Overall, the paper is an interesting report on a short measurement campaign at a measurement station with the aim of demonstrating that photoacoustics with clever evaluation algorithms can indeed be a sensitive and cost-effective alternative to expensive established analysers. 

The paper, however, gives the impression that it seems somewhat pieced together. This is noticeable, for example, in acronyms that were not introduced in time or the justification why the sample gas for the PA sensor has to be humidified.

The task for the authors are, from my perspective:

- Make the text more consistent overall, explain acronyms and special terms briefly, even when referring to corresponding papers, so that the reader gets all the important information without having to jump to other papers first.

- Please take into account the comments that I have included in the attached supplement (pdf). 

- Please use the typical format defined by copernicus.org for literature references.

1. Content

In the presented manuscript data of ambient trace methane measurements made with a photoacoustic spectrometer over period of several days at the site of a meteorological observatory is described. Applying previously published methods for photoacoustic signal correction, determined concentration levels are quantitatively compared to a high-end cavity ringdown reference instrument, showing agreement within +-100 ppbv over the whole measurement period.

1. General Comments & Impression

The overall impression of the manuscript is good to fair. While the comparison of a photoacoustic instrument to a cavity ring down instrument with comparable accuracy at trace levels over an extended period of time is well worth publishing, the manuscript may benefit from including some information about the applied signal corrections as well as some more details about the measurement location. Also, proper initial introduction of the main correction algorithm (CoNRad) with references, and shortening the general discussion about photoacoustic spectroscopy and the introductory limit of detection discussion, would help conveying the key points in the manuscript.

1. Specific Comments & Questions

• Line 16 & 19: Initially, the targeted accuracy of 2 ppbv is mentioned. After that, the 3 sigma precision of the cavity ringdown instrument is quoted, without going into detail about the device accuracy. What is the long-term stability/accuracy of the CRDS device and is the device calibration traceable to some standard?
• Line 17: What is the targeted measurement rated or maximum averaging time for the specified 2 ppbv accuracy?
• Line 69: This is the first mentioning of the algorithm CoNRad and without any reference or detailed explanation, making the reader wonder about the significance of this method to the presented manuscript.
• Line 75: The authors mention “ppbV-level-precise GHG monitoring”, while accuracy may be more relevant.
• Line 84 to 94: The general discussion about excitational relaxation losses for methane in ambient air is misplaced in the section about the photoacoustic sensor and should be moved to the introduction or included in a theoretical section.
• Line 103: What are the uncertainties of the concentrations in the reference gas cylinder? What is the specific reason for including 312 ppmv CO2 in the reference gas? Please specify the volume fractions for the components of “dry natural air”.
• Line 105: How were the seven reference gas measurements used to “avoid” sensor drifts? What were the differences in measured and true reference gas concentrations? (quantitatively) How high was the deviation of the measured concentration of the reference CRDS instrument to the concentration of the reference gas? This information would be beneficial also in Table 1.
• Line 135: What processing and corrections have been applied to the “raw PA data”?
• Line 138: Does CoNRad only compensate for the efficiency of non-radiative excitational relaxation? What other effects have been compensated for?

1. Technical Comments & Suggestions

• Line 8: As relaxational effects and relaxation time constants in photoacoustic spectroscopy and spectroscopy in general are manifold (hydrodynamic, excited state, etc.), I would suggest specifying the type of relaxation more precisely whenever possible.

• Line 16: Is there a reference publication for the specific CH4 instrument requirements agreed upon by ICOS?
• Line 23 & 25: The statements about methane concentrations “by up to 2.7 ppmv” and “by about 40 ppmv” are incompatible.
• Line 43: The cited equation (1) should describe the sound pressure amplitude for harmonic excitation.
• Line 48: Is N_i really the volume ratio or is it the volume fraction?
• Line 49: Is P_0 really the optical power or is it the optical power amplitude of the modulated light source?
• Caption of Figure 1: Abbreviation “DWD” is not defined.
• Line 103: Perhaps the term “reference gas” is less confusing than the term “target gas”.
• Line 108: Only white noise is mentioned. Is the argument deliberately limited to white noise?