the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 16 Feb 2024
Effect of surficial geology mapping scale on modelled ground ice in Canadian Shield terrain
Abstract. Ground ice maps at circumpolar or hemispherical scales offer generalised depictions of abundance across broad geographic regions. In this paper, the effect of surficial geology mapping scale on modelled ground ice abundance is examined in the Slave Geological Province of the Canadian Shield, a region where the geological and glacial legacy has produced a landscape with significant variation in surface cover. Existing model routines from the Ground ice map of Canada (GIMC) were used with a 1:125 000 scale regional surficial geology compilation and compared to the national outputs, which are based on surficial geology at 1:5 000 000 scale. Overall, the regional scale modelling predicts much more ground ice than the GIMC due to greater representation of unconsolidated sediments in the region. Improved modelling accuracy is indicated by comparison of outputs to available empirical datasets due to improved representation of the inherent regional heterogeneity in surficial geology. The results demonstrate that the GIMC significantly underestimates the abundance and distribution of ground ice over Canadian Shield terrain. In areas with limited information on ground ice, regional-scale modelling may provide useful reconnaissance-level information to help guide field-based investigations required for planning infrastructure development. The use of current small-scale ground ice mapping in risk or cost assessments relating to permafrost thaw may significantly influence accuracy of outputs in areas like the Canadian Shield where surficial materials range from bedrock to frost-susceptible deposits over relatively short distances.
-
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
(1822 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1822 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-68', Anonymous Referee #1, 05 Mar 2024
Review tc-2024-68:
The aim of this manuscript is to compare the results of a ground ice model using surficial geology input layers of greatly differing scales and to compare the modelled ground ice content to empirical datasets. The study area, located within the Slave Geological Province of the Canadian Shield has an abundance of exposed bedrock and thin till. The distribution of these materials results in the underrepresentation of minority component materials in national-scale mapping (Fulton, 1995). Because bedrock is assumed to have no ground ice content and till veneers minimal, the GIMC (O'Neill et al, 2022) underrepresents ground ice compared to the regional-scale modelling presented in this manuscript that better reflects real ground conditions due to increased surficial geology mapping resolution which allows for delineation of more minority constituent materials which are comprised of unconsolidated sediments.
General Comments:
The manuscript is well written, easy to follow and contains figures that effectively illustrate the findings. I am very happy to see the method developed by O'Neill et al (2022) applied at a regional scale as this is when the utility and efficacy of the models can begin to be tested properly and begins to have more real-world benefits.
This manuscript primarily explores the idea of the impact of the spatial scale of input data on modelling results through the comparison of a model run with inputs at 1:5 000 000 and 1: 125 000 scale. That generalizations of the landscape based on map scale will underrepresent certain elements, which may be of significance, is not a new concept. While this concept is not new, it is not commonly explored in relation to ground ice. While the GIMC presents an interesting and novel approach to modelling ground ice distribution in Canada, there are limited use cases for a 1:5 000 000 scale product. It is unclear that comparison of the regional scale modelling to a national-scale model using a surficial geology layer developed specifically to look good on a wall map (line 169) is useful in particular; however the exploration of the concept is useful in general. Only a single geological region is examined within this manuscript rather than randomized sites distributed throughout the whole GIMC. The extreme change in scale of mapping means minority materials are better represented in the RC vs the GIMC. In the Slave Province, this results in increased modelled ground ice abundance due to increased representation of unconsolidated materials. I expect that in an area where bedrock is the minority material, e.g. in Northern Yukon, the expected difference between the RC and GIMC would be the opposite, less ground ice than predicted by the GIMC based on surficial geology. In this vein, I believe it would benefit this paper to include some more general statements on how the scale of model inputs are likely to affect the results in both discussion and conclusion sections.
Throughout the paper many different terms are used to describe mapping scale and many appear to be used interchangeably, please be consistent and use only as many terms as are necessary. E.g. small, broad, regional, national, circumpolar, hemispherical, finer.
On the figures showing the regional-scale modelling there appears to be subtle rectilinear boundaries which likely represent 125 000 map boundaries (e.g. Fig 2). This suggests some inconsistency in delineation or attribution of surficial geology units between map sheets. This limitation is worthy of discussion, is it something that could have been rectified during compilation?
Specific Comments:
13: hyphenate “regional-scale”
14: can you be more precise than “greater”
15: remove “available”
32: the grainsize of the till is specified in general terms for thicker deposits but not for veneers
41: see also McKillop et al, 2019. Predictive Mapping of the Variable Response of Permafrost Terrain to Climate Change for Optimal Roadway Routing, Design, and Maintenance Forecasting in Northern Canada
49: hyphenate “national-scale”
58: remove the word “terrain”
86: Undifferentiated units are allowable in the Deblonde et al Surficial Data Model however the mapping of methods of Olthof et al are perhaps unorthodox and so their definition of “undifferentiated” may not be consistent with Deblonde et al.
90: It would also be useful to note the number of surficial material classes represented in the study area by the GIMC.
132: Specify the location within NT, Yellowknife?
149: instead of organic “terrain” use “materials” or “accumulations”
Table 1. The use of italics vs border thicknesses should be reexamined here. Perhaps move water to the top and have the sum of uncondsolidated materials (this can include organics) shown below a thicker border.
196: higher “modelled” ground ice abundance
198: suggest change “which is associated with” to “which results in”
229: re: incorporating line features. These line features do not have associated polygons because they cannot be delineated as such at the map scale. Discussion of including them is akin to saying that incorporating surficial geology data at a finer spatial scale would better represent the landscape and result in improved modelling. This is true and the main point of your paper but not made clear here.
293: “predicted” ground ice conditions
294: Slight distinction, the GIMC does not estimate the occurrence or abundance of sediments, but the ice content within the sediments.
314: Inaccuracy of the GIMC relative to RC will occur everywhere where terrain heterogeneity exceeds what can be represented a 1:5 000 000 scale. The direction of the effect will depend on the specific material present with the bias favouring whatever materials are dominant in the region.
317: The above comment applies to this line too.
320: Perhaps it would be a good idea to call attention to the need for detailed surficial geology mapping to support our understanding ground ice distribution at scales necessary for northern development.
Citation: https://doi.org/10.5194/egusphere-2024-68-RC1 -
AC1: 'Reply on RC1', Hugh O'Neill, 02 Apr 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-68/egusphere-2024-68-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Hugh O'Neill, 02 Apr 2024
-
RC2: 'Comment on egusphere-2024-68', Anonymous Referee #2, 19 Mar 2024
This paper aims at comparing ground ice abundance modelling output from different scale surficial geology products for a region of the Canadian Shield. It uses an existing modelling method used for creating the GIMC, which was proven to underestimate ground ice abundance. Difference between two surficial geology scales are presented and validated using ancillary data, and implications of accuracy of ground ice abundance modelling are discussed for the Canadian Shield and along a proposed infrastructure route.
General comments
This paper presents novel and valuable insights into permafrost ground ice abundance modelling and mapping. It highlights the importance of scale and landscape heterogeneity, which is a very relevant challenge/issue related to products such as the GIMC and IPA map that needed to be addressed. The purpose is clear and is effectively reached using adequate methods. The conclusions are well supported by the results presented.
Overall, this paper is well-written and concise, but I think a few sections could benefit from additional information (discussed below).
Specific comments
The introduction is a little bit short in introducing the subject of ground ice and ground ice modelling. First paragraph could contain more information on the influence of surficial geology on ground ice abundance; why are we using surficial geology as an input to ground ice abundance modelling? Could also refer to existing datasets and why it is important to improve them, or even why it is important to quantify ground ice abundance.
Also, at line 35, there is a jump from geology of the study area to modelling methodology. I suggest adding “used for modelling ground ice abundance” after “methodology” and before “was” in sentence “modelling methodology was developed by O’Neill et al. (2019)…”.
In the Study Area section, some locations are mentioned, but are not presented on Figure 1, which makes it harder to understand the geological and climatic context. Suggest adding locations of the Great Slave Lowlands and Lac de Gras.
In the Methods section, the different products used become obscured. I suggest reviewing and keeping a constant terminology for each product/group to avoid confusion:
- Line 84: What are these ten CGMs? Are they the RC surficial geology maps mentioned at line 80?
- Line 93: What is the meant by “at the national scale”? Is that the product used to generate the GIMC? What product is that?
- Line 90: “surficial material classes” are mentioned, but the term “units” is used at line 92-93. This occurs in other places throughout the text.
- Line 99: What is meant by the “other model”?
In the Results section, there is mention of “unconsolidated sediments and organic terrain associated with ground ice” (Line 149 and Table 1). Again, I think information on what makes certain types of surficial deposits susceptible to being more ice-rich than others is lacking. This could be addressed in the introduction, as mentioned above.
Technical corrections
I somewhat question the sectioning of the Results and Discussion sections:
- Results of the validation (Section 5.1) and infrastructure corridor assessment (Section 5.2.1) belong in the results section.
- I suggest keeping the discussion section for the implications of the results only (i.e., impact of homogeneity/heterogeneity of deposits, inclusion of linear features, model exceptions for ice-marginal deposits, limitations of wedge ice modelling based on imagery, etc.)
- These implications could also benefit from being further discussed, including within are broader context (E.g., impact of homogeneity/heterogeneity of deposits in other regions/publications, how can linear features be included in such modelling exercises, etc.)
Citation: https://doi.org/10.5194/egusphere-2024-68-RC2 -
AC2: 'Reply on RC2', Hugh O'Neill, 02 Apr 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-68/egusphere-2024-68-AC2-supplement.pdf
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-68', Anonymous Referee #1, 05 Mar 2024
Review tc-2024-68:
The aim of this manuscript is to compare the results of a ground ice model using surficial geology input layers of greatly differing scales and to compare the modelled ground ice content to empirical datasets. The study area, located within the Slave Geological Province of the Canadian Shield has an abundance of exposed bedrock and thin till. The distribution of these materials results in the underrepresentation of minority component materials in national-scale mapping (Fulton, 1995). Because bedrock is assumed to have no ground ice content and till veneers minimal, the GIMC (O'Neill et al, 2022) underrepresents ground ice compared to the regional-scale modelling presented in this manuscript that better reflects real ground conditions due to increased surficial geology mapping resolution which allows for delineation of more minority constituent materials which are comprised of unconsolidated sediments.
General Comments:
The manuscript is well written, easy to follow and contains figures that effectively illustrate the findings. I am very happy to see the method developed by O'Neill et al (2022) applied at a regional scale as this is when the utility and efficacy of the models can begin to be tested properly and begins to have more real-world benefits.
This manuscript primarily explores the idea of the impact of the spatial scale of input data on modelling results through the comparison of a model run with inputs at 1:5 000 000 and 1: 125 000 scale. That generalizations of the landscape based on map scale will underrepresent certain elements, which may be of significance, is not a new concept. While this concept is not new, it is not commonly explored in relation to ground ice. While the GIMC presents an interesting and novel approach to modelling ground ice distribution in Canada, there are limited use cases for a 1:5 000 000 scale product. It is unclear that comparison of the regional scale modelling to a national-scale model using a surficial geology layer developed specifically to look good on a wall map (line 169) is useful in particular; however the exploration of the concept is useful in general. Only a single geological region is examined within this manuscript rather than randomized sites distributed throughout the whole GIMC. The extreme change in scale of mapping means minority materials are better represented in the RC vs the GIMC. In the Slave Province, this results in increased modelled ground ice abundance due to increased representation of unconsolidated materials. I expect that in an area where bedrock is the minority material, e.g. in Northern Yukon, the expected difference between the RC and GIMC would be the opposite, less ground ice than predicted by the GIMC based on surficial geology. In this vein, I believe it would benefit this paper to include some more general statements on how the scale of model inputs are likely to affect the results in both discussion and conclusion sections.
Throughout the paper many different terms are used to describe mapping scale and many appear to be used interchangeably, please be consistent and use only as many terms as are necessary. E.g. small, broad, regional, national, circumpolar, hemispherical, finer.
On the figures showing the regional-scale modelling there appears to be subtle rectilinear boundaries which likely represent 125 000 map boundaries (e.g. Fig 2). This suggests some inconsistency in delineation or attribution of surficial geology units between map sheets. This limitation is worthy of discussion, is it something that could have been rectified during compilation?
Specific Comments:
13: hyphenate “regional-scale”
14: can you be more precise than “greater”
15: remove “available”
32: the grainsize of the till is specified in general terms for thicker deposits but not for veneers
41: see also McKillop et al, 2019. Predictive Mapping of the Variable Response of Permafrost Terrain to Climate Change for Optimal Roadway Routing, Design, and Maintenance Forecasting in Northern Canada
49: hyphenate “national-scale”
58: remove the word “terrain”
86: Undifferentiated units are allowable in the Deblonde et al Surficial Data Model however the mapping of methods of Olthof et al are perhaps unorthodox and so their definition of “undifferentiated” may not be consistent with Deblonde et al.
90: It would also be useful to note the number of surficial material classes represented in the study area by the GIMC.
132: Specify the location within NT, Yellowknife?
149: instead of organic “terrain” use “materials” or “accumulations”
Table 1. The use of italics vs border thicknesses should be reexamined here. Perhaps move water to the top and have the sum of uncondsolidated materials (this can include organics) shown below a thicker border.
196: higher “modelled” ground ice abundance
198: suggest change “which is associated with” to “which results in”
229: re: incorporating line features. These line features do not have associated polygons because they cannot be delineated as such at the map scale. Discussion of including them is akin to saying that incorporating surficial geology data at a finer spatial scale would better represent the landscape and result in improved modelling. This is true and the main point of your paper but not made clear here.
293: “predicted” ground ice conditions
294: Slight distinction, the GIMC does not estimate the occurrence or abundance of sediments, but the ice content within the sediments.
314: Inaccuracy of the GIMC relative to RC will occur everywhere where terrain heterogeneity exceeds what can be represented a 1:5 000 000 scale. The direction of the effect will depend on the specific material present with the bias favouring whatever materials are dominant in the region.
317: The above comment applies to this line too.
320: Perhaps it would be a good idea to call attention to the need for detailed surficial geology mapping to support our understanding ground ice distribution at scales necessary for northern development.
Citation: https://doi.org/10.5194/egusphere-2024-68-RC1 -
AC1: 'Reply on RC1', Hugh O'Neill, 02 Apr 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-68/egusphere-2024-68-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Hugh O'Neill, 02 Apr 2024
-
RC2: 'Comment on egusphere-2024-68', Anonymous Referee #2, 19 Mar 2024
This paper aims at comparing ground ice abundance modelling output from different scale surficial geology products for a region of the Canadian Shield. It uses an existing modelling method used for creating the GIMC, which was proven to underestimate ground ice abundance. Difference between two surficial geology scales are presented and validated using ancillary data, and implications of accuracy of ground ice abundance modelling are discussed for the Canadian Shield and along a proposed infrastructure route.
General comments
This paper presents novel and valuable insights into permafrost ground ice abundance modelling and mapping. It highlights the importance of scale and landscape heterogeneity, which is a very relevant challenge/issue related to products such as the GIMC and IPA map that needed to be addressed. The purpose is clear and is effectively reached using adequate methods. The conclusions are well supported by the results presented.
Overall, this paper is well-written and concise, but I think a few sections could benefit from additional information (discussed below).
Specific comments
The introduction is a little bit short in introducing the subject of ground ice and ground ice modelling. First paragraph could contain more information on the influence of surficial geology on ground ice abundance; why are we using surficial geology as an input to ground ice abundance modelling? Could also refer to existing datasets and why it is important to improve them, or even why it is important to quantify ground ice abundance.
Also, at line 35, there is a jump from geology of the study area to modelling methodology. I suggest adding “used for modelling ground ice abundance” after “methodology” and before “was” in sentence “modelling methodology was developed by O’Neill et al. (2019)…”.
In the Study Area section, some locations are mentioned, but are not presented on Figure 1, which makes it harder to understand the geological and climatic context. Suggest adding locations of the Great Slave Lowlands and Lac de Gras.
In the Methods section, the different products used become obscured. I suggest reviewing and keeping a constant terminology for each product/group to avoid confusion:
- Line 84: What are these ten CGMs? Are they the RC surficial geology maps mentioned at line 80?
- Line 93: What is the meant by “at the national scale”? Is that the product used to generate the GIMC? What product is that?
- Line 90: “surficial material classes” are mentioned, but the term “units” is used at line 92-93. This occurs in other places throughout the text.
- Line 99: What is meant by the “other model”?
In the Results section, there is mention of “unconsolidated sediments and organic terrain associated with ground ice” (Line 149 and Table 1). Again, I think information on what makes certain types of surficial deposits susceptible to being more ice-rich than others is lacking. This could be addressed in the introduction, as mentioned above.
Technical corrections
I somewhat question the sectioning of the Results and Discussion sections:
- Results of the validation (Section 5.1) and infrastructure corridor assessment (Section 5.2.1) belong in the results section.
- I suggest keeping the discussion section for the implications of the results only (i.e., impact of homogeneity/heterogeneity of deposits, inclusion of linear features, model exceptions for ice-marginal deposits, limitations of wedge ice modelling based on imagery, etc.)
- These implications could also benefit from being further discussed, including within are broader context (E.g., impact of homogeneity/heterogeneity of deposits in other regions/publications, how can linear features be included in such modelling exercises, etc.)
Citation: https://doi.org/10.5194/egusphere-2024-68-RC2 -
AC2: 'Reply on RC2', Hugh O'Neill, 02 Apr 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-68/egusphere-2024-68-AC2-supplement.pdf
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 405 | 89 | 26 | 520 | 15 | 16 |
- HTML: 405
- PDF: 89
- XML: 26
- Total: 520
- BibTeX: 15
- EndNote: 16
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 191 | 35 |
| Canada | 2 | 146 | 27 |
| China | 3 | 47 | 8 |
| Netherlands | 4 | 24 | 4 |
| Germany | 5 | 21 | 3 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 191
H. Brendan O'Neill
Stephen A. Wolfe
Caroline Duchesne
Ryan J. H. Parker
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1822 KB) - Metadata XML