CODEX (Connectivity Data Exchange)

Marine natural capital is of immense importance for humanity. 3 billion people rely on fisheries to supply them with food, and many others depend on the ocean’s living resources for recreation, culture, employment, and coastal protection. Despite this, many marine populations have been under decline for decades due to anthropogenic stressors including, increasingly, climate change. The sustainable exploitation and conservation of marine life is managed through marine spatial planning, for instance the implementation of Marine Protected Areas (MPAs). Approximately 3% of the global ocean is currently ‘highly protected’ using MPAs, but this leaves the overwhelming majority of most species’ ranges unprotected (Klein et al., 2015). There is broad recognition that at least 30% of the ocean must be protected to sustainably manage marine resources (O’Leary et al. 2016), which has been adopted as a goal by the International Union for Conservation of Nature (IUCN, 2016).

Many marine organisms are motile, through swimming or passive larval dispersal, resulting in connectivity across MPA boundaries. This connectivity is extremely important for effective MPA design. For instance, MPAs are often designated to protect larval sources that can reseed populations outside MPA boundaries, and to preserve genetic diversity by including multiple genetically connected populations. However, despite being recognised as a fundamental criterion for MPA designation (O’Leary et al. 2016), only 11% of MPAs actually incorporate connectivity data in their design (Balbar and Metaxas, 2019). 

Many scientists are actively working on evaluating the connectivity between marine populations: hundreds of studies are published annually. These studies frequently generate network-based connectivity datasets, quantifying the connectivity between a set of locations. Unfortunately, these valuable data are often published in a form that is written primarily for an academic audience. Data are often either not made publicly available at all, or require manual extraction from figures and/or technical file formats. Crucial meta-data are frequently lacking, and there is little data standardisation.

Based on a systematic literature review, we identified 142 studies published since 2022 that produced network-based connectivity datasets for fish, turtles, corals, or other invertebrates. Network-based datasets were at least partially available (without requiring manual extraction, further processing, or liaising with authors) from 73% of these studies, whilst they were available in full from 37%. The fact that neither of these figures is 100% indicates that a considerable proportion of (usually publicly-funded) research is unable to maximise its real-world impact. Further barriers are introduced by journal paywalls, which disproportionately impact non-academics, and academics from lower-income countries.

The information required to make decisions about the future of our global ocean therefore exists, but is not being effectively communicated to decision makers, and incorporated into marine management. This is clearly an unacceptable situation.

References

Balbar and Metaxas (2019) The current application of ecological connectivity in the design of marine protected areas

IUCN. (2016). Increasing marine protected area coverage for effective marine biodiversity conservation. Motion 53.

Klein et al. (2015) Shortfalls in the global protected area network at representing marine biodiversity

O’Leary et al. (2016) Effective Coverage Targets for Ocean Protection

CODEX (Connectivity Data Exchange) is a proposed fully open-access online data-sharing and visualisation platform for ecological connectivity data. CODEX aims to solve the present lack of connectivity data accessibility by:

1. Centralising connectivity data into a single, easily searchable repository, to avoid non-academics from having to trawl through opaque scientific literature.

2. Standardising data formats, to improve data interoperability and eliminate the need for end-users (such as marine managers) to manually extract or process data from academic papers.

3. Eliminating gatekeeping of connectivity data behind non-committal data availability statements and/or paywalls, by normalising the deposition of connectivity data to an open-access repository.

4. Automatically visualising connectivity data as georeferenced networks, facilitating the incorporation of connectivity data into marine spatial planning efforts, particularly in lower-income regions where advanced data analysis tools may not be available. 

5. Enforcing metadata requirements to ensure that data are genuinely useful for conservationists and other researchers.

We developed a prototype for the data visualisation module with Julia Janicki, which can currently visualise any directional or non-directional connectivity dataset, albeit requiring some manual rewriting. Examples for model-derived coral connectivity in the southwest Indian Ocean, and species assemblage-derived hydrothermal vent connectivity in Japan, can be found here and here respectively. With support from MIT Solve, we intend to expand on this prototype by allowing researchers to submit their own datasets, adding a search and download interface, and improving the flexibility of the data visualisation module to increase its usefulness for arbitrary connectivity datasets. 

Researchers generating network-based ecological connectivity data will interact with CODEX by submitting their datasets through a submission portal, including a mandatory set of metadata (including species type/s, geography, methodology, and associated publications), in line with FAIR principles. Submissions will be manually checked for appropriateness, completion, and compliance to data requirements by volunteer researchers, which is practical given the volume of datasets we expect to process. Valid datasets will then be automatically converted into a standardised network file format, and deposited into a repository hosted by an institutional partner. 

End-users (such as marine managers) will be able to search for connectivity datasets based on geography and/or tags (including species and methodology). The web-app will display automatically computed summary statistics for each dataset, and an interactive georeferenced visualisation (see the prototype). Users requiring the raw data will be able to download the full datasets in a standardised and fully documented file format for further work if required. The web-app will also include tutorial pages explaining how to use hosted datasets, and how major types of connectivity datasets are derived

Our solution ultimately serves everybody who relies on a healthy ocean. This includes the 3 billion people who rely on fish as a significant source of food (UN-FAO), the millions (particularly in low-income countries) who derive their income from diverse and healthy coral reefs (Sing Wong et al. 2022), and untold numbers of people in the future who stand to benefit from undiscovered products, including pharmaceuticals derived from marine species (Hossieni and Al-Jabri, 2022). 

More directly, CODEX will greatly improve the capacity for governments, conservationists, and marine managers to implement evidence-based marine protected areas (MPAs). For instance, the development of coastal resources, such as offshore windfarms, may result in the obstruction or alteration of connectivity among nearby MPAs (Bishop et al., 2017), so a quantitative consideration of connectivity is particularly important in these situations. The mismatch between the universal acknowledgement of the importance of connectivity for marine management (O’Leary et al. 2016) and the reality that only a small minority of existing MPAs incorporated connectivity data into their design (Balbar and Metaxas, 2019), and the significant shortfalls in data accessibility in the field that we have identified, demonstrates that effective data sharing is a bottleneck in effective marine management. The CODEX platform therefore represents an important step towards more effective marine management.

Aside from the primary purpose of supporting conservation, CODEX will also support researchers through improving data accessibility, and the visualisation tools targeted at non-scientists will also act as a useful science communication tool.

References

Balbar and Metaxas (2019) The current application of ecological connectivity in the design of marine protected areas

Bishop et al. (2017) Effects of ocean sprawl on ecological connectivity: impacts and solutions

Hosseini et al. (2022) Marine microbial bioprospecting: Exploitation of marine biodiversity towards biotechnological applications—a review

O’Leary et al. (2016) Effective Coverage Targets for Ocean Protection

Sing Wong et al. (2022) An assessment of people living by coral reefs over space and time

Dr Otis Brunner and Dr Noam Vogt-Vincent are both experts in marine connectivity (e.g. Brunner et al. 2022; Vogt-Vincent et al. 2023). They therefore have an excellent understanding of (1) the types of connectivity data being generated by academia, (2) the metadata required to make these datasets effective, and (3) how other researchers intend to interact with these datasets. The prototype for CODEX was developed by Julia Janicki, an expert data visualisation specialist with a diverse backlog of ecologically impactful projects. We have a highly-experienced, international network of collaborators (both from academia and NGOs) who will be contributing towards CODEX, including specialists in marine spatial planning, policy, and conservation. The scope of CODEX is global, and the development phase will involve an in-depth consultation with end-users to ensure that CODEX is genuinely designed for practical implementation in marine management. We are sensitive to the fact that developing countries (and, in particular, small island developing states) face severe barriers in terms of access to ecological data. The incorporation of connectivity data into MPA design has historically been concentrated in the US, Europe, and Australia, with vanishingly few projects in the South Atlantic and Indian Ocean (Balbar and Metaxas, 2019). As a result, our consultation process will be carefully designed to amplify voices in these underrepresented regions, so that CODEX is genuinely accessible and useful to marine managers and conservationists in all countries. 

References

Balbar and Metaxas (2019) The current application of ecological connectivity in the design of marine protected areas

Brunner et al. (2022) Species assemblage networks identify regional connectivity pathways among hydrothermal vents in the Northwest Pacific

Vogt-Vincent et al. (2023) High-frequency variability dominates potential connectivity between remote coral reefs

Working with data visualisation expert Julia Janicki, we have developed an interactive visualisation system for connectivity network data. This visualisation system can be adapted to different connectivity datasets, and we have already applied this to coral reef connectivity in the Indian Ocean, and deep-sea hydrothermal vent connectivity in Japan. These visualisations have been well received by academics, marine managers, and the public. For instance, the coral reef connectivity visualisation is featured as a resource for the Seychelles Marine Spatial Plan Initiative, to inform future marine spatial planning efforts in the island nation. We expect that this visualisation system will act as the basis for the visualisation module in CODEX.

The project leaders for CODEX are academics, so the leadership coaching, networking opportunities, and workshops offered by Solve will enable us to build capacity within the project team to make CODEX genuinely useful for all end-users, beyond just scientists. In particular, the need to develop effective marine protected areas is most acute in developing countries, which are the same regions that face the greatest barriers to accessing connectivity data (Balbar and Metaxas, 2019). By building bridges with more end-users, Solve will allow us to expand our network of marine managers beyond our existing connections, and support us in ensuring that CODEX is a globally accessible solution, not just those focused on the North Atlantic and North Pacific. Media exposure through Solve will also increase the visibility of CODEX, (1) ensuring that this tool is communicated to all relevant stakeholders, and (2) raising awareness across the scientific community, and conservation in particular, for the importance of accessible and open data. Finally, we hope that Solve will also support us in seeking funding for the development of CODEX as a scalable and future-proof solution, so it can benefit the marine environment for decades to come.

References

Balbar and Metaxas (2019) The current application of ecological connectivity in the design of marine protected areas

No satisfactory solution to the problem of marine connectivity data inaccessibility currently exists. General academic data repositories such as Dryad and Zenodo have existed for years. However, the continued lack of adherence to data-sharing best practices (e.g. our review highlighted that only around a third of recent academic papers in the field made all data publicly available, before even considering the usability of these data), and the limited incorporation of connectivity data in marine spatial planning, despite its acknowledged importance (Balbar and Metaxas, 2019), demonstrates that data accessibility remains a significant barrier in this field. Just as it has now become standard practice to deposit raw genetic data in standardised formats in repositories such as GenBank, we intend for it to become standard practice to deposit ecological connectivity data in CODEX. Since CODEX is designed specifically for ecological connectivity data, this will ensure that all datasets are findable and are associated with the necessary metadata (following FAIR principles), and will allow us to incorporate tools (such as our interactive network visualisation) to make these datasets easy to use for non-academic end-users. CODEX will therefore massively improve the incorporation of ecological connectivity data into marine conservation projects (particularly in developing countries, where the greatest barriers to accessibility exist), and will also normalise good practice in data sharing and accessibility within academia. Whilst we are focusing on marine connectivity for the time being, terrestrial connectivity data is often generated in a similar format, and expanding CODEX beyond the marine realm is a natural next step. 

References

Balbar and Metaxas (2019) The current application of ecological connectivity in the design of marine protected areas

As we have demonstrated in our review of recent literature, whilst a huge amount of ecological connectivity data is generated by academia (many thousands of studies in total), only around a third of recent studies make network data (which can be incorporated into marine spatial planning efforts) available in full. Data is nevertheless made available in part by around three quarters of recent studies yet, despite the acknowledged importance of these data for effective marine spatial planning, only around 11% of marine protected areas used connectivity data in their planning process (Balbar and Metaxas, 2019). This strongly suggests that data accessibility (as opposed to a lack of data) is a bottleneck in the process of evidence-based marine spatial planning. By providing an easy-to-use web platform, where connectivity data is comprehensively described and documented, easily and intuitively visualised, and with raw data available for download in a standardised format, CODEX will immediately facilitate the incorporation of these data into marine management. In its early stages, CODEX will be limited by (1) how quickly team members are able to transfer existing data from the literature onto the platform, and (2) uptake by those generating data (i.e. primarily academics). However, over time, we hope that making data available on CODEX will become routine, and end-users (such as marine managers) will be able to refer to CODEX as a comprehensive database of all connectivity data. 

References

Balbar and Metaxas (2019) The current application of ecological connectivity in the design of marine protected areas

There are two primary quantitative goals for CODEX:

1. What proportion of relevant datasets (i.e. network-based connectivity data) are available on CODEX?

2. What proportion of marine protected areas (and other marine management activities) are quantitatively incorporating connectivity data into their planning processes?

The first goal describes how useful CODEX has the potential to be, whereas the second goal describes how useful CODEX actually is (or, at least, one measure thereof).

The first goal is quite easy to assess, through a literature review. This figure is currently small (to date, two connectivity datasets have been incorporated into the prototype precursor to CODEX), but our literature review suggests that 73% of recent marine connectivity studies could contribute at least some data to CODEX, without requiring considerable additional effort, which will hopefully approach 100% in the future. 

For the second goal (the proportion of marine protected areas incorporating connectivity data into their design), Balbar and Metaxas (2019) found that this figure was approximately 11% (in 2017). We will be able to qualitatively assess this goal by engaging directly with marine managers, and discussing whether CODEX is fulfilling their requirements (in terms of the data that are available, and whether this is accessible and useful for their needs). Progress towards this goal could be quantified by performing another literature review, similar to that carried out by Balbar and Metaxas (2019). 

Of course, the ‘true’ end goal is to maximise the effectiveness of marine management activities (contributing most directly towards UN SDGs 13 and 14), rather than just maximising the incorporation of a particular type of data into their design. Quantifying this is challenging since different management activities have different goals, but we look forward to working with academics and marine managers to monitor the effectiveness of marine management. 

References

Balbar and Metaxas (2019) The current application of ecological connectivity in the design of marine protected areas

CODEX will be a free to use web-based app, combining a data repository hosted by an academic partner, with a search interface (based on geographic, biological, and methodological tags), visualisation interface, a download function, submission portal, and documentation (both for how to use CODEX, and how to use ecological connectivity data more broadly). The submission portal will include basic data cleaning functionality (to prevent misuse), but the platform will also rely on academic volunteers who will use their expertise to manually check submissions for relevancy and methodological quality before being published on CODEX. Given the current rate at which connectivity research is being generated (c. 100 relevant datasets per year, based on our review), there is more than sufficient volunteer capacity to process submissions.

The data visualisation prototype is publicly available and is already listed as a resource for the Seychelles Marine Spatial Plan Initiative. A related web app has also been highly successful in another branch of ecology. The COMPADREdatabase has a similar goal to CODEX, applied to a different application of demographic modelling. COMPADRE has been successfully maintained for several decades, and has transformed data accessibility for plant and animal demography. We are using COMPADRE as inspiration for CODEX, and COMPADRE's success demonstrates the enormous impact that this type of system has on data accessibility and effectiveness.

CODEX is not registered as an organisation and therefore has no staff. Dr Otis Brunner and Dr Noam Vogt-Vincent are currently leading the development of CODEX as a side project, as part of their ongoing connectivity research. We will contract software engineers to further develop the software behind CODEX, and have a concrete proposal and quote from a highly experienced consultancy.

Development on the visualisation prototype began in 2022, and was completed in late 2023. We began developing this prototype into the CODEX project at the start of 2024.

Since CODEX is a solely web-based platform, the development team is small and is currently led by two volunteer academics. It is unlikely that the core team will need to expand considerably in the near future, as we expect a small development team to be sufficient to meet the needs of the community. The diversity of the core team is therefore necessarily limited. However, if the CODEX development team does require expansion in the future, it is of course essential that the team remains inclusive and representative of the communities it supports. Our vision for diversity and inclusivity is aligned with that of One Ocean Hub, with whom we will consult to achieve our shared goal of inclusivity in marine spatial planning and capacity strengthening across the globe.

As an open-access community tool, CODEX will never require payment from any users or contributors. The expenses associated with CODEX will be dominated by one-time software development costs, with hosting expected to be supported by academic institutional partners, and maintenance managed by academic volunteers. The need for accessible connectivity data will continue as long as evidence-based marine spatial planning remains important, i.e. the foreseeable future. The continued long-term operation of CODEX will therefore depend on (i) grants, and (ii) academic volunteers. We do not anticipate any problems with recruiting academic volunteers, since we have already received considerable interest, given the importance of this project. Given the relatively low anticipated maintenance costs of CODEX (beyond the initial development costs), we expect that a grant-supported funding model for CODEX is sustainable. There is precedence for this model. A project in another branch of ecology (demographic modelling), the COMPADRE database, has a similar data accessibility goal and business model to CODEX, and has been supported through grants and institutional partners since 1989. 

It is unlikely that this business model will need to change in the foreseeable future, although the project will of course incorporate feedback from end-users to ensure it is as useful as possible. CODEX is designed with flexible data formats in mind (it is based on multidimensional structured arrays, i.e. not tied to any specific methodology), so will remain useful when new approaches for quantifying connectivity are developed. Indeed, if successful in the marine realm, we hope to extend CODEX to terrestrial connectivity data once this is financially and logistically feasible.

The CODEX prototype was developed through funding obtained by Dr Otis Brunner from the Sustainable Development Goal Pioneers fund from the Okinawa Institute of Science and Technology.

Aside from initial development costs (which we expect to be around $50,000-$100,000), the maintenance costs for CODEX will be largely in the form of hosting fees for (i) the website, and (ii) the datasets. Since CODEX aims to host high-level network data rather than low-level raw data, file sizes are relatively small (<1GB, and KBs-MBs in most cases), so hosting fees for the datasets will not be excessive (it is unlikely that the entire database will exceed 1TB in size in the foreseeable future which, for context, would cost around $20/mo based on current Google Cloud storage prices). The actual costs may be lower, if an academic institutional partner hosts the datasets, as is the case for COMPADRE. We expect hosting fees for the website to cost a similar amount. The end-users we expect to use CODEX are primarily marine managers, and secondarily academics. Although we hope that CODEX will also be useful as a science communication tool, the main uses for CODEX are high-impact and relatively low-volume (i.e. on the order of hundreds to thousands of human users per month at most), so it is unlikely that CODEX will receive enough traffic to warrant high hosting costs. In other words, beyond initial development costs, we do not expect that annual maintenance costs will exceed $1,000, which will be relatively straightforward to support indefinitely through academic funding.