From Aaron Cahill

This study describes an investigation that sought to evaluate the integrity of legacy oil and gas wells with a view to identifying leakage through fluxes of CH4 from soils into the atmosphere. The authors have used a powerful combined approach utilizing soil surface flux measurements, depth discrete soil gas concentration measurements with stable carbon isotope analyses and evaluated the soil microbiome to better understand the sources and cycling of CH4 at their sites which were dominated by peat/organic rich sediments. The authors describe how due to the presence of peat at their sites that a strong underlying methane cycling system was active that acted to make it very difficult to easily identify anomalous CH4 and potentially mask any leakage from the wells. That being said all data seem to indicate that the observed high levels of CH4 observed are almost certainly related to the peat rich settings and not from well leakage. This is important as >15% of the >20k onshore legacy wells in Germany are located in such peat rich settings. Consequently, being able to evaluate legacy well integrity against a backdrop of active methane cycling is of great value. It seems as though the study, which perhaps set out to evaluate well integrity and leakage, has pivoted to having to unravel the peat system and its methane cycling. Thus it seems to focus more on those aspects (as a necessity) rather than the wells and their integrity themselves. and in doing so is a very valuable case study for others to learn from where legacy wells are hosted in similar highly active methane cycling environments

Overall, the study has been well conducted with excellent fieldwork and analytical efforts. I commend the authors particularly for the nice addition and use of microbial data in trying to unravel the puzzle which is often missing in such studies.

Overall, the article is of good quality but some limitations in the written language which could do with improvement at multiple points throughout. The figures are of good quality, but a few tweaks could make them even better (See comments below).

I recommend publication after moderate revisions and addressing the following comments.

Comments to address as follows:

L36 -39: The discussion on what abandonment means is a little inaccurate and could be improved. In my understanding there are various definitions in different jurisdictions, but abandonment itself is a part of the decommissioning process where the well itself is first sealed/flow zones are isolated with plugs. In fact, the abandonment process (which is part of the decommissioning process) can then have several phases such as in the UK where there are 5 states of abandonment that are enacted depending  on the well and what is being done with it. See https://www.nstauthority.co.uk/media/1100/operator_work_instructions_v3.pdf . Wells can also be suspended which is a different status. Next there are other parts of the overall decommissioning process where the well head is removed, cut capped and buried (such as the wells you examine here) and in Canada for example the well is then subject to a reclamation certificate. So, you need to be more specific on the generalized process of decommissioning in a generalized sense recognizing its phases and different potential statuses. Even though the details can vary by jurisdiction, more if not all western regions follow a similar process now. Then after you do this generally it perhaps makes sense to use your study area (i.e. Germany) and describe in detail the process and definitions. In my experience no jurisdictions these days allow a well to have the head remain and call that decommissioned – it would be termed suspended or similar. I think all jurisdictions these days aim towards full decommissioning and reclamation of the land... I think your description needs improvement in general.

L32: Again, I think this is a little inaccurate as other countries (most western ones) do require full reclamation including cut and buried status. If you expand more accurately as instructed above it will improve the accuracy of the manuscript.

L44-48: The problem you are identifying is not just for methane rich environments, but it is a result of trying to detect leakage in a dynamic and active zone of the soil system. In shallow soils, many physical and bio-geochemical factors can act to alter/attenuate any signal being measured or sought after on timescales of hours/days, meaning it might lead to false positives, positive false (and/or just make it difficult to see what is being looked for). I think you should expand this a little to explain more clearly upfront. A subset or very specific issue with this involves systems that are methane rich which are more prone to generating false positives and that is something you are particularly examining here as you have peat rich soils. I feel your introduction needs some more general discussion around the factors that influence observations associated with soil gas and efflux in shallow soil systems and oil and gas well integrity...

L50 – 60: Here you almost switch gears from oil and gas wells to a very detailed description of wetlands and peat bogs. I understand that some of the wells you examine here are located in such settings, but I feel like it would be beneficial to contextualize how common this is globally (i.e. for oil and gas wells to be located with peat bogs/wetlands at surface). Needs some grounding to show this is a typical environment that wells are hosted in and so is likely to be a common complexity – also to point out all soil systems whether they are wetland related or not are complex living, breathing systems and all have processes that might make evaluating leakage difficult without wider characterization and contextualization of processes... I think being more general, but using peat bogs as a particular case study, is a better general approach..

L74-76: this is true for most/all settings – unless leakage is obvious! Which can often be the case in my experience (i.e. a leakage signal is massively overpowering and obvious to se with large fluxes and concentrations compared to reference sites). I think in general redrafting the intro to be more general and acknowledge the universal complexities would be a better way to frame this with the idea of peat bogs and their important in Europe/Germany (and elsewhere) properly described/acknowledged...It’s not clear currently how common it is for oil and gas wells to be hosted in such settings and so hard to gauge importance/significance...

L80 – 83 – here you now address some of the above clarifications around how common this is (i.e. 15% of wells in Germany) but I feel this needs to come earlier (i.e. before the description of peat bogs etc) to justify why you focus on this so much....

L83: “Such soils are highly likely to produce and emit methane” – this is true but perhaps a little too simplistic as how much is generated and how it is observed will also be variable depending on the seasons/soil conditions and other physical parameters – I feel this needs to be a bit better explained in the intro as per my main comment.

L84-85 – Isn’t a key focus also to determine the integrity status of the wells being examined? Seems like you should state that here.

L100 – 103: can you more clearly define the forest and meadow sites – some way to more rigorously quantify them perhaps with soil types or land use classification. Seems a little qualitative to state them this way – just wondering if you can be more accurate/robust and descriptive here

L106 – What do you mean “visible”? human eye visible? Likely the human eye cannot always capture all remnants of drilling. In my experience many drilling sites have altered and potentially contaminated soils (very old contamination perhaps not obvious) and so generally I would say that samples for lab analyses are needed to be more quantitative here. So, samples for TPHs, org c etc. and N to show the soil health and to compare this to the associated background locations. At the minute this is a very qualitative and loose statement. Can you be more quantitative?

Figure 1: Somewhat strange you provide results in the methods section and in figure 1 (i.e. methane sink or source) and seems like this occurs prematurely before you describe the methods for how measurements are taken etc. I suggest you need to remove these results and have figure 1 as a more generalized figure setting the scene for the whole study (without results included). I recommend a figure of all of Germany showing all wells and those which are in peat areas (i.e. the 15% you say that are) and then an inset of the area of interest showing no results and a combined inset of the generalized geological soil profile. Save these results for the results section..!!

Section 2.3: Flux measurements – did you determine an R2 value for the quality or stability of the flux measurements (i.e. how correlated with time the increase is – how smooth it is) – in my experience this is a very important parameter to evaluate and mention as it highlights how reliable the flux measurement is and potential methods of gas migration/source.... Also what about other weather parameters? Even if you didn’t measure at the site can you get local weather station data to allow contextualization of the general weather patterns during measurements (e.g. temp, wind speed and barometric pressure cycles) – these are important to contextualize... and nice to mention and include...Important to show a single example of flux measurement data

Section 2.6: It is not immediately totally clear to me how oxidation rate is calculated – can you specify exactly how the rate is derived with the equation?

Section 3.1: This seems very much like methods to me and I think it needs moving into the methods section—it is setting the scene of the site/location investigated and not actually results. Please move such sections to the appropriate place...The results seem to start after L294....

L294 - 296 – Firstly are these 206 measurements from both wellhead locations and reference sites? Please clarify this. Next, I have some concerns with how this is described in general that makes it hard to follow and potentially a little misleading. It should be expected that soils emit and take up methane naturally and that typically in natural soils we get close to (but not typically exactly) net zero (i.e. +/- 0.05 µmol/m2/sec CH4 flux) – good to be clear on that as what you see falls within natural ranges – in my view you need to describe and contextualise natural ranges better in the article to help the reader understand what is normal and what might be anomalous.... Next, if this is all measurements (i.e. 206 from wellhead and background lumped together) you need to differentiate these and then compare well head and reference location by site – I find not differentiating makes describing the results in this way limited. Also to refer to the lowest flux as –ve is not really correct. A negative result is CH4 uptake, not a small flux which would still be positive.

L298 – 300: These are all within typical natural ranges to my knowledge (i.e. are not at all indicative of anomalous processes or leakage) – so you need to better describe what natural ranges are and then state these are within them...

L301 – 305: Again – these all seem to be within natural ranges for such settings – you need to state that and better articulate typical natural ranges in the paper.

L306-312: The >500nmol value seems higher and potentially approaching the outer boundary of typical natural ranges – again you need to define better what are potentially natural ranges for peat bogs/similar settings. Here as well you refer to “high” fluxes – but what is this relative to? – this value is still quite low compared to other reported leakages at oil and gas wells (see Cahill et al Evaluating methane emissions from decommissioned unconventional petroleum wells in British Columbia, Canada for example). In fact, this also highlights a lack of review or identification of the types of rates that are observed when leakage is happening from a well. In general I think you need to better review and state this in the intro and also then refer to it here when comparing your results. Other good papers to mention/review that are very similar to what you do here (but are not acknowledged) are Samano et al Constraining well integrity and propensity for fugitive gas migration in surficial soils at onshore decommissioned oil and gas well sites in England and Forde et al Identification, spatial extent and distribution of fugitive gas migration on the well pad scale – among many others. I think what you see here is not indicative of leakage but to say that with confidence you need to robustly describe what leakage looks like as well as natural ranges...

Table 2 and figure 3 would benefit from inclusion of typical natural ranges for the soil/land types and also perhaps inclusion of reported values for leaking oil and gas wells to show how these values compare......

Figure 4 – there appears to be no differentiation between reference sites and well heads in this figure – suggest to show that for clarity. Also again seems like have some indicators of natural ranges (e.g. atm concentration also) on this figure would allow the reader to better judge how the observed results deviate from that.

General comment on the results – I am quite interested to see the high levels of CH4 in soils in the meadow but guess this is a result of the underlying Peat? Overall seems like this strong natural CH4 signature would make identifying leakage much more difficult than other settings. I also wonder what the mixture of potential thermogenic gas and natural peat derived gases would look like. It would be beneficial to create a simple mixing line to show this and then plot the observed gas concentration on the same plot to see if and where they fall in relation to this mixing line. Can this be done?

Figure 5: really nice microbial data. It seems you do not delineate between reference sites and well heads in this figure --- is this correct? So, this data is for both reference and wellhead locations? Seems like it would be useful to show somehow this delineation in the results.

Section 4.1: A good overall discussion. However, would point out a very recent study showed also the extent of methane oxidation that can occur in soils around leaking wells (in Canada and where it was also cut and buried) this study would be useful to mention as it supports everything you are saying here (Cahill et al. Evaluating methane emissions from decommissioned unconventional petroleum wells in British Columbia, Canada). As for factors affecting spatiotemporal variability – I think this discussion could be improved with more references that demonstrate these factors. For example, Forde et al “Barometric-pumping controls fugitive gas emissions from a vadose zone natural gas release” showed how barometric pumping controls surface manifestation of leakage and fluxes very robustly and this seems like a paper worth mentioning. Was it not possible to get weather station data to evaluate against your results in this study?

L630 – 632- I think some would disagree with this statement. For example, British Columbia Canada actually has very good regulations and most development occurred since the 1960s (with most since 2000) – so in that case they have mostly plugged and abandoned wells.

L644 -648 – Again this was also thoroughly examined in Cahill et al “Evaluating methane emissions from decommissioned unconventional petroleum wells in British Columbia, Canada”. It might be worth comparing your numbers to those observed in that study….

The authors present a field and lab study on the biogeochemistry of methane in peat soils  as related to potential leakage of deep, abandoned oil wells. The study is fine with respect to the field and lab measurements done and the authors show they are at home in the topic of biogeochemistry of methane. The peculiar aspect of the manuscript is its framing. The authors refer to the need to know background emissions from the shallow subsurface when it comes to potential leakage from the deeper subsurface. I fully agree with this need but several remarks must be made:

1. the title is misleading as it is very general. It should be indicated in the title that this study deals with peat soils as they have the highest potential for methane emissions (probably together with paddy rice fields)

2. the authors were unfortunate that no thermogenic methane was found at all. However, they stick to searching for thermogenic methane which makes the manuscript in a way forced in its scope. Section 4.2 is a peculiar read as well as other parts of the manuscript.

3. the study is relevant when it comes to greenhouse gas emissions and climate change. This comes around the corner at a rather late stage in the manuscript. The scope of section 4.3 should be introduced much earlier.

4. the authors studied peat soils above abandoned oil wells. This is a somewhat poor choice since oil reservoirs often have lower liquid pressures than the surrounding rocks due to the exploitation for oil. Hence, there is oil and/or water flow from surrounding rocks to the  reservoir (often not completely depleted). Depending on the exact composition of the field, there may be more or less natural gas involved in oil reservoirs which has a tendency to move upwards  because of buoyancy effects but this may be restricted depending on the resulting pressure field. It would have been much, much better when the researcher had studied peat soils above abandoned gas wells. It seems to me that the authors are not aware of this major difference. They however should indicate this essential difference. Here, I realise that local geology plays a role in the way reservoirs may be connected or not to above as well as potential leakage of gas or oil along or through (abandoned) wells.

Especially based on remarks 3 and 4, I suggest that the authors rewrite the manuscript where they present a new framework addressing both greenhouse gas emissions from peat lands and dealing with shallow biogenic methane emission when addressing potential leakage of thermogenic methane. If not, I must recommend to reject the manuscript.

In addition, I have the following general remarks in addition to individual remarks annotated in the manuscript:

1. carefully check on the use of single versus plural. Subjects and related verbs are frequently not in harmony.

2. the RESULTS become boring to read when presenting all kinds of numbers. These numbers should be presented in Tables (as done) and patterns should be discussed in terms of higher, lower, etc. Very importantly, it remains unclear whether the different sites differ statistically significant or not. I recommend that significance tests get added and box-whisker plots get presented (instead of means or so in tables that may be moved to supplementary material).

3. the  DISCUSSION presents a lot of results which should be avoided. Move more effectively to the discussion topics.

4. the CONCLUSIONS are poorly written. It reads more like a summary than conclusions