Development of a combined marine and terrestrial biodiversity indicator: research

Appendix 2: Consultation documentation

Appendix 2.1: Attendees at consultation workshops

Workshop 1, 23^rd April 2019

Botanical Society of the British Isles: Chris Miles

British Ecological Society: Maggie Keegan

British Trust for Ornithology: David Noble, Mark Wilson

Buglife: Craig Macadam

Highland Biological Records Centre: Ro Scott

James Hutton Institute: Robin Pakeman

Joint Nature Conservation Committee: Chris Cheffings

Marine Scotland: David Stirling

National Biodiversity Network: Jo Judge

RSPB: Mark Eaton, Ellen Wilson, Jeremy Wilson

Scottish Government: John Landrock

Scottish Wildlife Trust: Gill Douse

SNH: David O'Brien

UK Centre for Ecology & Hydrology: Nick Isaac

Workshop 2, 24^th April 2019

Cefas: Murray Thompson

Marine Biological Association: Dan Lear

Marine Conservation Society: Calum Duncan

Marine Scotland: Kirsty Bosley, Tom Reilly

RSPB: Mark Eaton

SNH (retired): John Baxter

University of Sheffield: Tom Webb

Workshop 3, 17^th July 2019

British Ecological Society: Brendan Costelloe, Maggie Keegan

British Trust for Ornithology: Ben Darvill

Buglife: Craig Macadam

Butterfly Conservation: Paul Kirkland

Highland Biological Records Centre: Ro Scott

James Hutton Institute: Robin Pakeman

Joint Nature Conservation Committee: James Williams

Marine Scotland: Kirsty Bosley

National Biodiversity Network: Jo Judge

NFU Scotland: Emma Bradbury

RSPB: Mark Eaton, Jeremy Wilson

Scottish Government: John Landrock, Sarah McCutcheon

Scottish Wildlife Trust: Gill Douse

SEPA: Scot Mathieson

SNH: John Baxter, David O'Brien

UK Centre for Ecology & Hydrology: Nick Isaac

Appendix 2.2: Document circulated for consultation with RAG:

This interim report outlines the different approaches, and associated issues, that could be used to create a high-level Scotland biodiversity indicator, while the associated data review highlights potential sources of information. We believe that the approaches employed elsewhere, and the availability of biodiversity data for Scotland, point towards an approach based upon trends in species status measured in both abundance and occupancy, available for both terrestrial and marine habitats. The report authors' will be meeting the Scottish Government's Research Advisory Group (RAG) on the 15^th May 2019 to discuss and narrow down the options for the next stage of the project, short-listing potential indicators. To help this process, a number of questions have been developed:

Indicator presentation

Given the issues raised above, can the RAG confirm that there is no option in final presentation other than a single index line presented without any estimate of error? This was confirmed.

Would the RAG prefer a single simple measure of a relevant biodiversity variable, or a composite from a narrow range of measures (e.g. bird trends), or would they be happier with a composite combining a wider spread of biodiversity data from disparate sources and potentially measured in differing currencies? The balance of opinion was in favour of a broader composite indicator.

Related to the question above, does the final indicator need to be easily understood/interpreted by the public (e.g. a measurable of change in abundance), or would a more complex measure – albeit one that still follows an 'up = good, down = bad' basis – be acceptable? No clear recommendation; although the former approach is preferable, it was accepted that some indicators are more complex (e.g. Natural Capital Index)

Does it need to be updated annually and what time lag in reporting (time since last year of data) will be acceptable? Annual update was preferred but not essential. Time lag of no greater than three years, although shorter would be preferable.

Data inclusion

We understand the focus on a current indicator, and that a start date in 2007 would be desirable, but how valuable would a longer time span be? There is potential for reporting of biodiversity trends for considerably longer which might enable an indicator to span a longer period and would provide valuable context to more recent changes. Options could include from as far back as 1970, from the change of government administration in 2007, or a more recent trajectory e.g. the last five years. There was considerable support for an indicator with as long a timeline as possible.

The composition of existing species-based indicators often changes through time as species data starts e.g. the existing UK species-based indicators start in 1970, but with new species entering the index at later dates as their monitoring began. Is the RAG happy with this for this indicator? Yes.

Depending on design, there is considerable potential for the indicator to become more robust in the future as new data becomes available e.g. to fill substantial gaps in the taxonomic representivity of the indicator, or as new methodological developments lead to improved analyses. Would the RAG be happy with the intention of future development, perhaps managed by periodic reviews? Yes.

In the event of such future revision, this may mean retrospective changes in the indicator – the addition of new data sources for previous years, or new analytical methods, may mean changes in index values for years already published. Would the RAG be comfortable with this? Yes.

We are disinclined to pursue indicator options based upon/incorporating proxy data e.g. on habitat extent or condition. Are RAG content with this? Yes.

Likewise, we are not convinced of the suitability of indicator options that are based upon species data in the form of diversity indices, measures of turnover, homogeneity, and Red List Indicators, with our preference being an average measure of species' status. Do the RAG have any thoughts on this, and any exceptions that they would like us to consider further? RAG were happy for the focus to be on a metric of average species' status.

Answers to the questions above notwithstanding, our current thinking is that the indicator will be a composite from species trend data. This will likely encompass the different currencies of abundance and occupancy data. Does the RAG have any opinion on the treatment of these metrics and whether they can be treated equally (e.g. combined straightforwardly into a single measure) or not (e.g. so may need to be treated separately to create sub-indicators that can then be merged, in an approach not dissimilar to the NCAI). Although some concern expressed, on balance RAG were happy for abundance and occupancy trends to be merged.

Alternatively, does the RAG have a preference for one of these metrics over another? Abundance regarded as better measure of change in species status, but the benefits of the larger sample of occupancy trends recognised; in effect, very difficult to choose one as preferable to the other.

For the consideration of marine biodiversity, the authors believe the indicator should consider the full reach of Scottish seas, therefore propose considering data from within the Exclusive Economic Zone (to 200 nautical miles) rather than territorial waters (12 nautical miles). Is the RAG happy with this suggestion? Yes.

Disaggregation

The RAG has previously expressed interest in indicator disaggregations, but we expect our ability to do this will be limited by data. Taxonomic disaggregations will be possible, and by data source/type, but others – habitat, spatial – will be either impossible currently or constrained by resources. How important is this to RAG, and will there be future interest in resourcing work to take this forward e.g. to identify disaggregations likely to respond to specific policy interventions? Our recent engagement with data stakeholders indicated a high level of interest in the provision of such contextual supporting information amongst this community (whose cooperation is essential for the creation of the headline indicator). RAG members recognise the restrictions around the headline NPF indicator, but are very supportive of the publication of disaggregations through other routes.

Should all the disaggregations be nested together, or would the RAG be happy with multiple types of disaggregation? For example, the Living Planet report presents disaggregations both by taxonomic group and by major habitat. No clear opinion – but interest in a wide range of disaggregations, not all of which are likely to be possible, at least in the short term.

Resourcing

What resources might be available beyond the scope of this current work in order to create/revise the final indicator in the future? For example, what is the scope of resources available each year to collate data outside the usual published data streams? Does this mean we should avoid the inclusion of data that may require annual processing outside the scope of other funded work streams, in order to minimise future resource requirements? No certainty over future funding, but advised that whilst future developments could be recommended, Scottish Government would not be able to commit to development funding at this stage; we are recommended to identify a draft indicator that can be produced using current data streams, with a limited amount of collation and processing time required.

The availability of terrestrial data, and the analyses required to produce species trends from this data, suggest that an indicator for terrestrial and freshwater biodiversity may be delivered within the scope of this project, albeit with recommendations for further improvements. However, the same can only be said for a proportion of the data required for a robust marine indicator (e.g. for demersal fish), with the time required to obtain and analyse other marine datasets being beyond the scope of current resources. We would wish for advice from the RAG as to how to approach this issue. No clear solution other than to do our best to incorporate marine data.

Appendix 2.3: Short note circulated to SBS SSG for input on indicator decision-making

Short note: progress on developing a Biodiversity Indicator for Scotland

Work to date

The following work has been conducted by the consortium working on the Scottish Government contract SPB/001/18, Development of a Combined Marine and Terrestrial Biodiversity Indicator for Scotland, by the project team from the RSPB, James Hutton Institute, Centre for Ecology & Hydrology and University of Sheffield.

• A review of the potential approaches for reporting on the state of biodiversity at a high level, drawing on practice in the UK, its constituent countries and further abroad.
• A review of the available terrestrial and marine biodiversity data. As well as species data, which often underpins biodiversity indicators, we considered a wider range of environmental data which might be used as a proxy for biodiversity.
• Collation of existing biodiversity to be used in the creation of draft indicators.
• Extensive consultation, with the Research Advisory Group and through three workshops with a wide range of consultees. Two workshops focused on data sources, for terrestrial and marine biodiversity separately; the third workshop focused on indicator creation.
• Inception and interim reports have been submitted to Scottish Government, describing the work listed above and the decisions made as a consequence.

We have, through the work described above, made decisions on the key elements of the biodiversity indicator which we will recommend to Scottish Government in a final report due to be submitted by the end of September 2019. This short note summarises this progress, and identifies in bold text the key issues which this work, including the consultation with relevant stakeholders, has made decisions on. Further to this we have highlighted the decisions on which we would most like to invite comment, to be considered in the final days of this project, although feedback on any other considerations would be welcomed.

Our conclusions

• The indicator will use the well-established approach, as used in the indicator on terrestrial birds which was part of the previous National Performance Framework, and in a number of the UK Biodiversity Indicators, of calculating the average trend from a wide range of individual species trends.
• The indicator should use as many species' trends as possible, to reduce various sources of bias e.g. taxonomic. In order to do so, the indicator could incorporate both trends in species abundance, as are available for species for which trends are derived from structured monitoring schemes, and trends in occupancy, a measure of distribution derived from Bayesian modelling of biological record datasets.
• A combination of robust standardised monitoring schemes, and recent CEH-led developments on the analyses of ad-hoc data, mean trends in either of species status are available for approximately 2,400 species native to Scotland and Scottish waters: an extremely impressive resource for the creation of biodiversity indicators.
• We will use established approaches, e.g. based upon sample size, to identify species' trends of sufficient robustness for inclusion in the indicator and filter out those of insufficient quality. For species which trends in both abundance and occupancy are available, the abundance trend would be used in preference to that in occupancy.
• Trends in abundance, and trends in occupancy are essentially different 'currencies' that are not directly comparable with each other. An indicator produced through combining these two will present some challenges in interpretation and communication. Nevertheless, we recommend that they should be treated as equivalent measures of changing species' status and combined in order to make best use of the available data, and produce an indicator as representative of Scottish biodiversity as possible.
• Data for marine biodiversity are considerably sparser, available for a much more restricted range of species than for the terrestrial realm. At present we have identified species' trends in abundance for groundfish (from bottom trawl surveys), marine mammals, and breeding seabirds for incorporation in the indicator. Trends for 88 species of marine groundfish are derived from bottom-trawl surveys from ICES, from the early 1980s onwards. Two versions of this dataset exist. Data are available from ICES, via the DATRAS portal in almost real-time, having received some filtering for data errors. Alternatively, Marine Scotland (MS) have created an improved dataset (Moriarty & Greenstreet 2017) with considerably more rigorous quality control, for use in OSPAR assessments. At a species level, species trends derived the two data sources vary in a non-consistent way i.e. they match for some species, and show marked discrepancies for others. When combined into multi-species indicators, the two versions show some consistency in overall pattern but differences in magnitude of change particularly in recent years. The MS data, with better quality control, is more robust, but to our knowledge there are no plans for its revision currently, and so is not available beyond 2017. Unless future updates of this are likely to be available, we will recommend the use of the ICES dataset.
• Species' trends suitable for inclusion in the indicator have start years from 1970 onwards, but the number and type of dataset increases through time as new monitoring schemes originated. Whilst the priority is a robust indicator going forward in time, and measurement of change over recent years, it has been acknowledged that an indicator encompassing as much historical timeline as possible would be preferable. We recommend a baseline year in the mid-1990s, aligned with the introduction of a large number of abundance trends for birds and mammals.
• We have identified a number of data sources which may prove to be valuable for inclusion in a future iteration of the indicator, thereby improving robustness. However, at present considerable further work (e.g. collation, analysis) is required to enable such datasets to be made available for use in an indicator, and such work is well beyond resources of this current project. We recommend, however, that further consideration is given to development work to enable future inclusion of such data. This is most pertinent for a number of marine datasets, e.g. data gathered on abundance of a range of intertidal species through the long-running MARCLIM project, as well as for terrestrial plants to address that gap in data availability.
• Existing workflows enable the production of annual trend updates from most of the data sources identified, enabling the indicator to be updated annually with relatively little effort. At present, we have permission from data-owners to use these trends for the purpose of this contract – to identify a suitable indicator – and not for a finalised, published indicator. However, we do not expect permission for a finished and published indicator to be withheld.
• There are a number of options for the creation of a final, single headline terrestrial and marine indicator using the species data as outlined above. Whilst each employs the same basic approach of calculating an average trend across all species trends, there are a number of options for stratification: employing weightings in an attempt to address biases in the availability of species' trends. Such biases include between terrestrial/freshwater and marine realms (we have far more trends for terrestrial/freshwater species than marine species); between taxonomic groups (there are considerable disparities in the representation of taxonomic groups e.g. species trends are available for a high proportion of birds, but a low relatively proportion of invertebrates); and in trend type (more species are represented by trends in occupancy than distribution). Note that we cannot address biases when data for desired underrepresented groups are not available at all e.g. fungi.
• We have, however, been unable to identify an objective manner in which to identify the appropriate approach to weighting e.g. by identifying which biases are most important to address, and how they should be addressed. Consultation to date has failed to identify a preferred option, and there is no published precedent to follow. In addition, further development work would be required in order to identify how to incorporate measurement of error within a weighted indicator. Therefore, our recommendation is that all available species trends should be incorporated in the indicator without weighting, each having an equal influence upon the indicator.
• Whilst fulfilling the requirements for a headline indicator for inclusion in the National Performance Framework indicators, the indicator recommended by this work will be difficult to interpret in terms of underlying biodiversity changes and the drivers of these changes. We strongly recommend the publication (which could be as supporting material to the headline indicator, or entirely separate) of disaggregated indicators to show patterns of change in constituent groups of the headline indicator which will aid understanding of change in the headline metric. Disaggregation could include by realm (terrestrial and freshwater, marine), by trend type (abundance and occupancy) and by taxonomic group. Further disaggregation, such as by region of Scotland or by habitat, might also be desirable but could require considerable additional work to achieve, or may not be possible given data constraints.