The referee kindly mentions the below paragraph at the beginning of the manuscript evaluation report.

The article provides a comprehensive overview of the current state of knowledge, detailing the methodology employed and presenting the obtained results effectively. I think that publishing the observations made in this study is crucial for improving our understanding of the interactions between the seismic cycle and fluid circulation. It deserves publication in a scientific journal.

We thank the referee for supportive and encouraging remarks. We also believe that publishing these observations are crucial as they are eye-opening and will lead to further studies by us and possibly others in near future; exactly as the referee suggests.

Here we would like to reply to what the referee has raised as a possible issue which is the possibility of seasonality of the AYRAN spring water the major ions content. 

In the manuscript, in addition to analyses of the available samples, we have mentioned - on Page 14. Lines 347-348 - that we have also utilized the information on major ion contents  for the Ayran water sample analyzed 10 years ago in 2012. we copied the section (Page 14. Lines 347-348) below. 

“We have also obtained a chemical analysis report on AYR water submitted with the business license application of the company dated 29 August 2012. The chemical analysis data of the samples collected more than 10 years ago include values only for Na⁺, Cl^–, and SO4^2– as 3.86, 3.12, and 8.37 mg/l, respectively. These values are very close to the analysis result of the AYR water sample dated 8 March 2022  (AYR 1 which is the oldest sample in our data set) and the AYR water samples collected after 15 February (Table 2); confirming that these samples represent background values for the AYR spring water.”

The analysis report dated August 2012 was taken also as a supporting evidence for our interpretation that the March 2022 and March 2023 samples represent chemical background before manifestation of pre-earthquake crustal deformation and after the February 2023 Earthquakes earthquake, respectively. Seasonal change in the major ion content does not seem to exist because we have sample both from March (2022 and 2023) and August 2012 which have comparable content for at least Na⁺, Cl^–, and SO4^-2.  The pdf of the business license issued on 29 August 2012 and chemical analysis on which this business certificate was issued are attached.

Therefore, we do believe  based on close similarity of major ion contents of samples from March (2022, 2023) and August (2012), that the changes in major ions are not due to seasonal effects are related to pre-earthquake crustal deformation as we have (and many others also have) shown in previous publications.   

Extending the time series cannot be backward as this is not possible ,but as we will continue our research for future publications, we will obtain future samples from the Ayran spring and conduct some analyses to further our understanding of the hydrogeochemical cycle.

At the moment our main aim is to publish these observations and our interpretations of these anomalies in light of seismicity. Once published, we believe that this paper will make a start point for future complimentary studies as the referee has suggested.

We appreciate referee's constructive criticism and open-minded approach.   

Dr. Vivek Walia

Thank you for your encouragement and constructive view and review on our manuscript. We will do review the manuscript and make all the corrections as you have suggested.

Thank you for kind consideration and scientific support you have provided for this manuscript.

General comments:

• The manuscript should be checked thoroughly and inconsistencies in grammar, tenses of verbs, prepositions, etc. should be corrected.
• There are some prominent repetitions of entire sentences or parts of sentences in Abstract, Conclusions, and the last paragraph Result and Discussion. This should be edited.

Specific comments are mentioned within the manuscript.

We thank Dr. Vivek Walia for his encouragement and constructive view and review on our manuscript. We have revised the manuscript and made all the style and linguistic corrections and eliminated repetitions as the referee has suggested.

Attached is the clean copy of the revised manuscript

We thank the anonymous referee for further comments and recommendations provided in a supplement file.

Below are our responses,

1) Table 2 lists the data and Figure 3 shows the plot of the data. This is conventional format to demonstrate.

2) Referee’s recommendations for utilizing additional data such as deep drilling, ground stress station data etc. are unfortunately unavailable in the study are. In the introduction Section of the manuscript we mentioned a multidisiplinary continuous monitoring network (including borehole tiltmeter stations, GPS stations, soil radon stations, cold and warm spring water stations, microseismology stations) that was established in 2006 (Inan et al., 2007) and continued until 2012. Unfortunately and as we regretfully mentioned in the introduction section, in 2012 the project was shut down and the montioring stations were removed. Therefore, we mentioned in the introdcution that “ the geosciences community was caught unprepared for the February earthquakes”. We the geoscience community llost the opportunity of obtaining very valuable pre-earthquake data for this giant earthquakes in the region.

3) The oldest sample before the earthquake was dated March 2022 (about one year before the earthquake and for this sample no anomalies were present in EC and/or major ions content). Moeover, as we mentioned the major ions of this sample was very similar to chemical analysis results of a sample that was analyzed and submitted for Business licence apllication in 2012 ! So the sample for March 2022 represent background (no sign of crustal deformation). However, the samples from September 2022 until February 2023 (covering a durtaion of six months) show undisputed positive anomalies (increase) in major ions. Therefore, we mentioned that pre-earthkuae anomaly lasted for at least six months (e.g., starting September 2022 and contiuning until the earthquakes of 6 February after which the mjor ion contents started to diminish and in mid February background levels were reached.). There is a gap in samples between March 2022 and September 2022. So the anomaly could have started some time after March 2022 and before September 2022; this would mean precursory anomaly more than six months. We have no way to speculate on this.

4) The water samples we analyzed were commercially bottled samples sold at markets/shops. Therefore, any variations in spring water temperatures are not possible to comment on. If the multidisiplinary monitoring network hasd not been removed, then contiunous water monitoring stations would have collected Ec, and temperature on hourly basis. For example for Please refer to results of continuous monitoring of water stations by the authors (Inan et al., 2010) cited in the manuscript. Unfortunately, before February earthquakes there was no and yet there is no continuous water monitoring in the region.

Thank you for recommending work of He and Singh (2018), we will certainlt refer to this in the introdcution. The paper discusses co-seismic response of the groundwater levels which is a well known phenomena reported to have taking place before, during, and after several earthquakes worldwide.

5) Two immediate possible mechanisms potentially causing chemical changes in the water samples discussed in the manuscript are quite different . One envisages mixing of different aquifers (pre-earthquake crusta dilataion theory) and the other mechanism suggests chemical corrosion of rocks by positive hole currents yielding more ions into the circulating water before earthquakes. The references given for each mechanism contain wealth of drawings therefore we did not want to repeat what others have already demonstrated.

We thank the anonymous referee for further comments and recommendations he/she has provided in a supplement file.

Below are our responses.

The work tested the changes in electrical conductivity and the ion concentrations in two springs near the epicenters before and after two earthquakes (Mw 7.7 and Mw 7.6, in Kahramanmaraş), suggesting the anomalies with an increase electrical conductivity and the major ions before the earthquake. The analytical methods used and the results obtained are reliable, but the conclusions are limited due to the limited samples (the number of spring holes) and the significant results from only one point. The manuscript needs major revision. I believe that the authors have done their best to obtain data on all possible spring holes, and it would be difficult to recommend that they test more samples from other possible springs.

Thank you for the understanding. We concur with the referee that at the moment due to lack of any ongoing project and no funding, it is not possible to collect more samples from the area for analysis. On the other hand, we are almost certain that we cannot find bottled samples dated before the February earthquakes.

Therefore, from other perspectives, I think the following possible improvements still exist, depending on the authors.

1. In terms of form, the contents of Table 2 and Figure 3 are somewhat repetitive, and it is suggested that the author could optimize or merge them.

Table 2 lists the data and Figure 3 shows the plot of the data for easy discussion and for the readers to follow.  This is conventional format to demonstrate.

1. Since the author only analyzed the relationship between the ion concentration changes and the earthquake from the perspective of time and location, this analysis is only correlation analysis and lacks causality analysis. Therefore, the author need try to find some data that can reflect the change of crust stress, such as deep drilling data, ground stress station data or satellite observation of surface displacement and deformation data (e.g., InSAR) to support the rationality of ion concentration changes from the perspective of time and space. I think this can greatly increase the reliability of the results of the work.

Referee’s recommendations for utilizing additional data such as deep drilling, ground stress station data etc. are very logical but unfortunately unavailable in the study are for pre-earthquake time period. In the introduction Section of the manuscript, we mentioned a multidisiplinary continuous monitoring network (including borehole tiltmeter stations, GPS stations, soil radon stations, cold and warm spring water stations, microseismology stations) that was established in 2006 (Inan et al., 2007) and continued until 2012. Unfortunately and as we regretfully mentioned in the introduction section, in 2012 the project was shut down and the montioring stations were removed. Therefore, we mentioned in the introdcution that “ the geosciences community was caught unprepared for the February earthquakes”. We the geoscience community llost the opportunity of obtaining very valuable pre-earthquake data for this giant earthquakes in the region.

1. From the time series observation data of the spring water, the increase in ion concentration began as early as one year before the M7.7 earthquake. However, since the author only presented the results of one year, the process of increasing ion concentration was not fully displayed, and the audience could not clearly see when the increase in ion concentration began, and whether it was in a low value stable state more than one year before the M7.7 earthquake. Please add this part of data, if not, please explain the reason.

The oldest sample before the earthquake was dated March 2022 (about one year before the earthquake and for this sample no anomalies were present in EC and/or major ions content). Moeover, as we mentioned the major ions of this sample was very similar to chemical analysis results of a sample that was analyzed and submitted for Business licence apllication in 2012 ! So the sample for March 2022 represent background (no sign of crustal deformation). However, the samples from September 2022 until February 2023 (covering a durtaion of six months)  show undisputed positive anomalies (increase) in major ions. Therefore, we mentioned that pre-earthkuae anomaly lasted for at least six months (e.g., starting September 2022 and contiuning until the earthquakes of 6 February after which the mjor ion contents started to diminish and in mid February background levels were reached.). There is a gap in samples between March 2022 and September 2022. So the anomaly could have started some time after March 2022 and before September 2022; this would mean precursory anomaly more than six months. We have no way to speculate on this.

1. In addition, I would like to know whether the author has obtained synchronized spring temperature data, which I think is also crucial to reflect the process and results of underground fluid migration. If so, it is suggested that the author add relevant content and make analysis. If not, I suggest that the author refer to the schematic diagram in Figure 10 of this paper (He A, Singh R P. Groundwater level response to the Wenchuan earthquake of May 2008[J]. Geomatics, Natural Hazards and Risk, 2018.), and combine the location and lithology of the spring in your work to analyze and discuss the reasons for the abnormal spring.

The water samples we analyzed were commercially bottled samples sold at markets/shops. Therefore, any variations in spring water temperatures are not possible to comment on. If the multidisiplinary monitoring network hasd not been removed, then contiunous water monitoring stations would have collected Ec, and temperature on hourly basis. For example for Please refer to results of continuous monitoring of water stations by the authors (Inan et al., 2010) cited in the manuscript. Unfortunately, before February earthquakes there was no and yet there is no continuous water monitoring in the region.

Thank you for recommending work of He and Singh (2018), we have referred to this in the introduction of the revised manuscript. He and Singh (2018) paper discusses co-seismic response of the groundwater levels which is a well-known phenomenon reported to have taken place before, during, and after several earthquakes worldwide.

1. Line 402, two immediate mechanisms was presented. Here the authors should elaborate them in detail as you can as possible, for example providing some schematic diagrams associating to the potential mechanisms. This is very important to understand the physical processes and increase the reliability of this work.

Two immediate possible mechanisms potentially causing chemical changes in the water samples discussed in the manuscript are quite different. One envisages mixing of different aquifers (pre-earthquake crustal dilatation theory) and the other mechanism suggests chemical corrosion of rocks by positive hole currents causing weathering of the rock surface yielding more ions into the circulating water before earthquakes. The references given for each mechanism contain wealth of drawings; therefore, we did not want to repeat what others have already demonstrated.

We thank the referee for kind approach and rightful comments that provided us an opportunity to discuss the issue a bit further. 

We agree that the precursory anomaly and (duration of this anomaly) that can be detected at a monitoring station depends on magnitude of earthquake and also the epicentral distance of the earthquake to the monitoring site. The closer the epicenter of an earthquake of a given magnitude will cause longer duration of anomaly at monitoring site. For example, taking Sultankhodhaev's (1984) empirical relation (log (DT) = 0.63 * M – b) which takes into account 1) earthquake magnitude (M), 2) distance of the earthquake epicenter to monitoring site (D), and 3) duration of anomaly (T), we can estimate duration of anomaly that can be expected in a monitoring site.   

Taking the example given by the referee, it can be said that an earthquake of magnitude 4.5 occurring at a distance about 100 km to the monitoring station can lead to duration of anomaly (T) as about 5 days before the earthquake (using log(100*T) = 0,63* 4.5 - 0.15) . On the other hand, the duration of anomaly (T) for an earthquake Magnitude 7.7 whose epicenter is 100 km away from the monitoring site can be estimated as 295 days  (using log(100*T) = 0,63* 7.7 - 0.15). So the greater the earthquake magnitude, the longer the duration of the precursory anomaly.

Obviously, we should keep in mind that the relations proposed by Dobrovolsky et al. (1979), Sultankhodhaev (1984), and Rikitake (1987) assume homogenous and isotropic crust where pre-earthquake stress and resultant strain propagates in all directions.  In fact, we know that this assumption is not totally correct as microplates and/or block boundaries hinder stress transfer (e.g., Inan et al. 2012b).   This issue should be seriously considered and care should be exercised.    

The referee's last comment is also very worth to comment on. "How to discriminate between a medium to small earthquake and a major one" based on precursory anomalies would definitely require an ideal network of distributed monitoring stations (preferably of multi-disciplinary nature) established at sufficient distance to each other. So that each monitoring station would detect precursory anomalies for different durations and that would ideally help the interpreter to use simple algorithms to detect possible location and magnitude of an approaching earthquake.  Here we should emphasize again that crustal heterogeneity and structural complexity is a must to consider. 

RC5: 'Comment on nhess-2023-133', Anonymous Referee #4, 12 Oct 2023 

The paper "Spring water anomalies before two consecutive earthquakes (Mw 7.7 and Mw 7.6) in Kahramanmaraş (Türkiye) on 6 February 2023" is of scientific relevance as it addresses the issue of hydrogeochemical precursors of major earthquakes. Although, unfortunately, two major earthquakes occurred this year in Turkey, within a short time of each other, causing massive destruction, the scientific community mobilized in an exemplary way in analyzing the situation and one of its efforts focused on precursors. A positive fact that could be deduced from these major occurrences is that it could calibrate the certain ionic anomalies that appear in natural spring waters about 6 months pre-earthquakes, as the authors point out in this paper.

As the article have very carefully considered by my fore-referees, I would like to draw attention to one aspect. In Subchapter 3.4, there is a discussion about the deduction of maximum distance from the epicenter where the anomalies could be detected versus magnitude. Actually, as it stated, an earthquake of Mw 4.5 could cause an ions anomaly at a distance of 100 km from the epicenter (Ayran Spring), but an earthquake of Mw 7,7, too.

I think there has to be a discussion in this paper or underlined that it would be addressed in the future, how to discriminate between a medium to small earthquake and a major one, if both could cause hydrogeochemical anomalies in the spring waters, because in the first case there really is no need to trigger an alarm.

We thank the referee for kind and open-minded approach and rightful comments that provided us an opportunity to discuss the issue a bit further. Parts of the reply below has been incorporated into the revised manuscript. 

We agree that the precursory anomaly and (duration of this anomaly) that can be detected at a monitoring station depends on magnitude of earthquake and also the epicentral distance of the earthquake to the monitoring site. The closer the epicenter of an earthquake of a given magnitude will cause longer duration of anomaly at monitoring site. For example, taking Sultankhodhaev's (1984) empirical relation (log (DT) = 0.63 * M – b) which considers 1) earthquake magnitude (M), 2) distance of the earthquake epicenter to monitoring site (D), and 3) duration of anomaly (T), we can estimate duration of anomaly that can be expected in a monitoring site.   

Taking the example given by the referee, it can be said that an earthquake of magnitude 4.5 occurring at a distance about 100 km to the monitoring station can lead to duration of anomaly (T) as about 5 days before the earthquake (using log(100*T) = 0,63* 4.5 - 0.15) . On the other hand, the duration of anomaly (T) for an earthquake Magnitude 7.7 whose epicenter is 100 km away from the monitoring site can be estimated as 295 days (using log(100*T) = 0,63* 7.7 - 0.15). So, the greater the earthquake magnitude, the longer the duration of the precursory anomaly.

Obviously, we should keep in mind that the relations proposed by Dobrovolsky et al. (1979), Sultankhodhaev (1984), and Rikitake (1987) assume homogenous and isotropic crust where pre-earthquake stress and resultant strain propagates in all directions.  In fact, we know that this assumption is not totally correct as microplates and/or block boundaries hinder stress transfer (e.g., Inan et al. 2012b).   This issue should be seriously considered and care should be exercised.    

The referee's last comment is also very worth to comment on. "How to discriminate between a medium to small earthquake and a major one" based on precursory anomalies would definitely require an ideal network of distributed monitoring stations (preferably of multi-disciplinary nature) established at sufficient distance to each other. So that each monitoring station would detect precursory anomalies for different durations and that would ideally help the interpreter to use simple algorithms to detect possible location and magnitude of an approaching earthquake.  Here we should emphasize again that crustal heterogeneity and structural complexity is a must to consider.