Correlation between Underwater Noise and Sea Level at Ieodo Ocean Research Station

Round 1

Reviewer 1 Report

· Line 66-67: "(...) in the frequency range of 3–30 kHz", please correct for "(...) of 3 Hz - 30 kHz". Because I imagine that this is the correct range of work for the hydrophone, and not between 3 kHz and 30 kHz.

· You should mention that if you are recording at 44.1 kHz of frequency sample, your maximum frequency registered is 22050 Hz. For this reason, your studies (and figures) are focused under 10 kHz.

· I miss more explanation about the big deviation in the measures. Can this point adultering the final results? Are these results reliable and applicable for other periods?

· About Figure 7, how is possible obtaining a negative correlation coefficient?

· I miss more text in the conclusions part about the results for the bigger than 50 Hz range. Because you are focused your results for 16 and 50 Hz frequencies, but you studied below 10 kHz. Comment in shortly about the results above 50 Hz frequencies, why are not stables and confidence?

 

 

Author Response

Dear reviewer...

Please see the attached file.

Author Response File: Author Response.docx

Reviewer 2 Report

Paragraph 2.3 (from Line 90)

• What type of Wavelet base was used (Morlet?Mexican Hat?...). Why this particular base was used?
• What type of wavelet transform was used (discrete/continuous)?
• How was wavelet spectrum and coherence calculated? How was wavelet phase defined?
• According to the plots you used in the results sections you should report here the essential mathematics that allowed you to derive your numbers

Line 99. Could you use a bigger marker and a more pronounced colour to mark the spring tides in Fig3? Here is barely visible (Note this is just a suggestion - Not mandatory change)

Line 150. What type of low-pass filter was used?

Line 158. I suppose that Fig7 is the spectral cross correlation coefficient. Again define this quantity. Also, how were the error bars calculated?

Figure 8. What are the black contours lines representing? What is the value of the colormap? i.e. it is wavelet intensity levels? if so this has to be stated in the text.

How was Fig10 computed? Is this the period averaged wavelet correlation coefficient?

 

Author Response

Dear reviewer...

Please see the attached file.

Author Response File: Author Response.docx

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

The work is interesting but needs some improvement in both discussion and interpretation. It is a nice confirmation of some results in very good conditions, but it can be more than that. Why is it important to correlate the underwater noise level with tide amplitude? You have a clue of it by the last sentence" studies are needed to predict the tidal current using underwater ambient noise levels". But it would help if you exploited it better in the paper.

I have a different question. While looking at hydrophone and OBS data in a very different frequency band, and while measuring the RMS and the RMSe, we found some of the periods you found. Do you have any hydrophone on the seafloor near IORS station? If yes, did you perform a similar analysis?

Some minor points below.

In the introduction, when you start talking about the IORS, I suggest that you briefly mention the type of sensors used to record ambient noise.

Still in the introduction, line 38, this sentence is not very clear. You want to mention possible ambient noise sources and their impact on noise levels?. So I would maybe say that underwater ambient noise level is controlled, or is affected, by nearby noise (ships, wind, biological noise, etc.).

Later, on line 47, can you be more precise on what you mean by long term? Ambient noise is now widely used in a broad range of frequencies. So long term can be different for different users, working at different frequency ranges.

Line 76 - 10 H -12,5 kHz should be 10 Hz -12,5 kHz

Line 80 - 10 May 2013 to 18:30 p.m. on 27 August (GMT+9). either you have 6:30 p.m. or 18:30. Have you compared the same time window in different years? Why did you choose May to August? Did you try other time windows?

Line 90 - 00:00 a.m. on 22 June 2013 - 00:00 on 22 July. I didn't you analysed the tidal curve at the same time window mention above, May to August? Why did you choose that window?

In section 3.1, As you said, the Average ambient noise level during spring tide is higher than during neap tide. Can you explain detail better?
Also, some of the thin lives of both tides overlap. Why?

In the text, When you mention the number of data is the number of points or the number of time series with a certain time-length?

Did you apply a linear cross-correlation? Before cross-correlation did you use any pre-processing to correct from local perturbations?

Reviewer 2 Report

The authors have not done enough to distinguish their data from flow noise, or non-propagating "pseudo sound" that is caused by the flow of turbulent eddies past the hydrophone (this is the "wind sound in your ear" when riding a bicycle).  There are many studies that discuss this effect.  The authors allude to one (Bassett, et al. ) but otherwise make no attempt to confirm whether they are actually measuring a genuine acoustic field or flow noise.   In other words, the spectrum measured at high currents is likely unique to the hydrophone itself, and would not be measured by a nearby "fish" or neutrally buoyant hydrophone that is drifting freely with the currents.   

This is hard problem.  One approach is to use two, nearby hydrophones and if the sound is correlated, then it must be real sound -- and not pseudo sound generated by the presence of hydrophone in high current.    Other strategies involve flow shields on the hydrophone to mitigate flow noise.  The authors seem to miss these point all together.

Beyond this key issue the authors have done some innovative things with wavelet processing, but otherwise the conclusion--that they observe a correlation of genuine ambient noise level with tidal variation--are not well-founded. 

Reviewer 3 Report

ABSTRACT:
· Parameter r not explained in the abstract, not use it
· Rewrite the abstract avoiding use r directly (you can talk about the "correlation coefficient"). Avoid information that the lector can't understand if the paper is not read. Include only the mot clearly and important conclusions.

KEYWORDS:
· Rethink the keywords, I suggest:
Wavelet Acoustic Analysis; Tidal amplitude; Underwater Environmental Noise; Ieodo Ocean Research Station (IORS);

INTRODUCTION:
· Line 34: Reference [3] at the end of the sentence. The information It can be taken from this same webpage
· Line 34: what is “(80 nm)”? nanometers? What does it mean?
· Lines 36-37: “which is rare in research around the world.” Is this sentence objective?
· Line 37: Remove “(since 2003)”, the information is repeat.
· Line 39: It is not better remove “wind” for “natural”?
· Line 39-40: add some more numerical information about “frequency ranges” and “long time” to clarify the concept
· Line 47-49: Reference about “the correlation between underwater ambient noise levels in the coast of Oregon in the United States as well as various elements have been studied.”?
· Line 47: Specify the concept “long term” in the article
· Line 51: You put “(< 100Hz)”, please correct for “(f < 100 Hz)”
· Line 53: Specify the concept “low frequency” in the article [13]
· In the figure 1b you indicate “weight” in the hydrophone, can you indicate something about it in the text? How much is the weight in kg? It means that the hydrophone is not fixed and it can be move-in strong sea currents? It can be affects to the acoustic recording?

DATA AND METHODS:
· I suggest you improve the coherence of the read using subsections in this chapter. For example 2.1 Acoustic data (explaining the acoustic receivers specs, the schedule data to recorded data, the process to analyze the data and a short introduction about the differences in noise recorded ambient for the red and green part) 2.2 Tidal data (explaining the measurement specs of the device, the particularity of the directions in its data, and its analysis using the low-pass filter) and 2.3 correlation analysis (explaining the methodology followed to obtain the results of this paper and the means/importance of r coefficient, due to you are talking about this parameter a lot of times in the text)
· Line 76: you put “10 H-12.5 kHz” correct by “10 Hz – 12.5 kHz” (if the sample rate is 48 kHz
· About acoustic data: How often it is recording? The receiver is installed permanently or only for specifics studies? The hydrophone is in movement produced by strong sea currents or it is fixed?
· Line 78: “minute” correct by “minutes”
· Line 79: “5233” complete it by “5233 samples”?
· Line 80: Indicates the year of the end recorded
· Equation 1: If you put IL in the formula, please use IL and not "overall level" in the text, the lector can mix the value. Another option changes IL in the formula
· Line 83: “intensity level” by “Intensity Level”
· About third-octave selection: why third octave? If it's not an audio application is the best option? Can you add a table specifying the third-octave values used?
· Line 86-87: What is an “RV Eardon”?
· Line 87: Indicates the time zone refer. CET?
· In this section, I miss a paragraph explaining the get of the correlation parameter. So, please explain your experiment and the method of obtaining your experimental results

RESULTS AND DISCUSSIONS:
· In figure 2a print in red and green the data in the red and green sections, then you can refer in the text to them like “red part” and “green part”. It is more visible to the lector.
· Line 96: “Figure 2, b” correct by “Figure 2b”
· Line 100: “1-day data”, are you refer to 24H of continuous data?
· Line 107: correct by “When”
· Lines 110-111: “The results for the underwater ambient noise level that were measured in the United Kingdom were compared with the levels that were measured for the ebb and flood current (1 day) [14].” This is not a comparison paper, please move this sentence to the conclusions section
· Line 119: How affects a low pass filter to the tidal data? Can you take more information about this procedure?
· Figure 2a and 2b, please correct the presentation of Xaxis label. They are in form to can mix the lector
· Figure 2d: You can add another subplot with the difference between both and their averages
· Figure 2: I suggest you show a spectrogram of the acoustic signal recorded (frequency vs time) to observe the variability of frequencies in the red and green part
· Figure 2: can you explain something about the deviation in the data? Talk in the text about if the deviation is strong or not to the article considerations and the reasons.
· Line 127: Why is Figure 5 mentioned before Figure 4?
· Line 129-130: About low-pass filter in acoustic recorded, order? frequency of cut?
· Line 133: “Figure 5c” correct by “Figure 4c”
· Line 140-143: “The low frequency of the underwater ambient noise level was affected by noise that occurred from the shipping vessels in the near distance and from sea surface disruption that occurred from the wind [4-8]. According to this, the correlation coefficient was low in the raw data of the underwater ambient noise level and tidal amplitude” please move this sentence to the conclusions section
· Figure 3: Extend the text size, they are not legible. You can remove the Fig.3a, it is “the same” like 3b. Suggestion: include another subplot with the value of correlation for the period shown
· Line 161-165: “In this study, the underwater ambient noise level also showed a high correlation with the tidal current at a frequency near 20Hz. Underwater ambient noise level showed a high correlation at a low frequency below 100Hz [11, 14]. This study also confirmed that the correlation was high at a frequency below 100Hz; in more detail, it showed a high correlation at a frequency below 50Hz.” please move this sentence to the conclusions section
· Lines 187-190: “The underwater ambient noise level at the Littoral zone of Korea was higher than in the deep ocean, because it had high traffic at the nearshore ([8-9]), and it considered the effect of the tidal current, which is an important current at the littoral zone.” This sentence is not relevant in this section, it is repeated from the introduction section
· Table 1: Correct the subindex in the symbols of the table. Include the units in the Relative Size column (%?)
· Figure 6: When you are talking about the overall level, are you using the signal amplitude recorded in the frequency domain or in the time domain? In any case, in volts, I imagine, right?
· Line 213: Why is the 1st time that you talk about sea current speed? Do you have data from a current meter? Can you show the measures for the experiment period (like in Fig2.a)?
· Figure 7: Units in the bar colors scale? What it is showing?
· Figure 8a: Can you increase the size like Fig.8b?

BIBLIOGRAPHY:
· As it is possible, please add its DOI numbers.

 

GENERAL:
· Add a space between numbers and units. Examples: "10Hz" and "20Hz" in line 50
· About figure labels, it is more common to indicate the figure (a or b) previous to its definition
· About acoustic data: Do you think is it a good process analyzing this type of data as audio signals? (I’m referring about the sample frequency and octaves used)
· The acoustic data is recorded 1min every 30 min, and the tilt data is registered every 10 minutes. How is the process to do they compare? The explanation is not clear to me.
· For me, it's not clear the obtention of the results in this paper. Not forget, that a scientific paper should be reproducible.
· It is not clear the reason to select data during one month period (figure 2) and then change to talk about 3 months of period (rest of figures)
· Why sometimes are you talking of absolute values for tidal and others not? can you clarify this mix? If your analysis uses only absolute values, please not mix showing the others every figure (figure 8a for example).