the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 06 Mar 2024
Modelling subglacial fluvial sediment transport with a graph-based model, GraphSSeT
Abstract. A quantitative understanding of how sediment discharge from subglacial fluvial systems varies in response to glacio-hydrological conditions is essential for understanding marine systems around Greenland and Antarctica and for interpreting sedimentary records of cryosphere evolution. Here we develop a graph based approach, GraphSSeT, to model subglacial fluvial sedimentary transport using subglacial hydrology model outputs as forcing. GraphSSeT includes glacial erosion of bedrock and a dynamic sediment model with exchange between the active transport system and a basal sediment layer. Sediment transport considers transport-limited and supply-limited regimes and includes stochastically-evolving grain size, network scale flow management and tracking of detrital provenance. GraphSSeT satisfies volume balance and sediment velocity and transport capacity constraints on flow. GraphSSeT is demonstrated for synthetic scenarios that probe the impact of variations in hydrological, geological and glaciological characteristics on sediment transport over multi-diurnal to seasonal timeframes. For steady-state hydrology scenarios on seasonal timescales we find a primary control from the scale and organisation of the channelised hydrological flow network. The development of grain size dependant selective transport is identified as the major secondary control. Non-steady-state hydrology is tested on multi-diurnal timescales, for which sediment discharge scales with peak water input leading to increased sediment discharge compared to steady state. With increasing application of subglacial hydrology models, GraphSSeT extends this capacity to define quantitatively the volume, grain size distribution and detrital characteristics of sediment discharge, and enables a stronger connection of models of the glacio-hydrological system with constraints from the sediment record and impacts on marine systems.
- Preprint
(15929 KB) - Metadata XML
-
Supplement
(63665 KB) - BibTeX
- EndNote
Status: open (until 30 May 2024)
-
RC1: 'Comment on egusphere-2024-274', Anonymous Referee #1, 10 Apr 2024
reply
This paper introduced the GraphSSet model, a sediment transport model driven by subglacial channelized water inputs and producing rates of sediment output at the grounding line including grain size and volume, among other properties. It’s really encouraging to see this work linking together catchment-level subglacial hydrology with sediment transport as it’s an area that hasn’t had much attention to date. I don’t have any major comments but I’ve made a lot of targeted suggestions to clarify the writing and raise areas where it’s less clear what the approach is.
In general the model seems quite complex for application to subglacial settings where there are so many unknowns about the basal system. For example, knowing the sediment base thickness above bedrock would be very difficult to establish for most glaciers. It would be good to include a sentence or two in section 5.5.1 acknowledging how this would be approached for someone who would want to apply your model to a non-synthetic system.
In the same section, it should also be noted that because GlaDS is operating with R-channels rather than canals or channels eroded into sediment, further development of hydrology models is needed to better represent the interactions between sediment and subglacial drainage systems.
Throughout the manuscript and for the figures you discuss the edge betweenness centrality. From the description of this as the frequency on the shortest path I don’t understand how it can represent channels as suggested in the figures (I’ve not come across this term before). Is there a way to describe this for those not so familiar with graph models.
I don’t think I’ve ever seen a 94 page supplement! Can you include a guide at the beginning telling the reader what’s included and the page range that those figures can be found.
Targeted line comments:
Line 11: “grain size dependent selective transport” is a tricky combination of criteria to understand. It’s the use of “selective” that I’m not sure about.
Lines 13-16: I’m not sure what this means. Do you mean that more people are applying subglacial models? “
Line 14 “quantitatively define”
Line 15: “stronger connection” could be rephrased to say “enables constraining glacio-hydrological models with sediment records”
Line 18: “impacts the delivery”
Line 21: I don’t think these are otherwise unobservable.
Line 23: “crucial for establishing”
Line 24-25: This statement of subglacial evolution is vague.
Line 24-30: These sentences are tricky to parse. Are you talking about Antarctic hydrology systems, Greenland or Alpine? There are lots of different drivers and timescales depending what systems you’re examining. I’m not sure how these fit into the sediment discharge arguments from the sentences above. What about supraglacial lake drainage? I’m not sure how water piracy fits into the ‘high flow’ argument.
Line 32: To determine the consequences?
Line 45: Can you specify which inputs are distributed and which are point source inputs.
Line 46: Can you include a citation for water that might be stored englacially.
Line 62: “These include”
Line 66: ‘coupled with ice flow’ is confusing phrasing. In general 66-69 could be shortened given that you then say you don’t focus on it.
Line 67: ‘move at the same speed as the ice’
Line 70: why glaciated margins?
Line 93: ‘basal water input’ usually means water produced basally in situ which is not what I think you mean here.
Lines 85-112: most of this content isn’t needed since you’re using GlaDS. Since there are already a lot of summaries of glacial modeling approaches (e.g. Flowers at al, 2015) it would be better to reference those and reduce this section to a couple of sentences.
Line 152: just saying you use pi is a bit misleading since S is a semicircle
Line 203: I think point 1 needs to be rephrased to illustrate this is grain size dependent (assuming I’ve understood correctly) rather than just preventing excessive velocities for all sediment sizes
Line 287: I don’t understand what you mean by ‘multiple remobilised sediment’
Line 302: what do you mean by ‘flow reaches the model boundary as grounded ice’. Wouldn’t grounded boundary ice be defined in the FEM boundary conditions?
Line 303: Could ‘lakes’ in the FEM model have hydraulic potential of zero?
Line 307: if you’re only using channel characteristics to drive GraphSSeT, why do you need information on the distributed system?
Line 313: What does ‘Key to our approach is the definition, from this main graph, targeted subgraphs’ mean?
Line 325: When you say the source nodes are a randomly selected set of input nodes there needs to be much more clarification. How can this represented distributed inputs? How many of these nodes are chosen?
Line 329: you mention model scenario A5 but we don’t know what this means yet.
Line 379: you say you run at least two default models then only list parameters for one. I’m not sure ‘reference’ and ‘default’ are clear descriptions for these various runs.
Line 395: specify whether this change is applied to GlaDS or GraphSSeT.
Line 437: the statement about the delay in the onset of bedrock erosion is interesting but I’m unsure how to read this in Figure 8. Is this from ‘volume derived from bedrock (never deposited)’? Is it the yellow dashed line you’re referring to? Very confusing here and in other figures what ‘total’ and ‘bedrock’ refer to.
Line 473: ‘have only small differences’
Line 510: Can you direct readers towards Figure 12 again when you start talking about the classes.
Line 546: Do you mean “Our GlaDs input model scenarios have the same basal ice velocity, and no basal topography…”?
Line 644: At the end of your conclusion it would be good to include a statement of where the model can be applied to next e.g. applied rather than synthetic scenarios.
Table A1: It would be helpful to include a column summarizing the main feature of the scenario e.g. moulins vs. distributed input (along the lines of Table 4 in DeFleurian et al, 2018)
Figure 2a/3. Can you make the node dots larger in the legend – it’s hard to see their color without zooming way in. Since you don’t have any moulin nodes there’s no point putting them in the legend.
Figure 4: I don’t understand what ‘total’ and ‘bedrock’ refer to here or ‘rerun’. This isn’t described in the text.
Figure 10. When you say ‘basement’ components do you mean bedrock? Consistent terminology would be good. I also don’t see this figure listed in the text.
Citation: https://doi.org/10.5194/egusphere-2024-274-RC1
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 185 | 52 | 14 | 251 | 18 | 12 | 6 |
- HTML: 185
- PDF: 52
- XML: 14
- Total: 251
- Supplement: 18
- BibTeX: 12
- EndNote: 6
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1