Switching LPS to LED Streetlight May Dramatically Reduce Activity and Foraging of Bats

Round 1

Reviewer 1 Report

This paper revisits a dataset from an earlier paper (Rowse et al 2016) and uses a different analytical technique to tease out the effects of streetlight spectrum and intensity on bat activity. While the paper is generally well presented, there were errors in grammar and in formatting that made it difficult to follow. In addition, there were some elements missing from this manuscript:

• Include in the introduction how each aspect of light intensity (power, illuminance) has been shown to affect wildlife. It’s easy to understand how illuminance would have an effect, but I don’t understand how to interpret power separately from illuminance in a biological context.
• Include a prediction with your hypothesis. What did you expect to find, and why?
• Address results in the context of that prediction.
• Explicitly compare your results to those presented by Rowse et al, highlighting exactly where you found differences. This might be most clearly presented as a table.
• Explain why your analytical approach uncovered those differences.

More specific feedback is included below.

Line 54: this high-level description of all the different European lamps is distracting and may not even be applicable to other areas. Consider moving it to the discussion.

Line 69: the K value given for “cool” LED lights is different from that given in the Rowse paper.

Consider adding an explicit description of the light variables, ideally as a concept box. While reading this paper, it took me awhile to understand that power and illuminance were two aspects of intensity, and the combined variable intensity was distinct from spectrum/wavelength/spectral composition. What is known about how the two aspects of intensity affect wildlife? For example, do the two sources cited for light intensity influencing wildlife (line 117) separate out these two aspects? Finally, make sure terms are used consistently throughout the manuscript.

Fig 1 refers to an inset figure, which does not appear. Also, if intensity consists of 2 different measures, what does the y axis represent? What are units of A.U.? What is the source of data for this figure?

Line 174: No figures show up in the supplementary materials document I received.

Line 194: I don’t understand this sentence, perhaps because the English is awkward. What does the co-variable have to do with reproducing the original analysis?

Line 198: How do you explain a difference in findings regarding feeding buzzes?

The supplementary materials were poorly organized and confusing. It was unclear how the authors decided what to put in the main methods vs. the supplementary methods; both seemed important to understand the methods since this is primarily a statistical methods paper. There is a reference to Table 3, in section Supplementary Material 2, which includes a Table S3-1 (the first table). In the caption for that table, what does “The control and the after period are placed in the intercept” mean? Again, none of the figures appear.  Section Supplementary Material 3 includes a table 2 (the second table). The heading of this section is incomplete. Was this run using the same R package? How was the Bonferroni correction calculated?  Why did you run these as separate regressions rather than multiple regressions, which is implied in the hierarchical partitioning analysis?

Line 226: If the random effect is date within pairs, should this equation have a vertical line rather than a slash after “Pair”? Also, I’m not sure why this shows up as a separate numbered section.  Given the conceptual model of variables, instead of “lamp type” here, consider using “spectrum”. If those terms are not equivalent here, then please explain somewhere why not.

In the conclusion, please explain explicitly why you recommend choosing warm vs cool LED lights. Also, please distinguish here between the two aspects of intensity.

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

The manuscript entitled: Switch from LPS to LED Streetlights: Disentangling the Relative Effects of Light Intensity and Spectrum on Bat Activity by Kerbiriou et al., deals with a relevant, very actual topic that is the effect of artificial light at night on the ecology of nocturnal life, evidenced by bat activity. The primary aim of the authors is to re-analyse the data and revise the conclusions of authors Elizabeth G. Rowse, Stephen Harris and Gareth Jones of their published paper: The Switch from Low-Pressure Sodium to Light Emitting Diodes Does Not Affect Bat Activity at Street Lights.

I fully resonate with authors opinion expressed in the paragraph starting on line 92, saying that every single study on this topic can be “viewed as a direct example to follow” and should be carefully processed. Yet, there are a few objections about the manuscript I would like to stress.

• I took me more than two hours of careful screening the methodical approach of the original (Rowse et al., ) and this manuscript to extract the primary information about the time interval between both measurements. Finally, I went to reference 35, cited by Rowse et al., to read that “Mean time interval between recording bat activity at each pair was 49 days (range 14–77 days).” As this is for readers crucial information to trust the data, it should be mentioned not far in the text than in the abstract. Otherwise, the bat activity between May and October changes dramatically, and these two seasons cannot be compared. The bats are seasonal animals, in GB latitude they have to adapt to a big difference in day/night ratio between summer and winter, after summer season, they will be adapted to the distinct spectral composition of sunset and sunrise, which may affect their sensitivity to artificial light, during mating and breeding period, there can be additive motivations to foraging that help them to overcome the higher light level. Which can, however, mean a higher level of stress, and compromised fitness ... I would like to summarise, that the activity between control and experimental groups should be done either as close as possible or the year after in the same season. I would suggest comparing at least the autumn and spring data if they are available to authors. Otherwise to formulate some study limitation in the discussion.

• I think that besides total activity analysis, the temporal distribution of the activity over the nocturnal period could be interesting as well. Unfortunately, as I look to the supplemental data of the original paper, the precise time of data collection does not seem to be available. The activity/rest cycles are primirily driven by the circadian clock which is sensitive to light and dictates the temporal phase to all processes in the organism, from the biochemical up to behavioural level. The time of foraging is dictated by the clock, which is synchronized by the light/dark cycle so well, as reliable is the contrast between day and night. The repulsivity or attractivity to light is the direct effect of light, which is secondary to the physiological impact processed via the circadian clock, although it is better visible and measurable in the field studies. If the data were available, one could expect the phase shift of the activity to late-night hours and the shrink of the activity period more to the middle of the night. The danger expected from such a situation can be, for example less available forage due to the night cold...I don't want to speculate too much, just want to illustrate the importance of the temporal resolution of the data over the circadian period. The effect of ALAN on the circadian system concerns insect as well as birds and human and plants. In my opinion, the range of adaptation limits of each organism to the level of nocturnal light is the question that should be asked by ecologists and physiologist as well, and they should cooperate. To start I recommend this short paper http://dx.doi.org/10.1098/rsbl.2016.0015

• The introduction should be significantly shortened. The authors deal with once-published data; this manuscript should be considered more as a revision as a new manuscript. All the information about the effect of ALAN on biodiversity are ok, but here it creates the message that “there MUST be the difference and we will make all efforts to find it”. In my opinion, the last paragraph in the discussion is sufficient to show the public impact of the results. However, the results of the analysis should be reflected in the title. The expression “disentangling the relative effects…” say nothing and may repulse an important part of the readers from further reading.

• In the discussion, the sentence starting on the lie 278 contradict the statements “LED that includes a lower amount of wavelengths attractive for insects” and a similar statement in the introduction
• Citation 44 does not correspond to the text, Lewanzik and Voigt did not study attractivity of LED lighting for insect - this information flow should be clarified.
• The conclusion should summarise the main findings.

Author Response

Please see the attachment

Author Response File: Author Response.docx