Misinterpreting proxy data for paleoclimate signals: A comment on Shukla et al. 2020

Abstract

Shukla et al. explored paleoclimatic signals from a ~8 m thick profile of a moraine-dammed lake in the central Himalaya exposed due to lake burst from a flash flood in 2013. The main objective of their research work is to understand the complex glacial-climate system during late-Holocene. They attempted a novel multi proxy approach for paleoclimate reconstruction but their work suffers from misinterpretation of various proxies and erroneous/misleading discussion. We therefore report following major points in this comment article. (1) Misinterpretation of magnetic parameters: Magnetic susceptibility (χlf) has been used to interpret changes in magnetic mineralogy rather than concentration of magnetic minerals. Susceptibility of anhysteretic remanence (χARM) has been used at several places to indicate presence of superparamagnetic (SP) and multi domain (MD) ferrimagnetic particles rather than single domain (SD) ferrimagnetic (magnetite) particles. Interpreting erroneous negative values of percentage of frequency dependent susceptibility (χfd%) for climate change. (2) Poor chronology: Overlaps in ages of glacial-lake sediments. (3) References: Several statements in paper have not been referenced and some of them have out of place citations. (4) Carefree writing: Authors have shown typical example of carefree writing of a research article, for example, giving units to dimensionless parameter S-ratio, and χfd%, differences in units of χlf in text and figure, different depths for the same age in text and figure. (5) Over interpretation: Authors at places have interpreted climatic variations based on only one sample. (6) Poor justifications: Authors did not provide any detailed justification for proxy data while interpreting climatic variations. (7) No data (results) on mineralogy and trace elements were given. Overall it is not only a problem of presentation and misinterpretation of proxy data but the study also fails to deliver the final message of climate change and glacier dynamics in the central Himalaya.

Keywords

chemical weathering environmental magnetism glacial-lake sediments Himalaya Indian summer monsoon

Shukla et al. (2020) investigated relationship between climate change and glacial system of the central Himalaya during late-Holocene using multi-proxy study of the Chorabari glacial-lake sediments in the Mandakini basin. There are several shortcomings in the article, which have been described section-wise.

Introduction

Shukla et al. (2020) wrote ‘numerous high-resolution palaeo-records of the Himalaya suggest that the temperature and precipitation conditions of the Himalaya are strongly controlled by variations in insolation on orbital and decadal time scales (Gupta et al., 2003, 2005; Shukla et al., 2018)’. Here, they have cited three references that include one self-citation from Himalaya and other two from Arabian Sea. Several high resolution paleoclimatic records from NW and Central Himalaya such as Demske et al. (2009); Kathayat et al. (2016, 2017); Kotlia et al. (2010); Leipe et al. (2014); Rawat et al. (2015a); Wünnemann et al. (2010) have recorded insolation induced changes in Indian summer monsoon (ISM) variability but have been missed by authors.

Materials and methods

Shukla et al. (2020) mentioned that they had carried out mineralogical analysis (also in abstract). However, no results and discussion on mineralogical analysis have been provided.

Environmental magnetic parameters

The authors, while writing formula for S-ratio, have mistakenly written IRM 300T, which should be in mT. The S-ratio formula is −IRM_bf/SIRM (Stober and Thompson, 1979a; Thompson and Oldfield, 1986). The IRM_bf (IRM measurement in backfield after first saturating sample in forward field) at 300 mT is mostly used for calculation of S-ratio and is expressed as −IRM_−300mT/SIRM (King and Channell, 1991; Liu et al., 2007). Shukla et al. (2020) have also incorrectly cited reference of Stober and Thompson (1979b) ‘Magnetic remanence acquisition in Finnish lake sediments. Geophysical Journal of the Royal Astronomical Society 57: 727–739’ for calculation of S-ratio. The correct reference of Stober and Thompson (1979a) in which they proposed S-ratio is ‘An investigation into the source of magnetic minerals in some Finnish lake sediments, Earth and Planetary Science Letters 45: 464–474’.

Shukla et al. (2020) explained the processes that control deposition of magnetic minerals in lakes. Authors stated in points 1 and 2 that magnetic mineral deposition in lakes are influenced by concentration of ferri/ferromagnetic iron oxides in the catchment and from their relative contribution. We wish to highlight here that ferromagnetism exists in pure metals (e.g. pure iron) and is rare in environment (Dunlop and Özdemir, 1997; Evans and Heller, 2003). Hematite is an antiferromagnetic mineral. Therefore, we think that authors want to state antiferromagnetic minerals in these two points rather than ferromagnetic iron metal.

Further, the authors, while explaining the processes that control lake magnetism, have missed to describe that the lake magnetism is also influenced by genesis of authigenic magnetic minerals (e.g. greigite) as well as bacterial magnetite in certain cases or environment that have been well reported from various lakes globally (Kim et al., 2005; Paasche et al., 2004; Snowball, 1991, 1994).

Shukla et al. (2020) wrote that ‘in periods of climatic transition, the supply of minerogenic material to lakes is reduced and the sediment with a lower organic content were deposited. A slow deposition rate could also lead to magnetite dissolution, resulting in a low magnetic concentration and high S-ratio. Conversely, during periods of glacial retreat, the minerogenic sedimentation rate increases and magnetite is preserved in the sediments because of rapid burial, which results in a high magnetic concentration and low S-ratio’. These statements are incomplete. First of all, it has to be clarified that whether these findings/statements on magnetite dissolution under slow depositional environment are author’s own observation from present analysis or just imaginary/presumed. If these sentences are authors own observation then they should have moved this paragraph to results and discussion section. If these are not author’s own observation from present analysis, then they have made gross scientific misconduct on ethics by not referencing any original research from where these ideas might have been taken (e.g. Snowball, 1993). Second, they stated that magnetite dissolves during slow deposition and is preserved during rapid burial. The post-depositional magnetite dissolution does not only depend on the slow deposition but also on the anoxic state and productivity/organic matter of lakes (Dearing, 1999a; Snowball, 1993; Stockhausen and Thouveny, 1999; Tamuntuan et al., 2015). Authors wrote that dissolution during slow deposition leads to low magnetic concentration and high S-ratio, whereas preserved magnetite in the sediment during rapid burial leads to higher magnetic concentration and low S-ratio. However, authors here gave incomplete information about the processes and thus interpretation regarding changes in S-ratio. The S-ratio parameter informs about the relative abundance of ferrimagnetic and antiferromagnetic minerals (Frank and Nowaczyk, 2008; King and Channell, 1991; Liu et al., 2007; Thompson and Oldfield, 1986). In absence of ferrimagnetic minerals due to dissolution, enhanced signal of antiferromagnetic magnetic minerals (e.g. hematite), generally unaffected by dissolution, is recorded by S-ratio (Dearing, 1999a; Wang et al., 2016). This does not imply that concentration of antiferromagnetic minerals have increased and should not be used for climatic interpretations (or only carefully).

Elemental analysis

Shukla et al. (2020) did not provide details of instrument or lab used. They also did not report the accuracy and precision of measurements for major and trace elements. They have cited two references Purohit et al. (2010) and Saini et al. (2002) for reporting the accuracy. Saini et al. (2002) have used Energy-dispersive XRF and SO-1 and GSS-4 soil standards for quality control of trace element measurements. They did not report any accuracy for major elements. Further, Shukla et al. (2020) wrote that they have analyzed several trace elements (Ba, Sc, V, Cr, Co, Ni, Cu, Zn, Ga, Pb, Th, Rb, U, Sr, Y, Zr, and Nb) but have not provided data or plots (results) for anyone of the trace element in the entire article. Shukla et al. (2020) also cited reference of Purohit et al. (2010) for accuracy determination using international standards. Purohit et al. (2010) studied C and O isotopes on Paleoproterozoic carbonate rocks of the Aravalli Supergroup and have only analyzed Mn and Sr elements using ICP-AES, and not by XRF. Purohit et al. (2010) used USGS rock standard G-2 for accuracy and precision determination of Mn and Sr elements, and did not report for any other elements such as Al₂O₃, Na₂O, and SiO₂, etc. The analytical measurement of Purohit et al. (2010) is totally different from the one carried out by Shukla et al. (2020) and therefore it seems that reference of Purohit et al. (2010) is out of place and has been cited unnecessarily. Hence, authors must have calculated and reported accuracy and precision of their measurements rather than citing references that are out of place in context of present analysis.

Shukla et al. (2020) calculated Chemical Index of Alteration (CIA) to understand chemical weathering of catchment rocks and their subsequent deposition in lakes in response to changing climate. However, authors did not provide the formula of CIA and neither has cited original research work of Nesbitt and Young (1982).

Results and discussion

Chronology of climatic events

Major doubts arise from OSL chronology of glacial-lake sediments. Authors have produced five OSL ages from ~8 m thick lake profile. The OSL ages are 2311 ± 117 yr (oldest) at 728 cm depth, 2134 ± 134 yr at 456 cm depth, 1842 ± 193 yr at 228 cm depth, 1515 ± 366 yr at 168 cm depth, and 1170 ± 214 yr at 125 cm depth (youngest). The errors (/uncertainty) in all OSL ages range from ~117 to 366 yr. When considering the upper and lower limits of errors of subsequent ages, we observed an overlap in ages. For example, the first (/oldest) age of 2311 yr at 728 cm depth has an error of ±117 yr. Therefore, upper and lower limits of this age will be 2428 and 2194 yr, respectively. The second age at 456 cm depth is 2134 yr which has an error of 134 yr. The upper and lower limits of this age are 2268 and 2000 yr, respectively. This shows that the lower limit of first/oldest age at 728 cm depth and upper limit of second age at 456 cm depth overlaps, which suggests almost similar ages (when considered uncertainty) at depth interval of almost 272 cm. Similarly, the lower limit of second age (i.e. 2000 yr) at 456 cm depth and upper limit of third age (i.e. 2035 yr) at 228 cm depth also overlaps. The same is observed with lower and upper age limits of the subsequent ages at 228, 168, and 125 cm depths. These overlaps itself casts a doubt on the reliability of OSL ages. Authors must explain why they have used ages that overlaps. They must have dated more samples at different depths as well as using different methods of chronology.

The depths reported for youngest and oldest ages in text and Authors’ Figure 7 are different. For example, depth for oldest age (~2.3 ka) in text is written as 763 cm whereas in Authors’ Figure 7 is 728 cm. Similarly, depth for youngest age (~1.2 ka) in text is 168 cm, whereas in figure depth is 125 cm. Authors must give a clarification for the discrepancies in reporting of depths in text and figure.

Shukla et al. (2020) did not provide sediment accumulation rate (SAR) for the intermediate depths and simply stated that lake strata preserved rapid sedimentation. We calculated SAR for the Sections I, II, and III (Figure 1). We considered year 2014 (exposure of lake sequence in 2013) as the calibration year for CE and BCE. For Section I, age ranges from 270 BCE to 260 CE and depth interval is ~688 to 212 cm. We found SAR of ~1.16 cm/yr for Section I (Figure 1). The Sections II (900–1260 CE) and III (1370–1720 CE) have depth intervals from ~119 to 81 cm and ~69 to 31 cm, respectively. Both Sections (II and III) have a similar SAR of ~0.11 cm/yr (Figure 1). It is important to note that top OSL age of lake profile is ~1170 yr (~125 cm depth) and ages for both Sections (II and III) are extrapolated based on this date. Therefore, two of the three studied sections have extrapolated ages and have not been dated directly. Problem arises from these extrapolated dates when they cover important periods of global/regional climate shifts (e.g. dry to wet climate) as no changes in sedimentation can be found during these shifts. For example, Shukla et al. (2020) referred Section II to warm-humid climate and Section III to cold-dry climate but sedimentation rate during these climate shifts are uniform (Figure 1).

Figure 1.

Depth plotted against sediment accumulation rate (SAR) for Chorabari glacial-lake sediments. Summary of SAR for three studied sections are provided in the boxes on right.

It is also important to note that sedimentation rate is almost 10 times higher for Section I compared to Sections II and III (Figure 1). Authors inferred cooler climate with reduced precipitation strength during Section I (270 BCE–260 CE). Therefore, authors also must explain why there is very high sedimentation during period of cooler climate with weaken precipitation strength (Section I) compared to very low sedimentation during period of warm and humid climate with strengthen precipitation (Section II)?

Interpretation of proxy data

Section I (270 BCE–~260 CE)

Shukla et al. (2020) misinterpreted environmental magnetic parameters. They wrote that ‘this section presents the sedimentation with low χlf values from ~260 CE to ~270 BCE, with the presence of antiferromagnetic minerals, except two peaks at ~170 BCE and~40 CE showing the presence of ferrimagnetic minerals’. This implies that period with low magnetic susceptibility (χlf) is dominated by antiferromagnetic minerals and peak in susceptibility at various places is characterized by ferrimagnetic minerals. χlf is a magnetic concentration dependent parameter that reflects contribution from all type of minerals (i.e. ferrimagnetic, antiferromagnetic, paramagnetic and diamagnetic minerals) in a natural sediment sample (Liu et al., 2012; Maxbauer et al., 2016; Peters and Dekkers, 2003; Thompson and Oldfield, 1986). Since, ferrimagnetic minerals such as magnetite has much higher magnetic moment than antiferromagnetic, paramagnetic and diamagnetic minerals, therefore, χlf is often interpreted to indicate the abundance of ferrimagnetic minerals (Maxbauer et al., 2016; Peters and Dekkers, 2003; Thompson and Oldfield, 1986). Further, in clay rich samples paramagnetic minerals can also dominate the magnetic susceptibility (e.g. Lanci et al., 1999; Yamazaki and Ioka, 1997). Therefore, simply assigning low susceptibility to antiferromagnetic minerals and high susceptibility to ferrimagnetic minerals in bulk samples rather than interpreting magnetic concentration variability is erroneous. Further, authors have not determined magnetic mineralogy of glacial-lake sediments either using high-temperature susceptibility (Curie temperature) or by other methods (e.g. SEM, TEM) which again disqualifies oversimplified assignment of susceptibility variations to changes in magnetic mineralogy rather than magnetic concentration.

Shukla et al. (2020) further wrote that ‘χARM follows the same trend as χlf and shows the presence of multi-domain (MD) grains along with antiferromagnetic minerals and SP grains with ferromagnetic minerals’. This implies that periods with low magnetic susceptibility dominated by antiferromagnetic mineral also corresponds to low χARM which suggests presence of MD ferrimagnetic particles and periods with peak susceptibilities has high χARM which indicates presence of SP ferrimagnetic particles. The χARM parameter is sensitive to single domain (SD) and small pseudo single domain (PSD) ferrimagnetic (magnetite) particle concentration (Banerjee et al., 1981; King et al., 1982). Therefore, low and high χARM does not inform about the presence or absence of MD ferrimagnetic particles and neither about the SP ferrimagnetic particles. The ratios of χARM/χlf and χARM/SIRM are often used to describe fine or coarse magnetic particle variations when certain conditions are met. For example, χARM/χlf ratio can only be used for magnetic grain size variations when samples are devoid of SP ferrimagnetic particles (Bloemendal et al., 1992; King et al., 1982). However, authors have not used any ratios and just randomly described χARM as a parameter for interpreting MD and SP particle variations. Therefore, interpretation of χARM in this section is completely wrong.

Authors wrote that maximum χfd% is +5.88 m³ kg⁻¹ in this section. First of all, we would like to suggest that χfd% does not have a unit and is presented in percentage. Therefore, it should be written as 5.88%. Overall on looking at Shukla et al. (2020) Figure 8, it can be seen that almost ~70% of samples have χfd% values in negative and only four to five samples show values above 3%. The χfd% is commonly used to measure the qualitative concentration of ultrafine (<0.03 μm) SD ferrimagnetic particles (Dearing et al., 1996). These ultrafine SD ferrimagnetic particles have strong but unstable magnetization due to thermal energies counteracting induced magnetization very quickly upon removal of applied field, a behavior similar to paramagnetism (Dearing, 1999b). However, such ultrafine SD ferrimagnetic particles have very high susceptibilities and are known as superparamagnetic (SP) particles (Dearing, 1999b; Thompson and Oldfield, 1986; Worm, 1998). In low frequency susceptibility measurement, SP particles close to SP and stable SD (SSD) boundary contribute fully to susceptibility, whereas in high frequency susceptibility measurement SP state shifts to SSD state (Dearing, 1999b; Liu et al., 2012; Thompson and Oldfield, 1986). This shift from SP state to SSD state causes a sharp decrease in susceptibility at high frequency (Dearing, 1999b; Liu et al., 2012; Thompson and Oldfield, 1986). Therefore, χfd% does not have a negative value and in natural sample ranges from 0% to ~14% (Dearing, 1999b; Dearing et al., 1996). The χfd% <2% indicate virtually no presence of SP ferrimagnetic particles, whereas ⩽5% indicate dominance of non-SP grains (Dearing, 1999b; Dearing et al., 1996). The χlf of Chorabari glacial-lake sediments is below 3 × 10⁻⁸ m³ kg⁻¹. The χlf values are low which implies that high frequency susceptibility measurements will be prone to very large errors. That’s why most of samples show negative χfd%. The high χfd% (~5%) in only two samples also seems to be unreal as high difference in χfd% can also be found for weak samples or samples with high amount of diamagnetic minerals and/or organic matter (Dearing, 1999b; Rawat et al., 2015b). The χfd% for most of the samples is erroneous and does not indicate presence of SP ferrimagnetic particles. Therefore, χfd% related interpretation is also unreliable in entire article.

In Shukla et al. (2020) Figure 8, the unit of χlf should be 10⁻⁸ m³ kg⁻¹. The χfd% does not have a unit and should be presented in %. The S-ratio is dimensionless and therefore does not have a unit. The χARM has a unit of 10⁻⁵ m³ kg⁻¹.

Shukla et al. (2020) reported that ‘sudden decrease in χlf and χARM values and opposite and positive trend in S-ratio between –150 BC and 40 CE indicate high input of detrital sediments under reducing environmental conditions’. The decreased/low χlf and χARM simply suggests that fine SD ferrimagnetic particle concentration has decreased, whereas increased S-ratio (values toward 0) indicate relatively increased contribution of antiferromagnetic minerals (i.e. hematite and goethite). However, authors interpreted these signals to show increase in detrital material to lake under reducing environment. This implies that antiferromagnetic minerals are the detrital magnetic minerals. Authors reported that catchment rocks in this region are metamorphosed banded calc-silicate gneiss, calc-schist, biotite psammitic gneiss, pegmatite, granite, porphyroblastic augen gneiss. The antiferromagnetic minerals (hematite and/or goethite) are not the primary constituents of these rocks. Antiferromagnetic minerals such as hematite and goethite are mostly secondary in nature and forms during alteration and pedogenesis of parent rocks. Authors interpreted that this interval corresponds to cold-dry climate with very low chemical weathering. It is therefore quite unlikely to have a mature soil/sediment rich in antiferromagnetic minerals in the catchment during this interval. The hematite rich soils develop mostly under warm-humid tropical environment where chemical weathering of rocks has reached to a maximum level (e.g. laterites). Further, authors did not provide any evidence of extensive pedogenesis of catchment rocks. Authors also inferred reducing environmental condition during this period and therefore it is implausible to have high concentration of oxic-minerals such as hematite during this interval.

As we mentioned above in the materials and methods section, it seems more appropriate interpretation here that fine SD ferrimagnetic particles might have dissolved in a lake anoxic environment during this period leading to low χlf and χARM, whereas enhanced signal of antiferromagnetic minerals unaffected by such dissolution have been recorded in S-ratio. However, considering very high sedimentation rate during this period, dissolution of magnetite at this scale seems to be unlikely (Figure 1). Authors have not provided enough data on the magnetic properties and organic geochemistry of lakes to determine actual cause of low concentration of ferrimagnetic minerals (Table 1).

Table 1.

Summary of proxy variations, interpretations and climate inferred by Shukla et al. (2020) and our observations.

Sections	Proxy variations reported by Shukla et al. (2020)	Proxy interpretations by Shukla et al. (2020)	Climate reconstructed by Shukla et al. (2020)	Our observations
Section I (~270 BCE–260 CE)	The low χlf and χARM in this sections showed presence of antiferromagnetic minerals and MD grains. The two peaks in χlf and χARM at ~170 BCE and ~40 CE indicated presence of ferrimagnetic mineral and SP grains. Deceased χlf and χARM and high S-ratio between −150 BC to 40 CE. Magnetic enhancement at 40 CE. Decreased χlf and χARM values from 50 to 250 CE Moderate to low CIA between 0 and 200 CE.	Higher detrital sediment input under reducing environment. Cold-dry climate. Absence of magnetic enhancement, higher detrital titanomagnetite concentration produced in-situ by low-temperature oxidation during a wet and cold climate condition. Low chemical weathering and transition from arid/warmer climate to cold and dry climate.	Overall climate was cool and dry during this period.	The χfd% for most samples are near zero and/or in negative indicating absence of SP ferrimagnetic particles. The negative χfd% is due to measurement errors as samples are too weak to be measured in two frequencies. The relatively low χlf and χARM between ~270 BCE and 260 CE indicated low concentration of magnetic minerals. The S-ratio is below (<−0.5) indicate relative abundance of antiferromagnetic minerals. The low magnetic mineral concentration could possibly due to several reasons, that is, (1) Lower supply of detrital ferrimagnetic material, (2) dissolution of fine ferrimagnetic particles, and (3) dilution from high organic matter. The dissolution of fine ferrimagnetic particles improves signal of antiferromagnetic minerals (less affected by such dissolutions). Further, S-ratio is also affected by magnetite grain sizes (Frank and Nowaczyk, 2008). The sedimentation in this section is also very high ~1.1 cm/yr. The authors have not provide enough magnetic data and total organic carbon data to evaluate exact reason for low magnetic concentration in this section.
Section II (~900–1260 CE)	Increase in χlf and χARM, SSD grains and silt fraction at ~935 CE. Low χlf and χARM and increased MD particles, high sand concentration, high organic matter at ~990 to 1000 CE. The gradual increase in χlf and χARM and silt fraction from ~1000 to 1200 CE Moderate to high CIA.	Abrupt shift from drier to wetter climate. Wet and Cold Phase. Warm and humid climate. Increased chemical weathering under warm and humid climate.	Overall warm and wet climate corresponding to Medieval Climate Anomaly.	The 2D smoothen data of χlf and CIA do not show any significant changes compared to period from 260 to 900 CE. Sand and silt fraction show minor variations with increased silt fraction and decreased sand fraction. Overall no gradual or abrupt/sudden change can be seen in the multi-proxy data of this section. Sedimentation in this unit is ~10 times lower than Section I.
Section III (~1370–1720 CE)	Gradual increase and decrease in χlf during this interval indicate presence of both ferrimagnetic and antiferromagnetic minerals. χARM follows trend of χlf indicating presence of SSD, SP and MD grains. S-ratio ranges from −1.08 to 0.84. χfd% is negative in this zone (except for one sample). Absence of magnetic enhancement in this section. Low CIA.	Magnetite mineral deposition in oxidizing environment. Warm and humid Phase. Wet climate and low temperature oxidation. Low weathering under cold climate.	Climate similar to Little Ice Age with small phases of cold and wet climate.	Increases and decreases in susceptibility of bulk samples does not inform about mineralogy but rather magnetic concentrations. χARM is a proxy for SD and small PSD ferrimagnetic (magnetite) particles and does not inform about the SP and MD grains. S-ratio above 1 is technically wrong possibly due to measurement error, that is, improper field inducing or sample orientation problem while demagnetizing. Further, calculation error is also a possibility. Negative χfd% is due to measurement error. No significant presence of SP ferrimagnetic particles based on χfd% is found, indicating invalid climatic interpretation. The 2D smoothen line for χlf and χARM data showed minor decrease in values indicating reduced influx of magnetic minerals due to lower erosion either in response to weak precipitation or decreased sub glacial melt supply. CIA variations are not distinguishable from section II.

Authors also stated that ‘A gradual increase at ~40 CE shows an increase in the magnetic enhancement during cold and dry periods, and then a sudden decrease is observed in the data which continues from ~50 to 250 CE, suggesting that χlf also decreased due to the absence of magnetic enhancement, increasing detrital titanomagnetite concentration, produced in situ by low-temperature oxidation during a wet and cold climate condition’. Authors reported here that decrease in χlf from ~50 to 250 CE is due to absence of magnetic enhancement and concluded such decrease in χlf to increasing detrital titanomagnetite concentration, produced in-situ by low-temperature oxidation during a wet and cold climate condition. However, there are several erroneous interpretation in this statement. First of all, titanomagnetite is a ferrimagnetic mineral and has very high χlf values (Peters and Dekkers, 2003). Therefore, if concentration of titanomagnetite has increased then χlf will also increase. Authors confusingly wrote that increased detrital titanomagnetite concentration has been produced in-situ by low-temperature oxidation during a wet and cold climate condition. If titanomagnetite is detrital then how it could have been produced in-situ by low temperature oxidation? The process of in-situ production of a mineral in lake is called as authigenesis. Titanomagnetite is a primary rock-forming mineral and its secondary origin, if any, may be very limited in such lacustrine environment (unknown to us). Authors also did not provide any evidence how they have identified titanomagnetite in the lake sediment? They did not carry out any mineralogical analysis or Curie temperature measurement for identification of titanomagnetite. Further, they have not even given any reference to explain the low temperature oxidation and genesis of secondary titanomagnetite. What was the primary/precursor mineral which has been oxidized into titanomagnetite?

In the geochemical result plot (Authors’ Figure 9), authors wrote at some places data in oxide mole and at some places in elemental mole. For example, they wrote SiO₂ mole and Ti mole not TiO₂ mole. It is unclear whether it is TiO₂ oxide mole or Ti element mole? Further, they have plotted both Al₂O₃ mole and Al mole in Authors’ Figure 9. Why both Al₂O₃ mole and Al mole were plotted? Are both having different significance in interpretation? Why not give the uniform presentation of major oxides? They must have also presented ratios in a uniform way, for example, Na₂O/Al₂O₃ not Na/Al. This non-uniform representation of data are not ideal for readers understanding.

The CIA in whole lake profile ranges from ~55 to 65 except for the samples from ~50 to 150 CE where CIA ranges between ~50 and 55 (Authors’ Figure 9). It has to be noted that rock types authors have described for the region (e.g. calc-silicate gneiss, calc-schist, biotite psammitic gneiss, pegmatite, granite, porphyroblastic augen gneiss) have CIA values from ~50 to 60, for example, CIA = ~50 for granite (Fedo et al., 1995). Therefore, it will be very unscientific to categorize the CIA values of lake sediments into low, moderate and high. The 2D smoothen line for CIA data lies at ⩽60 for all the sediments deposited in to lake indicating incipient weathering (Authors’ Figure 9; Fedo et al., 1995).

Section II (~900–1260 CE)

Shukla et al. (2020) wrote that ‘magnetic data of this section indicate a wet and cold phase during ~1000 CE when concentration of MD grains along with sand-mixed clay deposits increased’. It again becomes questionable what parameter the authors had used to infer the concentration of MD magnetic particles?

Authors wrote that ‘susceptibility values initially increased to 2.03 × 10⁻⁸ m³ kg⁻¹ at ~935 CE, showing the presence of ferromagnetic minerals which suddenly decreased to minimum level at ~1015 CE. This sudden increase in susceptibility is coupled with an increase in silt-sized particles, indicating a sudden shift from drier to wetter conditions. Also, the sudden decrease in susceptibility values at ~990 CE could have resulted due to an increased organic matter mixed with sediment and water’. Here, authors again linked increase in susceptibility to mineralogy that is, presence of ferrimagnetic minerals and not to the concentration. However, it is clear from the S-ratio (close to minus 1) that magnetic assemblage is dominated by ferrimagnetic minerals from 250 to 2000 CE. Authors suggested that increase in magnetic susceptibility coupled with an increase in silt-sized particles at ~935 CE indicate shift from drier to wetter conditions. This indicates that higher detrital material has been supplied to the lake due to more erosion in response to strengthen precipitation. In previous Section I, authors have reported input of higher detrital material during period of low susceptibility. Therefore, it seems contrasting as high detrital input in Section I was reported during period of low susceptibility, whereas in Section II during period of high susceptibility. Further, authors reported that sudden decrease in susceptibility values at ~990 CE could have resulted due to an increased organic matter mixed with sediment and water. Authors have not analyzed the total organic carbon. Therefore, how they have claimed that the decreased susceptibility is due to an increase in organic matter but not due to overall reduced concentration of magnetic minerals. Did they calculate organic matter from LOI? No information is provided how they have calculated LOI. Whether they measured LOI on 950°C or above or at 550°C? They also assigned the low susceptibility to water content. Had they analyzed wet samples for magnetic property? For magnetic measurements, standard protocol is to air-dry the samples before measurement.

Authors in this section linked high χARM values to SSD and low χARM values to MD ferrimagnetic minerals. As we have stated above, χARM is not a proxy for MD minerals. Authors here gave S-ratio values in units of 10⁻⁵ m² kg⁻¹. S-ratio is a dimensionless parameter and therefore does not have a unit.

Authors reported that ‘occurrence of very small magnetic grains of stable single domain (SSD) boundary gives rise to χfd%; however, the χfd% values are low throughout the section. A smaller fluctuation in χfd% values until ~1077 CE and then a gradual increase are observed towards positive values’. The SP ferrimagnetic particles close to SSD boundary gives high χfd%, not the SSD particles itself. As we stated above, χfd% in entire profile does not indicate any significant presence of SP particles and negative χfd% for almost 70% of the samples are due to measurement error. Therefore, χfd% related interpretation is not real.

A major concern arises that authors in this section mostly described variations in magnetic properties based on only one or two samples and interpreted these variations for climatic change. Authors wrote that ‘magnetic data of this section indicate a wet and cold phase during ~1000 CE, and during this time, the concentration of MD grains along with sand-mixed clay deposits increased and since then it gradually changed towards warm and humid phase of climate until ~1200 CE’. This interpretation is very confusing. Authors stated that during a wet and cold phase, concentration of MD grains (which they have not identified and have wrongly assumed based on low χARM) along with sand-mixed clay deposits have increased and then climate gradually changed towards warm and humid phase until ~1200 CE as magnetic susceptibility has increased. The variations in magnetic concentration in lakes is often linked with the supply of detrital magnetic minerals from catchment in response to the strength of precipitation/runoff (Dearing et al., 1981; Oldfield, 1991; Oldfield et al., 2003; Rawat et al., 2015b). From author’s interpretation of cold-wet and warm-humid climate, it seems that precipitation conditions were similar in both climate phases and major change came in terms of temperature that is, cold and warm conditions. Thus, it becomes very unclear on what basis or based on which environmental magnetic parameters authors have described cold and warm climate?

Authors reported that CIA is moderate to high during this period. The CIA values range between ~58 and 62 for this period and falls in category of incipient weathering, not in moderate or high weathering. Further, Na₂O also did not show much variation (Authors’ Figure 9). Overall, the interpretation of proxy data for this period has many shortcomings. It is again questionable that why authors did not interpret the proxy data for climate variations from 260 to 900 CE? Looking into 2D smoothen magnetic (χlf and χARM) and geochemical data from 260 to 900 CE and comparing with data from 900 to 1260 CE, any sharp/distinguishing changes cannot be seen as authors have suggested for selecting periods of interest (Authors’ Figures 8 and 9). All the data including χlf and CIA show almost straight 2D smoothened line from 260 to 1260 CE (Authors’ Figures 8 and 9). It is unclear that why authors did not interpret the proxy data from 260 to 900 CE? This shows selective nature of author’s interest to only highlight the periods of global climate change (e.g. MCA and LIA) which has already been well established in the Himalayan regions (e.g. Kathayat et al., 2017; Rawat et al., 2015a; Rowan, 2017).

Section III (~1370–1720 CE)

The interpretations of magnetic data in this section are also erroneous. Authors described alternating low and relatively high susceptibility values to presence of antiferromagnetic minerals and ferrimagnetic minerals, respectively. In whole article, authors have assigned magnetic susceptibility variations to changes in mineralogy and not to concentration. As we stated above, magnetic susceptibility is a concentration dependent parameter and therefore oversimplified assignment of magnetic susceptibility variations to mineralogy is altogether erroneous. Authors further made a statement that ‘χARM shows similar trend to χlf and shows the presence of SSD and super-paramagnetic (SP) and MD grains which are dependent directly on size and complex spin patterns of electrons and indirectly on time and temperature’. Here, authors interpreted χARM to indicate the presence of all magnetic domain sizes that is, SD, SP, and MD. However, χARM is only sensitive to the SD and small PSD ferrimagnetic (magnetite) particles and does not give information about SP and MD ferrimagnetic particles. Authors made this sentence more complex as they wrote that domains are dependent on size and complex spin patterns of electrons and indirectly on time and temperature, but have not provided any reference. Authors wrote ‘χfd% in this section shows more negative values ranging from ~−2 to −6 m³ kg⁻¹ except a positive value of +5.12 m³ kg⁻¹ depicts warm and humid phase of climate’. χfd% does not have a unit. As explained above, negative χfd% is possibly due to measurement errors. Further, only one sample has shown high positive χfd% (~5.12) in this section. The χfd% ⩽5 suggests dominance of non-SP grains. Therefore, how authors have come to a conclusion that χfd% (=5.12) for one sample in this section indicate warm-humid climate? No reasoning was given. Authors wrote that ‘magnetic mineral concentration was initially low during ~1470 CE and then gradually increased at ~1500 CE. The sudden decrease in the MD and single domain (SD) grains is also supported by sediment grain size data’. Did authors establish any relation between variations in grain size (e.g. sand, silt, or clay) and magnetic particle sizes? We did not find any such description in the whole article. Authors finally concluded magnetic properties of this section as ‘absence of magnetic enhancement during the wet climate and low-temperature oxidation’. This interpretation is also flawed in view of above discussion. We observed that 2D smoothen data of χlf and χARM showed a small decrease during this period which suggests low supply of magnetic minerals to the lake in response to either weakened precipitation strength under a drier climate and/or decreased sub-glacial melt supply under cold climate (Table 1).

Authors reported that CIA is low during this period. The CIA values range from ~55 to 65 during this period, which falls under the incipient weathering category as stated by authors. However, it becomes trickier when CIA values or trend of this section is compared with section II, where authors found moderate to high CIA. There is no significant difference in CIA values and also 2D smoothen line of CIA data show almost straight line from Section II to III (Authors’ Figure 9). Authors also reported low organic matter during this period. However, they have not provided any data for the same.

Shukla et al. (2020) compared the Chorabari lake climate record with regional glaciations of the central Himalaya. (I) Authors recorded cooler climate with weak monsoon precipitation during ~270 BCE to 213 CE interval in Chorabari glacial lake. They reported that majority of glaciations records during this period are from precipitation deprived regions for example, Kosa and Upper Dhauliganga valleys. (II) Authors have recorded warm and humid climatic conditions during Medieval Climate Anomaly between ~900 and 1260 CE. Authors reported that glaciations records during this period are also mainly from precipitation deprived regions of Kosa and Upper Dhauliganga valleys. Here, authors explained that monsoon enhancement in the central Himalayan region during this period has produced favorable environment for glaciations in precipitation-deprived regions. However, these observations seem to be self-contradictory when compared with the glaciations during ~270 BCE to 213 CE (period with weak monsoonal strength). From author’s hypothesis it appears that glaciations record from the precipitation deprived regions indicate independency on climate changes. Therefore, author’s hypothesis requires more explanation as temperature and precipitation conditions in the central Himalaya were quite different for glaciations in these two respective periods.

We think that a need to re-look into data in the wake of these comments is required along with proper justification for readers interest otherwise it’s just an article that gives no reasoning and provides unsubstantiated claims about climate change.

Footnotes

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Priyeshu Srivastava acknowledges funding from Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) Post-Doctoral grant 2019/11364-0.

ORCID iD

Priyeshu Srivastava

References

Banerjee

King

Marvin

(1981) A rapid method for magnetic granulometry with applications to environmental studies. Geophysical Research Letters 8(4): 333–336.

Bloemendal

King

Hall

, et al. (1992) Rock magnetism of Late Neogene and Pleistocene deep-sea sediments: Relationship to sediment source, diagenetic processes, and sediment lithology. Journal of Geophysical Research: Solid Earth 97(B4): 4361–4375.

Dearing

(1999a) Holocene environmental change from magnetic proxies in lake sediments. In: Maher

Thompson

(eds) Quaternary Climates, Environments and Magnetism. Cambridge: Cambridge University Press, pp. 231–278.

Dearing

(1999b) Environmental Magnetic Susceptibility Using the Bartington MS2 System. 2nd ed. Kenilworth: Chinese Publications.

Dearing

Dann

RJL

Hay

, et al. (1996) Frequency-dependent susceptibility measurements of environmental materials. Geophysical Journal International 124(1): 228–240.

Dearing

Elner

Happey-Wood

(1981) Recent sediment flux and erosional processes in a Welsh upland lake-catchment based on magnetic susceptibility measurements. Quaternary Research 16(3): 356–372.

Demske

Tarasov

Wünnemann

, et al. (2009) Late glacial and Holocene vegetation, Indian monsoon and westerly circulation in the Trans-Himalaya recorded in the lacustrine pollen sequence from Tso Kar, Ladakh, NW India. Palaeogeography, Palaeoclimatology, Palaeoecology 279(3–4): 172–185.

Dunlop

Özdemir

(1997) Rock Magnetism: Fundamentals and Frontiers. Cambridge: Cambridge University Press.

Evans

Heller

(2003) Environmental Magnetism: Principles and Applications of Enviromagnetics. San Diego, CA: Academic Press.

10.

Fedo

Nesbitt

Young

(1995) Unraveling the effects of potassium metasomatism in sedimentary rocks and paleosols, with implications for paleoweathering conditions and provenance. Geology 23(10): 921–924.

11.

Frank

Nowaczyk

(2008) Mineral magnetic properties of artificial samples systematically mixed from haematite and magnetite. Geophysical Journal International 175(2): 449–461.

12.

Gupta

Anderson

Overpeck

(2003) Abrupt changes in the Asian southwest monsoon during the Holocene and their links to the North Atlantic Ocean. Nature 421(6921): 354–357.

13.

Gupta

Das

Anderson

(2005) Solar influence on the Indian summer monsoon during the Holocene. Geophysical Research Letters 32(17): 1–4.

14.

Kathayat

Cheng

Sinha

, et al. (2016) Indian monsoon variability on millennial-orbital timescales. Scientific Reports 6(1): 24374.

15.

Kathayat

Cheng

Sinha

, et al. (2017) The Indian monsoon variability and civilization changes in the Indian subcontinent. Science Advances 3(12): e1701296.

16.

Kim

Kodama

Moeller

(2005) Bacterial magnetite produced in water column dominates lake sediment mineral magnetism: Lake Ely, USA. Geophysical Journal International 163(1): 26–37.

17.

King

Banerjee

Marvin

, et al. (1982) A comparison of different magnetic methods for determining the relative grain size of magnetite in natural materials: Some results from lake sediments. Earth and Planetary Science Letters 59(2): 404–419.

18.

King

Channell

JET

(1991) Sedimentary magnetism, environmental magnetism and magnetostratigraphy. Reviews of Geophysics 29(S1): 358–370.

19.

Kotlia

Sanwal

Phartiyal

, et al. (2010) Late Quaternary climatic changes in the eastern Kumaun Himalaya, India, as deduced from multi-proxy studies. Quaternary International 213(1–2): 44–55.

20.

Lanci

Hirt

Lowrie

, et al. (1999) Mineral-magnetic record of Late Quaternary climatic changes in a high Alpine lake. Earth and Planetary Science Letters 170(1–2): 49–59.

21.

Leipe

Demske

Tarasov

(2014) A Holocene pollen record from the northwestern Himalayan lake Tso Moriri: Implications for palaeoclimatic and archaeological research. Quaternary International 348: 93–112.

22.

Liu

Roberts

Larrasoaña

, et al. (2012) Environmental magnetism: Principles and applications. Reviews of Geophysics 50(4): RG4002.

23.

Liu

Roberts

Torrent

, et al. (2007) What do the HIRM and S-ratio really measure in environmental magnetism? Geochemistry, Geophysics, Geosystems 8(9): Q09011.

24.

Maxbauer

Feinberg

Fox

(2016) Magnetic mineral assemblages in soils and paleosols as the basis for paleoprecipitation proxies: A review of magnetic methods and challenges. Earth-Science Reviews 155: 28–48.

25.

Nesbitt

Young

(1982) Early Proterozoic climates and plate motions inferred from major element chemistry of lutites. Nature 299(5885): 715–717.

26.

Oldfield

(1991) Environmental magnetism: A personal perspective. Quaternary Science Reviews 10(1): 73–85.

27.

Oldfield

Wake

Boyle

, et al. (2003) The late-Holocene history of Gormire Lake (NE England) and its catchment: A multiproxy reconstruction of past human impact. The Holocene 13(5): 677–690.

28.

Paasche

Løvlie

Dahl

, et al. (2004) Bacterial magnetite in lake sediments: Late glacial to Holocene climate and sedimentary changes in northern Norway. Earth and Planetary Science Letters 223(3–4): 319–333.

29.

Peters

Dekkers

(2003) Selected room temperature magnetic parameters as a function of mineralogy, concentration and grain size. Physics and Chemistry of the Earth, Parts A/B/C 28(16–19): 659–667.

30.

Purohit

Sanyal

Roy

, et al. (2010) ¹³C enrichment in the Palaeoproterozoic carbonate rocks of the Aravalli Supergroup, northwest India: Influence of depositional environment. Gondwana Research 18(2–3): 538–546.

31.

Rawat

Gupta

Sangode

, et al. (2015a) Late Pleistocene–Holocene vegetation and Indian summer monsoon record from the Lahaul, northwest Himalaya, India. Quaternary Science Reviews 114: 167–181.

32.

Rawat

Gupta

Srivastava

, et al. (2015b) A 13,000 year record of environmental magnetic variations in the lake and peat deposits from the Chandra valley, Lahaul: Implications to Holocene monsoonal variability in the NW Himalaya. Palaeogeography, Palaeoclimatology, Palaeoecology 440: 116–127.

33.

Rowan

(2017) The ‘Little Ice Age’ in the Himalaya: A review of glacier advance driven by Northern Hemisphere temperature change. The Holocene 27(2): 292–308.

34.

Saini

Mukherjee

Rathi

, et al. (2002) Trace element estimation in soils: An appraisal of ED-XRF technique using group analysis scheme. Journal of Trace and Microprobe Techniques 20(4): 539–551.

35.

Shukla

Mehta

Dobhal

, et al. (2020) Late-Holocene climate response and glacial fluctuations revealed by the sediment record of the monsoon-dominated Chorabari Lake, Central Himalaya. The Holocene 30(7): 953–965.

36.

Shukla

Mehta

Jaiswal

, et al. (2018) Late Quaternary glaciations history of monsoon-dominated Dingad basin, central Himalaya, India. Quaternary Science Reviews 181: 43–64.

37.

Snowball

(1991) Magnetic hysteresis properties of greigite (Fe₃S₄) and a new occurrence in Holocene sediments from Swedish Lappland. Physics of the Earth and Planetary Interiors 68(1–2): 32–40.

38.

Snowball

(1993) Geochemical control of magnetite dissolution in subarctic lake sediments and the implications for environmental magnetism. Journal of Quaternary Science 8(4): 339–346.

39.

Snowball

(1994) Bacterial magnetite and the magnetic properties of sediments in a Swedish lake. Earth and Planetary Science Letters 126(1–3): 129–142.

40.

Stober

Thompson

(1979a) An investigation into the source of magnetic minerals in some Finnish lake sediments. Earth and Planetary Science Letters 45(2): 464–474.

41.

Stober

Thompson

(1979b) Magnetic remanence acquisition in Finnish lake sediments. Geophysical Journal of the Royal Astronomical Society 57(3): 727–739.

42.

Stockhausen

Thouveny

(1999) Rock-magnetic properties of Eemian maar lake sediments from Massif Central, France: A climatic signature? Earth and Planetary Science Letters 173(3): 299–313.

43.

Tamuntuan

Bijaksana

King

, et al. (2015) Variation of magnetic properties in sediments from Lake Towuti, Indonesia, and its paleoclimatic significance. Palaeogeography, Palaeoclimatology, Palaeoecology 420: 163–172.

44.

Thompson

Oldfield

(1986) Environmental Magnetism. London: Allen and Unwin Press.

45.

Wang

Song

Cheng

, et al. (2016) Mineral magnetism and other characteristics of sediments from a sub-alpine lake (3080 m a.s.l.) in central east China and their implications on environmental changes for the last 5770 years. Earth and Planetary Science Letters 452: 44–59.

46.

Worm

(1998) On the superparamagnetic-stable single domain transition for magnetite and frequency dependence of susceptibility. Geophysical Journal International 133(1): 201–206.

47.

Wünnemann

Demske

Tarasov

, et al. (2010) Hydrological evolution during the last 15 kyr in the Tso Kar lake basin (Ladakh, India), derived from geomorphological, sedimentological and palynological records. Quaternary Science Reviews 29(9–10): 1138–1155.

48.

Yamazaki

Ioka

(1997) Environmental rock-magnetism of pelagic clay: Implications for Asian eolian input to the North Pacific since the Pliocene. Paleoceanography 12(1): 111–124.