|Siegmund et al have done a number of revisions on the original version with a few interesting new aspects. Overall, however, I cannot recommend that the manuscript should be published in Biogeosciences for the following reasons.|
In their comments to the referees the authors elaborate on how coincidence analysis differs from ordinary correlation analysis. I think both reviewers have understood this difference. The authors did not convince me of a novel result derived from coincidence analysis that cannot be derived directly from past analysis based on correlation analysis. In essence their main result boils down to the finding that extreme temperatures in spring can have an effect on flowering dates, which is already known (many of the relevant work is cited by the authors). The use of Figure 1 in their reply as argument that coincidence analysis can give qualitatively very different results than correlation analysis is simply not convincing and in fact even incorrect. With correlation analysis one would conclude that higher spring temperatures lead to earlier flowering dates. This includes that also that extreme spring temperatures lead to extremely early flowering dates. Coincidence analysis doesn’t add anything new here. The argument that coincidence analysis can show the validity of the relationship for extremes is granted but has nowhere been shown explicitly in the paper. Further, the relevance of this result has to be evaluated and discussed (since most often such a result is implicitly already contained using correlation analysis).
Coincidence analysis uses only a subset of the distribution, in my opinion interesting results can thus be obtained in particular when they do not match previous analyses based on the entire distribution.
Neglecting reviewer’s requests with the argument that “all questions raised by the reviewers would provide much more results than can be meaningfully described in a single paper” does not give much credit to the reviewers who carefully went through the manuscript and made suggestions how to improve/obtain meaningful results. Not all analyses have to be put in a paper. If no interesting results have been found, spare the readers with details. For instance, I still see no point in including figures 7 and 8 in the manuscript. You cannot assess visually whether there is a clustering or not. Without a rigorous test, the conclusion that can be drawn by visual inspection are not of much value.
The authors also neglect current developments in the field. For instance, Laube et al. (2014) discuss the factors controlling budburst in spring mentioning, besides others, the impacts of cold spells and temperatures in the previous autumn. Consequently, an interesting question would be: How could the authors’ results contribute to model development in the field? All in all I believe the manuscript would be more suitable for a more technical journal.
The abstract lacks specific results that set this paper apart from previous studies on the subject.
L 6: “ecosystem resilience”, what do you mean by this term
L 8: “severe ecological disturbances”, what do you mean by that
L 11: incorrect, there is no robust single of drought trends in Europe. Generally the North gets wetter, the Mediterranean drier and the center stays more or less the same.
L 17: cite Seneviratne et al. (2012)
L 35: cite Frank et al. (2015), Reichstein et al. (2013)
L 68: I don’t think it can be called a “debate”, maybe use a different word. The effects are probably highly species specific.
L 150-154: Are the analyzed species the only ones that fall into the selection criteria? Please specify.
L 195: why was the data normalized? Using the original units might allow to derive statements such as “A change in T/P by X leads to a shift in flowering dates by Y”
L 338: “coincide significantly” is quite sloppy statistically. What can only be done statistically is to reject the hypothesis that a certain coincidence rate has occurred by chance.
L 348-359: does not add relevant information, can be omitted
L 374: This is incorrect. If two variables have a correlation coefficient of one, of course this implies that when one variable is extreme, the other is too. This also holds approximately when the correlation is high.
L 460: As in a few other places, here the difference between coincidence analysis and correlation analysis is discussed. Yet to really emphasize how both approaches differ a direct comparison should be made. This could be done, for example, by doing the correlation analysis, analyzing what that would imply for the extremes and comparing that to coincidence analysis. Otherwise it is not clear whether the authors’ arguments hold.
Fig. 3: There is a slight change in slope in the extremes for 3 species. However, to draw robust conclusions a thorough analysis should be done, comparing correlation and coincidence analysis (see above).
L 565: Specify which regions and species. Are there opposing results for the same species?
L 605: Specify which statistical properties the boxplot shows (line, box, whiskers). Discussing mean and standard deviations based on boxplots is a bit odd because usually it’s the median that is shown and the standard deviation cannot be derived by visual inspection.
L 636: Did the authors test for these dependencies? Otherwise omit.
There is a discussion missing about Section 4.5. What are the reasons for the shifts visible in Figure 6? What are its implications? How does the bivariate approach differ from what is expected from a univariate analysis (see e.g. Zscheischler et al., 2014)?
In all plots font sizes are much too small.
Frank, D. A., Reichstein, M., Bahn, M., Thonicke, K., Frank, D., Mahecha, M. D., Smith, P., Van der Velde, M., Vicca, S., Babst, F., Beer, C., Buchmann, N., Canadell, J. G., Ciais, P., Cramer, W., Ibrom, A., Miglietta, F., Poulter, B., Rammig, A., Seneviratne, S. I., Walz, A., Wattenbach, M., Zavala, M. A., and Zscheischler, J.: Effects of climate extremes on the terrestrial carbon cycle: concepts, processes and potential future impacts, Global Change Biology, 21, 2861-2880, 2015.
Laube, J., Sparks, T. H., Estrella, N., and Menzel, A.: Does humidity trigger tree phenology? Proposal for an air humidity based framework for bud development in spring, New Phytologist, 202, 350-355, 2014.
Reichstein, M., Bahn, M., Ciais, P., Frank, D., Mahecha, M. D., Seneviratne, S. I., Zscheischler, J., Beer, C., Buchmann, N., Frank, D. C., Papale, D., Rammig, A., Smith, P., Thonicke, K., van der Velde, M., Vicca, S., Walz, A., and Wattenbach, M.: Climate extremes and the carbon cycle, Nature, 500, 287-295, 2013.
Seneviratne, S. I., Nicholls, N., Easterling, D., Goodess, C. M., Kanae, S., Kossin, J., Luo, Y., Marengo, J., McInnes, K., and Rahimi, M.: Changes in climate extremes and their impacts on the natural physical environment, Managing the risks of extreme events and disasters to advance climate change adaptation, 2012. 109-230, 2012.
Zscheischler, J., Michalak, A. M., Schwalm, C., Mahecha, M. D., Huntzinger, D. N., Reichstein, M., Berthier, G., Ciais, P., Cook, R. B., El-Masri, B., Huang, M. Y., Ito, A., Jain, A., King, A., Lei, H. M., Lu, C. Q., Mao, J. F., Peng, S. S., Poulter, B., Ricciuto, D., Shi, X. Y., Tao, B., Tian, H. Q., Viovy, N., Wang, W. L., Wei, Y. X., Yang, J., and Zeng, N.: Impact of large-scale climate extremes on biospheric carbon fluxes: An intercomparison based on MsTMIP data, Global Biogeochem Cy, 28, 585-600, 2014.