At our “Fishing in the Data Age” event on 21 February 2023, we heard from a variety of speakers on the data collaborations underway involving the fishing industry, and the challenges and opportunities in ensuring effective and fair collection and use of fishing data.
Attendees heard from:
Peter Clark - Operations Director - Marine Management Organisation (MMO)
Tom Catchpole - Principal Fisheries Scientist - Cefas
Elena Balestri - Science Policy Officer - Scottish Fishermen’s Federation (SFF)
Steve Mackinson - Chief Scientific Officer - Scottish Pelagic Fishermen’s Association (SPFA)
John Reidar Mathiassen - Senior Research Scientist - Fisheries and New Biomarine Industry (SINTEF)
Jenny Stewart - Director of Planning and Change - Marine Management Organisation (MMO)
Rebecca Skirrow - Senior Fisheries Scientist - Cefas
A full recording of the online event can be found here:
Following the presentations, attendees put their questions to the panel. Written summaries of all questions and answers are provided below. Please note that these answers reflect the views of the panelists, rather than the APPG Secretariat or its Members.
Have we asked what data we need, and do we have the people who can process that data usefully to yield management advice?
Catchpole said that Cefas has well-established data monitoring programmes which have been designed over years to deliver evidence needed to drive current fisheries management processes, and that prioritisation has been undertaken to establish key data needs for the new Fisheries Management Plans. He said that Cefas has excellent data analysts, and with the skills needed becoming more specialised over time, the agency is always recruiting for these. Catchpole added that working closely with universities, which can be better placed to stay at the forefront of fast-moving technologies, is key for Cefas. He used the example of the University of East Anglia, which has an excellent team working on image analysis using AI.
Balestri said that, at least in the case of Scotland, it is increasingly difficult to find the funding and skilled people to maintain and indeed enhance fisheries research and monitoring; this is in contrast to faster-growing industries such as offshore renewables. She commented that research and monitoring should be more ambitious, which requires enhancing communication between policymakers, scientists, fishers and other stakeholders to help to ensure that the right data is collected.
Many existing data collection programmes focus on larger fishing boats with trained crew. What provision is there for under-10 metre boats without the capacity to measure fish or carry out other data collection?
As an addendum to this question, Sherryll Murray MP asked: Experienced fishers can make accurate estimations of the grading and weight of their catch. Could this be taken into account with smaller boats which can’t take on the equipment needed for taking measurements?
Peter Clark said that the MMO aims to take the practicalities of different fishing operations into account when designing its programmes. He used the example of the Catch App, which allows for experienced skippers to estimate rather than weigh their catches. In response to Sheryll Murray’s question specifically, he said that in some cases the MMO asks skippers for self-declarations, which are validated afterwards to account for practicalities such as the loss of weight as the catches dry out; where self-declarations are used, the MMO would only ask for additional information where any anomalous data arose. Clark acknowledged that the English fleet has a comparatively large proportion of small vessels, so it is important to include these to avoid missing out on significant swathes of data, which involves working in partnership to ensure solutions exist for all.
Tom Catchpole said that Cefas has also been working with smaller inshore vessels, for example combining AI and camera technology on crab and lobster potters. He said that this particular collaboration has demonstrated the feasibility of automatically generating data from potting vessels and transmitting it straight to a database. Catchpole agreed that there are different challenges for different fleets, and that lack of space, crew, and time on smaller vessels is an issue, although the potential for technology such as mobile phones for capturing images is being explored as a way to address challenges.
How has Cefas convinced vessels to take part in Remote Electronic Monitoring (REM) trials, given that fishers associate REM with compliance measures?
Catchpole acknowledged that the topic of compliance and enforcement is always one of the first to be raised in conversations around REM. He said that Cefas views REM as one of a number of tools to generate data, and that in some cases REM offers the best resolution. Catchpole gave the example of skippers in southwest England being willing to install REM systems to generate highly detailed catch data for scientific purposes. He added that it is essential to always discuss the intended use of data at the outset of any collaboration, to make clear to participants about what data would be used and how, and to have data sharing agreements in place.
Clark said that, as with any other source of data, REM data could be used to identify non-compliance. However, this is not the primary purpose of REM when it comes to the MMO’s activities, and the MMO would only use data from this source in a highly targeted way if for monitoring purposes.
Are there opportunities for fisheries to become involved with data collection projects such as EVERYFISH?
John Reidar Mathiassen said that in the specific case of EVERYFISH, there is scope for fisheries to join this ongoing project if they do not need funding to do so. Catchpole added that Cefas’ involvement in EVERYFISH centres on mid-scale demersal fisheries, which connects with Cefas’ other work including the voluntary scientific REM study; as such, there may be opportunities for UK fisheries to get involved with this or with other projects.
Are there mechanisms in development or in place for agencies such as Cefas and IFCAs to collaborate on the analysis of new types of data such as that from inshore Vessel Monitoring Systems (i-VMS)?
Jenny Stewart (MMO) said that as part of the i-VMS rollout, the MMO is currently focusing on building capability to merge that data across its suite of fisheries data, rather than partnering to share or merge the data more widely.
Catchpole added that the GeoFISH tool which links VMS data with logbook data is also able to incorporate i-VMS data, and given that datasets are held with the MMO but accessible by Cefas, some such data connections are already in place.
Is i-VMS data going to be made publicly available in an anonymised format for use in Environmental Impact Assessments or other consenting processes? If so, when might this be published?
Stewart said that the project to implement i-VMS is still in the early stages, and there are currently no plans to publish the data from this.
Could i-VMS be used in enforcement where there is no other legal evidence of non-compliance? For example, if a vessel was fishing in a Marine Protected Area?
Clarke said that it was not possible to make generalisations, and that each and any case would have to be considered on its own merits. However, he added that he would be surprised if i-VMS was ever used in isolation, and that it is much more likely that such data would need to form part of a suite of evidence, or else could be used to indicate where investigation should be focused.
Sheryll Murray MP commented that if i-VMS showed a vessel to be in an area, this alone could not provide proof of fishing activity. In response to Clarke saying that it might be possible to infer that fishing has taken place if the vessel showed a slower rate of knots, Sheryll added that there might be circumstances such as a vessel breaking down. Clarke replied that this is why it is unlikely that i-VMS data would ever be used exclusively in any investigation; rather, it could be a useful source of intelligence if combined with other sources to enable the evidential test for prosecution to be passed. He added that prosecution was always a last resort for the MMO anyway.
Would it be feasible for the fishing industry to contribute to the costs of data collection in the future?
Elena Balestri said that this could be considered in the future, but that the focus should be on involving the industry in co-management and enabling it to develop policies and processes that are appealing and practical for fishers. She added that it would be unfair to ask the industry to take on the costs of any initiatives that might be imposed instead of co-developed.
Steve Mackinson agreed with Balestri, saying that costs should be commensurate with roles and responsibilities and the degree to which the industry is passively or actively involved in collaborations. He added that different fleets within the industry have different abilities to shoulder costs, and so there could not be a one-size-fits-all approach. He also noted that the industry already pays for surveys in certain cases (e.g. in transferable quota systems) but that traditionally the public purse has covered costs relating to surveys with a conservation focus.
There is a lot of discussion of data, but who owns it?
Mackinson said that in the case of the Scottish pelagic industry-science data collection programme, the data owners are the data providers. The specific definition is as follows: “Data provider or owner: entity providing data (i.e. individual pelagic vessels and/ or factories), and responsible for agreeing data use”. He added that it is important to have in place a data policy and any data sharing agreements which define different roles and responsibilities.
Balestri said that for any kind of data collection, it is important to establish protocols on data ownership, data sharing agreements and clear understandings of the purpose and intended use of the collected data. She added that as data collection can have implications for confidentiality and business impact, it is essential to clarify what level of anonymisation will be applied to data before it can be published or released to mitigate any such impact.
Rebecca Skirrow said that the owner(s) of the data can vary between projects and programmes and is often determined before data collection begins. Data sharing and usage agreements can outline how the data can be used and who is the ultimate owner, but often the party who has commissioned the data collection and/or is paying for it ultimately owns the data. For example, data collected by Cefas on behalf of Defra is ultimately owned by the UK government. But Skirrow added that where data is collected in collaboration with partners (such as members of the fishing industry), those partners should also have a say regarding how the data is handled, hence the importance of having agreements around data use, storage, and deletion in place.
Much data is provided to the MMO that would be of real use to the seafood market. Has the MMO considered how this data could be provided to the seafood market in a way that would be GDPR-compliant and protect fishers’ commercially sensitive information (e.g. by providing at low-resolution/ ICES sub-rectangle level)?
Stewart said that making the best use of - and generating value from - the data it collects is one of the MMO’s strategic objectives, and that the MMO already makes a range of data publicly available. She added that the MMO would welcome suggestions on what further data provision would be useful.
Regarding the Remote Electronic Monitoring (REM) projects which Cefas is running or participating in, are the REM analysts AI or human?
Skirrow explained that REM analysis is currently undertaken by human analysts, but Cefas and many others are working towards automating elements of this process using AI. She used the examples of the Horizon 2020 SMARTFISH project and the more recent Horizon Europe EVERYFISH (Digital transition of catch monitoring in European fisheries) project, which Cefas has been involved in. Skirrow said that EVERYFISH aims to develop a suite of technologies that will automate catch accounting, which will include integration of AI algorithms into REM software.
Given that the same level of traceability expected of fish caught in the UK is not being applied to imports, does this disadvantage the UK fishing industry?
Clark said that some processes are in place for monitoring imports, including partnerships with port health authorities. However, the importance of seafood exports to many UK businesses means that ensuring export processes are in place has been the priority. He acknowledged that more work is needed on imports, and that Defra is keen for this.