Using new noninvasive technologies (accelerometry) to record motor activity, the authors examined the influence of environmental factors on the total amount of sleep, as well as its quality and efficiency in wild boars. Particular attention was paid to the variability of sleep parameters. The duration of recording was up to 1 year. The study has determined for the first time the correlations between total sleep amount (an estimate based on accelerometry data) and environmental parameters (temperature, humidity, precipitation). The introduction and discussion sections are very well written. They are easy to read. It seems to me that the description of the modeling method may need to be explained in some simpler language. Please see my specific comments and concerns.

It is a pity that the analysis of correlations was carried out for the total amount of sleep, without an attempt to subdivide sleep into stages or based on the sleep postures. SWS and REM sleep do not depend in the same way, for example, on the environmental temperature. In horses, cows and deer, REM sleep is usually recorded in lateral recumbency. If boars exhibit REM while in lateral recumbency, then it would be interesting to assess the sleep duration in two postures (sternal or lateral) and evaluate the correlation between sleep duration in the two postures with environmental parameters. Based on the figures, I suggest that the sensor’s accuracy in identifying the postures was high. Then, it could be done. In the MS REM sleep was mentioned only in Introduction and Methods... In any case, data on sleep amounts in different positions over the study period would be more interesting than for total sleep time. A discussion addressing limitations of this study, as well as potential future perspective, should be included. What I am talking about does not call into question the approach and established correlations and their discussion in this version of the MS.

Everything about modeling seemed very complicated to me. I couldn't figure out all information in Figure 2 and in tables s2-s10 in the Supplementary material. I focused on the numbers in the text of Results. I think I won't be the only one. For example, line 264, what the numbers in square brackets (-0.12 [-0.24, -0.01]) mean? I wish it was explained in simpler language for those who, like me, do not familiar with the approach.

Line 116. The word “Ideal” is no accurate. Accelerometry certainly provides new opportunities such as longer and noninvasive monitoring. The accuracy of sleep scoring depends on how the postures are associated with the sleep stages in a species (see my comments above). This association needs to be validated in a combined EEG and accelerometry recording.

Line 119. Please provide references to the EEG studies in pigs.

Lines 129-130. Sleep is characterized by depth through the arousal thresholds (and the EEG power for SWS). I wonder how the duration of episode alone could be a feature of sleep quality? Drowsiness and SWS in pigs as reported comprised 16% and 27% of 24-h, respectively if 24-h (Robert and Dallaire 1986). I don’t think the episode duration is selective to the drowsiness to SWS ratio. Some discussion on the limitation of this approach (duration of episodes as a measure of sleep quality) may be added.

There is no information pertaining to the accuracy of the posture identification based on accelerometry while Figure1, S4-S5 suggest that it was high.

African elephants in the wild have been reported to be on the move for up to 46 hours. Were there any periods when the boar did not sleep for more than a day?

The differences would be better visible if Figures s4 and s5 were parts of the same figure, and aligned horizontally next to each other, and the scale was same.

Figures 4 and 5 show accelerometry data for lateral and sternal postures, although time in different postures was not separated. Why show the accelerometry for 2 postures if the time spent if these postures were not quantified?...

The authors responded to my comments.