This paper uses satellite and ground-based measurements of methane columns to estimate emissions from three landfills in Madrid.  It is innovative in its approach and in the data sets that it uses.  But I found the paper to be far too long relative to its information content, I was very confused by the presentation, and I found the results unconvincing.  I cannot support presentation in present form and would be willing to review a revised manuscript only if it is considerably shorter.

1. The Introduction takes a very long time to get to the point, and at the end we still don’t have a clear statement of the problem. Is the paper to quantify emissions from Madrid, or from three landfills, or is this the same thing? Why should we care about Madrid? What is actually new here?
2. Section 2.1.2: give reference for the specific TROPOMI product that you are using. TROPOMI has a very low success rate (3% globally), is that an issue here? Or is Madrid sunny enough?
3. Section 2.1.3: this combined IASI+TROPOMI TXCH4 product is probably totally dominated by TROPOMI information in the PBL, which is what matters here. So how is it independent from TROPOMI?
4. Figure 1 is very difficult to read.
5. Section 2.3: this section is very confusing because it is not clear what the authors are trying to optimize. Are the ‘daily plumes’ for the individual landfills? Are they summed over the three landfills? Are the three landfills treated as a single plume? The cone model is surely wrong for instantaneous plumes but is reasonable for time-averaged plumes, which is what is fitted but it takes the paper a while to explain this.
6. Line 206: I don’t get the point about seeking an analogy with NO2. The landfills don’t emit NOx, NOx is an area source, and the decay of the NO2 plume is by oxidation rather than dilution into background.
7. Wind speed is denoted  v in the text, w in Figure 2.
8. Equation (9): not clear how you get y_BG
9. Section 2.3: there are many uncertainties in the procedure for inferring emissions. How can it be validated? An obvious way would be to use the independent COCCON observations to evaluate the posterior concentrations resulting from the TROPOMI inversion.
10. Figure 3: the agreement between TROPOMI and COCCON in that Figure strikes me as very poor, despite the authors’ claim to the contrary. I’m not surprised by this in view of the known TROPOMI biases, but it undermines confidence in the results of the TROPOMI inversion. The paper goes on about the problems on Sept 25 and Oct 4 but that seems anecdotal and those two days don’t seem any worse than the rest of the population in Figure 3.
11. Lines 329-330: how do we know that the ‘COCCON instruments show a very good ability to detect the source’? No specific results or data from COCCON are shown.
12. Figure 8 is cryptic. What domain is shown? What are we learning from it?
13. Table 5: I don’t see the relevance of this Table to the paper.

Comment on

“Quantification of CH4 emissions from waste disposal sites near the city of Madrid using ground- and space-based observations of COCCON, TROPOMI and IASI” by Q. Tu et al.

The Work uses Ground based measurements using 5 EM27/Sun spectrometer during a two weeks intensive field campaign in 2018 together with  3 years of Satellite borne measurement from  TROPOMI and IASI instruments to determine the emission of Methane from waste deposals sites near Madrid.

NO2 measurements retrieved from TROPOMI was used as tracer of urban emissions from Madrid, to characterize the daily dispersion and fit a simple plume model, valid for the region around Madrid at a specific day.

TROPOMI and IASI retrieval from three years have been combined to an improved product. Which allowed separating upper tropospheric methane and the methane in the lower Troposphere, which is influenced by the local and regional emissions.

Ground based measurements show enhancements up to 4% during certain wind directions in their time series and give clear evidence of huge CH4 fluxes from the disposals and sources nearby.

For the quantitative reconstruction of the emissions of Methane from columnar CH4 enhancement downwind of the sources, a plume model is applied and fitted to the Satellite born CH4 measurements.

The wind direction and average wind speed is obtained from a reanalysis model, while the broadening of the plume is estimated by an angel of 60º, a geometrical factor which has been determined from the space born NO2 observation of the same satellite.

The ground based measurements validate the space born measurements and the strategy to use the wind speed at 10m to calculate the CH4 flux and permit to understand that the enhancement measured by a single TROPOMI measurements is typically an average of strong enhancements and no enhancements in the same pixel-area viewed by the TROPOMI-instrument. The use of IASI data help to increase the sensitivity of TROPOMI in the lower atmosphere and therefore improve the estimation..

The paper contains several new findings, new techniques, finally quantifies CH4 emissions in a clever “Top-Down” approach based only on remote sensing measurements from space and grounds. The Top Down measured/estimated emissions are significant higher than the official bottom up estimation.

In my opinion the paper is well written and should be published in ACP in its actual form after addressing minor corrections and the specific comments given by reviewer 1 and here below.

All new developed methods are necessary to reach this goal, but as they are new it is a little bit work to understand them and the paper is getting a bit large, as mentioned by the Reviewer 1. The Introduction gives extensively credit on former works and is therefor also an important part, which should not be missing even it might not help the reader in understanding of the new technics.

Therefore I would ask for a kind of road map at the beginning of second section ”method”, before 2.1 which gives an idea, why and how the measurements and models are combined, or maybe it could be included directly at the end of introduction where the different sections are already mentioned.

Specific comments:

TROPOMI+IASI:

You report two estimations in the abstract, which are slightly contra dictionary. The error in the estimation is higher than the error estimation from the combined retrieval, I imagine that this is more or less the effect that partial columns are constructed with more errors (Error of TROPOMI + error of IASI), but there should be another systematic error, which is reduced if the l tropospheric partial column anomaly is sub or over estimated. It would be more easy for us readers, if the authors conclude on a single best estimation, and report the other option using just TROPOMI just in the results and discussion of the article.

The combination of IASI and TROPOMI assumes measurements of the same air mass, and therefore at the same time, but IASI comes around 10:00 and TROPOMI around 13:00, does this matter here, o IASI characterize more the background CH4 and is therefore less critical. Maybe it would be interested in a discussion how the time difference between IASI and TROPOMI of some hours, might affect the combined retrieval.

Please also describe a bit more how much the estimation improves due the combination with IASI, I could imagine, that this advantage will increase if you use long term averages, so that the random error loses its importance in relation to improved sensitivity in the lower atmosphere.

Plume model.

The authors realize their proper idea and do not have to cite other works, which might be partly similar ideas, but it would help some reader to understand it easier, which is similar to other approaches (up,downwind Rotating method, Gaussian's Law).

Do I understand correctly the point, that the plume model is retrieved using the City centre of Madrid as central source, but applied to the source located at a different location (landfills). So orography should be similar or not be important. I would have been interested in the requiered conditions on the regional topography for this approach to  be transfer also to other cities or waste deposals sites. And I  also would like to know  the  typical mixing layer heighin the area around Madrid .

The test using NO2 from TROPOMI results in a lower emission rate compared to estimation from the literature. The authors identify the lifetime as the reason and maybe also the value in the literature might have errors, or might not be comparable, different time ....  However a validation of the wind speed estimation from a 10 m altitude would be nice.

For CH4 and its long lifetime the anomaly in the column should be definitively given by 1/d, as described by the equation 3.

The forward model contains the velocity and actually the quotient emission/velocity determine the concentration, or vice versa in the inversion. That is why the wind speed estimation might be so important. The sensitivity study using the wind at 10 m is very helpful and important, never the less I did not get completely get the point how you decided that 10 m is the best. I think the wind velocity you look for is the velocity you can multiply with the column anomaly to get the total CH4 flux. Ad hoc I would take the average windspeed in the atmospheric boundary layer.

The COCCON sites at “ Jose Echegaray “ and “Barajas” have a distance of more or less around 8km, on 25 September 2018 ( Figure 4) we see that anomalies en CH4 arrives around 1-2 hours later in Barajas than it appears in “Jose Echegaray”.

So the velocity of propagation of this plume might be 8 km/2 h to 8 km/h or 1.11m/s to 2.2 m/s respectively, this fits actually very well to the assumed and modelled wind velocity in the plot below. Surely that is the intention showing the plot, but maybe it should be also explicitly be mentioned, that you can also use the 5 FTIR sites to validate the wind estimation strategy you have chosen, as here the interest in the effective flux of CH4 and it might be the best validate the velocity using the FTIR EM27 measurements of columnar CH4, so that you are independent of the vertical distribution of CH4.

EM27 measurements take typically a little bit less than a 1 minute, maybe you could do a cross correlation to retrieve the delay after interpolate data to a common 1 minute grid, and then include a point in the figure showing the wind speed or projection, but it is sufficient just to mention it.

Page 2:

Line 37:

“The wind-assigned plume method is also applied to the tropospheric and upper tropospheric/stratospheric column averaged CH4 mixing ratio products (in the following referred to as TXCH4 and UTSXCH4) derived from a-posteriori merged Infrared Atmospheric Sounding Interferometer (IASI) profile and TROPOMI total column data.”

Maybe you could split the sentence into 2  and find somehow a different description, which is easy to understand:

For the CH4 emission estimation, the wind-assigned plume method is applied to the lower tropospheric methane /dry air column ratio (TXCH) of the combined TROPOMI Infrared Atmospheric Sounding Interferometer (IASI) Product.

TXCH4 and the upper tropospheric/stratospheric column averaged CH4 mixing ratio (UTSXCH4 ) are derived from a-posteriori merged Infrared Atmospheric Sounding Interferometer (IASI) profile and the TROPOMI total column data.

Does it make sence to apply the method to the UTSXCH4 Product ?

As mentioned ealier I would just use TXCH4 in the abstract and conclusion.

Line 40:

The first sentence is a bit redundant, maybe you could drop here the sentence “Based on the NE and SW wind fields, we developed a simple plume model locating the source at three waste disposal sites east of Madrid for CH4.” or move it to line 37.

Line 44:     “day and. All em…

Maybe you could rewrite the sentence which is also not clear .” The COCCON observations indicate a weaker CH4 emission strength of around 3.7×1025 molec s-1 from local source (near to the Valdemingómez waste plant) in accordance with observations in a single day and. “

Please write very clear if COCCON is used to estimate an independent “local source”, if this source is part and included in the Tropomi estimation given quantitatively in the line above. Please state if TROPOMI and COCCON based estimations are contradiction, complementary and/or consistent.

If you want, you might move the last sentence to line 25, as it is finally the most important finding.

Figure 8,9:

Please use a fixed color scale in Fig 9 and maybe also in Fig 8. The modelled XCH4 for total column, lower and upper troposphere seems to be exactly the same in fig 9 b,e,h. …. .

Maybe when you talk about emission rates line 474 and line 477 you cloud use CH4 and not the dry air mol fraction.

XCH4-> CH4