Annual and Interannual Variability in the Diffuse Attenuation Coefficient and Turbidity in Urbanized Washington Lake from 2013 to 2022 Assessed Using Landsat-8/9
Round 1
Reviewer 1 Report
The authors seek to model light attenuation and turbidity (water clarity) using several in situ data sources in combination with Landsat 8 and 9 data products (in the visible spectrum). The authors explain the importance of this work in the context of predator/prey visibility. Indeed, water clarity is a very important variable with regards to predator/prey visibility in aquatic species. That said, with the quantity and perceived quality of the in situ data available at Lake Washington, I’m surprised of the lower-quality modeled outputs presented. I think that a bit more effort could be applied in order to increase the accuracy of the output metrics from the Kd and FNU models. This would allow the authors to derive more meaningful time series attenuation and turbidity curves. Which would allow them to better explore their ultimate goal of getting at pred/prey visibility and get a sense for juvenile salmon decline.
Some specific comments and general thoughts are below:
- Lines 105-108: Why do you think the quality of the measurements in the upper level of the water column are so poor? Is it actually safe to assume that the water column from 0 – 2 m of depth is uniform here? Is this a common assumption. Any sources to back that up?
- Line 147: I think you mean four bands here instead of five?
- Line 180: One of the deepest.
- Line 199: I’m not sure what a “wave band” is. Is this a typo? If not, I suggest using a better descriptor.
- Lines 202-203: This sentence structure is hard to understand/digest. I suggest adjusting accordingly.
- Line 213: Is C unitless? Need to better explain this equation. E.g., AT, BT, and C.
- Line 222: I don’t see any correlation coefficients listed. Perhaps I missed them?
- Line 240: I’m not sure “d” is universally understood as days. I suggest being explicit. It’s only an extra two letters to write “days”.
- Lines 250-251: “The algorithm was unable to reproduce in-situ Kd(PAR) for spectral shape two pixels (Figure 3a), so these data were removed from the analysis”. More explanation on why that is would be beneficial to the reader.
- Line 290: There are only two spectral shapes here. Also, in the above paragraph the author states that “No statistically significant model could be derived for 286 spectral shape three pixels, so these data were removed from the analysis.” I presume this is just a typo and should read as “two spectral shapes”?
- Line 298: The months are not displayed on Figure 5. Also, point the reader to where they can get this information (assuming it exists in a figure). Actually, I see the differences between the modeled and in situ FNU data are plotted in Figure 6b. This plot shows that months 3, 7, 9, and 5 have the greatest discrepancy. And not June, July, and August as is mentioned here in Line 298. Unless I am missing something?
- Figure 6a/b are missing a meaningful x-axis label (i.e., Month).
- Line 350: “thus” what? Adjust accordingly.
- Lines 418-419. You could expand your dataset by including Sentinel 2 data products. Thoughts?
- There wasn’t really an explanation of how the three spectral shapes were derived or how often they occurred in the data.
- Why not develop a model for each spectral shape and apply it? Low sample size?
- Did the authors think of using random forest, CNN, or other methods to derive the models? As it stands now, it appears they only looked at two band-differencing methods for the regional models. There are many papers out there exploring different band combinations and deriving turbidity. The authors are in a great position as there appears to be a wealth of in situ turbidity data to train a meaningful model/s.
- Section 2.5 is quite succinct and could be expanded. For example, to which “datasets” are you referring?
- It’s generally good practice to expand acronyms prior to their first use. e.g., PAR
The paper could use some minor syntax revisions. See the review for a few examples.
Author Response
Letter attached.
Author Response File: Author Response.docx
Reviewer 2 Report
The article titled: Annual and inter-annual variability in the diffuse attenuation coefficient and turbidity in an urbanized Washington lake from 2013-2022 assessed using Landsat-8/9. This paper evaluates existing semi-analytical models for Kd (PAR) and turbidity, develops a regional turbidity model based on spectral spectral shape, and evaluates spatial and temporal trends in Lake Washington between 2013 and 2022 using Landsat-8/9 OLI. 2022 using Landsat-8/9 OLI.
General comments: The manuscript is well written. The results are as indicated in the objectives. A great job of data processing is appreciated. Although the manuscript contributes to the study of the clarity of inland water bodies it does not relate to other factors, both climatic and anthropogenic, and performs a detailed spatial analysis at the lake level, without taking into account the basin input. However, due to the importance of local studies and the increasing use of satellite imagery as valuable data, I believe that, after review, it should be considered for publication.
Specific Comments:
1. I suggest adding more bibliography, the introduction section does not reach 20 citations (suggested for this section).
2. Add bibliography of the last 5 years, there are more than 1o bibliographic citations of almost 20 years old.
3. write the months of the year in small letters
4. add map elements such as north to figure 1
5. for international readers it is suggested to add where in the continent or within the united states the lake is located, the figure is very specific.
6. the first time PAR is defined write (photosynthetically active radiation)
7. in figure 2 bring the values of the statistics to 2 significant figures after the decimal point
Comments for author File: Comments.pdf
Author Response
Letter attached.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
Thanks for addressing these issues. No further comments. Looking forward to seeing the future work using the nighttime light data.