«Dear Editor, we are submitting our revised version hereby. In the reviews, the topic marine ice was strongly pronounced. We would kindly like to ...»
we are submitting our revised version hereby.
In the reviews, the topic marine ice was strongly pronounced. We would kindly like to
emphasize that the manuscript has the title ‘The evolution of the western rift area of
the Fimbul Ice Shelf, Antarctica’ and not ‘Marine ice in Fimbul Ice Shelf’ – which
would be an interesting study itself. The topic is thus primarily the discussion of the
rifts and not the generation of marine ice.
Both reviewers criticized the structure of the manuscript. We have worked on an improvement, but kept the structure ‘Intro-Database-Synthesis-DiscussionConclusion’. Thus we shifted all presentation of data, including the radargrams, in the synthesis section, concentrated the description of the data and figures in this section.
Consequently, we reordered the figures as well.
The manuscript covers a lot of details, which we believe is beneficial for the reader.
Papers with a structural glaciological approach have taken a similar route with success. We admit that it is at some instances lengthy, but there is not much potential to tighten the description without loosing information. Where we could tighten the manuscript we did that. We did our best to balance between presenting detailed information and focusing on the important issues.
We thank the reviewers for the helpful comments that lead to an improved version of the manuscript that we submit hereby. We answer all reviewer comments point by point below. You will find the reviewer comment in gray and our answer in black letters.
Sincerely, Angelika Humbert and Daniel Steinhage ___________________________________________________________________
Anonymous Referee #1 Received and published: 6 May 2011 This paper takes a detailed look at the structure of part of Fimbul Ice Shelf, including an analysis of satellite imagery and airborne ice-penetrating radar data. The study divides the western area of Fimbul into sections, providing a detailed commentary on the structure of the ice thickness and isochrones observed in each section.
My overall impression of the paper is that it probably deserves to be published in The Cryosphere eventually but the current manuscript is an unexciting read and perhaps not the greatest advance in knowledge.
We apologize that we could not enchant the reviewer in the same way as the topic enchanted us and we hope that the efforts we put into the revised version lead to a sharpened manuscript, as well as highlighted more of the crucial points.
I was left thinking that the paper provided neither new theoretical insight into rifts (the final hypothesis for downwarped isochrones, a very interesting observation, was unconvincing) or even a well-motivated and comprehensive overview of this particular location (why should we be interested? what about the east of Fimbul?
what about the modelling?).
“What about the modelling?” is a provocative statement at this location.Indeed one of the authors performed extensive numerical simulations, however, we decided that the manuscript would be totally overloaded with a description of the modelling, that would result in about the double of pages than we presented here. Thus we decided to split this and submit the modelling in a separate manuscript. We think that the reviewer should owe us this freedom of decision.
The east of Fimbulisen is simply not the topic of this manuscript. We picked out the western rift zone, because it is a prominent area in the Fimbul ice shelf, whereas the East is rather unexciting from the structural glaciological aspect. We were particularly interested in understanding how such massive rifts could form and what are the driving factors for that. We hope to satisfy the reviewer with the incorporation of this brief section about the East.
‘The eastern part of Fimbulisen, visible in Figure 1 as well, consists basically of ice with a smooth surface structure. The only exception is the area north of Tsiolkovskiy Island, where the ice flow diverges and forms a rifted zone northwards. The flow of the ice shelf in this eastern part is decoupled from the fast central part and does furthermore not affect the western rift area. Therefore, our study does cover this part.’ There are also significant deficiencies in the interpretation of the data, particularly relating to interaction of Fimbul rifts with the ocean.
We discuss this point in detail further below, as this topic appeared also in the section ‘Larger points’. Here we only want to state, that we improved the discussion in the manuscript by focusing more on the references provided by the reviewer and took up points made by this reviewer.
The use of English is imperfect.
Obviously both authors are not native speakers. We have worked on an improvement of the English and will authorize the copy-edit service offered by The Cryosphere.
I would suggest major revisions at the very least.
Larger points The paper contains no motivation. The introduction launches straight into the detail of previous work on Fimbul, without explaining sufficiently well why we might be interested in it, and in particular only its western section.
We have deleted the text about the thermal structure of Fimbulisen (1090/23 to 1091/4), which might have been some kind of distraction. The rest of the introduction guides into the area of Fimbulisen (1090/18-1091/4), explains that the western area is considerably different from the eastern, is untypical and large (1091/4-1091/13), that this rift zone influences calving of giant icebergs (1091/14-24) and what approach we chose to investigate the origin. (1091/25-1092/3). It seems to us that the reviewer’s perspective is focused on the marine ice, which was never meant and never claimed to be topic of this paper.
The ‘evolution’ of the rift area (as alluded to in the title) is not discussed.
We disagree with this statement. A look onto the conclusion proofs that we explain the evolution of the rift zone: the first half of the conclusion summarizes, the second half discusses the evolution and even ends with ‘We infer that this process is the origin of the western rift zone.’ We would call this a discussion of the evolution of the rift area.
The observation of downwarped isochrones coinciding with basal crevasses is fascinating– I’m not aware of such an observation elsewhere?
We could not find any similar observations in the literature as well.
However, it is not mentioned in the
or conclusions and is not satisfactorily explained in the paper. It’s not particularly well displayed in the figures either.
This is not true.
Abstract: 1090/8-10 ‘Downstream of the rumple we found down-welling of internal layers and local thinning, which we explain as a result of basal crevasses due to the basal drag at the ice rumple.’ Conclusions: 1110/ 7-8 ‘Although the vertical structure exhibits strong deformation of the internal layers and also hyperbolas throughout the thickness’ Figure 4a+b(new3a+b) shows down-wrapped isochrones.
However, we have highlighted this in the new figures now, so that it is easier visible in the Figures. The best explanation offered is that there is increased melting within crevasses, but it is an observed fact that the ocean freezes in Fimbulisen crevasses, rather than melting them (e.g. Khazendar & Jenkins JGR 2003).
We like to emphasize, that the area where freezing is observed and modelled is an ice melange (Zone22/new1) and not a zone with single crevasses. It is definitely not comparable to the area where we observed down-welling of the layers. Figure 4b(new: 3b) shows where the down-welling is observed and the comparison with Figure 7(new:5) where the ice melange is located.
It is possible, I suppose, that higher melting occurs on the sidewalls of a rift than under the ‘flat’ ice outside crevasses, and this might downwarp the isochrones. If this is the case, then ‘older’ crevasses (further downstream from the rift) should have more downwarped isochrones, since the melting has been going on for longer – it doesn’t look like that is the case but it is hard to tell from the figures?
In case melting on the steep sidewalls of a rift would simply remove mass but not alter the position of isochrones in vertical, as the thickness of the ice shelf does not change. Whereas melting on the base would cause a change in the vertical of isochrones, because of the buoyancy of the ice.
It is also possible that simple ice dynamics are responsible for the downwarping – e.g. Leysinger-Vieli et al Ann. Glac. 2007.
This is a very interesting point (whereas ‘simple ice dynamics’ is a somewhat wide term, it is definitely ice dynamics that is responsible, what else?). A definite answer can likely only be given with a 3D full-Stokes model, including a linear-elastic and linear-elastic fracture model in the vicinity of the rumple, for an area of about 30x30km2 around the ice rumple. The sliding+melting examples in Leysinger-Vieli and others (2007) are comparable to some extent to the situation here, as the plug flow that these authors denote as ‘sliding’ is the flow regime in an ice shelf. However, we neither have a pure sliding + melting, nor a pure ‘channel’, we have ‘internal deformation’ over the rumple, ‘sliding’ around and a ‘channel’ in the lee zone of the rumple (to keep these authors terminlogy). However, profile A to A’ is relatively far away from the ‘channel’ and it seems to us unlikely that the channel effect is relevant at this location. That the removal of material at the base causes the layers to plunge is exactly what we have suggested in the manuscript. This is indeed simply ice dynamics – nevertheless not comparable to Leysinger-Vieli et al., as that ice is not floating, which is an additional contribution here.
Can the opening of a crevasse cause a downwarping due to ice divergence in the lower part of the ice column in the same way as the transition from sticky to slippery basal drag causes a downwarping, I think due to horizontal divergence of ice that increases with depth? That would give a downwarping that happens only on crevasse formation, so the downwarping remains constant with crevasse ‘age’ (distance downstream) which is agreement with the observations as far as I can see. In the reference mentioned by the anonymous reviewer, Leysinger-Vieli and others (2007) investigate influence of various scenarios (changes on flow mode, basal melting and flow convergence) on isochrones using a simple model for grounded ice.
The model deduces the velocities based on the mass fluxes. As basal melting at the base of the ice shelf removes mass, the model is not applicable to an ice shelf, even though the model provides a valuable insight on the response of isochrones of grounded ice on varying basal conditions.
Another apparently key observation is that the isochrones converge vertically. I can’t see this in the figures.
We provide the reader now in Figure 4b(new: 3b) with two tracked layers in one radargram and display the distance between two layers along the flight-direction.
Similarly, what are the hyperbolae high up in the ice column?
We suppose that they are the tips of basal crevasses.
Do the figures show an example of one anywhere?
Yes, Figure 4a(new: 3a), the radargram from A to A’ shows hyperbolas at a high location in the ice column, as well as the map in Fig.3b(new: 4b).
I would suggest that example figures need to be added showing detail of downwarping, isochrone convergence, and hyperbolae at different depths.
The isochrone convergence is now clearly visible in Fig.4b(new: 3b), profile B to B’, even highlighted in color. In this figure the downwraping is also well visible, as now two of the layers are colorized. The internal hyperbolas at different depth are superimposed on section A to A’ now and a colored bar denotes where basal hyperbolas exist.
The existence of ocean-sourced marine ice in Fimbul (Khazendar & Jenkins) is never mentioned in the paper, despite its ability to explain many of the observations.
We don’t see how marine ice explains our observations and the reviewer does not clarify which observations it would explain neither. It is hard to believe that the reviewer claims that marine ice explains the stress state that creates the rift system, nor the formation of single rifts and surely also not the subsequent deformation and propagation of the rifts. ‘[...] despite its ability to explain many of the observations.’ is thus in our point of view an assertion. Therefore, we would like to emphasize here again, that this study aims to investigate the evolution of the rift system.
Although we did not cite this reference in the original version (we do this in the revised version) Khazendar & Jenkins refer to a site called Jutulgryta (former Zone 22, new Zone 1), an area formed when the ice stream just passed a confining valley and became afloat. This area consists partially in summer even of open ocean, fragments of meteoric ice and sea ice, which was used by Orheim et al. (1990a,b) for an easy access to the ocean, after the first drill attempt through the main, thick, part of the ice shelf failed. This area (#22 in the original version) is not comparable to the area where we observed the downwelling of the layers. Using common terminology (as we did in the manuscript), we would call this area an ice melange.
Although Jutulstraumen has a deep grounding line and was previously suggested to experience high melt rates (for a quick overview see Humbert, 2010 which displays her basal melt rates from steady state assumption, the basal melt rates from Smedsrud et al., 2005 from ocean modelling and Smith, pers comm, 2009, steady state), there is no evidence for a thick marine ice layer like the large ice shelves exhibit. The estimation of the mean density from hydrostatic equilibrium using the ICESat DEM and the various ice thickness measurements, including the extensive survey of Nost, 2004 and the one presented here, give not indication for a considerably thick marine ice layer. If marine ice exists, it is unlikely to play a role in the evolution of the western rift zone, which we investigate here.