average soil moisture availability” do you mean the “high soil moisture availability”?

• Fig 14, please add an explanation of why this selection of warm and dry years.

• Section 6 should be section 5.

• Paragraph starting in line 384, this is just the justification for using ERA5 in the analysis but not the discussion of how uncertainties in ERA5 like the overestimation of LH (Martens et al., 2020) could be affecting these results. This paragraph should probably be placed in section 2.1.

REFERENCES:

Dirmeyer, P. A., and Coauthors, 2018: Verification of Land–Atmosphere Coupling in Forecast Models, Reanalyses, and Land Surface Models Using Flux Site Observations. J. Hydrometeor., 19, 375–392, https://doi.org/10.1175/JHM-D-17-0152.1.

Martens, B., Schumacher, D. L., Wouters, H., Muñoz-Sabater, J., Verhoest, N. E. C., and Miralles, D. G.: Evaluating the land-surface energy partitioning in ERA5, Geosci. Model Dev., 13, 4159–4181, https://doi.org/10.5194/gmd-13-4159-2020, 2020.

Schwitalla et al. present a study on the coupling strength of moisture and energy fluxes, and atmospheric characteristics over the European continent. The goal is to understand interanual variability of the LA coupling sign and strength in different climatic regions under varying moisture and energy conditions in summer. Focus is placed on the drought conditions. As a reference they chose the climate period 1991 to 2020, while the time series investigated extends to 2022. In the quantification of the coupling sign and strength, they use standard indices and correlation coefficients based on linearity assumptions.

A plethora of work has been on published on the sign and strength of LA coupling in the past couple of decades. The study corroborates previous findings; in my opinion there are no surprises, or perhaps I have missed them. In this case, the authors need to revise the manuscript and clearly point out the new findings.

Schwitalla et al. present a study on the coupling strength of moisture and energy fluxes, and atmospheric characteristics over the European continent. The goal is to understand interanual variability of the LA coupling sign and strength in different climatic regions under varying moisture and energy conditions in summer. Focus is placed on the drought conditions. As a reference they chose the climate period 1991 to 2020, while the time series investigated extends to 2022. In the quantification of the coupling sign and strength, they use standard indices and correlation coefficients based on linearity assumptions.

A plethora of work has been on published on the sign and strength of LA coupling in the past couple of decades. The study corroborates previous findings; in my opinion there are no surprises, or perhaps I have missed them. In this case, the authors need to revise the manuscript and clearly point out the new findings.

In the summary and discussion section, the authors touch on the main the goal of the study and many interesting questions. For example, the authors state that “the hydroclimatological conditions during each summer drive considerable interannual variability in LA coupling…”. This would indeed be an interesting finding indeed. However, in my opinion, the analyses does not show this in a rigorous way. The indices of different years are presented, without further analysess. They also suggest “a growing influence of soil moisture variability on the meteorological conditions…” in the second half of the study period, which was drier than the first half. Again, the presentation of the coupling indices and linear correlations for individual years does not afford this conjecture in my opinion.

Two additional points that caught my attention are the categorization of the different years and application of the ERA5 data set (which was also brought up by the other reviewer). In the former, the classification appears to be rather arbitrary. The years 2021 may server as an example, which is categorized as a warm and dry year in the table, but exhibits a wet anomaly and is referred to as warm and wet in the text, if I am not mistaken. This type of confusion does not lend confidence in the results. In the latter, the issue of data assimilation in ERA5 in the diagnosis of LA coupling has to be discussed further. Also in my opinion, reanalyses are of limited value in feedback studies, which leads to the challenge of identifying feedbacks in simulations while reproducing real world weather conditions. Perhaps one has to make a choice and accept that in case of feedback studies in order to identify mechanisms, internal model consistency is more important than reproducing past weather. It would be interesting to understand the perspective of the authors in a more in depth discussion.

In my opinion, the study requires much more work beyond major revisions in order to contribute new and interesting results addressing the important issue of interannual variability of LA coupling in summer.

In the summary and discussion section, the authors touch on the main the goal of the study and many interesting questions. For example, the authors state that “the hydroclimatological conditions during each summer drive considerable interannual variability in LA coupling…”. This would indeed be an interesting finding indeed. However, in my opinion, the analyses does not show this in a rigorous way. The indices of different years are presented, without further analysess. They also suggest “a growing influence of soil moisture variability on the meteorological conditions…” in the second half of the study period, which was drier than the first half. Again, the presentation of the coupling indices and linear correlations for individual years does not afford this conjecture in my opinion.

Two additional points that caught my attention are the categorization of the different years and application of the ERA5 data set (which was also brought up by the other reviewer). In the former, the classification appears to be rather arbitrary. The years 2021 may server as an example, which is categorized as a warm and dry year in the table, but exhibits a wet anomaly and is referred to as warm and wet in the text, if I am not mistaken. This type of confusion does not lend confidence in the results. In the latter, the issue of data assimilation in ERA5 in the diagnosis of LA coupling has to be discussed further. Also in my opinion, reanalyses are of limited value in feedback studies, which leads to the challenge of identifying feedbacks in simulations while reproducing real world weather conditions. Perhaps one has to make a choice and accept that in case of feedback studies in order to identify mechanisms, internal model consistency is more important than reproducing past weather. It would be interesting to understand the perspective of the authors in a more in depth discussion.

In my opinion, the study requires much more work beyond major revisions in order to contribute new and interesting results addressing the important issue of interannual variability of LA coupling.