Review of 'Publicly available data sources to compile an urban natural capital account according to the SEEA EEA: A London case study'

The paper describes a very comprehensive assessment of data sources for urban SEEA EA accounting in London for the years 2007-2018. Results of an extensive data assessment effort are shown in Appendices 1-2, and should be very useful for future urban SEEA EA researchers in London and beyond. The paper is well written.

However, I have three larger concerns in addition to minor comments below. First, the paper reflects a dated version of SEEA EA (which granted has quickly evolved in the last few years). It would be good if the paper were updated to reflect the latest state of SEEA EA – particularly regarding the revision process (accepted as an international standard in March 2021) and regarding elements of ecosystem extent (IUCN GET) and condition (SEEA ECT), as well as referencing physical and monetary supply and use, as described in detailed comments below.

Second, the authors set a very high bar for themselves of “comprehensive urban SEEA EA accounts” covering diverse ecosystem condition metrics and ecosystem services, that are calculated every year. I feel the authors are in some way too hard on themselves and setting too high an expectation in today’s data conditions. Sections 3.2 and 5 of course point out that some ecosystem accounting practitioners feel annual updates are not necessary, at least early currently. While the data assessment exercise is a useful one, starting with “incomplete” accounts that cover only a subset of ecosystem condition metrics and services for selected years may still give a very useful picture of urban sustainability trends and issues!

Third, I was troubled with the focus on the number of datasets with little mention paid to the important issue of data quality. Granted there are multiple dimensions to quality and it’s harder to consistently and concisely report on it vs. raw numbers of datasets. But I do feel the introduction and discussion need to pay more attention to the data quality issue. This is important to make sure that readers and future accounts compilers don’t for instance get overly impressed by a large number of datasets with low quality (which could produce a poor-quality ecosystem account), vs. a single dataset with very high quality (which might produce a brilliant piece of an ecosystem account).

Minor comments:

Obviously, replace “SEEA EEA” with “SEEA EA” throughout, to reflect SEEA EA’s new standard as of March 2021, and rework the second half of paragraph 3 of the introduction to reflect these changes.

Abstract: “development of inclusive natural capital accounts”

Introduction: throughout, a comma goes before “which” and after e.g. or i.e.

“Degradation of nature-on which society and the economy…” – this sentence is confusing as written – it sounds as if society is dependent on degradation. This can be fixed with some light rephrasing.

Introduction, final paragraph: While the UK ONS effort has been the most comprehensive urban SEEA account to date, it would also be appropriate to cite and discuss urban ecosystem accounting efforts in Norway (https://www.nina.no/english/Fields-of-research/Projects/Urban-EEA), the U.S. (https://www.sciencedirect.com/science/article/pii/S2212041620301686), and any other countries the authors are aware of. Further, can you describe in more detail what the limitations of the UK ONS urban ecosystem accounts are, since those are arguably the most comprehensive urban ecosystem accounts in the world?

Methods, first paragraph. You mention “ecosystem accounting areas”. This is a natural concept for readers familiar with SEEA EA, however when writing for a more general journal such terms need to be defined and referenced (this applies to any other terms that could be considered “SEEA jargon” by readers of a generalist journal). Also, can you explain why you chose the years 2007-2018 (i.e., what’s special about those years as your start/end year)? The answer for the end year is obvious, but not so for the start year.

Section 2.1 is a bit dated, not referencing at all the IUCN GET v 2.0 that is now the standard for ecosystem accounting (and was basically accepted as such by summer 2020). While the problem remains that IUCN GET still has only a single class for urban ecosystems, it proposes to deal with this heterogeneity through finer tiered descriptive modifiers. I suggest consulting GET 2.0 and updating this section accordingly. This point also applies to the conclusion, which talks about land cover classification from a dated (i.e., pre-IUCN GET) perspective.

The same problem applies to section 2.2. The SEEA ECT is not referenced at all; it would be very useful to place the ecosystem condition work in the context of recent publications (http://opus.sanbi.org/bitstream/20.500.12143/7467/1/Czucz%20et%20al%202021%20Typology%20for%20ecosystem%20characteristics%20and%20ecosystem%20condition%20variables.pdf, https://oneecosystem.pensoft.net/article/58216/).

Section 2.3: while measuring ES flows is important for SEEA EA, it would be a better framing for SEEA EA to describe how ES flows enter in physical and monetary supply and use tables.

Section 3.1 is a bit limited, and would benefit from additional details. You mention for instance that CCI has poor spatial resolution for urban applications – which is very true. But what is the spatial resolution of the UK CEH data? And do any high-resolution land cover datasets exist e.g., as a result of drone data collection? These are becoming available in many cities in the Global North, with, for instance, 1 m resolution (though often only for single time periods). Further, you note a tradeoff between spatial and temporal resolution. But these aren’t the only tradeoffs faced when choosing between land cover datasets. Accuracy is of course quite important, as is thematic resolution (# of land cover classes). It would be helpful to expand this section to address these issues.

Section 3.2 – it would be helpful to frame this section in terms of the SEEA ECT – how comprehensive are your indicators across the six ECT categories? The example you give about species records at the end of this section is great – you could really make this point hit home by saying explicitly (instead of implicitly) that data need to track actual trends and not be a relic of varying data collection methods or intensities (this particular indicator would seem to fail this test).

Figure 2D – Since you organize section 3.3 around provisioning, regulating, and cultural ES it would be helpful to indicate in the figure which ES are in each of those three categories – you could just have (P), (R), or (C) after each ES’ name for instance.

Section 3.3 – again, the description is fairly thin. You give the numbers of ES by broad category (provisioning, regulating, cultural), but which ES have more or less information needed for their quantification? And what about the distinction between physical and monetary terms?

Section 4, “We did not find publicly available data for key provisioning, regulating and cultural ES” – for which services is data totally lacking? Also, in the discussion about engineered natural assets like green roofs, an additional useful reference may be: https://www.francoangeli.it/riviste/Scheda_rivista.aspx?IDArticolo=64067.

Also, it would be very useful for the discussion to reflect on how the London data assessment experience can inform urban ecosystem accounts in other UK cities. London is of course the UK’s biggest city, but ecosystem accounting is often done on a national scale and even if you can’t directly translate your findings to other cities it would be good to mention this as a goal/future work.

Finally, the issue of spatial resolution in urban EA should be at least briefly mentioned in the discussion. This is briefly touched on in section 3.1, but because cities are so heterogeneous high-resolution data are essential for accurately measuring ecosystem extent and certain processes that underlie ecosystem services (and probably less important e.g., for more homogeneous atmospheric data on climate and air pollution). The sensitivity analysis in the U.S. urban SEEA EA paper is one good example of considering the importance of spatial resolution of data (https://www.sciencedirect.com/science/article/pii/S2212041620301686).

Section 4.1, paragraph 2: while land use data can be problematic for ecosystem extent accounting, they are useful in for example identifying beneficiaries of different ecosystem services (in physical and monetary use tables). This would be useful to point out, so that it’s not wrongly implied that land use data are not valuable for SEEA EA accounts.

Section 4.2: a period in the third sentence should be removed. Further, when talking about the challenges of coordination and links between ecosystem extent, condition, and flows, it would be useful to at least mention the problem of fragmented governance in the urban context. The pollination example is a good one but in many cases natural capital accounts-based resource management will require coordination between different political bodies (towns in the Greater London area) and landowners. SEEA EA can help inform this, but it’s a key challenge even with good information.

Appendix Tables 1 and 2: These tables contain a wealth of information and indeed are the core results of the paper. Please consider making these appendices available to readers as Excel spreadsheets. This format would be much more valuable to readers than as a PDF, which is difficult to navigate due to the size of the tables.

Appendix Tables 1 and 3: Clarify what greyed-out cells mean in these tables. Does grey or white indicate “available” (Table 1) or “linked” (Table 3)?