the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 23 Sep 2024
The visible and hidden climatic effects on Earth's denudation
Abstract. Denudation is the opposite process of mountain uplift and plays a major role in the Earth system. Despite the research to constrain its environmental control, uncertainties remain about which are the dominant physicochemical processes at play. Here, the 10Be-derived denudation rate, encompassing time windows from 102 to 105 yr, was modelled in over a thousand basins across the Earth. The results suggest that water and associated life have a positive effect across their whole range, which is regulated by topography due to processes such as the energy expended by rivers on their beds, the feedback between erosion and weathering, and the transport and production rate of soils. Consequently, bioclimatic influence is weak in flat landscapes, but it could vary denudation forty times in mountain settings. It was also observed that other things being equal, water availability steepens basins, so climate also has an indirect effect acting on geological timeframes. The results can be useful for the landscape's numerical modelling and highlight the importance of climate on denudation.
- Preprint
(1052 KB) - Metadata XML
-
Supplement
(3359 KB) - BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on esd-2024-27', Anonymous Referee #1, 27 Nov 2024
This paper by Vergara et al. presents empirical models that attempt to quantify the controls on millennial-scale denudation rates using 47 candidate controlling factors.
I will focus my review primarily on the climatic controls on denudation reported in this study, because the title of the paper emphasizes these controls. Previous studies have shown that weathering rates are positively correlated with water availability throughout the full range (hyperarid to humid). Erosion by overland flow on hillslopes and by confined flow in channels has generally been shown to have a negative effect on transport rates (i.e., less vegetation cover, all else being equal, tends to result in higher transport rates, e.g., Acosta-Torres et al., 2014). Since denudation is controlled by the combination of weathering and transport, I would have expected the positive correlation of weathering with water availability and the negative correlation of transport with increasing vegetation cover to result in a complex (i.e., non-monotonic) relationship between denudation rates and water availability similar to the “humped” Langbein-Schumm (1958) relationship for short-term erosion rates and the similarly humped relationship for millennial-scale denudation rates documented by Schaller and Ehlers, 2022, https://doi.org/10.5194/esurf-10-131-2022, which Vergara et al. do not reference. I welcome any study that seeks to tease out the climatic controls on denudation. But after reading this study, I did not come away with any enhanced understanding of how climate effects denudation, nor whether this study is consistent or not with the papers referenced above.
Major concerns:
1) It is clear from the presentation that the variable “clim” has a positive correlation with denudation rates. This fact may be the basis of the statement “water and associated life have a positive effect across their whole range” (line 19). What I cannot tell from the information provided is how “clim” depends on the bioclimatic variables input to it. “clim” is described as the first principal component of an analysis that includes 10+ inputs (described on lines 264-282). Lines 264-282 list 13 input datasets used to create “clim”, but lines 278-282 state that 3 of these 13 datasets (i.e., NDVI, AI and paleoprecipitation) were used to derive a first principal component that was used as input. As such, the variable “clim” seems to be the first principal component of at least 13 datasets, including at least one variable that is also a principal component of multiple other datasets. Please quantify how sensitive “clim” is to each input variable. How does “clim” correlate (positively or negatively) with each input variable? Without this information, it is impossible to even begin to figure out what can be learned from this analysis or to comprehensively review this preprint. I acknowledge that some of this information may be present in one of the 16 supplementary tables. If so, that information still has to be at least summarized for the reader in the main text.
2) Whether an empirical model is superior to another is not only a function of its goodness of fit. Models with more degrees of freedom will tend to fit any dataset better than one with fewer. Akaike’s Information Criterion (Akaike, 1974) was developed to address this issue. Before the authors claim that their model is superior to others in the literature, they must report a metric that includes both the goodness of fit and the number of degrees of freedom for both their model and the alternative models. The number of degrees of freedom reflects the number of coefficients in the multivariate regression that are varied to fit the data, and also the degrees of freedom associated with the PCA used to construct the “clim” variable. I noticed that the authors reported AIC values in their Supplementary Tables, which is laudable, but how they determined the number of degrees of freedom is unstated. There are no AIC values reported for any alternative model. As such, the authors should avoid stating that their model is superior to others in the literature based only on goodness of fit.
3) I did not see any support in the paper for the conclusion that “other things being equal, water availability steepens basins” (line 24). It’s possible that this text is referring to the fact that, in equation (3), slope varies with precipitation to the 0.2 power. However, one cannot conclude from that relationship that water availability causes basins to steepen. It could just a likely be the case that steeper basins have more orographic precipitation and water availability has no causal effect on basin steepness.
Small issues:
1) Some of the wording needs to be improved throughout. For example: line 212: “To use only reliable average denudation rates, were discarded measurements on sediment higher than 1mm and in basins smaller than 100km2 or that have lakes area plus their upstream area larger than 25% of the basin surface.” I think the word “were” should be “we”. I don’t know what “sediment higher than 1 mm” means. Does it mean sediment with mean diameters larger than 1 mm? There are many other examples of awkward phrasing and missing words. Another example: “This erodibility index has a well relationship with uniaxial compressive strength at a regional dimension.” I think “well” is supposed to be “well-defined” in this sentence.
2) Please provide an equation for equations (1)-(3) (not just a proportionality). Also, I don’t know what some of the variables are because many are undefined. For example, in equation (1), I can guess that “den” is denudation rate and “lit” is a lithologic hardness index, etc. I searched for “PGA” and saw that it was introduced on line 285, but please make it easier for the reader by defining variables (with units) as they are introduced. I tried to figure out what some of the variables represent by going to the supplement, but that often made me more confused. For example, in trying to figure out the variables in equation (3), I went to Table S6. But some of the variable names differ between equation 3 and Table S6. For example, the lithologic hardness index seems to be referred to in Table S6 as “soft”. And I couldn’t figure out which variable related to precipitation in Table S6 (is it “int_daily”)?
In summary, I find this analysis intriguing, but I could not, in a reasonable amount of time, determine from the information provided how the different aspects of climate (including the various aspects of precipitation (mean, seasonality, extremes) and vegetation cover) influence millennial-scale denudation rates.
Citation: https://doi.org/10.5194/esd-2024-27-RC1 - AC1: 'Reply on RC1', Ivan Vergara, 02 Jan 2025
-
RC2: 'Comment on esd-2024-27', Stefan Hergarten, 22 Dec 2024
This letter presents a regression model for denudation rates based on publicly available data derived from cosmogenic nuclides. In a nutshell, my impression is that trying to compress it into a letter destroys a nice piece of work. I read the manuscript several times, permanently switching between the main text, the appendix, the supplementary text and figures and the supplementary tables. Finally, I think I understood the main ideas, but many questions remain open. To be honest, having agreed to review the manuscript was my main motivation not to give up earlier.
Written in a clear and reproducible way, this study might become a very good research paper. Since this would require a complete rewriting, I focus on a few points in the following and do not go deeper into details.
(1) The results section starts with some kind of promise that the model proposed here "would be the best physically plausible denudation prediction for a planetary scope so far." While some references are given, I did not find any serious discussion about this aspect (number of adjustable parameters, removing basins, ...). And given that the relation provided here is really the best one, what would we do with it?
(2) The parameters introduced in Eqs. (1) and (2) are not explained at their first occurrence and are partly not defined completely later. So it is practically impossible to recognize which of the covariates used here have the strongest effect on denudation. It is also not clear why the peak ground acceleration is taken into account in a different way than the other covariates.
(3) The discussion section mainly addresses the limitations. After reading it, I was left with the impression that those aspect that could really deepen our understanding cannot be addressed.
(a) In the first section, it is admitted that the correlation of vegetation with other properties does not allow for a separation of its effect. In turn, it is stated in lines 122-123 that "positive sign of vegetation's factor loading in PC1clim suggests that vegetation resulting influence is positive for the analysed temporal windows." I do not understand why this is the case. To my knowledge, the loads only refer to the variability in the climatic components and not to their relation to denudation rates.
(b) The second section discussed the contributions of rivers and hillslopes. In lines 145-146, it is stated that "The higher predictivity of variables related to hillslope suggests that for most basins the majority of denuded mass comes from there." I would agree that most of the denuded mass comes from the hillslopes, just owing to their larger area. Concerning the predictivity, however, it could also be that the properties used for characterizing rivers are more uncertain than those for slopes.
(c) It was very difficult for me to follow the rest of this section (about the slope) and I am not sure what to learn from it.
(4) Finally, I would suggest to compare the relate to those obtained by Harel et al. (2016, doi 10.1016/j.geomorph.2016.05.035). As far as I can see, these authors used basically the same data on denudation rates. As a major difference, these authors already tried to predict the parameters of the stream-power incision model for the rivers draining the respective basins. In its spirit, however, it is very similar, although the recent study is obviously more comprehensive concerning the covariates.
Overall, I think this nice work will be lost when published in its present form as a letter. So I would suggest to rewrite it and submit is as a research paper.Citation: https://doi.org/10.5194/esd-2024-27-RC2 - AC2: 'Reply on RC2', Ivan Vergara, 02 Jan 2025
Status: closed
-
RC1: 'Comment on esd-2024-27', Anonymous Referee #1, 27 Nov 2024
This paper by Vergara et al. presents empirical models that attempt to quantify the controls on millennial-scale denudation rates using 47 candidate controlling factors.
I will focus my review primarily on the climatic controls on denudation reported in this study, because the title of the paper emphasizes these controls. Previous studies have shown that weathering rates are positively correlated with water availability throughout the full range (hyperarid to humid). Erosion by overland flow on hillslopes and by confined flow in channels has generally been shown to have a negative effect on transport rates (i.e., less vegetation cover, all else being equal, tends to result in higher transport rates, e.g., Acosta-Torres et al., 2014). Since denudation is controlled by the combination of weathering and transport, I would have expected the positive correlation of weathering with water availability and the negative correlation of transport with increasing vegetation cover to result in a complex (i.e., non-monotonic) relationship between denudation rates and water availability similar to the “humped” Langbein-Schumm (1958) relationship for short-term erosion rates and the similarly humped relationship for millennial-scale denudation rates documented by Schaller and Ehlers, 2022, https://doi.org/10.5194/esurf-10-131-2022, which Vergara et al. do not reference. I welcome any study that seeks to tease out the climatic controls on denudation. But after reading this study, I did not come away with any enhanced understanding of how climate effects denudation, nor whether this study is consistent or not with the papers referenced above.
Major concerns:
1) It is clear from the presentation that the variable “clim” has a positive correlation with denudation rates. This fact may be the basis of the statement “water and associated life have a positive effect across their whole range” (line 19). What I cannot tell from the information provided is how “clim” depends on the bioclimatic variables input to it. “clim” is described as the first principal component of an analysis that includes 10+ inputs (described on lines 264-282). Lines 264-282 list 13 input datasets used to create “clim”, but lines 278-282 state that 3 of these 13 datasets (i.e., NDVI, AI and paleoprecipitation) were used to derive a first principal component that was used as input. As such, the variable “clim” seems to be the first principal component of at least 13 datasets, including at least one variable that is also a principal component of multiple other datasets. Please quantify how sensitive “clim” is to each input variable. How does “clim” correlate (positively or negatively) with each input variable? Without this information, it is impossible to even begin to figure out what can be learned from this analysis or to comprehensively review this preprint. I acknowledge that some of this information may be present in one of the 16 supplementary tables. If so, that information still has to be at least summarized for the reader in the main text.
2) Whether an empirical model is superior to another is not only a function of its goodness of fit. Models with more degrees of freedom will tend to fit any dataset better than one with fewer. Akaike’s Information Criterion (Akaike, 1974) was developed to address this issue. Before the authors claim that their model is superior to others in the literature, they must report a metric that includes both the goodness of fit and the number of degrees of freedom for both their model and the alternative models. The number of degrees of freedom reflects the number of coefficients in the multivariate regression that are varied to fit the data, and also the degrees of freedom associated with the PCA used to construct the “clim” variable. I noticed that the authors reported AIC values in their Supplementary Tables, which is laudable, but how they determined the number of degrees of freedom is unstated. There are no AIC values reported for any alternative model. As such, the authors should avoid stating that their model is superior to others in the literature based only on goodness of fit.
3) I did not see any support in the paper for the conclusion that “other things being equal, water availability steepens basins” (line 24). It’s possible that this text is referring to the fact that, in equation (3), slope varies with precipitation to the 0.2 power. However, one cannot conclude from that relationship that water availability causes basins to steepen. It could just a likely be the case that steeper basins have more orographic precipitation and water availability has no causal effect on basin steepness.
Small issues:
1) Some of the wording needs to be improved throughout. For example: line 212: “To use only reliable average denudation rates, were discarded measurements on sediment higher than 1mm and in basins smaller than 100km2 or that have lakes area plus their upstream area larger than 25% of the basin surface.” I think the word “were” should be “we”. I don’t know what “sediment higher than 1 mm” means. Does it mean sediment with mean diameters larger than 1 mm? There are many other examples of awkward phrasing and missing words. Another example: “This erodibility index has a well relationship with uniaxial compressive strength at a regional dimension.” I think “well” is supposed to be “well-defined” in this sentence.
2) Please provide an equation for equations (1)-(3) (not just a proportionality). Also, I don’t know what some of the variables are because many are undefined. For example, in equation (1), I can guess that “den” is denudation rate and “lit” is a lithologic hardness index, etc. I searched for “PGA” and saw that it was introduced on line 285, but please make it easier for the reader by defining variables (with units) as they are introduced. I tried to figure out what some of the variables represent by going to the supplement, but that often made me more confused. For example, in trying to figure out the variables in equation (3), I went to Table S6. But some of the variable names differ between equation 3 and Table S6. For example, the lithologic hardness index seems to be referred to in Table S6 as “soft”. And I couldn’t figure out which variable related to precipitation in Table S6 (is it “int_daily”)?
In summary, I find this analysis intriguing, but I could not, in a reasonable amount of time, determine from the information provided how the different aspects of climate (including the various aspects of precipitation (mean, seasonality, extremes) and vegetation cover) influence millennial-scale denudation rates.
Citation: https://doi.org/10.5194/esd-2024-27-RC1 - AC1: 'Reply on RC1', Ivan Vergara, 02 Jan 2025
-
RC2: 'Comment on esd-2024-27', Stefan Hergarten, 22 Dec 2024
This letter presents a regression model for denudation rates based on publicly available data derived from cosmogenic nuclides. In a nutshell, my impression is that trying to compress it into a letter destroys a nice piece of work. I read the manuscript several times, permanently switching between the main text, the appendix, the supplementary text and figures and the supplementary tables. Finally, I think I understood the main ideas, but many questions remain open. To be honest, having agreed to review the manuscript was my main motivation not to give up earlier.
Written in a clear and reproducible way, this study might become a very good research paper. Since this would require a complete rewriting, I focus on a few points in the following and do not go deeper into details.
(1) The results section starts with some kind of promise that the model proposed here "would be the best physically plausible denudation prediction for a planetary scope so far." While some references are given, I did not find any serious discussion about this aspect (number of adjustable parameters, removing basins, ...). And given that the relation provided here is really the best one, what would we do with it?
(2) The parameters introduced in Eqs. (1) and (2) are not explained at their first occurrence and are partly not defined completely later. So it is practically impossible to recognize which of the covariates used here have the strongest effect on denudation. It is also not clear why the peak ground acceleration is taken into account in a different way than the other covariates.
(3) The discussion section mainly addresses the limitations. After reading it, I was left with the impression that those aspect that could really deepen our understanding cannot be addressed.
(a) In the first section, it is admitted that the correlation of vegetation with other properties does not allow for a separation of its effect. In turn, it is stated in lines 122-123 that "positive sign of vegetation's factor loading in PC1clim suggests that vegetation resulting influence is positive for the analysed temporal windows." I do not understand why this is the case. To my knowledge, the loads only refer to the variability in the climatic components and not to their relation to denudation rates.
(b) The second section discussed the contributions of rivers and hillslopes. In lines 145-146, it is stated that "The higher predictivity of variables related to hillslope suggests that for most basins the majority of denuded mass comes from there." I would agree that most of the denuded mass comes from the hillslopes, just owing to their larger area. Concerning the predictivity, however, it could also be that the properties used for characterizing rivers are more uncertain than those for slopes.
(c) It was very difficult for me to follow the rest of this section (about the slope) and I am not sure what to learn from it.
(4) Finally, I would suggest to compare the relate to those obtained by Harel et al. (2016, doi 10.1016/j.geomorph.2016.05.035). As far as I can see, these authors used basically the same data on denudation rates. As a major difference, these authors already tried to predict the parameters of the stream-power incision model for the rivers draining the respective basins. In its spirit, however, it is very similar, although the recent study is obviously more comprehensive concerning the covariates.
Overall, I think this nice work will be lost when published in its present form as a letter. So I would suggest to rewrite it and submit is as a research paper.Citation: https://doi.org/10.5194/esd-2024-27-RC2 - AC2: 'Reply on RC2', Ivan Vergara, 02 Jan 2025
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 311 | 54 | 14 | 379 | 19 | 9 | 8 |
- HTML: 311
- PDF: 54
- XML: 14
- Total: 379
- Supplement: 19
- BibTeX: 9
- EndNote: 8
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1