the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Relationships between Andean Glacier Ice-Core Dust Records and Amazon Basin Riverine Sediments
Abstract. Dust particle studies in ice cores from the tropical Andes provide important information about climate dynamics. We investigated dust concentrations from a 22.7 m ice-core recovered from the Quelccaya Ice Cap (QIC) in 2018, representing 12 years of snow accumulation. The dust seasonality signal was still preserved with some homogenization of the record due to surface melting and percolation. Using a microparticle counter, we measured the dust concentration from 2–60 µm and divided the annual dust concentration into three distinct groups: fine particle percentage (FPP, 2–10 µm), coarse particle percentage (CPP, 10–20 μm) and giant particle percentage (GPP, 20–60 μm). Increased dust was associated with the warm stage of the Pacific Decadal Oscillation index (PDO) from 2014–2017 with significant increases in FPP and a relative decrease in GPP. There was a positive correlation between PDO and FPP (r = 0.68, p-value < 0.02). CPP and GPP were dominant during the PDO cold phase (2005–2013) and were more strongly associated with the Tropical Northern Atlantic index (TNA), which was positive from 2005–2017. The relation between TNA and CPP was r = 0.60 (p-value < 0.05) and that with GPP was r = 0.59 (p-value < 0.05). We also revealed a potential link between QIC dust and Madeira River sediments and runoff. Sediment concentration decreases at Porto Velho station were correlated with %GPP (r = 0.67, p < 0.02) from 2005–2017. This relationship contributes to a better understanding of the effects of PDO oscillations on both parameters. The %GPP and sediment decreases were potentially linked with the PDO phase change from negative to positive. We also noted a strong negative correlation between FPP and runoff (r = −0.80, p < 0.002) from 2005–2016, which was understandable due to the relationship of FPP to wetter conditions while runoff decreases were associated with increasing dryness in the southern part of the Madeira Basin. Assessing dust record variability by distinct size groups can help to improve our knowledge of how the Pacific and Atlantic oceans influence atmospheric oscillations in the QIC. In addition, the association of dust variability with dynamic changes in sediments and runoff in the Madeira River system demonstrates the potential for future investigation of linkages between QIC dust and Amazon basin rivers.
- Preprint
(8955 KB) - Metadata XML
-
Supplement
(56 KB) - BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on tc-2021-186', Anonymous Referee #1, 27 Sep 2021
This paper tries to establish linkages between dust deposition on the Quelccaya Ice Cap (QIC) in Peru, its grain size distribution and tropical Atlantic and Pacific modes of variability. In my opinion the paper fails to convincingly do so. The entire analysis is exclusively based on correlation analyses, without providing a causal mechanism that could support and explain the suggested relationships. The abstract alone lists 6 r-values, yet does not include a discussion as to why these correlations occur. Discussing correlation coefficients is fine, but in the end scientific inquiry requires understanding or at least investigating the mechanisms that underpin these statistical relationships. The paper falls short in this regard and the proposed relationships remain conjecture. There are other problems with this paper as well, including the seeming lack of awareness of prior work, or inadequate statistical methods employed, that further lower the quality of this study. I can therefore not recommend this paper for publication, but I have tried to outline a few avenues for improvement that may help the authors to reconfigure their analyses.
The discussion of prior studies in section 2 focused on snowfall and climate on QIC is inadequate. There is a lot of work that has been performed understanding snowfall and related circulation mechanisms on the QIC, yet most recent studies are ignored. For example, Hurley et al. (2015) analyzed snowfall on QIC and related atmospheric circulation mechanisms and tied snowfall events to cold air incursions. Perry et al. (2017) also analyzed snowfall events at the same site and tied them to northerly and easterly airflow using back-trajectory analyses. Hurley et al. (2019) compared the climatic conditions on QIC during Pacific cold and warm events and documented through which pathways tropical Pacific SST influence climate on QIC. All these studies are highly relevant for the work presented here, yet none of them are even mentioned in the paper.
Line 62-63: Rabatel et al. (2013) make no such statement that all glaciers will disappear in the tropical Andes by the end of the 21st century.
Lines 128-132: The influence of ENSO on QIC is indeed profound but it manifests itself in changes of the mass and energy balance, rather than direct retreat of the ice margin, as large ice sheets such as QIC respond with some delay in adjusting their extent to climatic forcing.
Lines 150-152: I don't understand the rationale for defining the hydrological year as April-March. April still very much belongs to the prior wet season as snowfall on QIC usually ends by the end of April or in early May (see Fig. 3 in Hurley et al., 2015). Defining the hydrologic year from July to June would therefore make much more sense.
Calculation of dust concentrations. There is no discussion of how snowfall amount and snow loss due to sublimation and wind scour (both of which are significant on QIC) factor in when calculating the actual dust flux. Concentrations are sensitive to both dilution by snowfall and increasing in concentrating via snow loss. This aspect requires a thorough discussion, but is completely ignored in this paper.
Statistical approach in Figs. 5&7: I have serious concerns about the statistical relationships and their significance derived from only twelve data points. The PDO is a slowly evolving multi-decadal index and establishing relationships with this mode of variability would require much longer time series, that cover at least one full warm and cold phase (i.e. at least 50 years). Furthermore, both the FPP and the GPP show clear trends in their data. For a robust statistical comparison these trends would have to be removed prior to the calculation of correlations, as otherwise the relationship may simply hinge on common trends in the data, rather than reflect actual year-to-year causal mechanisms. The same comment also applies to Madeira River sediments and runoff in Figure 7.
How exactly the ice core chronology was determined, needs to be explained in much more detail. The year 2015/16, for example, was marked by an extreme El Niño, with almost zero net accumulation on QIC. Yet the chronology presented in Fig. 2 assumes a normal year with ~ 1.5 m weq snow accumulation, which is hard to reconcile with on-site accumulation measurements for that year (see Fig. 1c in Hurley et al., 2019)
Fig. 6: These correlations are strongly influenced by one outlier. You should repeat this analysis using only data that fit a normal distribution (i.e. without the outlier) to confirm that your relationship still holds.
Minor edits:
Line 32: there are no ‘atmospheric oscillations in the QIC’
Line 79: It is a ’Stampfli’ drill (not ‘Stampli’)
Line 117: “Atmospheric circulation over the Amazon basin’ may be influenced by, but does not ‘come from the tropical Atlantic Ocean’
Lines 202-203: This sentence is incomplete: “To explore the relationship between total dust concentration in different size ranges (Figure 4)”.
Line 375: capitalize ‘Cordillera Vilcanota’
Lines 400-403: You reference a discussion paper that was rejected after peer-review. Please delete this reference and refrain from citing it in the text.
References cited in this review
Hurley, J.V., et al., 2015: Cold air incursions, d18O variability and monsoon dynamics associated with snow days at Quelccaya Ice Cap, Peru. J. Geophys. Res., 120, 7467-7487, doi:10.109/2015JD023323.
Hurley, J.V., et al., 2019: On the interpretation of the ENSO signal embedded in the stable isotopic composition of Quelccaya Ice Cap, Peru. J. Geophys. Res. 124, 131-145, doi:10.1029/2018JD029064.
Perry, L.B., et al., 2017: Characteristics of precipitating storms in glacierized tropical Andean Cordilleras of Peru and Bolivia. Ann. Amer. Assoc. Geogr., 107(2), 309-322, doi:10.1080/24694452.2016.1260439.
Rabatel, A., et al., 2013: Current state of glaciers in the tropical Andes. A multi-century perspective on glacier evolution and climate change. The Cryosphere, 7, 81-102, doi:10.5194/tc-7-81-2013.
Citation: https://doi.org/10.5194/tc-2021-186-RC1 -
AC1: 'Reply on RC1', Rafael Reis, 17 Nov 2021
We thank the Anonymous Referee #1 for his constructive comments and suggestions. The reviewer comments are summarized in bold, and our point-by-point responses are listed below in italic format. Our responses document and the data available are in the file attached.
-
AC1: 'Reply on RC1', Rafael Reis, 17 Nov 2021
-
RC2: 'Comment on tc-2021-186', Anonymous Referee #2, 26 Oct 2021
This paper demonstrates correlations between ice-core dust records in Peru and Amazon Basin river sedimentology. Glacier melting in the tropical Andes and hydrology of the Amazon Basin is well studies in recent years. Even though the established correlations are convincing, a detailed discussion on the underlying mechanisms that lead to such correlation is missing. I recommend Major revisions.
I strongly recommend analyzing precipitation (snowfall) and snowmelt during the study period. How about albedo?
I understand from figure 3 that the dust concentration is higher particularly when both El Nino and warm-PDO coincide (1997/1998 and 2015/2016). I don’t see it in the discussion.
I’m curious about the delays between the causes and effects. Examples: delay between PDO/TNA and glacier changes, delay between glacier melting and changes in sediments in the Amazon Basin. How these delays between causes and effects were considered while establishing correlations?
Minor comments
Line 180: PDO and TNA SST anomalies may influence precipitation (not indices).
When PDO and ENSO in phase, what would influence on snowline? Temperature? Precipitation?
Citation: https://doi.org/10.5194/tc-2021-186-RC2 - AC2: 'Reply on RC2', Rafael Reis, 17 Nov 2021
Status: closed
-
RC1: 'Comment on tc-2021-186', Anonymous Referee #1, 27 Sep 2021
This paper tries to establish linkages between dust deposition on the Quelccaya Ice Cap (QIC) in Peru, its grain size distribution and tropical Atlantic and Pacific modes of variability. In my opinion the paper fails to convincingly do so. The entire analysis is exclusively based on correlation analyses, without providing a causal mechanism that could support and explain the suggested relationships. The abstract alone lists 6 r-values, yet does not include a discussion as to why these correlations occur. Discussing correlation coefficients is fine, but in the end scientific inquiry requires understanding or at least investigating the mechanisms that underpin these statistical relationships. The paper falls short in this regard and the proposed relationships remain conjecture. There are other problems with this paper as well, including the seeming lack of awareness of prior work, or inadequate statistical methods employed, that further lower the quality of this study. I can therefore not recommend this paper for publication, but I have tried to outline a few avenues for improvement that may help the authors to reconfigure their analyses.
The discussion of prior studies in section 2 focused on snowfall and climate on QIC is inadequate. There is a lot of work that has been performed understanding snowfall and related circulation mechanisms on the QIC, yet most recent studies are ignored. For example, Hurley et al. (2015) analyzed snowfall on QIC and related atmospheric circulation mechanisms and tied snowfall events to cold air incursions. Perry et al. (2017) also analyzed snowfall events at the same site and tied them to northerly and easterly airflow using back-trajectory analyses. Hurley et al. (2019) compared the climatic conditions on QIC during Pacific cold and warm events and documented through which pathways tropical Pacific SST influence climate on QIC. All these studies are highly relevant for the work presented here, yet none of them are even mentioned in the paper.
Line 62-63: Rabatel et al. (2013) make no such statement that all glaciers will disappear in the tropical Andes by the end of the 21st century.
Lines 128-132: The influence of ENSO on QIC is indeed profound but it manifests itself in changes of the mass and energy balance, rather than direct retreat of the ice margin, as large ice sheets such as QIC respond with some delay in adjusting their extent to climatic forcing.
Lines 150-152: I don't understand the rationale for defining the hydrological year as April-March. April still very much belongs to the prior wet season as snowfall on QIC usually ends by the end of April or in early May (see Fig. 3 in Hurley et al., 2015). Defining the hydrologic year from July to June would therefore make much more sense.
Calculation of dust concentrations. There is no discussion of how snowfall amount and snow loss due to sublimation and wind scour (both of which are significant on QIC) factor in when calculating the actual dust flux. Concentrations are sensitive to both dilution by snowfall and increasing in concentrating via snow loss. This aspect requires a thorough discussion, but is completely ignored in this paper.
Statistical approach in Figs. 5&7: I have serious concerns about the statistical relationships and their significance derived from only twelve data points. The PDO is a slowly evolving multi-decadal index and establishing relationships with this mode of variability would require much longer time series, that cover at least one full warm and cold phase (i.e. at least 50 years). Furthermore, both the FPP and the GPP show clear trends in their data. For a robust statistical comparison these trends would have to be removed prior to the calculation of correlations, as otherwise the relationship may simply hinge on common trends in the data, rather than reflect actual year-to-year causal mechanisms. The same comment also applies to Madeira River sediments and runoff in Figure 7.
How exactly the ice core chronology was determined, needs to be explained in much more detail. The year 2015/16, for example, was marked by an extreme El Niño, with almost zero net accumulation on QIC. Yet the chronology presented in Fig. 2 assumes a normal year with ~ 1.5 m weq snow accumulation, which is hard to reconcile with on-site accumulation measurements for that year (see Fig. 1c in Hurley et al., 2019)
Fig. 6: These correlations are strongly influenced by one outlier. You should repeat this analysis using only data that fit a normal distribution (i.e. without the outlier) to confirm that your relationship still holds.
Minor edits:
Line 32: there are no ‘atmospheric oscillations in the QIC’
Line 79: It is a ’Stampfli’ drill (not ‘Stampli’)
Line 117: “Atmospheric circulation over the Amazon basin’ may be influenced by, but does not ‘come from the tropical Atlantic Ocean’
Lines 202-203: This sentence is incomplete: “To explore the relationship between total dust concentration in different size ranges (Figure 4)”.
Line 375: capitalize ‘Cordillera Vilcanota’
Lines 400-403: You reference a discussion paper that was rejected after peer-review. Please delete this reference and refrain from citing it in the text.
References cited in this review
Hurley, J.V., et al., 2015: Cold air incursions, d18O variability and monsoon dynamics associated with snow days at Quelccaya Ice Cap, Peru. J. Geophys. Res., 120, 7467-7487, doi:10.109/2015JD023323.
Hurley, J.V., et al., 2019: On the interpretation of the ENSO signal embedded in the stable isotopic composition of Quelccaya Ice Cap, Peru. J. Geophys. Res. 124, 131-145, doi:10.1029/2018JD029064.
Perry, L.B., et al., 2017: Characteristics of precipitating storms in glacierized tropical Andean Cordilleras of Peru and Bolivia. Ann. Amer. Assoc. Geogr., 107(2), 309-322, doi:10.1080/24694452.2016.1260439.
Rabatel, A., et al., 2013: Current state of glaciers in the tropical Andes. A multi-century perspective on glacier evolution and climate change. The Cryosphere, 7, 81-102, doi:10.5194/tc-7-81-2013.
Citation: https://doi.org/10.5194/tc-2021-186-RC1 -
AC1: 'Reply on RC1', Rafael Reis, 17 Nov 2021
We thank the Anonymous Referee #1 for his constructive comments and suggestions. The reviewer comments are summarized in bold, and our point-by-point responses are listed below in italic format. Our responses document and the data available are in the file attached.
-
AC1: 'Reply on RC1', Rafael Reis, 17 Nov 2021
-
RC2: 'Comment on tc-2021-186', Anonymous Referee #2, 26 Oct 2021
This paper demonstrates correlations between ice-core dust records in Peru and Amazon Basin river sedimentology. Glacier melting in the tropical Andes and hydrology of the Amazon Basin is well studies in recent years. Even though the established correlations are convincing, a detailed discussion on the underlying mechanisms that lead to such correlation is missing. I recommend Major revisions.
I strongly recommend analyzing precipitation (snowfall) and snowmelt during the study period. How about albedo?
I understand from figure 3 that the dust concentration is higher particularly when both El Nino and warm-PDO coincide (1997/1998 and 2015/2016). I don’t see it in the discussion.
I’m curious about the delays between the causes and effects. Examples: delay between PDO/TNA and glacier changes, delay between glacier melting and changes in sediments in the Amazon Basin. How these delays between causes and effects were considered while establishing correlations?
Minor comments
Line 180: PDO and TNA SST anomalies may influence precipitation (not indices).
When PDO and ENSO in phase, what would influence on snowline? Temperature? Precipitation?
Citation: https://doi.org/10.5194/tc-2021-186-RC2 - AC2: 'Reply on RC2', Rafael Reis, 17 Nov 2021
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
841 | 268 | 53 | 1,162 | 103 | 56 | 52 |
- HTML: 841
- PDF: 268
- XML: 53
- Total: 1,162
- Supplement: 103
- BibTeX: 56
- EndNote: 52
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1