Early-Warning Signals of Drought-Flood State Transition over the Dongting Lake Basin Based on the Critical Slowing Down Theory
Round 1
Reviewer 1 Report
Interesting study, important contribution to applying EWS-analysis to real-world cases.
The one concern I have is that the actual reason why floods and droughts should be stable equilibria between which the dynamic system of the Dongting Lake basin shifts, does not become clear (respectively, is not addressed at all). This would be needed to justify the application of EWS-analysis, since - according to theory - autocorrelaton and variance are signals sent by a system, which bifurcates (i.e. changes the quality of its equilibrium from stable to unstable or back). May be you can elaborate a bit more on why the non-local climate and weather conditions (as described section 3.2) seem to cause such different equilibrium states.
Minor comments:
IMO Figure 1 is not really needed. the concept of rolling windows is well-known, and variance and autocorrelation are standard EWS
page 5, line 170: sentence is unclear
page 10, line 342: repeats preceding sentence
The assertions on page 10, line 337 and page 11, line 371 should perhaps be formulated more weakly. In any case, the signals do not appear as clear as the interpretation claims. And the assessment is only based on two empirical cases (1993 and 2003).
Author Response
Thanks to reviewers for their valuable suggestions which are very helpful to improve the quality of this paper.
In fact, this article doesn’t mean drought and flood should be two stable equilibrium state, as shown in Figure 2 in the manuscript. Before 2003, the durations of dry and wet states of the Dongting Lake basin are roughly the same. No one state last for a long time. Namely, the precipitation in the Dongting Lake basin fluctuates around a value (the climatic mean) stably. This is the so called "stable equilibrium state" in this paper. However, after 2003, the dry or wet state of Dongting Lake basin tended to last for a long time, which is another "stable equilibrium state" mentioned in this paper. The purpose of this paper is to investigate the early warning signals of the transition between these two "stable equilibrium states".
In order to correspond with the circulation analysis in section 3.2, we modified the original description of "drought-flood-drought" into the description of "less precipitation - more precipitation - less precipitation".
After responding to the comments from several other reviewers, we decided to retain Figure 1 to enhance the readability of the paper, since it seems that not all readers can immediately understand the algorithm from the text description.
Revisions in the manuscript have been marked in red.
Reviewer 2 Report
My comment is on the drought index. Why authors chose SPI as the drought index? Incorporating the role of temperature, as in the SPEI, could better characterize drought scenarios.
Author Response
Reply: Thanks to reviewers for their valuable suggestions which are very helpful to improve the quality of this paper. As one of the drought-flood monitoring indicators recommended by WMO, SPI has proved its effectiveness in different regions of the world. In general, the use of SPEI is more representative when evaporation in the study area or period has a significant impact on the dry and wet states. According to existing studies, SPEI has a better applicability in northern China, while SPI has a better applicability in southern China (Yang et al., 2017). The purpose of this article is to study the early warning signal of drought-flood transition. When a region is in a dry or wet state in a period of time (as shown in Figure 2), precipitation is the most direct reason for this stable state. So in this paper, SPI should be more representative, and the use of SPI could also facilitate follow-up studies on the related atmospheric circulations.
Yang, Q.; Li, M.; Zheng, Z.; Ma, Z. Regional applicability of seven meteorological drought indices in China. Science China Earth Sciences 2017, 04:147-162. doi: 10.1007/s11430-016-5133-5
Reviewer 3 Report
Authors should better discus along the text that an Early Warning System (EWS) includes hazards which manifest rapidly, such as tsunamis, as well as creeping hazards which manifest slowly, such as drought. (see Kelman and Glantz, 2014, at https://link.springer.com/chapter/10.1007/978-94-017-8598-3_5 ) Consequently, prediction in advance regards both large and short (i.e. in real time) time windows
Moreover, Authors should discuss (concerning their future developments for the proposed methodology) the possibility to evaluate probabilistic and/or deterministic SPI thresholds, which could be very useful for and efficient EWS in this context (see, for example, the methodology proposed in De Luca and Versace 2017).
References
- Kelman I., Glantz M.H. (2014) Early Warning Systems Defined. In: Singh A., Zommers Z. (eds) Reducing Disaster: Early Warning Systems For Climate Change. Springer, Dordrecht. https://doi.org/10.1007/978-94-017-8598-3_5 (see also https://link.springer.com/chapter/10.1007/978-94-017-8598-3_5)
- De Luca D.L.; Versace P. (2017). Diversity of Rainfall Thresholds for early warning of hydro-geological disasters. Advances in Geosciences, 44: 53-60, ISSN: 1680-7359, DOI: https://doi.org/10.5194/adgeo-44-53-2017
Author Response
Reply: Thanks to reviewers for their valuable suggestions which are very helpful to improve the quality of this paper. In fact, this article is not to predict when and where a particular drought or flood will occur, and what their effects will be, but is to find the early warning signals of a climate shift (drought-flood transition). The climate shift is not necessarily a danger per se, but it can have a huge impact on the frequency,intensity, duration and other characteristics of other hazards. Early warning signals of climate shift can remind people of the frequent occurrence of drought or flood in a certain area in the future, so as to encourage people to prepare for the possible consequences of this change. For example, we can do researches on some specific disasters and determine the disaster threshold. In addition, it is also possible to determine the high risk sites of disasters in the region by referring to the previous disasters, and to estimate the probability of future disasters by considering the current exposure and vulnerability.
Under the premise that the prediction accuracy of climate model needs to be further improved, the method in this paper can provide a new technology for early warning of climate transition by using existing observed data. In addition, the influence factors such as atmospheric circulation and sea surface temperature during climate transition can be analyzed in time to provide comparison and reference for improving climate models. Revisions in the manuscript have been marked in red.
