Flaring efficiencies and NOx emission ratios measured for offshore oil and gas flaring facilities in the North Sea – Review

Summary of Paper

            The authors present analysis of emissions efficiencies and ratios for four different gases from North Sea oil and gas flaring. They have used data from two flight campaigns and measured 58 flaring plumes using the FAAM aircraft between 2018 and 2019. By using combustion efficiency and destruction removal efficiency calculations the authors show that previous assumptions of emissions from this sector broadly agree with field measurements. The calculated combustion efficiencies are between 94 – 100% and the authors demonstrate that including ethane in combustion efficiency calculations has a very small impact on the final result. They also show no statistical correlation between wind speed and combustion efficiency but higher combustion efficiencies were measured in the Norwegian sector of the region. Comparisons with two emission datasets show their calculated emission ratios to be 30 times greater than that of the ECLIPSE inventory resulting in 30 times less methane emissions when compared to the emission dataset. They conclude the paper by extrapolating their results to a global scale, based on coarse assumptions.

Comments

            I think this is generally a good paper and is worth publication. The results presented are not ground breaking or particularly controversial but are non the less important to back up assumptions of emissions from flaring. Below are some comments:

Line 37 – This would be more readable as 142 billion rather than 149 x 10^9

Line 73 – I think ‘shipborne-based’ should either be shipborne, or ship-based

Line 74 – I do not understand why the sentence about isotopic ratios is important here as there is nothing about it in the rest of the paper. What is the context of this statement?

Figure 1 – Lat/lon grid lines would be useful on this map to be able to get a sense of scale when comparing to the model resolution

Section 2.3 – It is unclear how you decide what/where a plume is in the data. It seems as if you decide a plume if its elevated from the background, and decide the background if its not a plume which seems a bit circular. 50 neighbouring measurements either side of a plume seems ok but clarification on if there is a measurement limit for when the plume starts and stops is needed here.

Lines 210 - 216 – What is the variation of fuel composition data you do have? Is the median value representative or is there a large spread of values and do you have an indication on whether this value would impact the results in major way or not?

Section 3.4 –It need to be clearer as to why the large disparity between ECLIPSE inventory and the observations is not a resolution problem. One the face of it would seem that a few pinpoint measurements in a 0.5 x 0.5 degree inventory is never going to match but I think (after a bit of head scratching) that it shouldn’t matter. But it would be useful to have more explanation so the reader doesn’t have to work this out themselves. Are your measurements going to be representative of the large grid square?   

Appendix figure D1, D2 & D3 – I don’t think these maps are useful. The regional ones are but the global ones don’t provide any useful information for this particular study

Appendix E – I may have missed this but where is this VIIRS data referred to in the text? Is this data used at all?

Review of ‘Flaring efficiencies and NOx emission ratios measured for offshore oil and gas facilities in the North Sea’

The authors present flaring efficiencies and emission ratios for 58 plumes measured during aircraft campaigns to investigate emissions from hydrocarbon production in the North Sea. They present combustion efficiencies, methane and ethane destruction removal efficiencies, and NOx emission factors. They find their estimates are roughly consistent with results from previous flaring studies, of which there are few. There has been little previous assessment of offshore flaring based on in-situ measurements. As a result, I find the work to be an important contribution to further the understanding of the full climate and air quality impacts of flaring during hydrocarbon extraction. The manuscript is well written and logically presented, however, there are few areas that I believe require further context and/or clarification. I detail these areas in my comments below.

Specific comments:

Lines 18-20: The authors provide combustion efficiency both with and without ethane. Is one thought to more accurate than the other?

Lines 119-120: ‘However, as we used enhanced C2H6 mole fractions (background subtracted) in this work, the systematic altitude-dependent biases were effectively removed,’ How does the use of enhanced C2H6 remove the altitude artifacts? Or, are the enhanced C2H6 mole fractions measured over a constant altitude, making the altitude-dependent biases irrelevant?

Line 128: What constitutes a ‘small temporal’ discrepancy? <1s? 10s? If it is large, is there a chence the plumes might be misaligned for other reasons?

Lines 149-157: The existence of correlated enhancements are used to select the flare plumes. Are expected signals such that you would be sure to see them given your instrument detection limits? Put another way, would it be possible that there are small signals you cannot see, and would this potentially bias your results to only larger flares?

Line 158 (and throughout): Why do you use the median and not the mean here and throughout your analysis? It does not seem wrong but is there a reason why you do not use the mean?

Line 164: What is ‘enough’ data?

Line 192: The calculation of combustion efficiency here assumes the fuel is 100% CH4 and no CO2 is present in the fuel gas. Later you say the gas is on average ~85% CH4, so how does this assumption of 100% CH4 affect your results. You say there is a ‘slight overestimation’ but what is slight? Some sort of test case would help provide context here. 

Line 200: ‘Eq. 3 will still overestimate the true combustion efficiency by some amount.’ Similar to the previous comment, what is meant by ‘some?’

Table 1: Is the assumption of 50% NO and 50% NO2 commonly used? I have seen some NOx studies use only NO2 when converting to mass, but I have not seen this 50/50 split before.

Line 358: Your NOx:CO2 ratios (0.003 ppm/ppm) are an order of magnitude larger than the values in Torres et al. (0.0002 ppb/ppb). What do you think explains this large difference?

Lines 384-389: Are the values from other works cited here specific to flaring or total emissions? If these other values are total emissions ratios, what does the comparisons mean?

Section 3.4: I am assuming that the inventory data shown in this section (and Figure 9) is for only flaring, but that is not explicitly stated anywhere.

Section 4: Do you have a sense of why the ECLIPSE inventory overestimates flaring methane emissions by such a large factor? The size of the discrepancy warrants a bit more discussion as to potential causes.

Lines 429-431: I do not see this global extrapolation number (7.6 Tg) in Plant et al., 2022. They do state their DRE_CH4 is ~91%, which is similar to the 92% used by the IEA to arrive at 8Tg of methane from global flaring.

Lines 479-482: In this concluding paragraph and previously in the results section, you discuss the skewed distribution of combustion efficiencies, but the median and mean values are close to expected. If I understand correctly, your emission estimate uses only this median value. So what does the skewness lead to, if anything?

Figure D1-2: It is interesting that no flaring emissions show up in the North Dakota, USA region. There are high flaring rates there. Does this suggest some other error in how ECLIPSE estimates flaring emissions?

Figure D3: Similar to the previous comment, there are no flaring emissions in the USA. That seems odd given it is one of the highest flaring nations.