A Demonstration of the Capability of Low-Cost Hyperspectral Imaging for the Characterisation of Coral Reefs

Round 1

Reviewer 1 Report

This is a competent piece of work that will be of no use to anybody. The last thing we need is ANOTHER monitoring tool.

Author Response

I have included the following in my discussion "The strength of the technique is in allowing for non-invasive physiological assessments of organism health which will help us understand how marine habitats are impacted by stressors such as heatwaves and crucially how they recover. This will enable us to better understand the impacts and to identify pathways to mitigate and adapt."

This encompasses why I believe hyperspectral imaging is of use, especially as it is a versatile tool that enables lots of different types of data to be gathered in one data collection event thus saving valuable resources and time.

Reviewer 2 Report

Interesting study by Teague et al. on the application of hyperspectal imaging techniques to coral reef monitoring. I like the fact that it is a honest paper with a good description of the limits and what need to be added (z) to solve the equations. I seriously doubt this will be sued in routinely by any monitoring program, even well founded, but I have no doubts this may lead to specific interesting applications in the context of research projects.

I have only minor comments as it is very clear to me.

- Even if it appears in reference 12 from the same authors, an illustration on how the DSLR is operated is helpful to make the paper self-contained. A good photo of the device is useful too. 

- add a scale bar to  Fig. 2 and 3, or  specifiy the lenght of the PVC pipe.

- Now, I think the main point that is missing is how the benthic spectra are going to be interpreted. Hence, the paper falls short of showing the processing required to reach a useful 'monitoring' goal. No need to show a real case, but it is necessary to explain in Discussion how spectra will have to be statistically processed to identify species/genus. There is of course a somewhat large coral reef literature on this, some about 20 years ago, from airborne or in situ spectral work. See Hochberg (one of the co-author)’s work for instance for corals but there are examples on algae, etc,. Indeed, data acquisition is only, often, the tip of the iceberg to reach the final goal. Post-processing for thematical interpretation is huge and this need to be acknowledged also.

Author Response

- I have included a diagram adapted from my previous paper to quickly show the method and the imager.

- I have added a scale bar to the figures for clarity.

- I have also added a section to clarify how hyperspectral imaging can be used to identify species/genus but in short we can only really classify to a group for example "blue corals" as shown previously by Hochberg using hyperspectral data alone not currently to specific corals. There might however be ways to complement spectral data with 3D data and using a combination of both coupled with machine learning. See text 536 for added section.  

Reviewer 3 Report

Hello,

Thanks for the article. It describes the proposed technique well.
However, the depth at which the study was conducted is not mentioned (and it is essential information since in the introduction you mentioned that the maximum depth of aerial photography is 25 meters). The measurement and calculation of H and Z are missing as well.

Also, I think you should explain more about the advantages (and disadvantages) of this technique over the others you describe in the Introduction. It is evident that the technology requires remarkably high computer memory resources. The introductory and the discussion sections did not fully answer the questions of feasibility and needs that the technology is supposed to answer, and only after I read your review published in 2022, I got a better understanding. 

Good Luck

The article is well written.

Author Response

Added clarity to depth in section 2.5 line 175. Just to note this is within the range of aerial photography but as I explained in my previous paper you can get much higher resolution underwater and you don't need to worry about the effects of the waters surface and atmosphere when correcting the data. I have added in areas about depth measurement (section 2.2, 2.3, 2.5), it isn't much of the focus as highlighted later more accurate measures are required. 

I think adding the advantages and disadvantages would be stepping on the toes of my previous paper too much. For me this paper is a follow on from the previous paper and is about demonstrating some of the use cases I described in my optical techniques review paper. 

As for the high computer memory resources its no more that of the needs of photogrammetry software but its important for everyone to understand that this as with photogrammetry is not an instantaneous technique from data collection to analysis (at least in my low-cost solution). 

I have added various sections especially in discussion to answer some of the concerns around the needs the tech answers. 

Round 2

Reviewer 2 Report

No comment.

Reviewer 3 Report

Thanks for the corrections. I have no further comments. Please do another edit check after receiving all corrections.