This study builds upon the Integrated Methane Inversion (IMI) v1.0 framework, advancing it to IMI 2.0 for monitoring total methane emissions with high resolution worldwide using TROPOMI satellite observations. The major advances involve the use of new observation data, improved consideration of inversion parameters (including state vectors and errors) and processes. These updates incorporate the outcome from a series of studies on methane inversion from the community. The capability of this new tool is demonstrated by a case demo of US methane emission. Overall, the manuscript is well-structured and well-written. It can be published once the following points are clarified.

Line 262: “we modified GEOS-Chem to simulate multiple methane species as separate transported tracers” It appears to me that the GEOS-Chem methane simulation only includes one tracer (methane), what does “multiple methane species” mean here? Are you referring to mean methane from different source?

Line 272: “Precompiling emissions with HEMCO yields an additional 2× speed-up, for an overall 10× decrease in CPU cost and a net decrease in wall time. U” Could you explain what does “precompiling” refer to?

Section 3.4. I am a bit confused about the term “super-observation”. It looks to me that the number of observations is reduced, but “super” mispleads as “more and stronger”.

Line 304. It mentions “three” updates but there are indeed four processes.

Line 335: I am curious about the extent to which boundary conditions would be adjusted in the inversion process if smoothed BCs were employed. Have the you tested this in the demo inversion?

Line399：Could you provide a test demonstrating how OH optimization influences the regional inversion results? Why was OH optimization not utilized in the US demo?

Table 2: Could you provide a suggested values for the regularization parameter? Is it sensitive to the number of observations and state vector?