 Well, very, very good morning, everyone. Thank you very much for joining today. This is part of the sixth series of the webinars that we're running as part of the Constructing a Digital Environment program. Details of the earlier webinars we've run are all on our website and on the YouTube channel, which we have, links to both of those will be popped onto the chat in just a moment. There's a Q&A section in the footer of the page and if you have questions for the colleagues today, please pop them in and we'll do our best to bring those in. So I'm Steve Halles, I'm one of the Constructing a Digital Environment, Digital Environment champions of the program. And the focus of this program really looks at supporting the development and the application of digital approaches in environmental science. And in this series, we're focusing on the demonstrator projects that we are supporting through the program. So there's a number of the demonstrator projects and each of the webinar looks at each one in turn. And this area that we're covering is a very important one because we live in a complex world and the human and natural worlds intersect, of course, even in ways that are unimaginably complex. And so in effect, we need improved and better tools to enable us to tease apart and understand the processes that are going on to understand the implications to our society and also both at a local city, regional and indeed national and global scale. And these are all very much key issues that were discussed even in COP 27 recently, of course. And so one of the things to notice the focus in these webinars is very much on the concept of the digital environment. And this is an approach whereby we are trying to represent natural processes and the world around us in digital forms and both methodologically and of course philosophically developing a digitally enabled environment for the benefit of all scientists, policy makers, businesses, communities and the public. Trying to understand these complexities that we have and we touch on an arc of technologies from capturing data with sensors through to the handling and transport of that data, the storage of that data, the analyses, the analytics that go without the processing and the visualization and decision support activities. And many of our demonstrators, as today will hear, pick off across that arc. And very much the other focus of these talks is the interdisciplinary nature of the projects that we are supporting. And that's one of the reasons why we're delighted to have the whole team from the Retina project with us today to give that full spectrum of insight into what it is to take part and to put together a project like Retina. And I'm very pleased to welcome all of you and Jagadish as the lead for this project. Jagadish, I wonder if I could ask you in a moment to introduce all of the colleagues on the team. And then I believe you're going to give us a bit of an overview with PowerPoint and then we'll get on with the discussion. So if that's okay, Jagadish, over to you. Thank you very much, Tiffan. Thanks for inviting us. And I'm very pleased to talk about Retina project and introduce our team this one. And before I ask the people to introduce themselves to our team, I would like to say that this project was led by Jane Sutton Institute with collaboration with Aberdeen Institute and CH. We have a huge number of people working in this project who are participating in this webinar as a basically subset of them. And also we have a really good mix of experts here including both from File Science, Computing Science, Biogy Chemistry, Modeling, Social Science and even the app development. It's a real diverse group. We are excited to show all these different perspectives how they come together to produce something innovative in this project. So I would like to introduce Pete Smith here. Hello, I'm Pete Smith from the University of Aberdeen. Keith Macalarch. Hi, I'm Keith Macalarch based at James Sutton Institute outside Aberdeen. Elizabeth Cowdery. Hi, I'm Elizabeth Cowdery. I'm a postdoctoral researcher at the James Hutton Institute. Thanks, Maxi. Damian. Hi, I'm Damian Bienkowski. I'm also a researcher at the James Hutton Institute based just outside Dundee. Thanks, Damian. David. Hi, I'm David Donnelly. I'm another Hutton scientist also based up in Aberdeen. And finally, Becky. Hi, I'm Becky Smith. I'm a research software engineer for the James Hutton Institute, also based near Dundee. Thanks, Ethan. I would like to share my presentation just to give you an overview of the project then you can have a discussion. Please, the floor is yours then. Let's give us a short overview. Hope everybody is able to see the flight. Yeah. And first of all, thanks for inviting us again. And I'm very pleased to talk about the digital project and which focus on developing a digital, monitoring the replacement reporting system. Just to give you a background, this is the UK greenhouse gas emissions by sector. As you can see from 1990 to 2020, there are substantial reductions were achieved in some of the sectors, especially energy sector, focusing on renewables. But I would like to highlight here the agriculture sector, which is almost state flat, more like. And that means there is a huge possibility that just the greenhouse gas emissions, but there is little progress was made in agriculture sector. So it is the same thing in Scotland. Scotland agriculture stands at the third position in terms of the source. If you see the change in the net emissions from last several decades, there is substantial reductions were achieved in energy sector and land use sector, especially increasing the forest areas. But when we come to the agriculture sector, there is a very little progress was made even though there is a huge potential in this sector. So if we want to achieve net zero by 2050, we need to make substantial reductions in every sector. As we can see in this graph, 1990 to 2018, there is reduction for achieve, but in order to hit 2050 net zero, the streak of the curve indicates that we need to make substantial reductions, including the agriculture sector. But to just go into the background of this one, agriculture sector Q is a huge potential in terms of harvesting the carbon into the soil. And there's a wider consensus on the scientific community that there is about this potential and the crop line worldwide could speak best between 0.9 to 1.85 petagram carbon per year. But soil organic carbon, but the complication here is that soil organic carbon content and the soils and greenhouse gas emissions cannot be easily measured, which is one of the key barriers for implementing programs to achieve net zero. So there is a need for credible, reliable monitoring verification reporting platform to reduce uncertainties in national reporting and also emission trading, especially operationalize the carbon market in agriculture sector, and also track toward net zero, because we need to inform the farmer where they are in terms of their journey towards net zero and make them make a conscious decision something. So the overall objective of the retina is to create an iterative near real-time digital monitoring and reporting system that forecast crop yield, greenhouse gas emissions and carbon sequestration at the farm level. So this will also create an end-to-end data pipeline from collection to the processing and understanding the data and to improve the way we visualize the data advancing modeling and nation making. So this will give a conceptual diagram how the entire retina works. We, in this project, we first look at the farm level, we look at the baseline, what is actually the baseline, what the farmers are doing. In order to do that, we need to have an, we doubt an automatic land parcel detection in the app. That means once you choose a land parcel, we integrated all the information and also see that a land parcel in the background. That means the farmers will be able to see what is with the existing databases, what is the soil carbon stock, what is the soil carbon levels, what is the nitrogen and all the set of information. And also the climate data with the existing databases. And then we don't play smart sampling. That means we guide the farmers to take the soil samples based on the local information, which is available, which is pulled all the information from the existing databases. So this is more tailored to the local condition. That's what we call it as a smart sampling. So then we deployed the sensors. We are capturing the information at three different scales. One is at the plot scale, where we are deploying the soil temperature, moisture and other sensors into the fields. And also at the landscape level, we are running the drone to look at the crop growth and biomass estimation so that we can estimate how much is the soil carbon input into the soil. So again, we are using a remote sampling to track the fallow period and also plowing even and also the planting and harvesting. We are trying to track before and validate what is really going on in the field. So we are integrating all these information and also we are in order to, we are feeding into the model basically. So produce the output and pass on to the end user. So in order to run the model, there are three basic information we need. One is biophysical information, soil climate condition and the climate and the management information. So by biophysical information, we are collecting through the real-time monitoring with sensors, whereas the management information is still missing. That is the reason we have developed an app, which is called Retina app, which is used to collect the information and also provide the output to the farmers. So the app will, the all the information that is going to the app that will be utilized by the farmers for relation making on the field. So the app will, the farmers will enter the information, the management information to the app that will feedback into the model, into the system. So the model will run even for the future conditions and then predict what is the best management practice for the particular location and advise the farmers once they adapt that one, again, the feedback will continue. So this will be useful for two stakeholders, there are major stakeholders, one is the farmers and agri-food businesses. The farmers can utilize for the carbon farming and also better management, they journey towards the next zero. And in future, they can use it for carbon trades because it's a more digital monitoring and verification system where we can provide a digital reference to every activity so that it will provide more credibility to the system. And also it will substantially reduce the cost of monitoring and verification. So because we are using the digital system where it's currently in the market, they need to go physically and monitor and verify which is very expensive. So, and also agri-businesses if they want to go towards the next zero, like Netflix and many other companies which are rely on the supply chain, where the supply chain emissions are very high if they want to reduce the product emissions, they need to focus on supply chain. So they can use the system to monitor their emissions and track their emissions through the system because it produces a real-time update to that. So what are the benefits of regional technology? Yeah. So it's an iterative real-time model prediction that gives and also a complete data traceability. This is very important because it will provide a lot of credibility to the carbon market if you want to go back. And also data collection directly from the field without any input from the farmers. There are many tools existing that currently in the market which are very cumbersome, needs a lot of input from the farmers and the way you process and input to the farm information actually that will change the outcome. So we are bypassing that actually so that the farmers need not input much information here. And real-time data driven science bring more credibility and transparency and real-time data visualization for information at the farm level. For example, if the farmers have made any activity in the field. For example, if today the farmers love the land, the system immediately show what is the impact in terms of greenhouse gas emission and how the carbon is going to change in the field. So they can make an information. And we are using tier three models instead of tier two factors. That's also enhancement in terms of net quality and high quality soil and climate data such as an integration. And also this will pay a way for the digital space as we go along. So one of the three requisites for the disciplines is to real-time system, real-time system. So this is what actually we are trying to do. So in terms of pretty outreach, what we have done so far, we have focused on four areas. One is that we have an engagement with the industry in the scientific community and policy and public. We have presented several outcomes in the Royal Highland Show and Aroble Scotland where we have interacted with several farmers there. So that's one engagement we have and several are coming up this year too. The project runs up to this year all get sold. We have several events are planned. And also we have an engagement with the industry. We have several discussions with Singenta, Nestle, Yoval Yogurt, Agni Carbon. And also in fact, we put a bit to the depth of farming features in order to scale this technology. It's in the stage two, which we have, it went up to the stage two. So they have a lot of interest from the industry on this project. And also scientific community, we have presented some of the outcomes of this project in various international forums, soil against matter dynamics in CO and other conferences. And we have two scientific papers published and several are in the pipeline. And actually the technology we have developed in this retina led to several projects for our follow-on projects. Some of them were EU SENS project, 1.4 million and SOAR project, NECSOAR project, another 1 million and RESIS transition projects, which is 1.5 million, which are all based more or less on the things which we have done in retina. And also we have a lot of policy engagement during this last three years. And we are interacting with the government and also DEFRA have given several presentations in the DEFRA. And also we are working very closely with the Scottish government. And this is one of the picture we can show that this is the, I'm explaining to the, about the project to Mary Gusion, who's the Cabinet Secretary of the Scottish government. Thank you. So finally, the one thing I would like to say is that retina spin out. And this is what we have developed in the retina. We want to scale it up through the spin out that's called the carbon experts, which is in the pipeline. Thank you very much. I will stop here. Jack, thank you very much for that overview, really very comprehensive overview of the project. And it's good to get that sort of feeling for some of the impact that you're clearly having with the Scottish government as well. I mean, you've mentioned a number of challenges really. And I was thinking in terms of maybe Pete, I could sort of turn to you here. Some of the negative things, the emissions from this sector are growing and there needs to be ways to address that. And on the other hand, side of the coin, some of the positives in the sense of the opportunities for sequestration and the mention of things like credits. I'm really interested in that. What's your take, Pete, on the sort of size of the challenge and what that challenge is? Yeah, so the challenge is to get into net zero. And we know that there are going to be some emissions that are very, very difficult to abate, and we must offset those with sinks. So we already know that. And soil carbon sequestration offers the potential for a really big sink, about five gigatons of CO2 per year globally, which is about 10% of our current global emissions. Now that's a technical potential, not an economic potential, but it shows the size of the potential mitigation benefits that we can get from carbon sequestration. The biggest barrier, why it's not happening, is it's difficult to monitor, report and verify those changes. When you plant a forest, you can go and put your tape measure around trees and you can demonstrate that it's growing every year. And you've got a good estimate then of the increase in biomass. You can't do that with soils. You can't see the organic matter. So MRV is the biggest challenge. So what's stopping it being implemented more widely and stopping the farmers from getting the carbon credits or the benefits from it is the lack of MRV. So that's what the RETINA project aims to address. And if we can crack that nut, it could unlock the potential that soils have to form that real, real big sink that we know they have potential to do. So you see a way of actually carbon credits extending to the soil in this way through monitoring and that's really interesting. Yeah, okay. I mean, I sense then that one needs digital approaches to monitor the soil conditions. You've talked about sensors going in the ground. And of course, one of the thrust of the program that's constructing a digital environment is this notion of digital environment. I'm just wondering what the context is for your sort of interpretation team as in terms of the digital environment. Maybe, Kate, you have a view on the, how do you deal with digital environment in your work and how have you addressed this? Yeah, no, thanks, Steve. And I see it as a sort of a gradient like many others. We've been carrying out environmental modeling for 40, 50 plus years. And this was highlighted in the sort of the report by John Sidborn and others about the information management framework for environmental digital twins. This is a gradation from environmental models to maybe real-time digital twins. So picking up from where Pete left off to have these MRV systems, then it just needs to be digital to bring together the information and knowledge to provide those robust and transparent ways to enable that all parties have got confidence in the system. And Jagadish might want to say a bit more because his expertise in carbon MRVs. Yeah, what do you think, Jack? Yeah, I would certainly add to this. What actually we have been in the modeling for the last 20 years, at least myself. One of the limitations we see is that lack of data that is coming from the field, the models can learn about the system only if there is a data flow from the system. A continuous data flow really helps the model to actually learn about the system better and really mimic better. So this advent of digital technology, especially the sense of technology really help us because that's what actually data we will streamline the data from the field to the model. And then that will make it really powerful because the model field is capable of learning about the system and get better and better over a period of time. So that gives the, you know, are there particular areas where you could apply these approaches in other contexts or other areas that you think perhaps wouldn't be appropriate for this sort of approach and what are those and why? No, actually I see much more at the system level even though we are focusing here on the carbon and greenhouse gas emissions, but these models can be used for nitrogen optimization in the field and many other things related to the agriculture management actually, this can be extended and many models can be bring in into the system and into the framework actually, even the biodiversity elements of it, which can be included in the lack of part of it so that we can comprehensively look it into the environment and record, monitor and verify. This is very three elements which are lacking even in the government policies, many policies now targeting towards the results oriented approach. So when we want to look at the results, we need to know, we need to monitor and verify those results. So in terms of those objectives, those sort of monitoring and so on, we clearly the role of digital is key here. Let's turn our focus a little bit to the technology that you're using. I mean, it'd be interesting to know from you what sorts of technology are you using? You've mentioned sensors in the ground. You showed the photograph of a rather interesting looking sensor with the minister. Maybe some of the, who are the sensor people here? I'll jump in. It's Damia and myself kind of do a lot of the things with the sensors. So I have come on board to manage the, or to help manage the Laura-WAN enabled sensors. So we've got two sites. One is arable and one is upland pasture and they both have the same sensors across them. So each site has a weather station and a CO2 sensor. And then both sites have a differing number of soil, moisture and temperature profiles, they're called, which stick into the ground and take readings. One of the things that we've sort of had to manage is being able to lower works on line of site technology, which sounds great because it can go up to 10 miles and things unless there's a hill or some trees or a building. And so it can be a bit more complicated and working out what sensors can actually be seen by each gateway, which is the receiver for the sensors. We've also had issues in terms of the some of in the upland pasture, the ground gets more solid higher or closer to the surface. So we've had to use shorter probes, which means that they don't have as many measurements. Those ones were also attacked by some sheep. Sheep found the sensors very interesting. We've now built them little protective cages, which you can see, but to begin with, they were very interested in our sensors and did want to eat them. On the arable farm, we've had a similar issue in the attractor versus the sensor is bad. The sensors don't like that very much. So our original placement decision didn't take into account where the tractors would have to drive. And that's I think that's been a key learning point because we don't want those sensors run over and something that we would definitely do differently. Next time is not put them in the tram lines and yeah. So those are the sort of the Laura sort of radio wave technology sensors, the sensors that we've got in place at the moment. And Damien does some cool things for drones and stuff. So I'll let him talk about that. Well, I like the idea of having less tasty sensors. I think that's a key knowledge that I don't think other teams have mentioned at this point. So it's a new one for us. Damien, yes, I mean, you're using aerial platforms as well to capture to get there. Because on the one hand, you have a specific sensor at a point. And I guess, Becky, the challenge would be, how do you know that that point is representative of the area that you're trying to capture? But then on the other hand, Damien, you're using drones and aerial platforms to look at the whole sort of whole farm or whole enterprise. And how have you approached that? And what sort of challenges have you had with that? Yeah, so thanks very much, Stephen. Yeah, so we're writing that. I've been conducting aerial surveys of those two experimental sites that Becky was just mentioning. And the ambition was to produce maps of the distribution of above ground biomass. And to do this, I was using off the shelf UAV with multi-spectral sensors. And originally, so I find multiple surveys at each site over the growing seasons. And originally, because I'm quite new to the area of aerial imaging, I had sort of made an assumption that existing models of biomass that other people had produced and published might would be applicable. But we discovered quite quickly that the models that other people have produced based on their data, their sensors collected under different conditions, couldn't be applied. To the data that we were collecting. And so in tandem with the aerial surveys, we've also been collecting ground truth measurements of biomass. So, you know, cutting quadrats and drawing them at the arable site and using a GPS pressure plate meter at the upland pasture one. And so actually we are sort of still in the process of constructing our own biomass models to map the biomass over the fields to feed into those MRV. So that's kind of been one of the big sort of challenges. What was the reason why you couldn't use the models with the sensors you had? What did you find? They just gave bizarre results as to how much biomass. It's just that the situations that they had been trained on, those models weren't directly applicable to what we were doing, even though, you know, pasture grass and things like that. But it perhaps only been trained on a subset of situations of the heights of grass and things. And we were looking at a broader kind of range of data. And so areas where the grass is obviously very thin, it's still giving quite high values and things. So it just, it was clearly incorrect. And so we're having to develop our own ones. And those are sort of still under construction and to be fed into the MRV models going forward. Re-evaluate those models as well? Yeah, well, I mean, those models work for, you know, of course, we're using a slightly different, you know, set of sensors and, you know, different geographic situation. So, yeah. And we have other sort of challenges that I guess will come into play moving forward. My collection of data and the processing of it is very sort of manual, has a lot of manual steps. And I think for the future automation of that is probably going to be a bit of a challenge if you, if there was an expectation that a particular landowner would have sort of surveys done of their land, when to do it best time. I'm interested, and Jagdish mentioned that we're sort of paving the way I think was the phrase used for digital twins will be interesting to come back to that in a bit and see, you know, what would need to happen. You've mentioned automation being one thing. You know, so, you know, back in day when you're referring to sensors in the ground and obviously at aerial platforms, but of course there's other ways of capturing data as well. And, you know, there's a link that I suppose needs to be captured where people and technology interact as well, particularly if you're trying to build citizen science type approaches and, you know, I know, David, you've got your camera either. Yes, David, you've built a number of these sort of tools, and I think that's been factored into the project. And, you know, in our discussion before, I think you mentioned there was a short clip that you might you might even show some of the apps that you built, but what's your take on that sort of interface between people and technology in projects like this? I think the key thing is, and it's really no complicated, understanding that you're making the app for somebody else. You're not making the app for yourself. You're not making the app for research scientists. In this case, we're making the app for a farmer, so it's got to be suitable for a farmer. Even to the extent that, well, certainly this part of the world is, your archetypical farmer is a kind of son of the soil. So an app on a phone with tiny little buttons just isn't going to work. It has to be actually physically usable. So that's one aspect. The other aspect is, well, actually, what I should say, we did have plans for effective stakeholder engagement, but unfortunately COVID got in the way of those and they've been delayed. They had to be abandoned at last minute. So, but we have some more planned for the next few weeks and we'll try and get those, the outcome of those into the app. And we have a number of active farmers, actually the institutes, colleagues, either the spouses or farmers. Yes, mentioned the video. Do you want to run the video? And I can talk to the video, Jagadish. I think that would be good if you've got it though. We can, I mean, there was a very, while Jagadish is just doing that though, it was a very interesting talk. The last webinar all about the idea of co-creation of citizen science with apps. I mean, some of the concepts that the team were drawing could be of interest as you develop the thoughts in that area. Indeed. As I say, unfortunately, COVID really got in the way of the stage of the project and we were doing that. Yes, unfortunately. Jagadish, it matters not whether you run it or not, but if you wish to share the screen instead of going. Just talking to it. Yeah, so this is a whistle stop. So three minutes. So the app includes registration sections so you can request to be a user of the app. The information stored in a server straight through to site configuration. If you like to set up the digital farm, this is using an interactive map as you can see. Click on the GPS button to take it to your location. This is simulated and it's actually taken to the research farm in Glensaw. Change the background so we can see our fields. Click to pick your field. Now, that's communication and survey database. So grabs the field boundaries from your database in real time, records it remotely in our server and also on the device. Now, this is showing a map of soil carbon. Click anywhere and you can see the top soil carbon. And this is very rushed. I'm really sorry, but Jagadish mentioned sampling. So this is where we generate sample locations based on the environmental conditions at the site. The app also helps you to measure the samples by using the device sensor, so GPS, but I'll just skip on past that. And this section, we can look at a bit more detail. Jagadish mentioned extracting information from existing databases. So I can show an example of this here. Zoom into one of the fields we've selected. This is a part of our boundary research farm. So you can get this all stopped just by clicking on the polygon. Just in the computation. And there we have this total soil carbon for that field. Topsoil and subsoil. And so nitrogen. Calculated in real time from the not in solid database. We also record farm activity. So. I'm sure. So this is a historic timeline. Spring barley potatoes and winter wheat. We also have a Jagadish mentioned submit the farm activity to the database. So it gets included in the models. This is a little tool of God for recording the date and crop and activity associated with it. This all feeds into the database on the server, which is then used to improve the model and produce new outputs for the. The farm. I won't press send, but this would send the information to the database as well as storing information and database on the device. And then we'll go back to the time of sensor data that Becky's been talking about. It's not quite real time at the moment. I'm not sure whether the sensor is down, but that's the last recorded. Records for the. So you saw moisture. So temperature. On the site. Becky's looking confused. We'll come back to that. She began, isn't it? Probably. And finally, just to show some sample results. And. Carbon sequestration. Over a period of 50 years. And. This will quickly show to, yeah, too quick. And that's my three minute. Show of the app. We can, we can go back to it. But. One of the main challenges with this is using. Multiple technologies mean this is using. Ordinary survey data in real time. Grabbing is using mapping services from ourselves is using. Our data to compute special. And it's also using three different types of databases. Three or four different types of computer languages. Tying all these together is one of the big challenges because there are no, no standards. So you have to write the glue. It ties one thing to another thing to another thing. And make it all useful and meaningful. But one of the, one of the sort of observations here is that the, the beauty of systems like this is, is that the farmer should be familiar in a sense if they've been using Google maps and other other packages like that, they should take to what you've done very nicely. The idea of bringing up a map, changing the background, selecting things and so on. So it's, there's a familiarity. I think that's good. Another advantage presumably is that you don't need your stakeholders to have expensive software and other packages because they, all they have to have is the app and then, and then they're often away, I guess. Yeah. As long as we retain it as being free. Yeah. We can continue to use these services because all these services are predicated on not charging. Right. If you start to charge your product, then you have to enter into different agreement, see what the ordinance survey into a commercial agreement, which the ordinance survey would be happy to do. Yes. But then you have to start charging the farm and charge. I think as long as you can keep these things free upon of use, it does make things much more straightforward. You're putting a tremendous amount of power in people's hand and all that information that they can bring up for that, that location. Thanks very much, David. That's great. I mean, I'm just Elizabeth, if I, if I may turn to you, keep your videos going, David. Don't, don't, don't disappear. Elizabeth, you know, you've obviously been working on the, the technology of the rest of the project as well. I'm just wondering whether you have any thoughts about, you know, things that you've learned as the, as the project has progressed in your role and. Implementing the different technologies and the old, that's not it. Would you do things differently if you were starting again? Yes. Well, um, so, uh, I could talk about a couple of things actually. One thing I could talk about a bit is the automation side of things. Yeah. Which actually could, um, I think tie a little bit into, um, design as well. And actually, I don't know. Um, it might, since, since we sort of have been talking a little bit about automation, you did sort of ask questions about that. It might be useful to show that slide as well, because it does sort of tie in, uh, the things that David and, and David and Becky talked about as well. I think maybe, um, I don't know if Jagadish, if you want to, if you want to share it, I can control your screen. And flip through the things. This is never going to work. It's never going to work. Maybe I should, maybe I should share it then. Uh, shall I, shall I do that? Yeah, go for it. Okay. Um, one second. Just while Betsy's doing that, um, certainly from, from our perspective with the sensors and things, one of the things we have found is that one location of the sensors, which David highlighted with his video and had me very confused because I fixed that issue last night. Um, so. Um, the data is up and I checked the live screen and I can see the data is up and then I remembered it was a video you were using. You weren't demoing live. Um, is, yeah, I think having more time to play with the sensors and the technology before putting it into the project would have been nice. Um, and would definitely be something we would do. I would do if I was doing it again next time, because this is the first time I've worked with water sensors. Um, and we've custom built all of the sort of, um, the servers are custom configured all of the servers and things in the back end. So we control everything. Um, but it was a steep learning curve and that would have been good to come before the project. I think this, um, Becky and Elizabeth, you know, this is all about what are the learning points ready to share with colleagues and other folks who might be thinking of doing projects like this in the future. How are you doing with the video? So. Oh yeah. No, I actually, I suppose that's an interesting point is that, you know, when you are using open, open software, open data, open services like that, often you are your own maintainer, um, which has its benefits, um, because you have full control and open access, but often then you are also responsible for, uh, your own maintenance. Uh, and also, you know, uh, understanding the services yourself. And of course now so many things have such wonderful documentation that it is, is very easy. Well, I should say the learning curve is a lot less steep, but at the same time, um, for example, we did put a lot on Becky's plate, uh, asking to get up and running with the, the lower one, uh, servers, um, or, uh, uh, uh, I think sensors I should say, um, because that, that was, yeah, complete, completely new technology. Um, but, uh, you know, there, there are other, there are other, you know, paid services, uh, that you can, can ask for. Um, but it is, it is a different, it is a different ballgame if you, if you take that, if you take that route. I'm just looking, um, um, thank you very much, um, Edward Darling, who popped a quest, uh, point into the, the chat. It looks like the red list revival, if I read that right, and their work with a life map, they're proposing having a discussion about potentially mutual interests, that sounds good. So something to follow up following this. And anyone else who'd like to pose a question to the team, please pop questions in. But if I may, perhaps, I think we've had a chance to discuss the technology, but one of the things that's interesting really is the policy implications of tools like this, and the way that technologies support the evolution and the prosecution of the policies, agricultural policies at the national level. Greg, if I may turn back to you, I'm just wondering what the scope of the policy engagement you've been able to have and demonstrate in the project is, and how that's gone? We have a couple of rounds of discussion with both the courts of government and DEFRA, and this one, and still we are in touch and plan several events in the next couple of months. The issue is that, as I mentioned before, even the DEFRA and also the courts of government, looking for the payments in future, they want to link for the results. But we need to provide the technology solutions. The reason why they can't do it at this point is, there's no credible way they can monitor and verify these things at the ground level. So this is one of the limitations. I think that's what I see the role of digital technologies can play a big role, especially projects like this, bringing different technologies together and make this work and monitor and verify, and that will allow the policy to roll on. One of the things we are looking potentially at this moment is courts of government is, courts of government already pledged 51 million for baseline soil monitoring. So the farmers can enroll, and that will be, they can do the soil testing freely that we paid with the government. And we want to utilize the data into this, where they can use the retina setup in their farm, so that they can be informed about the net zero, how they travel towards the net zero. And the government can use this one for monitor how the progress that is happening in terms of soil carbon stock across, once we scale it up. So there is a huge potential, but there are few limitations too. That's what we are discussing with the courts of government and also with the DEFRA. Thank you. David, you're showing the current view, I think that. Yeah, I was just to show the live stream from the... Oh, very good. Yes, yes, yes. One of the interesting things in these projects is the juxtaposition. And we've spoken, David, we spoke about the people and technology coming together. But, Pete, if I could turn to you as well, is the juxtaposition of the science and the policy objectives. And how have you sort of pitched that in this project? How has that gone? What are your sort of learnings from that that you could share? Well, I think the policy pull is there because as Jagadish has already mentioned, you know, the Scottish government and the UK, the Westminster government, is keen to pay for... You know, there's the Elm scheme in England is going to pay for public money for public goods. And there's an intention to pay for that on an outcome basis rather than an activity basis. So this issue of monitoring and verifying and reporting changes in soil carbon is important not only for carbon credits but for the private voluntary carbon markets, but also for the national inventories and for tracking our progress towards net zero in 2045 in Scotland and 2050 in the UK. So there's... You know, if we can get it right, governments I think will bite our hand off to use this system. So that's why it's so exciting to be involved in it because it's got a bunch of policy customers and government customers just waiting to use it if we can crack it and we're getting close to that. It could be a step change forward. Yeah, absolutely. Thanks, Pete. I mean, Kit, what are the challenges of actually doing that, of taking a project like you've got here and actually rolling it out in that sense? What are the sort of challenges that might exhibit in doing what Pete's just laid out? As I speak, highlighted that there is a demand and that's key because that's what's always highlighted, but as Mark Esno sort of often summarizes from the digital twin hub, which is the industry aspect of the digital twin activity, it's learning by doing, progress by sharing and it's actually getting into the details of specific projects like this one and finding what works but also trying to learn from others because sometimes others not only within research but also industry and the tech sector have solved a lot of these issues and I think one of the challenges, especially for researchers is being aware of, what are some of those options in terms of technologies and choosing what to use and what works where? Absolutely, yeah. I just want to add to what Kit mentioned here. One of the challenges we see is that we demonstrated this project in two parts at this point and there are so much diversity in the firm as we go along when we scale it up. We need to live up to that. We need to design the system where the system will cope with that diversity and also the scalability of it, the impact that we can see when we are able to scale it up. Can we scale it up to 100,000 hectares? Can we scale it up to 1 million hectares? So that is the next phase we want to go in whether what actually works we demonstrated here, we want to scale it up to really high land area so that we can see and produce a real-time system that will change the whole paradigm in this. And there are huge challenges in scaling up because for farm level operations, there are many things which we are using that works at the farm level. We don't need such an intense scaling software or even when we deliberately use the Docker system here so that we can multiply in future that one. So we need to see how this will operate when we scale it up to 100,000 hectares. So that's a challenge actually. Yeah, certainly is a challenge. Pete, as a distinguished scientist and the environmental scientist, how are these sort of digital approaches supporting the advancement of scientific understanding in projects like this? So what's the role of digital in helping further our understanding of environmental processes and so on? Well, as a modeler, I would say this, but we can use models to run different scenarios and to examine the science. Our best understanding is encapsulated in these models. It takes many years to set up and run a field experiment. I'm not saying we don't do those, we still have to do those, but they take five years and an awful amount of money to run. So we can use models, we can use the digital environment to explore the possibilities, to shortlist the things that we might do and we might want to look at in more detail and to base policy priorities on the outcomes of the models. But that has to be grounded in good data and we can then go and do longer-term experiments, also funded by UKRI through the standard grant mode to fill in our gaps and the information. But the digital environment allows us to explore much more possibilities quickly and cheaply. Interesting, yeah. And I mean, are there other... Elizabeth, if I could just come back to you. Are there areas where it doesn't work having digital environment approaches? So what's the learning for where it's appropriate and perhaps isn't appropriate? Well, I think it's interesting when I got this question, I wanted to turn it on its head and say, I think that I really benefited from going to the digital environment conference that happened a couple of months ago and seeing all the different versions of digital environment projects that are happening and seeing the interpretations of what that meant. And seeing that even though we can sometimes accomplish a lot, we're still not necessarily always very good at communicating it. And so I think that there's still a lot to be accomplished on the science communication side of these things, which is a goal point of this science project. And so although there is an automation side of this to accomplish the policy and management side of this and also to actually make the science worthwhile, the digitization won't actually, and also to be able to incorporate the new technology in from our peers. I saw lots of people trying to do the same things but in lots of different ways and we weren't necessarily communicating it very well with each other. And so I found that to be a major challenge that we all need to work on. And hopefully now that COVID is over, we can have more conferences, more communication amongst ourselves and we really need to hold ourselves to higher standards. I think communication is absolutely critical. And thanks for mentioning the conference. Yes, it was a fascinating time. We hope to run another conference this coming summer, perhaps in July, but so watch our website for that when we look for some more fascinating talks and that. We're beginning to run out of time but I have a sort of a last question. I'd just like to come back to perhaps Kirt and Damon for you. I mean, we've talked about this project like retina paving the way for digital twins. And I'd just like to get your take and others as well. What, where things stand with digital twins? What the opportunities are? What would need to perhaps develop in projects like yours to have a digital twin? What's the gap that we're trying to fill if we're paving the way towards a digital twin? And what's your perspective on digital twinning? So if I said, I mean, others have said like Mark, Esna, that federation is key and that's a social technical issue. It's about people, but also enabling it because especially one business isn't going to take it forwards. I think two other key aspects, one is privacy stroke security. You know, that's becoming more and more important where data isn't how it's looked after, but also how we do it sustainably. You know, there are different decisions we can make in terms of architecting sort of applications. And some of them have got lower carbon footprints than others. And what we do need to do is trying to do this work as sustainable as possible. Yeah, quite. Damian, what are your thoughts on digital twinning? And how projects like this can... Sorry, digital twinning isn't perhaps my background, but certainly one thing that springs to mind is, you know, what's the minimum sort of level of data you need to collect to, you know, accurately reproduce that? You know, you don't want to make your data collection for your user too onerous. So you need to work out, you know, what's the... When do you start to kind of saturate in terms of, for example, in the drone flying? How often does that need to be done? Can you get away with it during once in the season? Does it need to be three or four times to get accurate representation for your digital twin? I suppose that's a key challenge to make sure that you're collecting enough data to have a realistic representation. I'm not too much. But maybe, thank you very much, Damian. Maybe time for one last quick question before I turn to you, just for sort of your closing thoughts. But Becky, if I could, we've heard about the app that David's shown the app, but citizen science is something that we hear about a lot. And a lot of the projects, so the demonstrated projects we have are looking at citizen science. And I just wondering what, you know, what you're thinking is about the role that citizen science, for citizen science in gathering data and what might be some of the issues that you might anticipate in citizen science-type applications? I think in this area specifically, particularly when it comes to the sensors and things, we can't obviously expect everybody to be able to afford a drone, but even our sensors and the gateway and things, I think it's quite cost prohibitive for just anybody to be able to help us out. And so we need to find a way that the general public can assist, but without, you know, spending a lot of money. And I think that's sort of one of the issues in terms of what Damian was saying as well, in terms of, you know, about how much data to collect. So also how much of this is absolutely necessary if we can save them, you know, do they need four sensors per field or five or, you know, honing that to make it as accessible as possible to get as many people as possible involved in the collection of data? Yeah. I'll also add to that that one of the things that we're trying to show through the app, well, through the entire project is that it's possible to show things like simulations for one's farm, like Damian said, with essentially as little personal data as possible. So, for example, we have regional meteorological simulations that we have. We have regional soil maps from the James Hutton. And so we can start out with baseline data and do simulations without input from the user and then hopefully engage the user, show them where they can, you know, show them the potential that they can add to and then how the predictions of things like crop yield can improve with their input. And maybe just with personalized field management, you know, them telling us how they manage their crops, then maybe they're willing to invest in sensors for their soil or maybe they're willing to put up a tower to measure meteorological data, something like that. So it can be steps of investment and they can see how that improves their personal predictions. So it doesn't have to be one large investment and they can see that data and they can see how it's changing the model predictions. But they can also see something before they've actually submitted any data to begin with. So it's about enriching the data landscape. Exactly, enriching is a fantastic word. Yeah, yeah. I think that's a good observation. Thank you for that. So you've talked, Jagadish, about the history and you mentioned several decades of modeling leading up. And if we consider Restina as being on some kind of roadmap where that's behind you and it's led to what you're doing now, what's the sort of the roadmap going forward really just to finish off our discussion today? Where next with this? Yeah, we identified two streams of development in Restina. One is the technical development of Restina where we produced a functional prototype. The next step we want to do is the data really feed in, now the data is feeding into the model. Now what we want next step is to, the model should learn from the data and modify themselves to really mimic what actually predicts well in the field. That's the next step, which we haven't done in this project. That's going to the next technical development. And also the other aspect which we are looking at is the scalability aspect, which we are trying through the carbon extract, the spin out. Whatever we have developed, how we can scale it up to very large areas and produce it when in real time. So these are the two streams of work development which we are doing and several projects are already in the pipeline in that line. Thanks a lot, yeah. Well, look forward to hearing more in the coming months and years. Well, I think unfortunately that's all we have time for today. We've come to the end of our allotted period of time and I sincerely would like to thank all of the speakers today. Thank you so much for coming and joining us. I think to truly understand a complex project like Ratano, it's wonderful to have everyone here and it's clearly a project that draws on a number of disciplines and it's that juxtaposition of all of the different skills coming together, which makes it so fascinating. It's been a very useful discussion indeed about the role of digital environment in approaching these complex challenges and we've heard a lot about that. All the videos that we will take and including today will be popped onto our YouTube channel and Cameron has put the link to that, I think on the chat, thank you for that. So do come along and look at that and do subscribe to the channel if you wish as well. And what remains then is to thank you, the audience for listening today and to inform you that the next webinar in this series will be on Friday, the 17th of March, which is speaking about the project Sentinel and Dr. Paul Brown at Ferra and Team, the Sentinel of Treescapes for Plant Biosecurity and Risk. So I look forward to seeing you there and thank you very much for joining today. We'll call matters to a close now. Thank you very much. Thank you, Chief. Thank you for that opportunity. Thank you.