The revised submission is much improved and both reviewers recognise the work done in this regard. The reviewers remain concerned about the novelty, especially in relation to the inundation modelling. Recommendations regarding publication are split, so I have additionally reviewed the manuscript myself and include some comments below. I share the reviewers concern around the novelty of the inundation methods and don’t believe the modelling method per say is novel. However, three of the four conclusions relate to the big data architecture and implementation of the modelling methods within this. In that regard I think it is possible for a novel contribution when looking at the work as a whole, with some further revisions to the manuscript needed to bring this out. Please respond to all of the comments below and those from both reviews

Comments:
Line 60-67 need references for examples of these approaches
Line 67: needs evidence to support the view that the majority of recent developments for large scale studies are of the simple conceptual type models.
Line 70: What do you mean by a class of mode in this sentence, please be specific. “A class of model which uses the output of a more complex model as a means of calibrating a relatively simpler model is also gaining popularity (Oubennaceur et al., 2019).”
Line 81: You are missing a class of parallelisation that uses shared memory threading on CPU’s.
Line 88-92: very long sentence, consider splitting
Line 92-101: “One such simple conceptual inundation model…” Given that you are implementing a simple conceptual model this section is very brief. The logic in the introduction breaks down here because you initially state that the simple conceptual models have seen the most progress in recent years and then suggest that conceptual flood models have remained largely uninvestigated. For example why not mention GLOFRIS
Assessing flood risk at the global scale: model setup, results, and sensitivity - IOPscience
HESS - A framework for global river flood risk assessments (copernicus.org)
Or
A rapid urban flood inundation and damage assessment model - ScienceDirect
I think this is why the reviewers are disputing the novelty claimed here. In my opinion there is a lot of work on simple inundation models that can inform the approach taken here.
Line 140: As a non-expert in alternative spatial data models could you state what the traditional models would be in this context. The novelty here was lost on the reviewers and I’m still unsure how significant this is.
Line 241: Is it channel or floodplain manning’s n being determined? Are both the same? Obviously the link to n is more direct for the floodplain. Channel n is often lower than the floodplain and more associated with channel morphology than adjacent landcover.

Line 232: “Regional hydrological frequency analysis at ungauged sites is also studied by few researchers” Is this true? Prediction in ungauged basins is one of the more fundamental and widely studies areas of hydrology. Perhaps this is the case in Canada or there is some more nuance to this point?
Line 283: check wording of sentence.
Lie 294: This is a very long sentence, consider if this can be split and reordered for readability.
Around line 359: OK 1D hydraulic models require cross-section to be defined. But HAND would also need channel bathymetry to capture the channel flow components of the discharge I assume.
How is channel conveyance represented in your HAND method? Is the bathymetry in the DEM and if it was I assume that’s been interpolated from cross section data or estimated from a design discharge? So, I agree that on the floodplain HAND avoids the problem of defining cross sections (as a 2D hydrodynamic model also would), but for the channel it will depend on what has been done here and this is not clear to me. I think you oversell the benefits of HAND here.
Line 368: Backwater effects are not only due to flood mitigation structures but will occur on all rivers with subcritical flows conditions (which will probably include the vast majority of those with significant floodplains). The issue is more ubiquitous than suggested here.
Line 685, sentence needs to be specific regarding what the “direct comparisons” are. I know its obvious from the earlier text but the sentence on its own is truncated in my view.
Line 694: Perhaps I missed it, but I didn’t get a good understanding of what a traditional raster based methods are in this context. This links back to earlier comments around line 140 where I didn’t really appreciate what the innovation being brought by DGGS was reference to.
Line 717: I’m happy with principal behind the approach taken to the inundation modelling, but I’m not convinced the use of catchment integrated Manning’s is really a novelty of this work. The other novelties listed are stronger in my opinion. Rather than reinvent HAND as novel, would it be more accurate to say an existing method has been implemented with a new data set and/or location and computation framework. I don’t know what has been done previously in the region, but in my opinion it is the application rather than the inundation modelling that is more likely to be novel. Furthermore, the results show how well the inundation model reproduces historical events and design extents, at no point do you analysis the physical process directly, so I would not claim that the inundation extents are physically justified but instead point to the accuracy of the extents.

Dear Chiranjib,
I'm very happy to accept your revised manuscript for publication. The revised version is much improved and I hope the novelties will now be clear to a general audience. I'm happy you have made all the revisions requested by the reviewers.

There is one technical correction I would make. The text on line 385 (tracked document) is much improved, but the real issue with using interpolated bathymetry as describe here is that the channel bed is not observed and the overall conveyance of the river is biased towards underestimation as a result. You assume this is negligible, which might be the case, but its an assumption that should be clear.

In my view the assumption that the DEM includes the relevant channel bathymetry is going to be more, or at least as, important as the challenge of drawing cross-sections in a 1D model. But , I think its sufficient to point out the assumption suggested in my previous paragraph.

Best wishes,
Jeff