The article is devoted to measurements of d¹⁸O and O₂ concentrations in the atmosphere by absorption of DFB laser emission at 760 nm. The excellent sensitivity is presented in short (20minutes) and long (several hours) times. The results of measurements are in nice agreement with that of IRMS. The methods of calibration of the device are suggested for continuous monitoring of O₂ concentration and d¹⁸O.

At the same time many details of the experiments are missing.

• Experimental setup is very useful for understanding further.
• It seems the device long time stability obtained owing to high stability of cuvette (temperature of ~mK fluctuation). Nevertheless, no info in the text about the actual reasons for the stability was provided (precision of measurement, response time of feedback).
• Far wings of absorption line contour are determined by Lorentz (apart from pure Doppler) in all models (Voigt, Rautian etc.). The difference in the applied model in the “main body” of line contour (O¹⁶O¹⁸) should be shown. One can estimate using cavity parameters the number of points (cavity modes) on that are not more than 10. (It is possible to see also the fluctuations around O¹⁶O¹⁸ line contour in fig.1.) The question is – is this enough to get the difference in calculation? In text lines 187-203 it is very difficult to understand how it was done. Apart from temperature influence on cavity length mechanical instabilities (for instance by outer pressure fluctuations) should be mentioned.
• Lines 206-215. Why the reflectivity of the second configuration is absent? “Less parasitic fringes” at lower finesse configuration means better spline/averaging of the signal (FSR cavity modes did not change). If it is so, single scan time of the laser frequency and time constant of the detector provided.
• Line 304. “An influence of O₂ concentration on d¹⁸O of O₂ was expected.” Sound like a general rule, but is only valuable for the method applied.

This paper reports the development of an optical gas analyzer for high temporal resolution and high precision measurement of δ18O and atmospheric O2 concentration based on the optical-feedback cavity-enhanced absorption spectroscopy (OF-CEAS). The results were compared with the isotope-ratio mass spectrometry (IRMS) and showed good agreement.

General comments:

1, Although this work was based on a commercial device (AP2E) and similar experiments have been performed elsewhere, the experimental details still require detailed explanation.

2, The conclusions of this paper need to be supported by solid experimental data, rather than just stating the conclusions.

3, A detailed review and comparison with other methods, especially spectroscopic methods, should be made.

Specifical comments:

1, Page 2, line 32, the argument should contain high accuracy.

2, Page 3, for O2 concentration measurement, some other spectroscopy can also achieve ppmv level detection limit, such as Opt. Express 20, 2927, 2012. More discussion and comparison of other high sensitivity laser spectroscopy method is needed.

3, line 80, the difference between two models and parallel measurements is unclear.

4, line 99, first OF-CEAS in the visible range, incorrect. Salter et al. (Aanlyst, 137, 4669, 2012) performed optical feedback measurement with 635 nm diode laser.

5, Page 4, a detailed description of the experimental setup is required.

6, line 136, more discussion about the tuning coefficient and the intensity noise is needed.

7, Page 7, page 8, a detailed description of data processing methods is required. The corresponding experimental data need to be presented.

8, line 258, is a data acquisition time of 10 to 20 minutes fast enough for concentration and isotopic measurements?

9, the influence of water vapor requires experimental data.