Climate Change and Extreme Weather Drive the Declines of Saline Lakes: A Showcase of the Great Salt Lake
Round 1
Reviewer 1 Report
Thanks to the author for providing replies to my comments and for modifying the manuscript. Still, I don’t find the manuscript convincing. Sure, finding more positive extremes for the recent time period compared to an earlier one in an overall (slightly) decreasing time series is a bit contra-intuitive and reasons for this behaviour should be discussed. However, the argument which the author uses to disprove water use as a main contributor for this long-term trend, namely that there are more positive extremes in the later time periods than in earlier, would be the same for climate. The author assumes (without presenting any data) that water use and consumption increase in time (which may very well be the case) and concludes that this conflicts with the increased frequency of positive extremes. But how is this different from climate effects? The author shows the increase in temperature, concludes from this that evaporation would be increased as well (which again I would agree to), but apparently sees no conflict with the recent positive lake level extremes in this case? Why is that? Please also note that water use is not necessarily constant just because population density and industry did not change. Water is consumed if needed and typically in wet years also e.g. agricultural water use is lower.
Furthermore, the shown time period may play a role in this case. Looking at figure 2 (https://www.nature.com/articles/ngeo3052/figures/2) in the related article by Wurtsbaugh et al, there are 25 years of lake level variation data available prior to the period shown by the author. Looking at the USGS website, there are even 50 more years of data which are neglected in this analysis. Why are they not included? As the earlier years are much wetter, the number of positive extremes are reduced (see attached figure). Defining extremes with percentiles (5 and 95) instead of standard deviation (which are much more robust against outliers) there are more negative extremes as would be expected from a time series with a decreasing trend. Subtracting the trend, the positive extremes are concentrated again in the 1980’s and can probably be related to an exceptional wet period (as the extremes also show up in precipitation and river flow) which recharge the lake in spite of climate change and/or human impacts. Still, this does not prove anything in terms of which effect is responsible for the trend.
In order to make this study (and conclusion) suitable for publication, the author still needs to demonstrate that water use plays only a minor or even negligible role in the water balance. As the author claims that there is not water consumption data available, the only possibility I see (and already recommended last time) is to compute the overall lake water balance (which is not done in spite of claiming so at the end of the manuscript) and demonstrate that it can be closed without the need of additional fluxes like water consumption. And even then, the conclusion is only valid for the lake itself because the decrease in river discharge could still be connected to upstream water consumption. I’ve seen from the replies that the author feels the compilation of all relevant water flow would be impossible, which seems to be strange as precipitation and lateral inflow are already shown in the manuscript and evaporation could be estimated using e.g. a reference evaporation computation or reanalysis data. Sure, it would only be a coarse estimate, but better than none.
Going to the details, there is still need to improve the manuscript:
Removing “Global” from the title did not change its message. Just because an increase in evaporation might be responsible for the desiccation of a given salt lake, it does not rule out another dominating processes for another salt lake. As is don’t agree with the authors view for this lake anyway, it is even harder for me to see how this would be a showcase for salt lake desiccation in general.
the term “extremely changing point” seems to simply refer to an extreme value and there is no need to define a special name for it. The whole extreme value analysis seems to be nothing else than identifying extreme values using a 2 standard deviation criteria. This method is not novel at all. How can the author claim novelty while at the same time citing two papers (among many) which already identify extremes using 2 sigma?
no, in a normal distribution the z-score is not limited to +- 3. These values refer to the 0.1% and 99.9% percentiles in the distribution and (depending on the length of the time series) there can be values below and above this points. Anyway, why not use percentiles instead of standard deviation?
I appreciate that the author made an effort to add an additional time series (river discharge) to the analysis, but in summary I don’t find my concerns adequately addressed. With the provided data and analysis I don’t see any hard evidence in this paper to support the authors conclusion (let alone to extent it to other salt lakes) as I don’t agree with the extreme frequency versus trend argument of the authors for the already explained reasons.
Comments for author File: 
Comments.pdf
Author Response
Response:
It is weird that this reviewer overlooked all his/her previous comments and my answers but directly listed other questions. Unfortunately, comments input this time by this reviewer were directly answered and rejected this reviewer’s questions (see blow please).Comments and Suggestions for Authors
Thanks to the author for providing replies to my comments and for modifying the manuscript. Still, I don’t find the manuscript convincing. Sure, finding more positive extremes for the recent time period compared to an earlier one in an overall (slightly) decreasing time series is a bit contra-intuitive and reasons for this behaviour should be discussed. However, the argument which the author uses to disprove water use as a main contributor for this long-term trend, namely that there are more positive extremes in the later time periods than in earlier, would be the same for climate. The author assumes (without presenting any data) that water use and consumption increase in time (which may very well be the case) and concludes that this conflicts with the increased frequency of positive extremes. (1) But how is this different from climate effects? The author shows the increase in temperature, concludes from this that evaporation would be increased as well (which again I would agree to), but apparently sees no conflict with the recent positive lake level extremes in this case? Why is that? Please also note that water use is not necessarily constant just because population density and industry did not change. Water is consumed if needed and typically in wet years also e.g. agricultural water use is lower.
Responses:
(1) The reviewer self has answered his/her above questions/comments asked by his/herself, and the reviewer also summarized his answers in this reviewer’s last two sentences above “Please also note that water use is not necessarily constant just because population density and industry did not change. Water is consumed if needed and typically in wet years also e.g. agricultural water use is lower”. It means that the reviewer agreed that climate and weather (e.g., wet years, which directly are driven and caused by climate and weather) matter and playing dominant roles, and human water use is secondary, which is/coincides my conclusion. Therefore, there is NOT any needs to address this reviewer’s above questions/comments.
This reviewer: Please carefully read this communication and then interpret it. I do NOT use “constant”, “no change”, or other similar terms. I did not state that “water use is constant just because population density and industry did not change”. Please stop imposing your belief onto others and Misleading the author and the editors.
Furthermore, (1) the shown time period may play a role in this case. Looking at figure 2 (https://www.nature.com/articles/ngeo3052/figures/2) in the related article by Wurtsbaugh et al, there are 25 years of lake level variation data available prior to the period shown by the author. Looking at the USGS website, there are even 50 more years of data which are neglected in this analysis. Why are they not included? As the earlier years are much wetter, the number of positive extremes are reduced (see attached figure). (2) Defining extremes with percentiles (5 and 95) instead of standard deviation (which are much more robust against outliers) there are more negative extremes as would be expected from a time series with a decreasing trend. Subtracting the trend, the positive extremes are concentrated again in the 1980’s and can probably be related to (3)an exceptional wet period (as the extremes also show up in precipitation and river flow) which recharge the lake in spite of climate change and/or human impacts. Still, this does not prove anything in terms of which effect is responsible for the trend.
Response:
(1) I started from about 1900, because the only continuous climate/weather data (precipitation, temperature, and snow fall) are available from about 1900. I do not think it’s needed or good to do data manipulations/predictions for any missing observations in water levels, climate, and weather data.
(2) This ‘percentiles’ (5 and 95) again. This reviewer thinks percentile is more robust than standard deviation. However, this is her/his opinion, and Why is 5% or 95%? Why not 3% or 97%, or why not 10 and 90 as used/referenced in short communication or others?
Unfortunately, this opinion is too subjective, although I added 5 and 95 percentiles in the Table 1 and a paragraph on page 4 to show the differences, which showed 5 and 95 percentiles do not work.
Standard deviation as the references in climate already showed its power that had been addressed in the first revision and the two added references. It is not necessary to repeat it again.
Again, you dived into your “wet period” fallacy. But, in fact, as you showed/commented here again, climate and weather matter first, and it is climate/weather that makes the so-called ‘wet’. It is totally misleading to subtract any “trend”, because the trend explains the story.
(3) This reviewer pointed out a “wet period” here again, which is certainly and directly caused by the climate/weather. Otherwise, how could it be named a wet period? Therefore, again, climate/weather matter first, and human water use is secondary. Again, this reviewer’s comment already directly answered his/her questions raised early.
It is weird why this reviewer cannot understand this simple/apparent point? May this reviewer just want to mislead the author or the editors?
In order to make this study (and conclusion) suitable for publication, the author still needs to demonstrate that water use plays only a minor or even negligible role in the water balance. As the author claims that there is not water consumption data available, the only possibility I see (and already recommended last time) is to compute the overall lake water balance (which is not done in spite of claiming so at the end of the manuscript) and demonstrate that it can be closed without the need of additional fluxes like water consumption. And even then, the conclusion is only valid for the lake itself because the decrease in river discharge could still be connected to upstream water consumption. I’ve seen from the replies that the author feels the compilation of all relevant water flow would be impossible, which seems to be strange as precipitation and lateral inflow are already shown in the manuscript and evaporation could be estimated using e.g. a reference evaporation computation or reanalysis data. Sure, it would only be a coarse estimate, but better than none.
Going to the details, there is still need to improve the manuscript:
Removing “Global” from the title did not change its message. Just because an increase in evaporation might be responsible for the desiccation of a given salt lake, it does not rule out another dominating processes for another salt lake. As is don’t agree with the authors view for this lake anyway, it is even harder for me to see how this would be a showcase for salt lake desiccation in general.
Response:
Please read my revision carefully. The “Global” had been deleted last time in the 1st revision. Again, Do Not Misled the author and editors.
Do Not Impose your own thought onto others again and again: “Just because an increase in evaporation might be responsible for the desiccation of a given salt lake“. It is your (this reviewer’s) opinion, Please read this short communication and then comment. Again, climate and weather (e.g., precipitation and local warming) matter first.
the term “extremely changing point” seems to simply refer to an extreme value and there is no need to define a special name for it. The whole extreme value analysis seems to be nothing else than identifying extreme values using a 2 standard deviation criteria. This method is not novel at all. How can the author claim novelty while at the same time citing two papers (among many) which already identify extremes using 2 sigma?
Response:
Do Not repeat same comment again and repeat percentile.
First, this reviewer commented to address “extremely changing point” in his/her first review, but in the 2nd review this reviewer totally changed his/her opinion. How weird is this type of reviewing!
Second, the reviewer did not read the revised Method section and did not know any about the two references, but directly made a too subjective statement about “extremely changing point”.
Yes, two references are cited, but these two references emphasized the importance of both mean and standard deviation in Temperature measurement and its distribution. However, the usage of them in this communication is totally different and is for identifying extremely changing observations. Thus, the reviewer’s above statement cannot stand.
Can this reviewer tell what is novel? Can this reviewer provide me a research using the method I designed in this short communication for climate-hydrological analysis? If not, what is the reason for this reviewer to stop others trying different research approaches just because this reviewer know the definition of mean and SD or regression?
At last, this reviewer needs to understand that we all know and even our college students know well about descriptive statistics (mean, sd, percentile, quantile, and so on) and regression models, but how to apply them does make different significance in science and in different applications. I understand what you want to say, but as a reviewer please first be Reasonable and Responsible.
no, in a normal distribution the z-score is not limited to +- 3. These values refer to the 0.1% and 99.9% percentiles in the distribution and (depending on the length of the time series) there can be values below and above this points. Anyway, why not use percentiles instead of standard deviation?
Response:
The short sentence related to z score is modified: When Z score is related to a normal distribution [8, 9], the Z score ranges from -3 to +3 covering almost the whole distribution approximately 99.7%.
Again, my early revision and response to your 1st review have answered the reason of using standard deviation as for your above comment too.
Ok, I still would like to answer again your repeated question “why not use percentiles”, I added 5 and 95 percentiles as you commented above/early, and please read Table 1 and the related paragraph added on Page 4. That shows that 5 and 95 percentiles do not work for identifying extreme observations.
When you review a revision, please read your comments and the author’s response first, and then move onto you “new” comments if needed. Also, please do not repeat the same simple question again and again.
I appreciate that the author made an effort to add an additional time series (river discharge) to the analysis, but in summary I don’t find my concerns adequately addressed. With the provided data and analysis I don’t see any hard evidence in this paper to support the authors conclusion (let alone to extent it to other salt lakes) as I don’t agree with the extreme frequency versus trend argument of the authors for the already explained reasons.
Overall Response:
This reviewer overlooked his/her (this reviewer’s) first-time reviewing comments and author’s responses, overlooked many facts (i.e., Landsat images and related analysis) and the added data/facts and related analyses (i.e., radiation and wind impacts, current years’ GSL water level monthly change analysis, and decadal river discharge analysis, and the water balance based equation analysis) that the author had carefully provided in the 1st revision, and did not have any answers to the author’s response and revisions. However, this reviewer directly and repeatedly imposed his/her ‘new’ opinions/questions onto the author and the editors, while this reviewer already answered and rejected his/her questions with his/her own comments.
This type of review is not reasonable, and it is nonresponsible either.
Reviewer 2 Report
I found that Dr. Meng’s rebuttals in the response letter are overall reasonable, and some editorial improvements have also been made in this revised manuscript.
The decline of the Great Salt Lake is a complicated and controversial topic. It is one example illustrating the complexity of human-nature interactions and how they relate to our water managements. So extra cautions are always good, to avoid misleading conclusions to the scientific community.
Here, Dr. Meng opted to investigate the causes using a simple bottom-up approach (by comparing basic climate forcing, mostly in a qualitative or statistical way). If I were doing the same thing, I may still like to try a standard budgeting method that synthesizes various variables for a more quantitative attribution. I admit that no model is perfect. But the point of modeling is to take into account the physical processes where climate/hydrological variables are not independent.
I understand this is a short communication, with a purpose of re-drawing our attention to the decline of some saline lakes worldwide. So upon the acceptance of the other reviewer(s), I am generally okay for this manuscript. However, I kindly request that Dr. Meng acknowledges his method limitations, and points out that conclusions should not be completely ‘closed’. There is room for continued improvements.
Meanwhile, I am recommending a new paper that may be helpful for some of Dr. Meng’s discussions. The authors of this paper discussed the recent water loss in the Great Salt Lake (as well as some other saline lakes), and how it may relate to recent climate conditions (e.g., precipitation deficit and warming) and human water managements. Please see: https://www.nature.com/articles/s41561-018-0265-7.
Line 33: “Although the current consumptive water uses in agriculture, salt pond mineral production, and industry are much larger than that in THE 1960s”. Please also provide references to support this statement.
Line 42: “Human water USE”, not “uses”.
Line 47: grammatical error. “was added” not “were added”.
Line 48: grammatical error again. “are little”, not “is little”.
Line 50: “in A short term”.
Line 76: Typo: “using the equation” not “quation”.
I won’t provide editorial corrections after this. But I encourage a thorough proof-reading from the author’s side.
Line 264. I am not sure about this statement.
Author Response
I appreciate the reviewing comments, which I think are useful not only for this short communication but also indicates this reviewer’s reviewing merit/attitude and contribution to the journal Climate.
Accordingly, adding the new reference provided by the reviewer below (i.e., (2)), I have revised some words in the manuscript and more added discussions/limitations in the last paragraph (page 10) to address the comment (i.e., (1) below).
Comments and Suggestions for Authors
I found that Dr. Meng’s rebuttals in the response letter are overall reasonable, and some editorial improvements have also been made in this revised manuscript.
The decline of the Great Salt Lake is a complicated and controversial topic. It is one example illustrating the complexity of human-nature interactions and how they relate to our water managements. So extra cautions are always good, to avoid misleading conclusions to the scientific community.
Here, Dr. Meng opted to investigate the causes using a simple bottom-up approach (by comparing basic climate forcing, mostly in a qualitative or statistical way). If I were doing the same thing, I may still like to try a standard budgeting method that synthesizes various variables for a more quantitative attribution. I admit that no model is perfect. But the point of modeling is to take into account the physical processes where climate/hydrological variables are not independent.
I understand this is a short communication, with a purpose of re-drawing our attention to the decline of some saline lakes worldwide. So upon the acceptance of the other reviewer(s), I am generally okay for this manuscript. However, (1) I kindly request that Dr. Meng acknowledges his method limitations, and points out that conclusions should not be completely ‘closed’. There is room for continued improvements.
A paragraph discussing the limitations of this short communication is added on page 10.
(2) Meanwhile, I am recommending a new paper that may be helpful for some of Dr. Meng’s discussions. The authors of this paper discussed the recent water loss in the Great Salt Lake (as well as some other saline lakes), and how it may relate to recent climate conditions (e.g., precipitation deficit and warming) and human water managements. Please see: https://www.nature.com/articles/s41561-018-0265-7.
Thanks. This new reference is added.
Line 33: “Although the current consumptive water uses in agriculture, salt pond mineral production, and industry are much larger than that in THE 1960s”. Please also provide references to support this statement.
The reference is added, i.e., Trends in estimated water use in the United States, 1950 – 2015, https://water.usgs.gov/watuse/wutrends.html.
Line 42: “Human water USE”, not “uses”.
It changed accordingly.
Line 47: grammatical error. “was added” not “were added”.
Edited accordingly.
Line 48: grammatical error again. “are little”, not “is little”.
“is” is changed into “are”.
Line 50: “in A short term”.
“a’ is inserted.
Line 76: Typo: “using the equation” not “quation”.
It was “using the equation”, may be caused in displaying.
I won’t provide editorial corrections after this. But I encourage a thorough proof-reading from the author’s side.
The proof-reading has been done carefully, and changes are marked also in this revision.
Line 264. I am not sure about this statement.
This sentence is deleted in this revision.
Reviewer 3 Report
The author investigates trends in Great Salk Lake (GSL)’s water level between 1904 and 2016 using simple statistics. The author claimed that the changing GSL water level is associated with climate variability and change. While this is an interesting study, the research design and presentation of the materials should be improved to meet the publication standard published in journal Climate.
Major comments
1. The reviewed literature is not up to date. Considering recent advances in the literature of hydroclimatology and climate impact studies, I suggest that the author review the following reports and papers and consider adding them in relevant sections.
USGCRP, 2017: Climate Science Special Report: Fourth National Climate Assessment, Volume I[Wuebbles, D.J., D.W. Fahey, K.A. Hibbard, D.J. Dokken, B.C. Stewart, and T.K. Maycock (eds.)]. U.S. Global Change Research Program, Washington, DC, USA, 470 pp, doi: 10.7930/J0J964J6.
McPhillips, L.E., Chang, H, Chester, M.V., Depietri, Y., Friedman, E., Grimm, N.B., Kominoski, J.S., McPhearson, T., Méndez-Lázaro, P., Rosi, E. J., Shiva, J.S. (2018) Defining extreme events: a cross-disciplinary review. Earth’s Future. https://doi.org/10.1002/2017EF000686
2. The author needs to state the source and quality of the data in more detail. It is unclear how many weather stations were used for the analysis. While the author mentioned that the precipitation, temperature, and snowfall data were obtained from NCDC, the author needs to be explicit about the location and quality of the weather data. If there were any missing data, how were they filled in? If the weather stations got moved around over time, that needs to be reported as well. Alternatively, since there were gridded climate data available, the author could potentially make use of them. In Table 1, why was the Kelvin scale used for temperature reading?
3. The manuscript needs a tight organization. The author describes data and methods in the results section. Additionally, there are a few repetitive sentences, which should be avoided for the brevity of the presentation. I marked such areas below in detailed comments.
4. While the trends of precipitation, temperature, snowfall, discharge, and GSL water level were well-presented, the relation between the water level and climatological factors were less clear. Why not showing simple correlation coefficients between individual climatological factors and water level? The authors might find some of the figures in the following paper for presenting such information.
Chang, H., Bonnette, M.R. (2016) Climate change and water-related ecosystem services: Impacts of drought in California, USA, Ecosystem Health and Sustainability 2: e01254
5. The author can the context of the research better. While the author discussed other studies from other climates, it would be much better if the author used more regional or local studies that drain into the Great Salt Lake. I was surprised to see no previous studies were cited in the paper. See the following references.
Gillies, R.R., Chung, O.Y., Wang, S.Y.S., DeRose, R.J. and Sun, Y., 2015. Added value from 576 years of tree-ring records in the prediction of the Great Salt Lake level. Journal of Hydrology, 529, 962-968.
Mohammed, I.N. and Tarboton, D.G., 2012. An examination of the sensitivity of the Great Salt Lake to changes in inputs. Water Resources Research, 48(11).
Wang, S. Y., Gillies, R. R., Jin, J., & Hipps, L. E. (2010). Coherence between the Great Salt Lake level and the Pacific quasi-decadal oscillation. Journal of Climate, 23(8), 2161-2177.
6. The authors claimed that human consumptive water use is not a significant factor affecting GSL water level. Given that USGS has such data, why not using water use data to back up their claim?
7. The conclusion section does not convey the most interesting information in a persuasive way. It could be shorter. The last two paragraphs may be condensed into one paragraph without losing too much information.
Other minor comments
1. Page 1, Line 11. “if the regression is statistically significant”. You mean “slope of the regression coefficients are statistically significantly different?
2. Page 1, line 14. Insert a comma before “respectively”
3. Page 1, line 17. Insert a comma before “including”
4. Page 1, line 29. Remove “in order”
5. Page 1, lines 33-34. This statement needs correction as US freshwater withdrawals have declined since 1980, which is clearly shown in the URL the author cited.
6. Page 2, Fig. 1 Can the author add a scale bar and a northern arrow in the map?
7. Page 3, line 97. IPCC 2001. I suggest that the author replace it with IPCC 2014.
8. Page 3, line 103. Consider adding a new reference such as McPhillips et al. (2018).
9. Page 5, Figure 2. Remove redundant decimal points.
10. Page 5, lines 205-207. This sentence belongs to data and methods.
11. Page 5, lines 213-215. Remove this sentence as the same information was already presented in methods.
12. Page 6, line 223. Report unit for temperature change.
13. Page 6, lines 225-226. How did the author estimate the average annual evaporation in two different periods? Are they based on linear interpolation from the regression equation?
14. Page 6, Figure 3. Label B. I presume it is “Precipitation”. Remove redundant decimal points in Y axes.
15. Page 7. Section 3.2 I do not think this section add too many new insights. Could be omitted.
16. Page 8. Line 336. Please check the cited reference [13] as I think it does not mention water use in the Great Salt Lake area.
17. Page 8. Lines 340-350. This section could be shortened.
18. Page 9. Lines 377-378. This is an important point to be stressed, which implies that warming could potentially be a more driving factor than precipitation in the future?
19. Page 10. Line 388. “significant”? could be “substantial”?
Author Response
See the attached file Response_3rdRview.docx
Author Response File: Author Response.pdf
Round 2
Reviewer 3 Report
Thank you for having me the opportunity to review the revised manuscript. The authors adequately addressed the reviewer's concern.
