article

General comments: This paper, written by a team led by someone with considerable experience in this field, concerns natural emissions to the atmosphere of methane and other hydrocarbons. As methane is a more potent Greenhouse gas than carbon dioxide, it is important that studies providing detailed evidence of natural methane emissions are made available to the scientific community. ACP is clearly a suitable vehicle for this paper.

The paper describes a method of long-term monitoring emissions to the atmosphere of methane derived from natural petroleum reservoirs in shallow waters off the coast of California. This is a valuable method, and could be applied to other natural or anthropogenic methane sources. The Introduction also provides a detailed explanation/ description of natural seabed gas seeps in general, citing numerous relevant sources. Similarly, the fate of methane released from the seabed is also discussed. Whilst this provides valuable context, this material is covered by numerous other papers. There is a danger of the ‘context’ overwhelming the detail of the method, its application and the acquired results.

Whereas the authors have endeavoured to integrate all aspects of a complex study, the paper does not flow well; there are several inconsistencies and other shortcomings that would benefit from more detailed editing. Furthermore, although some important conclusions have been drawn, there is no discussion of long-term temporal emission variations, as might be expected from the title.

Specific comments:

Lines 138-9: It should be pointed out that, whilst the bubbles from the COP seeps may be oil coated, this is not the case in many (?most) other seeps areas worldwide.

Line 212 - Section 2.1: Would it not be appropriate to identify the type(s) of equipment used to acquire these data (especially the THC)?

Section 4.2.2 Methane and non-methane hydrocarbon emissions. This section is particularly interesting in the light of the implication by Hmiel et al (2020) that pre-industrial natural geological contributions to atmospheric methane are practically insignificant. Can any comment be made about emission trends over the three decades covered by the data reported here? It would be interesting to compare emission trends with petroleum production - is reservoir depletion reflected by a reduction in seepage emissions? Such a trend is mentioned (Lines 755-778) but only for the Seep Tent Seep.

Technical corrections:

Line 146: “…. with dissolved plume concentrations decreasing with time ….”. Time or distance - or both?

Line 151: what is “water-side turbulence”?

Lines 173-4: “COP seep field sources from the South Ellwood oil field whose primary source rock is Monterey Formation, which is immature to marginally mature.” This could be re-cast as ”The source of the methane of the COP seep field is the South Ellwood oil field, which contains petroleum from the immature to marginally mature Monterey Formation.”Line 191: wet season (singular).

Line 230: There is no verb in this sentence.

Line 250: “The plume inversion model is a three-step process”. Surely this should be “The plume inversion modelling is a three-step process”.

Line 252: What is C’ ?

Line 255: “is fit” should be “was fitted” - although present and past tense seem to be interchangeable in this section.

Line 328: For the benefit of readers, it should be noted that the ‘Seep Tent’ refers to an installation whereby the natural seabed seepage was captured and utilised along with gas produced from nearby petroleum fields. Suggest moving text from Lines 765-769.

Line 367: “almost due south to the”. Surely WCS is almost due north of Coal Oil Point?

Line 373: “allowed far higher values of C and u” - add “ to be measured”

Line 437: “C and u for the seep field direction, useep, and Cseep, respectively” should be: C and u for the seep field direction, and Cseep, and useep , respectively.

Line 450: replace “largely” with “mainly”

Line 579: “A range of approaches are available” - a range … is available!Lines 665-680 is clumsily worded, and units are mixed (tons - should be tonnes and could be abbreviated to ’t’ - Gg, Mg and nmol). All previous multi-authored papers are cited using multiple names, except Römer et al. 2017. e.g. Line 666-8: “e.g., summary Römer et al. (2017) where emissions for 12 different seep areas including for sites in the North Sea, Pacific north west, Gulf of Mexico, etc., were 2-480 tons yr-1, multiple orders of magnitude less than seabed emissions for Coal Oil Point. Römer”. Poorly worded. Suggest: ‘For example, Römer et al. (2017) identified emissions from 12 different seep areas (in the North Sea, Pacific north west, Gulf of Mexico, etc.) of 2 to 480 tonnes yr-1.’ [N.B. the last 12 word duplicate the previous sentence. Römer et al. likely used metric tonnes rather than US (Imperial) tons - why not abbreviate to ’t’?]

Line 670: Tommelieten should by spelled Tommeliten.

Line 693: “emissions were” estimated as …

Section 4.4: to conform to section 4.2.2, emissions should quantified by mass rather than volume.

Line 771: If the Seep Tent Seep is post 1978, how come it was observed in 1970? (Line 768).

Line 783: “WCS seep emissions” - surely you mean measurements, not emissions.

Line 1199: If C is defined, then C’ should also be defined.Fig. 4: Why are the rose diagrams plotted with 0 (presumably representing North) at the bottom, South at the top, and therefore East and West reversed from their intuitive places? Also, it would make interpretation of the rose diagrams easier if they were superimposed on the map; this would enable correlations with seep locations more intuitive. 

In this manuscript the authors use a simplified inverse technique to constrain marine emissions of hydrocarbons from a seep area offshore of southern California. The authors draw on a 30-year record of total hydrocarbon concentration (THC) from a ground-based monitoring site to estimate THC emissions from a defined seepage area off the coast of Santa Barbara, California. This estimation is achieved by assuming that THC outgassing is well-described as a Gaussian plume (effectively treating them as point sources). The study region was gridded and each grid was treated as a separate plume. The authors offer a validation of this approach through comparison with shipboard atmospheric measurements from a single day. 

Major comments: This manuscript has major flaws. It is overly long, convulted, and would benefit from a major rewrite, focusing on clarity and brevity. The Introduction contains a large amount of extraneous detail, as do the Methods and Results. Most of the key information on the model is in the Supplemental. 

The authors' approach is presented as a novel technique, but upon reading the paper I think it is better described as a crude approximation. There is nothing wrong with simplifying assumptions or simplicity of approach per se, but the authors' description of it as novel is misleading, in my opinion, since there is nothing new about the plume model they employ (they cite a text from the 1980s for the relevant equations). The paper, somewhat incredibly, lacks any details on how the measurements of THC were conducted that I could find. I am not sure, after reading it, what compounds are included in this bulk tracer. Without any information whatsoever it is difficult to properly evaluate the manuscript further. 

Another serious shortcoming of the study is the treatment and discussion of uncertainty, which is insufficient. They have done some sensitivity tests of the model, but not attempted to account for, or at least discuss, the uncertainty of their approach. The authors claim that a Monte Carlo simulation was not feasible due to computational constraints. I realize no one has infinite resources, but they could propagate the uncertainty through the model using quadrature sums.

The assumptions underlying the model also require significant justification, which is not presented to my satisfaction. For instance, the authors claim they remove terrestrial emissions from the model based on wind sector, but this is hardly sufficient--atmospheric transport is complex and this can only account for emissions in the very near field. Why do the authors not use boundary layer height and wind fields from a reanalysis product? There are many such products which are freely available. I am still unclear how they defined a background THC concentration. I do not think the large number of simplifying assumptions made in this approach would be valid over the 30 year record. The potential for biases and errors is too great to make the emission estimates with any confidence. I am doubtful that outgassing relating to hydrocarbon seep ebullition is well-described as a point source. During transport across the sea surface, the hydrocarbons will be diluted and also exchange with the sea surface in vastly different proportions based on their solubilities. Lumping all of these compounds into a single tracer, THC, makes no sense, because it is not a conservative tracer. The plume model assumes that the emission rate is uniform and constant, which is not a valid assumption for seepage ebullition. Even assuming the seepage was completely constant, the state of the sea surface (roughness, surfactants, skin temperature, etc) changes constantly, which will greatly impact the exchange of CO2, CH4, HCs, etc. Calculating the seabed emissions from 50:50 paritioning is so crude as to be useless. 

A better place to start, in my opinion, would be to calculate a footprint for the WCS station and to convolve this with THC anomalies against a background. Without additional observations of some other conservative tracer for marine air, however, I am still skeptical that they can separate marine emissions from urban/terrestrial emissions. Even in a state of the art inversion, this would be challenging. 

The validation study using a field survey is unconvincing, since they are not actually validating through observation that the THC enhancements measured at the station are indeed coming from the seafloor; they are simply comparing additional atmospheric measurements of CO2, CH4, and alkanes measured in the atmosphere to the station measurements on a single day. Measurements of dissolved THC would have been more convincing. 

The deficiencies here require more than just rewriting or reanalyzing the data, and so I am afraid I cannot recommend publication.