Linking Glacier Retreat with Climate Change on the Tibetan Plateau through Satellite Remote Sensing

Zhao, Fumeng; Gong, Wenping; Bianchini, Silvia; Yang, Zhongkang

doi:https://doi.org/10.5194/egusphere-2024-1083

Preprints

https://doi.org/10.5194/egusphere-2024-1083

Preprints

04 Jun 2024

| 04 Jun 2024

Linking Glacier Retreat with Climate Change on the Tibetan Plateau through Satellite Remote Sensing

Fumeng Zhao, Wenping Gong, Silvia Bianchini, and Zhongkang Yang

Abstract. Under global climate change, glaciers on the Tibetan Plateau are experiencing severe retreat, which significantly impacts the regional water cycle and the occurrence of natural hazards. However, detailed insights into the spatial-temporal patterns of this retreat and its climatic drivers remain insufficiently explored. In this study, an Adaptive Glacier Extraction Index (AGEI) is proposed based on analysis of multispectral Landsat images integrated with the Google Earth Engine, and comprehensive and high-resolution mapping of glaciers on the Tibetan Plateau is realized at five-year intervals from 1988 to 2022; subsequently, the ERA5-Land air temperature and precipitation data are downscaled to a finer 1-km resolution; finally, impacts of the annual and seasonal change of downscaled meteorological factors on the glacier retreat are quantified. Results demonstrated a rapid yet heterogeneous pattern of glacier retreat across the Tibetan Plateau between 1988 and 2022, with retreat rates ranging from 0.14 ± 0.07 % to 0.51 ± 0.09 % annually. A notable trend was observed where most glacier areas experienced a decrease from 1990 to 2000, followed by a slight increase. From 2010, a majority of the glaciers exhibited either a static state or minimal retreat. The most pronounced impact of annual temperature on glacier retreat is observed in Zone VIII, with a value of -9.34 × 10³ km²/°C, and the most restraining impact of precipitation on glacier retreat reaches 261 km²/m, which is observed in Zone VI for the spring season. Further, it identifies the influence of debris thickness on the glacier retreat rate. These insights are pivotal in comprehending the temporal and spatial heterogeneity of glacier retreats, and in understanding the effects of climatic variations on the state of glaciers on the Tibetan Plateau.

Received: 10 Apr 2024 – Discussion started: 04 Jun 2024

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.

Download & links

Preprint (PDF, 4535 KB)

Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
Preprint (4535 KB)

Download & links

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Journal article(s) based on this preprint

03 Dec 2024

Linking glacier retreat with climate change on the Tibetan Plateau through satellite remote sensing

Fumeng Zhao, Wenping Gong, Silvia Bianchini, and Zhongkang Yang

The Cryosphere, 18, 5595–5612, https://doi.org/10.5194/tc-18-5595-2024,https://doi.org/10.5194/tc-18-5595-2024, 2024

Short summary

Fumeng Zhao, Wenping Gong, Silvia Bianchini, and Zhongkang Yang

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2024-1083', Connor Shiggins, 02 Jul 2024
The manuscript presented by Zhao and co-authors put forward a glacier mapping tool using Landsat collections and reanalysis data over the Tibetan Plateau (TP).
The manuscript is generally well written and mostly well structured, however I have made some suggestions to try and improve the flow of the manuscript.
My main and minor comments can be found below:
Main:
Why is RGI 6.0 being used instead of RGI 7.0 which was published in September 2023? I appreciate if analysis had been carried out before the most recent release, however I think it is important the most up to date data products are used in current day studies as I know there can be large differences in glacier extents between version 6.0 and 7.0 and therefore may impact your results significantly. I will leave this to the decision of the handling editor, but I would suggest using RGI 7.0 to ensure at the time of publication, the results reflect the most current version.

There is no consideration of the poor ‘completeness’ value in Table 2, particularly for 2010. The F-1 scores for glacier mapping are moderate, I would not consider them strong metrics in support of the method. It is of course fine to have these scores, however, there needs to be consideration of why the scores are this low and how this may impact the results in the discussion.

Each section of the paper (i.e. results, discussion) does not need to have an introductory section of how it is structured and I would suggest deleting them and going straight into the text of the section.

Minor:
L17: Would just note what time period this 'slight increase' was

L19: Would consider stating geographically where these 'zones' are as out of context as they don't mean very much as no lead in from the abstract (i.e. NE of TP)

L35: I get what you mean, but I would rephrase the opening on this paragraph saying 'Glacier change can be measured using variations in...' to be clearer

L42: Would put the reference for each method after it was referred to i.e., spectral analysis (ref), object-based (ref), etc.

L46: needs a reference for debris-free glaciers before full stop

L51: are 'quality' and 'resolution' not the same thing? Also, what type of resolution, spatial or temporal? If both I would suggest stating the limitation of spatio-temporal resolution over the TP.

L58: Climate change seems a bit broad here, is it the increasing air temperature? Would clarify

L77: If you state 'numerous studies' I would suggest citing a handful of them as examples

L82-85: not convinced this is required for such a specific structure, think you're fine with just the aims of the paper being highlighted

L86: Just call it Study Area

L117: Would call it data and methods instead of materials

L118: This text is not needed - would go straight to 3.1

L124: I would just be careful saying images via GEE catalog are 'open-access' - while they are for individuals, there is commercial cost to access the platform

L124: Would merge these sentences, suggest at start of sentence saying Landsat data is used via GEE

L158: Figure 2 caption - Is this the total number of Landsat images or the total number used in the study with less than 60% cloud as defined by your study? Would clarify

L162-164: basically, the same as previous, would suggest merging and stating previous studies chose 0.4 as a threshold and therefore it was chosen here

L167-169: same applies about thresholding values NDWI, just merge them and say you chose 0.4

L171: 'Therefore' does not fit here as does not follow the previous sentence

L173-174: the final paragraph is just repetition of values you've defined - suggest deleting

L183: What holes? What do you mean 'filled'? Interpolated? Would clarify

L185: sentence does not make sense - what do you mean 'validated using reference debris-free glaciers'?

L253: Can Fig 3 be made bigger?

L255-258: Delete introductory

L275: Figure 4. Would suggest having the map as a) at the top of the figure then having panels b-i stacked. The lettering order seems a little confusing

L287: What do you mean the glaciers mapped between 2000 and 2005 'exhibit greater consistency'?

L300: Table 2 metric scores – see main comment 2

L400: I think Fig.10 and above text is more results than discussion. Would suggest moving to results. Do you have any numbers for the total area difference between the two methods? Would add to the argument of the AGEI method

L405: Would refrain from starting discussion sentences 'Fig X', if wanting to directly refer to the figure, place in brackets at the end of the sentence
Citation: https://doi.org/10.5194/egusphere-2024-1083-RC1
- AC1: 'Reply on RC1', Fumeng Zhao, 28 Jul 2024
  
  We thank the reviewer for the careful review and valuable suggestions. Hereby we give a point-by-point reply to address the comments in the attached pdf.
  
  Citation: https://doi.org/10.5194/egusphere-2024-1083-AC1
RC2:
'Comment on egusphere-2024-1083', Joseph Mallalieu, 17 Aug 2024
GENERAL COMMENTS
This manuscript investigates how changes in debris-free glacier extent in the Tibetan Plateau between ~1990-2020 are related to changes in climate (temperature, precipitation) over the same period. The authors: i) outline a novel technique (the AGEI) for mapping glacier extent, and use it to quantify changes in glacier area at 5-year intervals for 8 subzones of the Tibetan Plateau; ii) downscale ERA5-Land reanalysis data (specifically 2 m temperature and total precipitation) from 9 km2 to 1 km2 using ancillary variables from MODIS and the SRTM DEM) in a random forest regression, and; iii) use linear regressions to investigate the relationships between changes in glacier extent, temperature and precipitation at annual and seasonal scales. The authors conclude that total (debris-free) glacier area in the region has reduced over the survey duration, that temperatures have typically risen, that precipitation has increased in some subzones and reduced in others, and that in some subzones (particularly those in the Karakorum) reductions in glacier area correlated strongly with increases in temperature, but that the severity of glacier area loss may have been mitigated by corresponding increases in precipitation.
The manuscript presents a significant body of work, including novel mapping methodologies, extensive glacier and climate datasets, and insights into regional glacier and climate change in the Tibetan Plateau, and should therefore be of interest to others mapping glacier change generally, and particularly researchers investigating glacier change in central Asia. The addition of a paragraph discussing the implications of the glacier - climate relationships identified in this work with regard to existing climate projections for the region could help add impact. The manuscript is largely well-written, structured and presented, and could be further improved with some edits to make it more succinct (see comments for specific guidance).
My main comments relate to the methods adopted, several of which require additional clarification and/or detail to by fully comprehensible and transparent. I think this is of particular importance given the novelty of the mapping method used and the borderline nature of some of the glacier - climate relationships presented (e.g. see confidence intervals in Fig. 9). A caveat to my comments is that I am not an expert in Himalayan/Tibetan glaciation, or well-versed in random forest or time-series analyses, and have thus taken the methods adopted in section 3.2 (specifically lines 211-250) in good faith.

SPECIFIC COMMENTS
The accuracy and consistency of the novel AGEI mapping technique is integral to the results of this analysis. Consequently, a more detailed evaluation of the technique would be welcome, both to aid replication and assess its effectiveness. In particular, Fig. 10 does not have sufficient resolution to assess the accuracy of the AGEI mapping. Instead, a figure with multiple high-resolution panels (including some features that present challenges to accurate glacier mapping, such as lakes, shadow, cloud etc.), would be useful to illustrate the accuracy of the AGEI glacier outlines, particularly if the reference RGI 6.0 glacier outlines can be overlaid on them. Given its importance to the results, I think such a figure would be better situated and discussed in Section 3.1 or Supplementary Information, rather than waiting until the discussion. In addition, I struggled to fully understand the mapping evaluation indices presented on lines 189-198, particularly the Correctness (differentiating between A_cg and A_tg) and F1 scores (e.g. what does a high F1 score indicate?). If possible, illustrations of high and low scoring glaciers/regions/test sites (e.g. in Supplementary Information) could aid interpretation of the indices and the demonstrate the accuracy of the AGEI technique.

An additional section should be added to the end of the methodology to describe the linear regressions used to investigate the relationships between temperature/precipitation and glacier change (the linear model is mentioned at the beginning of Section 4.3, but not described fully). For transparency it would be useful to write out the model equation(s) used in the methods. It is unclear whether numerous independent models were generated (e.g. one for each for each climate variable, zone and season, totalling ~80 independent regressions), or if these were all included in one multiple regression model (and, if so, how multicollinearity between the seasonal climate was addressed). The results of the regressions should also be presented in a table that includes the coefficients, confidence intervals, standard error, t-values and p-values, and a summary of model diagnostics (R²/Adjusted R², RSE etc.). Although the coefficients and confidence intervals are presented in Fig. 9 their numeric values are difficult to determine, hence a table of the regression outputs would be useful to support this figure, aid interpretation of the data, and provide more robust conclusions.

A major caveat of this work is that the analysis presented only considers debris-free glaciers. Greater context regarding the prevalence and distribution of debris-covered glaciers in the region would therefore be useful (in the introduction, or sections 3.1 or 5.3) to understand the potential implications of this methodological choice on the results and their interpretation (e.g. are debris-free glaciers typically representative of glaciers across the region; could there be some sub-zones where the exclusion of debris-covered glaciers from the analysis significantly reduces the sample size and therefore the representativeness of the results?).

Given the relationships identified between changes in glacier area, temperature and precipitation from ~1990-2020, I think a paragraph could be added to the discussion to explore the implications of these results in respect of existing climate projections for the region. For example, how is regional climate likely to change by the end of the century and what could be the implications for the glaciers in the different sub-zones given the trends identified between 1990-2020? A short paragraph addressing this could help to add impact to the manuscript, and potentially identify future research priorities.

Throughout the manuscript the authors refer to ‘glacier retreat’ which (to me at least) implies a measure of distance retreated by a glacier terminus, particularly the terminus of a valley or outlet glacier. Given that the work presented is concerned with changes in glacier area (not just recession at termini) ‘glacier extent’ may be a more intuitive term to use.

The manuscript could be made more succinct by removing several sections of text outlining the structure (which is self-evident), particularly lines 82-85, 118-119, 255-257, and 380-381.

Figures 4 & 5: Both figures show an increase in glacier area in zones II, V and VII between 2000 and 2005. This seems unusual given the overall regional trend, so I’m interested to know whether this is an artefact of the mapping methodology, or whether this brief increase in glacier extent in these regions has been documented in any other studies/literature? The same question also applies to zones VI and VIII, which appear to show notable increases in glacier area between 2005-2010.

TECHNICAL COMMENTS
Abstract: Here, references to the zones used in the study would be better substituted with the names of the corresponding regions (e.g. Karakorum, southern Himalaya etc.) because the zones are only intelligible to those who have already read Section 2.

Lines 209-210: Is there a precedent or justification for the use of the 3 ancillary MODIS and SRTM DEM variables to downscale the ERA5-Land data?

Line 247: A significance level of 0.1 seems unusually low, and unnecessary given that the 0.05 significance level is also employed. Is there a good rationale for the inclusion of the 0.1 significance level?

Lines 262-263: This description (‘most pronounced retreat’) does not appear to be entirely consistent with the panels in Fig. 4 which show rates of area reduction being greatest in zones I, III and IV, with more modest losses in zones II and VIII.

Lines 263-264: The term ‘glacier advance’ is a little misleading here, given that regions VI and VIII (encompassing the Karakoram) show overall trends of area loss in Fig. 4. Presumably, ‘advance’ here refers to some individual glaciers that are acting counter to the overall trend in these subzones. If so, they are not easily visible in the main map in Fig. 4. Consequently, an enlarged inset of these subzones (or a separate figure) may be required to better illustrate their presence.

Lines 268-269: States that glacier advances are seen from 2000-2005 ‘in all zones except for VI and VIII’. However, zones III and IV also do not show evidence of glacier advance over this period (as illustrated in both Figs. 4 & 5).

Figure 4: To aid interpretation of the figure it would be worth noting in the figure caption that the values presented below each zone number represent the annual change in glacier area (and not the total change over the duration of the study).

Lines 284-286: What might explain the much higher percentage differences between the AGEI and RGI outlines in the other zones? It is not good practice to only report the lowest differences and use them as validation of the technique.

Lines 290-291 & Table 2: More detail regarding these indices would be useful. My (possibly incorrect) interpretation of this table is that only around 60-70% of the RGI 6.0 glacier areas have been mapped by the AGEI technique, which seems rather low (also see Specific Comment 1 above).

Lines 305-306 & Figure 6 caption: more information regarding the training and validation samples would be useful. These do not appear to be mentioned in the methodology.

Lines 327-329 & 337-338: It is unclear what ‘glacier advance’ is referring to specifically in these instances (e.g. individual glaciers in the subzones, or overall advances in some zones in particular time periods such as 2005-2010?). Also see comment regarding Lines 263-264.

Lines 334-335: Avoid the use of ‘respectively’ in this instance (the list is too long), and just include the name of each period (annual, fall etc.) next to each numeric value.

Figure 8: It was not immediately apparent to me that the colour key underneath the figure was universal to both sides of the figure (a & b). A note in the figure caption could help to clarify this.

Lines 354-355: Presumably ‘variations’ here should be ‘increases’ to specify the direction of the relationship between glacier area and temperature? In addition, it states that ‘all the subzones’ have a negative correlation between glacier area and temperature change, but Figure 9a (annual) only shows zones II, III, VI and VIII to be clearly below 0 on the y-axis (thus presumably indicating glacier area loss). The coefficients and confidence intervals of the remining zones is difficult to discern. A table containing these values could be useful here (see Specific Comment 2).

Lines 361-362: See previous comment. A visual interpretation of Figure 9b (annual) suggests that only zones III, V, VI and VIII have confidence intervals in excess of 0 on the y-axis, and that therefore only these 4 zones can be confidently interpreted as exhibiting a positive relationship between glacier area and total precipitation at an annual scale. Equally, if the confidence intervals for zone II also cross the y-axis and it cannot be said with confidence that the relationship is negative. I think it would be beneficial for the authors to re-write this entire paragraph with greater consideration given to their confidence intervals, particularly where they cross 0 on the y-axis.

Line 363: Should ‘zones III, V, and VIII’ read ‘zones III, VI and VIII’? It also looks like zone V could be added to this list.

Lines 363-365 and Figure 9b: Units are inconsistent here, the text states km²/mm, but the y-axis on Figure 9b states km²/m. The abstract also states km²/mm.

Line 383: Which earlier study; this study or Huang et al. 2021?

Lines 405-406: For consistency, it would be useful to also include zone numbers when referring to locations in this sentence.

Lines 409-411: See previous comment.

Lines 432-436: Given the scarcity of data points I think the authors should be cautious of giving too much weight to the data presented in Figure 11. For example, although zones I and III have lower median debris thickness and relatively high rates of glacier area loss, zone 8 has a similar level of debris thickness, but much lower rates of glacier area loss, and zone 4 has the highest median debris thickness, but still has the 3rd highest rate of glacier area loss. This paragraph could benefit from a discussion of the differences between the regional glacier area losses presented here (for debris-free ice) and any studies of regional glacier area losses for debris-covered glaciers (should the data be available).

Line 451: Clarify that the ‘total glacier area’ is the debris-free total glacier area.

Line 457-458: I think this sentence should be rephrased because the total glacier area in these regions has typically reduced over the duration of the study (see trends on Figure 4), rather than advanced. Given the overall reduction in glacier area and the correlations presented in Fig. 9b it may be more appropriate to state that annual and summer precipitation may have helped to mitigate reductions in glacier area in the Karakorum.

Lines 461-464: I think this conclusion could also be revised with regard to the earlier comment about confidence intervals (see comment for Lines 361-362).
Citation: https://doi.org/10.5194/egusphere-2024-1083-RC2
- AC2: 'Reply on RC2', Fumeng Zhao, 07 Sep 2024
  
  We thank the reviewer for the careful review and valuable suggestions. Hereby we give a point-by-point reply to address the comments in the attached pdf.
  
  Citation: https://doi.org/10.5194/egusphere-2024-1083-AC2

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2024-1083', Connor Shiggins, 02 Jul 2024
The manuscript presented by Zhao and co-authors put forward a glacier mapping tool using Landsat collections and reanalysis data over the Tibetan Plateau (TP).
The manuscript is generally well written and mostly well structured, however I have made some suggestions to try and improve the flow of the manuscript.
My main and minor comments can be found below:
Main:
Why is RGI 6.0 being used instead of RGI 7.0 which was published in September 2023? I appreciate if analysis had been carried out before the most recent release, however I think it is important the most up to date data products are used in current day studies as I know there can be large differences in glacier extents between version 6.0 and 7.0 and therefore may impact your results significantly. I will leave this to the decision of the handling editor, but I would suggest using RGI 7.0 to ensure at the time of publication, the results reflect the most current version.

There is no consideration of the poor ‘completeness’ value in Table 2, particularly for 2010. The F-1 scores for glacier mapping are moderate, I would not consider them strong metrics in support of the method. It is of course fine to have these scores, however, there needs to be consideration of why the scores are this low and how this may impact the results in the discussion.

Each section of the paper (i.e. results, discussion) does not need to have an introductory section of how it is structured and I would suggest deleting them and going straight into the text of the section.

Minor:
L17: Would just note what time period this 'slight increase' was

L19: Would consider stating geographically where these 'zones' are as out of context as they don't mean very much as no lead in from the abstract (i.e. NE of TP)

L35: I get what you mean, but I would rephrase the opening on this paragraph saying 'Glacier change can be measured using variations in...' to be clearer

L42: Would put the reference for each method after it was referred to i.e., spectral analysis (ref), object-based (ref), etc.

L46: needs a reference for debris-free glaciers before full stop

L51: are 'quality' and 'resolution' not the same thing? Also, what type of resolution, spatial or temporal? If both I would suggest stating the limitation of spatio-temporal resolution over the TP.

L58: Climate change seems a bit broad here, is it the increasing air temperature? Would clarify

L77: If you state 'numerous studies' I would suggest citing a handful of them as examples

L82-85: not convinced this is required for such a specific structure, think you're fine with just the aims of the paper being highlighted

L86: Just call it Study Area

L117: Would call it data and methods instead of materials

L118: This text is not needed - would go straight to 3.1

L124: I would just be careful saying images via GEE catalog are 'open-access' - while they are for individuals, there is commercial cost to access the platform

L124: Would merge these sentences, suggest at start of sentence saying Landsat data is used via GEE

L158: Figure 2 caption - Is this the total number of Landsat images or the total number used in the study with less than 60% cloud as defined by your study? Would clarify

L162-164: basically, the same as previous, would suggest merging and stating previous studies chose 0.4 as a threshold and therefore it was chosen here

L167-169: same applies about thresholding values NDWI, just merge them and say you chose 0.4

L171: 'Therefore' does not fit here as does not follow the previous sentence

L173-174: the final paragraph is just repetition of values you've defined - suggest deleting

L183: What holes? What do you mean 'filled'? Interpolated? Would clarify

L185: sentence does not make sense - what do you mean 'validated using reference debris-free glaciers'?

L253: Can Fig 3 be made bigger?

L255-258: Delete introductory

L275: Figure 4. Would suggest having the map as a) at the top of the figure then having panels b-i stacked. The lettering order seems a little confusing

L287: What do you mean the glaciers mapped between 2000 and 2005 'exhibit greater consistency'?

L300: Table 2 metric scores – see main comment 2

L400: I think Fig.10 and above text is more results than discussion. Would suggest moving to results. Do you have any numbers for the total area difference between the two methods? Would add to the argument of the AGEI method

L405: Would refrain from starting discussion sentences 'Fig X', if wanting to directly refer to the figure, place in brackets at the end of the sentence
Citation: https://doi.org/10.5194/egusphere-2024-1083-RC1
- AC1: 'Reply on RC1', Fumeng Zhao, 28 Jul 2024
  
  We thank the reviewer for the careful review and valuable suggestions. Hereby we give a point-by-point reply to address the comments in the attached pdf.
  
  Citation: https://doi.org/10.5194/egusphere-2024-1083-AC1
RC2:
'Comment on egusphere-2024-1083', Joseph Mallalieu, 17 Aug 2024
GENERAL COMMENTS
This manuscript investigates how changes in debris-free glacier extent in the Tibetan Plateau between ~1990-2020 are related to changes in climate (temperature, precipitation) over the same period. The authors: i) outline a novel technique (the AGEI) for mapping glacier extent, and use it to quantify changes in glacier area at 5-year intervals for 8 subzones of the Tibetan Plateau; ii) downscale ERA5-Land reanalysis data (specifically 2 m temperature and total precipitation) from 9 km2 to 1 km2 using ancillary variables from MODIS and the SRTM DEM) in a random forest regression, and; iii) use linear regressions to investigate the relationships between changes in glacier extent, temperature and precipitation at annual and seasonal scales. The authors conclude that total (debris-free) glacier area in the region has reduced over the survey duration, that temperatures have typically risen, that precipitation has increased in some subzones and reduced in others, and that in some subzones (particularly those in the Karakorum) reductions in glacier area correlated strongly with increases in temperature, but that the severity of glacier area loss may have been mitigated by corresponding increases in precipitation.
The manuscript presents a significant body of work, including novel mapping methodologies, extensive glacier and climate datasets, and insights into regional glacier and climate change in the Tibetan Plateau, and should therefore be of interest to others mapping glacier change generally, and particularly researchers investigating glacier change in central Asia. The addition of a paragraph discussing the implications of the glacier - climate relationships identified in this work with regard to existing climate projections for the region could help add impact. The manuscript is largely well-written, structured and presented, and could be further improved with some edits to make it more succinct (see comments for specific guidance).
My main comments relate to the methods adopted, several of which require additional clarification and/or detail to by fully comprehensible and transparent. I think this is of particular importance given the novelty of the mapping method used and the borderline nature of some of the glacier - climate relationships presented (e.g. see confidence intervals in Fig. 9). A caveat to my comments is that I am not an expert in Himalayan/Tibetan glaciation, or well-versed in random forest or time-series analyses, and have thus taken the methods adopted in section 3.2 (specifically lines 211-250) in good faith.

SPECIFIC COMMENTS
The accuracy and consistency of the novel AGEI mapping technique is integral to the results of this analysis. Consequently, a more detailed evaluation of the technique would be welcome, both to aid replication and assess its effectiveness. In particular, Fig. 10 does not have sufficient resolution to assess the accuracy of the AGEI mapping. Instead, a figure with multiple high-resolution panels (including some features that present challenges to accurate glacier mapping, such as lakes, shadow, cloud etc.), would be useful to illustrate the accuracy of the AGEI glacier outlines, particularly if the reference RGI 6.0 glacier outlines can be overlaid on them. Given its importance to the results, I think such a figure would be better situated and discussed in Section 3.1 or Supplementary Information, rather than waiting until the discussion. In addition, I struggled to fully understand the mapping evaluation indices presented on lines 189-198, particularly the Correctness (differentiating between A_cg and A_tg) and F1 scores (e.g. what does a high F1 score indicate?). If possible, illustrations of high and low scoring glaciers/regions/test sites (e.g. in Supplementary Information) could aid interpretation of the indices and the demonstrate the accuracy of the AGEI technique.

An additional section should be added to the end of the methodology to describe the linear regressions used to investigate the relationships between temperature/precipitation and glacier change (the linear model is mentioned at the beginning of Section 4.3, but not described fully). For transparency it would be useful to write out the model equation(s) used in the methods. It is unclear whether numerous independent models were generated (e.g. one for each for each climate variable, zone and season, totalling ~80 independent regressions), or if these were all included in one multiple regression model (and, if so, how multicollinearity between the seasonal climate was addressed). The results of the regressions should also be presented in a table that includes the coefficients, confidence intervals, standard error, t-values and p-values, and a summary of model diagnostics (R²/Adjusted R², RSE etc.). Although the coefficients and confidence intervals are presented in Fig. 9 their numeric values are difficult to determine, hence a table of the regression outputs would be useful to support this figure, aid interpretation of the data, and provide more robust conclusions.

A major caveat of this work is that the analysis presented only considers debris-free glaciers. Greater context regarding the prevalence and distribution of debris-covered glaciers in the region would therefore be useful (in the introduction, or sections 3.1 or 5.3) to understand the potential implications of this methodological choice on the results and their interpretation (e.g. are debris-free glaciers typically representative of glaciers across the region; could there be some sub-zones where the exclusion of debris-covered glaciers from the analysis significantly reduces the sample size and therefore the representativeness of the results?).

Given the relationships identified between changes in glacier area, temperature and precipitation from ~1990-2020, I think a paragraph could be added to the discussion to explore the implications of these results in respect of existing climate projections for the region. For example, how is regional climate likely to change by the end of the century and what could be the implications for the glaciers in the different sub-zones given the trends identified between 1990-2020? A short paragraph addressing this could help to add impact to the manuscript, and potentially identify future research priorities.

Throughout the manuscript the authors refer to ‘glacier retreat’ which (to me at least) implies a measure of distance retreated by a glacier terminus, particularly the terminus of a valley or outlet glacier. Given that the work presented is concerned with changes in glacier area (not just recession at termini) ‘glacier extent’ may be a more intuitive term to use.

The manuscript could be made more succinct by removing several sections of text outlining the structure (which is self-evident), particularly lines 82-85, 118-119, 255-257, and 380-381.

Figures 4 & 5: Both figures show an increase in glacier area in zones II, V and VII between 2000 and 2005. This seems unusual given the overall regional trend, so I’m interested to know whether this is an artefact of the mapping methodology, or whether this brief increase in glacier extent in these regions has been documented in any other studies/literature? The same question also applies to zones VI and VIII, which appear to show notable increases in glacier area between 2005-2010.

TECHNICAL COMMENTS
Abstract: Here, references to the zones used in the study would be better substituted with the names of the corresponding regions (e.g. Karakorum, southern Himalaya etc.) because the zones are only intelligible to those who have already read Section 2.

Lines 209-210: Is there a precedent or justification for the use of the 3 ancillary MODIS and SRTM DEM variables to downscale the ERA5-Land data?

Line 247: A significance level of 0.1 seems unusually low, and unnecessary given that the 0.05 significance level is also employed. Is there a good rationale for the inclusion of the 0.1 significance level?

Lines 262-263: This description (‘most pronounced retreat’) does not appear to be entirely consistent with the panels in Fig. 4 which show rates of area reduction being greatest in zones I, III and IV, with more modest losses in zones II and VIII.

Lines 263-264: The term ‘glacier advance’ is a little misleading here, given that regions VI and VIII (encompassing the Karakoram) show overall trends of area loss in Fig. 4. Presumably, ‘advance’ here refers to some individual glaciers that are acting counter to the overall trend in these subzones. If so, they are not easily visible in the main map in Fig. 4. Consequently, an enlarged inset of these subzones (or a separate figure) may be required to better illustrate their presence.

Lines 268-269: States that glacier advances are seen from 2000-2005 ‘in all zones except for VI and VIII’. However, zones III and IV also do not show evidence of glacier advance over this period (as illustrated in both Figs. 4 & 5).

Figure 4: To aid interpretation of the figure it would be worth noting in the figure caption that the values presented below each zone number represent the annual change in glacier area (and not the total change over the duration of the study).

Lines 284-286: What might explain the much higher percentage differences between the AGEI and RGI outlines in the other zones? It is not good practice to only report the lowest differences and use them as validation of the technique.

Lines 290-291 & Table 2: More detail regarding these indices would be useful. My (possibly incorrect) interpretation of this table is that only around 60-70% of the RGI 6.0 glacier areas have been mapped by the AGEI technique, which seems rather low (also see Specific Comment 1 above).

Lines 305-306 & Figure 6 caption: more information regarding the training and validation samples would be useful. These do not appear to be mentioned in the methodology.

Lines 327-329 & 337-338: It is unclear what ‘glacier advance’ is referring to specifically in these instances (e.g. individual glaciers in the subzones, or overall advances in some zones in particular time periods such as 2005-2010?). Also see comment regarding Lines 263-264.

Lines 334-335: Avoid the use of ‘respectively’ in this instance (the list is too long), and just include the name of each period (annual, fall etc.) next to each numeric value.

Figure 8: It was not immediately apparent to me that the colour key underneath the figure was universal to both sides of the figure (a & b). A note in the figure caption could help to clarify this.

Lines 354-355: Presumably ‘variations’ here should be ‘increases’ to specify the direction of the relationship between glacier area and temperature? In addition, it states that ‘all the subzones’ have a negative correlation between glacier area and temperature change, but Figure 9a (annual) only shows zones II, III, VI and VIII to be clearly below 0 on the y-axis (thus presumably indicating glacier area loss). The coefficients and confidence intervals of the remining zones is difficult to discern. A table containing these values could be useful here (see Specific Comment 2).

Lines 361-362: See previous comment. A visual interpretation of Figure 9b (annual) suggests that only zones III, V, VI and VIII have confidence intervals in excess of 0 on the y-axis, and that therefore only these 4 zones can be confidently interpreted as exhibiting a positive relationship between glacier area and total precipitation at an annual scale. Equally, if the confidence intervals for zone II also cross the y-axis and it cannot be said with confidence that the relationship is negative. I think it would be beneficial for the authors to re-write this entire paragraph with greater consideration given to their confidence intervals, particularly where they cross 0 on the y-axis.

Line 363: Should ‘zones III, V, and VIII’ read ‘zones III, VI and VIII’? It also looks like zone V could be added to this list.

Lines 363-365 and Figure 9b: Units are inconsistent here, the text states km²/mm, but the y-axis on Figure 9b states km²/m. The abstract also states km²/mm.

Line 383: Which earlier study; this study or Huang et al. 2021?

Lines 405-406: For consistency, it would be useful to also include zone numbers when referring to locations in this sentence.

Lines 409-411: See previous comment.

Lines 432-436: Given the scarcity of data points I think the authors should be cautious of giving too much weight to the data presented in Figure 11. For example, although zones I and III have lower median debris thickness and relatively high rates of glacier area loss, zone 8 has a similar level of debris thickness, but much lower rates of glacier area loss, and zone 4 has the highest median debris thickness, but still has the 3rd highest rate of glacier area loss. This paragraph could benefit from a discussion of the differences between the regional glacier area losses presented here (for debris-free ice) and any studies of regional glacier area losses for debris-covered glaciers (should the data be available).

Line 451: Clarify that the ‘total glacier area’ is the debris-free total glacier area.

Line 457-458: I think this sentence should be rephrased because the total glacier area in these regions has typically reduced over the duration of the study (see trends on Figure 4), rather than advanced. Given the overall reduction in glacier area and the correlations presented in Fig. 9b it may be more appropriate to state that annual and summer precipitation may have helped to mitigate reductions in glacier area in the Karakorum.

Lines 461-464: I think this conclusion could also be revised with regard to the earlier comment about confidence intervals (see comment for Lines 361-362).
Citation: https://doi.org/10.5194/egusphere-2024-1083-RC2
- AC2: 'Reply on RC2', Fumeng Zhao, 07 Sep 2024
  
  We thank the reviewer for the careful review and valuable suggestions. Hereby we give a point-by-point reply to address the comments in the attached pdf.
  
  Citation: https://doi.org/10.5194/egusphere-2024-1083-AC2

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision | EF: Editorial file upload

ED: Publish subject to minor revisions (review by editor) (01 Oct 2024) by Nicholas Barrand

AR by Fumeng Zhao on behalf of the Authors (10 Oct 2024) Author's response Author's tracked changes Manuscript

ED: Publish as is (15 Oct 2024) by Nicholas Barrand

AR by Fumeng Zhao on behalf of the Authors (16 Oct 2024) Author's response Manuscript

Journal article(s) based on this preprint

03 Dec 2024

Linking glacier retreat with climate change on the Tibetan Plateau through satellite remote sensing

Fumeng Zhao, Wenping Gong, Silvia Bianchini, and Zhongkang Yang

The Cryosphere, 18, 5595–5612, https://doi.org/10.5194/tc-18-5595-2024,https://doi.org/10.5194/tc-18-5595-2024, 2024

Short summary

Fumeng Zhao, Wenping Gong, Silvia Bianchini, and Zhongkang Yang

Viewed

Total article views: 465 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
305	115	45	465	20	15

HTML: 305
PDF: 115
XML: 45
Total: 465
BibTeX: 20
EndNote: 15

Views and downloads (calculated since 04 Jun 2024)

Month	HTML	PDF	XML	Total
Jun 2024	119	39	16	174
Jul 2024	65	28	8	101
Aug 2024	40	7	5	52
Sep 2024	42	19	15	76
Oct 2024	20	13	1	34
Nov 2024	18	9	0	27
Dec 2024	1	0	1
Jan 2025	0

Cumulative views and downloads (calculated since 04 Jun 2024)

Month	HTML	PDF	XML	Total
Jun 2024	119	39	16	174
Jul 2024	65	28	8	101
Aug 2024	40	7	5	52
Sep 2024	42	19	15	76
Oct 2024	20	13	1	34
Nov 2024	18	9	0	27
Dec 2024	1	0	1
Jan 2025	0

Viewed (geographical distribution)

Total article views: 501 (including HTML, PDF, and XML) Thereof 501 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 14 Jan 2025

Download

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Preprint (4535 KB)
Metadata XML

Short summary

Glacier retreat patterns and climatic drivers on the Tibetan Plateau are uncertain at finer resolutions. This study introduces a new glacier mapping method covering 1988 to 2022, with downscaled air temperature and precipitation data. It quantifies the impacts of annual and seasonal temperature and precipitation on retreat. Results show rapid and varied retreat, with annual temperature influencing retreat in the west and spring precipitation in the northwest.


Total:	0
HTML:	0
PDF:	0
XML:	0