 David Mills is going from the Marine Centre in Wales is going to be talking about maximising the effects of marine monitoring to support the development of underwater marine renewable energy. Great to be here, fantastic to be here, pretty fantastic survey, great to be within the smithing which has been so pertinent to much of my least recent well and earlier work with careers and earlier a doctor of digital technologies. Matt was talking about smart boys yesterday, I would conduct a smart boy programme at CFAS probably about 20, 25 years ago so really an early rock doctor of solid state single channel recorders attached to single sensors left in sea and actually for me that started a complete journey in my working life and data rich from data ports data rich information for information. After working at CFAS for 25 years or more I then moved to a post funded by the European Commission within Ibanga to support the development of the safe data systems that in turn were delivering data and information to de-risk business decisions to independently the blue growth within the marine domain of Wales specifically focused on marine renewable energy and particularly subsurface systems tidal stream way potentially tidal lagoon but this work in particular which began last year is to support the specific development of the marine demonstration zone off the coast of Anglesey designated by the Crown Estate but where my slide has changed on the laptop but not on the, what do we need to do here, try to get a smooth link from there, try to get next slide with the remote thing so there's a remote here yeah all right well I try to do that okay yeah great okay so on the right hand side of the side you can see the sort of devices that are being deployed think of a windmill turned upside down placed in areas where there are strong tidal streams as I say there are two zones designated by the Crown Estate from testing subsystems waves off temperature tidal stream devices off the coast of Anglesey a number of these around around the UK so the work I'm just trying to spend about the EU we had the last largest significant funding from the European Regional Development Fund into Wales last year about 30 million to support the development of this capability the Crown Estate is also putting funding in as well why why is this important we have a climate emergency we need to decarbonise the economy we need a different mix of technologies to deliver the energy to the UK grid and clearly we need to understand whether these devices are are capable of addressing that challenge you'll see here digital technology is paying a key role in the work that I'm going to describe from gathering data right the way through of course to ensuring that data is reusable and trustable and provides the evidence-based division required we're deploying a land-based cyber infrastructure specifically to meet that data challenge and I'll finish up with some summary and create a say he can see the advice of some of the companies that we're working with so this is very much industry focused piece of work so what I'm going to describe to you then is the the marine characterisation search project this has been funded partially by the EU and by the Crown Estate to deliver something called the environmental mitigation and management plan this is a requirement of the marine licence required by the regulator national the NRW the national regulator within Wales it's comprised of a number of projects through which we're taking field surveys baseline data gathering continued monitoring through the presence of development building an operation of this scheme there will be mitigation measures required by the regulator and there's a cross cutting data management capacity which we which I will describe in team so sitting behind this scheme is an advisory board it comprised of the conservation bodies domain experts and the developers of these systems an evidence base is required which will be assembled around a number of questions these questions in turn need to a set of indicators which we need to populate with data the evidence must be for example the tidal state may influence the behaviour of sensitive species I haven't mentioned the sensitive species here primarily marine mammals diving seabirds migratory fish so those are the target species that we're interested in and we want to find out how their behaviour may change in relation to the building the operation of these underwater structures we're interested in the potential for changes before so baseline data what happens before what happens afterwards what happens in terms of distribution and specific behaviour of the target species there are specific metrics that are being put in place and so what you'll see this program began last year not all the observational systems are in place by far but it's a piece of work which is evolving so this really is a huge data challenge I don't think the actual funds of the work realised quite the challenge involved we go from everything on the left hand side from drone surveys looking for surface hydrographic features looking for observations of marine mammals for example surveys from small boats are within the demonstration zone again looking for both the birds and marine mammals you can see the sorts of annual data challenges annual estimates through these different observational strategies we move across to tagging this is done by RSPB making use of IoT and Loroan networks and really cutting-edge technologies deployed here photographic surveys from the shore interested in potential impacts on breeding of birds in the synergy off the coast of Pollyhead we will be gathering data using particularly the pan sand traps these are acoustic passive acoustic microphones deployed in a variety of different ways drifting boil rays fixed points over long periods of time towed behind vessels and range of that imaging and multiple we estimate about 270 terabytes of data a year which will be forming the basis of the evidence base upon which decisions will be made as to whether this technology be can be developed and implemented at scale so the other infrastructure in a way was sort of already in place it was in place because because earlier ERDF funded work to develop data solutions for the marine renewable energy sector and so we've been able and are in the process of adapting this to the particular challenges of this of the MRCP is the encapsulation so much of this is in place we have a metadata system we upload metadata we developed across the templates one of the challenges here is that battle for hearts and minds we cannot simply allow data to be stored on us bees and car drives and drawers this data can be legally challenged it needs to be in a place where we can find it and trust it and even reuse it so we're trying to make it as easy as possible for the researchers to engage with this system so we developed templates to work with the data gatherers we will upload the metadata to our metadata management system alongside the location data for the data itself the data itself would be placed in Amazon cloud storage the metadata was then can be interrogated in our Mac base web portal this portal was developed earlier it was developed co-developed with the industry we sat down worked with them not following perhaps the highly detailed schedule that we saw for example in workshop 2 yesterday but nevertheless embedding the principles of co-design from the outset was critical to make sure that this portal was fixed for purpose needs to be adapted for example to enable us to to bring in the new data types that we would command once there people can interrogate the method the searchable data can locate where the data is in your match we extract and interrogate the data in a variety of different ways and then download the overall process or other data provide some analytical we provide analytical facilities alongside the work portal in order to then to generate the outputs which will underpin the evidence base that we're creating this next slide then so it begins to show how we work with across the different data collectors across the different work packages I notice that there's a poster out there from the c-mammal research unit we're working very closely with a couple of the the main the named authors on that on that particular poster so for example one of the big data challenge which is this around of course the the c-mammal soundtracks the acoustic monitoring that's undertaking and so we're working with St Andrews to develop a new pipeline ensuring that the data once collected is secured and stored but also to embed for example the standard data analysis package which is pangar within the cloud so we can scale up the processing of this data and deliver outputs either as a standalone in terms of sightings of marine mammals identification abundance and distribution or to pass that down into what we're calling an EEMMMP reporting tool which is under development at the moment so that we can bring in data across all work packages there's an exciting opportunity here really to bring in data from multiple sources including abiotic and biotic data from other sources to enrich the evidence base with all available data or relevant data for this location to bring it together to visualize it to analyze it and then filter it and generate outputs agreed with the end users with the advisory board that will form the basis the the building blocks of the evidence base which must be sufficiently robust to enable us to withstand the scrutiny potentially as I'm saying with an early setting this final slide gives an example of the way that we're beginning to think how we may deploy AI to enhance our capacity can we for example reduce uncertainty in the processing PAM data using AI as a tool to let's say reduce the time to insight in these very large data sets that we gather a challenge of course for AI here and was picked up and by the last speaker whose name first name I would struggle to prevent but really hit the ball on the head this clearly needs to be an acceptable way of generating evidence that can be challenged is it we need to consider the way that AI may be explicable if we have a white box solution is this going to be fixed for purpose are there white box solutions so called that we can perhaps think about making making AI acceptable in this sort of to address these sorts of challenges and we clearly see the potential as we draw together all this data for really letting the picture tell the story to explore these data look for new relationships in the data that perhaps would not have been envisaged at the outset so to finalize and finish then that we have we've got the capacity to ingest very large amounts of data we will be working to the standards repeated so often here it's really interesting to see how frequently now we can come across the desire to embrace fair principles absolutely essential we can integrate the stage across all my work packages with this opportunity to generate new insights that were not anticipated that the outset also interesting I think there was a reason there is a real opportunity to create a forum here which now regulates the domain experts to work together to explore that data to reach a common understanding so it's clear the only feasible solution to this challenge is through the application of digital technologies both from data collection its analysis its long-term storage and latency and here we you know our purpose was to develop a monitor fit for purpose monitoring assessment framework and for me that challenge of what is fit for purpose is key that means we need clarity about purpose at the outset so we don't design a solution in search of a problem and finally that's the real challenge for us this issue I must have within space what does it look like what are the challenges it's going to encounter to ensure that we can determine whether these technologies are indeed a sensible strategy for the UK to adopt thank you Excellent timing got one minute for question I'm trying to pick up on your last point really about sustainability or legacy I think you said how do you imagine your portal that you co-design with users with spoke functionality and all that you know terabytes of data does that eventually make it into a data center absolutely and and in particular the functionality because you design the user interface and we've got mechanisms for ingesting data but not necessarily for the long-term legacy of the functionality I guess most top of my mind was legacy for data absolutely we're working to recognize standards we've found the marine metadata standards for the UK which is comparable for Europe and yes that data could be held in recognized data centers at the moment makes sense to keep it locally because we do things here that we couldn't do elsewhere and so yes we have some legacy we know what the costs are we're in any discussions about it's really I think at the end of the day what I always feel is the challenge for things that I've constantly the best things that slice bread is it has to earn the right to be sustained what's the next question is this the appropriate solution data is came we need to make sure as far as the functionality it's concerned I suppose I'm less concerned about that and I actually haven't thought about legacy for that aspect but I do feel at the end of the day it's good to be doing it you know what it's got to do something is we have an operational system here designed to do a specific job that may not be the job that's required in the future and so perhaps it's right but it should be written on the vine and maybe the design the concepts that you acquire during the co-design could also be packaged up somehow and when you donate to the long with the data so a new application thank you