 to welcome all of you. Thank you for joining the panel today. And Paul Brown, the leader of the Sentinel Project. Perhaps, Paul, we could just quickly introduce colleagues around the table and then after that if I may, just call on you to give a short overview of the project to set the scene before we start our discussions. So, Paul. No problem. Thanks very much. So yeah, I'm Paul Brown from Ferrer Science Limited, leading on the Sentinel TreeScape project. I'm a remote sensing scientist, a geometric surveyor. Personally concerned with monitoring the tree scape using remote sensing technology to monitor tree scape. So, thanks to the management of our current tree scape also to monitor the new tree scape as we have policy objectives to plant trees, but how we monitoring those objectives. That's a very brief introduction to me. I'll pass over to Rachel Galton from Newcastle. Rachel Galton, I'm a senior lecturer in remote sensing at Newcastle University and my research areas in remote sensing predominantly and other ground based sensing approaches for monitoring vegetation health and dynamics. And we'll go to Barbara, my colleague at Ferrer. Hi, Barbara Axner. I'm a social scientist at Ferrer working in the land use team together with Paul. And I've been working on work package for the exact name of which I forgot but essentially is all about citizen science aspect and working with tree wardens to make sure that we document their experience. I'm director of tree science and research at the Tree Council. Tree Council is the umbrella organization for tree groups across the UK dealing particularly with the non wood trees outside woods in all sorts of habitats. And we brought the citizen science to the mix to assist Paul and the team to collect some data into the ground truth that was real people. Fantastic. Very good. Well, thank you for your introductions. Paul, perhaps you could sort of set the scene for us and tell us a bit about the project and what you've achieved. No problem. I'll just share my screen. Yeah, I'm just going to introduce the central tree scapes for plant biosecurity and risk management multiple threats to give it its full title project. So let me give a very brief overview. Firstly capturing the multidisciplinary team and mentioned it before. We're lucky to have scientists and collaborators from Ferrer University, University of stress Clyde UK sensory ecology and hydrology. The tree council, along with the tree council broadened tree warden network, and then Norfolk County Council. Why monitor tree health trees are under increasing stress from pest diseases climate change as implications for conservation biodiversity ecosystem services, and also public safety. A little timeline just looking at some threats and incursions over the past few years. Just want to highlight obviously the major one we've we've all heard of in 2012 of ash die back decimating the species in the UK. Project aims. First thing was to develop and demonstrate a monitoring system to detect stress from multiple causes, which is able to be deployed at a landscape regional scales, we're not looking at any specific cause of stress we're just looking at general condition of the tree scape and looking at deviations from the baseline. Were we combining sits and scientists with technology and modeling methods, hopefully provide a blueprint for future deployment throughout the UK. This little diagram constructed by my colleague Rachel Galton just gives a really nice overview of the project. I'm not going to go through it here but it's a nice pausing point in the video if you have a look at it in the future. So the IOT network the Internet of Things Network we've deployed and we deployed we commencing the 10th of May 2021. This is one of the hubs the main hubs that all the sensors transmit their data to which further transmits back the database at Farrah. So we have the tree talkers directly mountains trees have 60 of those deployed across three sites. We have 12 soil sensors in groups of two different depths deployed across sites and then we also have to open sky spectrometers allowing us to correct for transmit mittens to the canopy. We have three sites, long plantation, which is a long thin plantation next to a road. We have a woodland site belt plantation and we also have an agricultural site that Lingwood with trees along the railway line. So just look at those sites in a bit more detail the main points here are the red and the green spots. The red show the locations of the tree talkers where we have those sensors mounted on trees and the green show further observation sample trees. So the citizen scientists will be monitoring their observations on both the tree talk trees and the further 30 trees per site of the sample trees. So looking at a tree warning engagement has been a massive part of the project the citizen science aspects of the project. The initial training was conducted on the 12th of June in 2021 a month after deployment of the sensors, and there's been continued engagement activities with thanks to the tree council. Looking at seasons in the woods so we've done workshops of spring in the woods summer in the woods autumn in the woods winter in the woods. So we're down with the volunteers the wardens and we have IOT technology discussions. We have observation discussions but we also have tree walks looking at species identifications and fungal identification just really good days in the woods that keeps that motivation going. So we've conducted further tree warden network workshops last summer, bringing in wider East Anglia looking at the different tree warden groups in the area, and how the motivation will be there to scale up the project beyond the And throughout the project Barbara and a team have been conducting surveys of volunteer motivations and what's keeping these volunteers motivated. And one key thing that's came up when we were looking at preparing for this was the technology really driving engagement and the wish to get involved with the tech, as well as the tree health observations from the volunteers was was really key coming out of this. And to that end we created a tree talk a data hub, which allows the tree wardens volunteers to go online and look at the condition of the batteries the battery level for each individual sensor. And they've taken ownership of changing the batteries as and when they need changing during their observations, saving the core project team, having to travel four hours each way to change those batteries so that's been a huge help to the project. Remote sensing data we've captured drone imagery throughout the course of the two years. We've also acquired a satellite image from the archive in September 21, and we test another satellite in September 22, allowing for the remote sensing upscaling and modeling part of the project. They're just going to briefly go through some data streams that coming from the sensors. So this is from one of the soil. Well, this is from some of the soil sensors, and looking at how they capture environmental events. So this is in July last year, and in Norfolk suffered quite severe droughts in that period so we can see the change in the soil temperature as it rises throughout the months peaking around the 21st 22nd of July and then slowly falling with a with a small secondary peak. Unexpected finding was trees falling. Storm Dudley and Eunice passed through our sites in February last year. We can just highlight on this particular tree the yellow the red and the orange areas where it's like normal movement within the tree. And then on the 17th of the green area there's there's a bit more movement happening and then suddenly on the 18th the tree falls down and we can see a huge amount of movement in that sensor until it reached its final resting place so we're really capturing these events in the data. We can see it like transmission through the canopy behaving as you'd expect less transmission in the summer months more transmission in the winter months, but very different on individual trees so it's really important to understand what we're looking at what the sensor is looking at so some are looking at more than one tree, and some of Ivy on the tree on the trunk so it's decreasing that transmission also in the summer months. It's very important to understand what you're looking at before you can interpret the data. And lastly just looking at some key findings to date so major learning outcomes is citizen science observations can capture variability in tree health. They can be subjective, obviously individuals look at things differently, but we try to constrain that as much as possible by creating an app which gives you, excuse me, drop down menus. Excuse me it gives you drop down menus and list the questions as you go through them. There's also a thing that would potentially be difficult when scaling up the project it hasn't really been an issue in this project we've got a core set of volunteers, which have been extremely engaged, excuse me, which have been extremely engaged throughout the project. And then but then there are preferences on areas to monitor like some woodlands are nicer than others, so it's nicer to walk in those woodlands. There's an ownership of the project and self organization from the citizen scientists really wanting to take those the technology and look after their technology and they're still wanting to do that even further by going further in battery changes and repairing the tech in the field. So this challenges with the technology that has been hub failures, which causes the data hub the battery level and monitoring system to not be updated that has caused issues. And the time it takes to get a new hub and go down and change that hub is a problem with the it when I work. It's not one we have one we have managed to overcome. The tree talks themselves a rich, high temporal resolution, but quite a noisy data source in some senses, a need for extended long term data to establish the baselines. The current objective is what working on the moment is to integrate the data streams observations by modeling to predict tree health status analysis of wider volunteer survey and stakeholder workshop to scale up the system. So I want to make data available for the NERC data centers after the project and find ways to continue to support the broadband tree warden network into continuing the project beyond the June deadline when this project comes to an end. That's a very brief summary of the project for me. Thank you. Paul, thank you. That's great. That's fantastic. I think you raised a very relevant example of ownership of the project and people coming in changing the batteries. I think there's nothing more practical about that than that. No, I think I think that's very, very good overview and you had these three sites I'm just wondering if any of the you mentioned some websites and so on other resources that could be shared with hope today as a link link so you can put the websites aren't public they're very constrained to the network at the moment just because you know we didn't want to advertise the locations because one key thing was vandalism and stuff are these are these internet things just going to be smashed out the woodlands after two weeks of deployment. Thankfully that never happened never had one issue with that at all so that was a really good learning outcome. But yeah, we've tried to keep it private at the moment but we will make it. Yeah, it's always a challenge when you put Santa's what's going to happen to those. But I mean john if I could if I could just turn to you I'm interested in the, the representativeness of this and what what's the sort of size of the challenge that that you're addressing here and the, you know you've got these these these sites. And how the issues that you are addressing really a fat society and wider wider implications. So obviously from the perspective of of the environment trees are a keystone species in that mix and obviously there's been a lot of focus over the last, over the last, particularly the last five years about the role they play in carbon and carbon and the greatest threat that we face to our treescape is that is that a rising of new pest and diseases that might that might alter the balance of species, particularly in the case of the work that we've been doing. Ash dieback which there's, there's the potential, there's about four billion ash trees in the UK of all sizes from little seedlings up to big mature trees. And if we lose all of those then then having an understanding of the changes that will make the landscape so understanding what's going on in the tree and helping to understand and model that and be able to understand that is a really vital part of going forward. And even, even that element of just being able to spot a new pest or disease arising bit to be on that cusp of of being aware that there's a problem before it may show in the, in the data because obviously once it's, once it's got hold it might be too late. So, so for us, this project was a really exciting way of bringing real people into study something that was local to them but which had, which had new technology and new devices involved in it that meant that way could see that they were part of something wider, something that was addressing national problems and if we get it right, that would be something that rolls out elsewhere. And one of the biggest issues we faced was that of risk management on the roads. You know if we if we have lots of dying ash trees by the roads. How we predict is there ways to be able to suggest that we know that this is coming. So although Paul said it was a surprise that that we had trees falling over in the data. And it was a bit of a spire to having trees falling over when we when we set up the, the grid and the fact that it happened and that we were able to identify that and start to be able to predict before and after what the conditions were in the tree. I think was truly, truly remarkable. And as you remember the huge consequences that Dutch Elm disease had on the landscape will will quickly realize that ash dieback is a major, major threat I guess, I guess now as well. I'm looking forward to hearing about the technology particularly about the tree talk as I'm very intrigued by that. But maybe just before we come to the sort of tech that you've been using. One thing we've been asking all of the all of the demonstrator teams. The constructing a digital environment program is all about the digital representation of the world around us and there've been all the different themes that the projects have been working with and you're working with treescapes. I'm just just interesting what you're what you're you and your colleagues perspective of is a digital environment. You know, how, how have you, how have you used digital in your work to, you know, to, to, to conduct the science that you're, you're, you're developing what is it about digital that's lent itself very well to your, your project. I think the digital I mean it allows for remote monitoring. There's a potential for full coverage across the landscape, combining the, the IoT tech train train machine learning algorithms for extrapolation across that landscape. So it allows for a potential targeting of ground survey you're never going to replace ground survey and expertise but it provides a targeting of that as opposed to just sample survey. So it will take time saving and being in the right place at the right time, essentially. And I think particularly now we're in an era where technologies are maturing at the same time. So we've got data tech we've got the technology about these sensors you can put on trees that measure trees. But we've also got the analysis the machine learning the AI and the computer station or cloud computing the power that's able to analyze this data. So that's an area that is reaching a level of maturity but then also that is a challenge because technology is changing all the time. You buy one sensor. Two years ago it's then replaced by a new sensor. So how do you integrate that new technology. That's a, that's a very common theme that a lot of the products have been grappling with actually a different interpretation. Yeah, so it's fascinating point. You mentioned the technology develops and we'll come to that in a moment but I'm just interested really how do you think this, what you've done here could be scaled up or I mean how easy would it be to reproduce taking a bit of an overview of the project what you've done in other environments and other wood treescapes. Yeah, I think scaling up it's always been in the back of my mind it's always in my mind for other projects as well is like doing these remote sensing these digital projects on the pilot study sites and small scales as always. Always the question is, is how do we do that across the whole of the UK treescape. The broader network is is a geographically spread network I mean we've we've created workshops. We've conducted workshops that bring in the other networks within the her wider East Anglia area the wider North Vicaria not because the broadened area and look at those motivations so the technologies are the apps there to measure the trees it's just and the motivation is there it's just the logistics of bringing those people together. And then the remote sensing data is there to scale up you know we can, we can take those ground observations look at the spectral signatures and the canopy and see how, how we far we can extrapolate before we need another sensor network and then extrapolate beyond that. And what's the expense of that data, you know that there's another challenge is the expense of that remote sensing data so I think the potential is definitely there is just overcoming those, those issues of expense and technology availability. So you don't need to instrument every tree it's a question of representativeness of the data you have. I mean, let's turn to the technology now. Before we before you go there I think, I think what Paul said is the technology is now maturing to a point you can do it the really interesting part is for us was was the community would they would they engage with it would they take it into their heart. What we've learned from this experience is is if you give the citizen scientists and active voice in the program, it becomes less of a, we're doing it at you rather than we're doing it with you if they feel that they're part of the conversation. They definitely took ownership and that that simple one of replacing batteries. And that will be started as a, as the problem the community saw that as an opportunity they were part of the conversation. So it's, it's scaling that will require. It's eminently possible but it will require that community engagement that makes people want to be part of the journey and trees are one of those aspects that people seem to really relate to so I think it would be utterly scared level. Thanks, John. I think in a way you've just helped address the first of the question that's currently open on the chat which someone's asked, what are the post projects activities following on from the project and will and indeed can it be scaled, scaled up across the country so I think I think you've both given some thoughts thoughts on that. Thank you for that. Rachel, I believe you're the other tech. The person here tell us a bit about the sensors and the technology that you're, what is a tree talker. So thanks. This project sort of being quite like it's quite complex and we're actually deploying quite a lot of different technologies within the project so ranging from the sort of field app that the wardens are using to collect the data. And they're accessible sort of dashboard that's helping them to monitor the network through the internet things tree talkers they'll come back to in a minute and then the layers of drone and satellite imaging that we hope will help us scale that up so there's a whole sort of new tech in that I think that's really exciting in this project probably is the tree talkers so these are wireless sensor network. And that's all online communicate through through hopes that they can then stream straight back to us at Ferra. And they measure quite a wide range of things. These are sensors that were developed in Italy originally and and supplied by company called nature for and they measure a whole range of canopy parameters tree parameters tree physiology but also some basic climate parameters so for example measure humidity and temperature as well which lets us really try to relate some of the tree physiology to what's going on in the environment which is quite important. So within the little boxes tricks if you like, and they have an accelerometer which so we get this tree movement data that will show briefly. And they have a radial growth sensor that attaches to the trunk the tree and this meant measure the stem growth. And they have SAP flow probes but they also have a capacitance sensor on those probes that measure the stem water content. So it's interesting for you for me as a remote sensing scientist is they have an upwards looking spectrometer that's measuring light transmittance through the tree canopy in 12 different wavelength region so that's giving us really rich data set of what's the canopy condition like but also potential to unpick some of the spectral changes that might occur for example if leaf pigments are changing due to disease or phenology as well. So it's taking a measurement every three hours so we get a very rich data set have been running for about a year and a half now. So it's quite an extensive data set observations, I, I'm not going to try and remember or quote how many measurements we now have, but it's a lot of data across all of those things. The other aspect of the project that gets really counts tech is the the machine learning aspect of trying to bring this together so we've been developing a or fairer been developing a hidden Markov chain model which is a type of model that can hopefully infer the current status of an object based on probabilities and also previous and current observations. And what are the challenges of our projects how do we bring all these observations together we've got the sort of non tech citizen science observations we've got the tree talker ones we've got the drone and satellite but they're all at different temporal frequencies. They all cover different areas they're all scattered in different ways across our study site and so on. And so the idea of that machine learning model is to try to combine those different observations and start to predict the health status of trees based on those. And, and currently that's been parameterized based as well also on the long term research of trees that's been conducted by an offer countercouncils we're also bringing in some wider tree data sets there as well. Well, I mean yeah thanks thanks I mean it's interesting to hear about the Italian sense that I don't know if it's possible to share a link to the to those for people to be fascinated and follow that follow that up. We've heard about batteries running out that's that's not really unexpected but I mean I think of every time I've tried to use sensors and woodland you find that GPS signals can't be attained and, and I just wondering about communication and getting data back out of the out of the dense canopy forest back to base I mean what what are some of the other challenges that you've had getting these sensors in the ground. I mean the main technology challenges that they've been all of those things you described I think we've probably encountered, and I think we're going to touch a bit more in a minute on the sort of environment, the actual sensors of the tree talk and network. But in a broader sense I think the biggest challenge for us technologically really has been how to interpret the data that comes from those sensors because we're working with sensors that whilst they've been deployed in a few small sites in other countries haven't really been widely deployed and they haven't been used for tree health monitoring extensively specifically, and although for some of the sensors like that flow there's sort of established, obviously we've been measured in trees before there's some baseline data to go on. But other things like the open looking spectrometers that really isn't there so how do we separate out the noise how do we establish what's down to the species or site versus what's down to the health status. So I think a lot of the challenges have actually been in what to do with the data set and then how to scale that up and how to relate it to the remote sensing for example that's viewing the forest from the top so seeing understory change as well as the tree canopy, and so on. And I'll let Paul may recover this pitfalls of the actual sensor network. Paul what are the pitfalls of the sensor network thanks Rachel. It's been a learning 2021 was a learning curve. Like you mentioned transmission of data just simple things like there's the sensors that orientated on the north side of the trees and then we had a hub on the south. And actually, the data was, it was falling behind and we were wondering why and I just moved it to the north and it picked straight back up so it's just little learning things like that. Obviously at the beginning of the projects we had logistical issues around COVID-19 that was a major thing for us to deploy that network. I've just mentioned transferring to different climates, you know, having water ingress and some of the in some of the sensors so learning that we're learning okay a bit of a silicon sealant around the around the edge of silicon grease around the o-rings just eliminates that completely. So just that sort of learning. And then little things like that is like that. Little things. Yeah, like you say it's so important so huge and animals. It's chewing through wires of soil sensors, you know things like that. But that's when the citizen sciences have also been involved a lot more is this fixing wires like that they notice when they're observations on this why has been generally to this this fix it up and they do. And then it's completely working against that's been another again going back to the ownership of the project the citizen sciences. In the presentation we had she beating the sensors. Yeah. We've got an upward facing spectrometer and a bird lands on the sensor, you know, do what they do and then covers the sensor so again, citizen scientists going around checking that sensor that windows clear and giving it a good clean every time they go around as just helps. Did you have to have duplication in the census to ensure. So if something goes wrong you can sort of see that it's gone wrong or how do you deal with spotting errors like that. I guess it's just with analyzing the data to be honest there's no real real time there is with the, like if the census only stops working you'll, you will see it because we monitor every morning we have a script that comes through and saying is it okay well how many sensors are working which ones aren't. And but individual sensors within the unit, like the spectrometer, you have to really at the moment analyze the data doesn't come through near real time. The moment that sort of thing, but that's a good point. We've had met stations with tipping rainfall gauges and one one of them stops the other one doesn't you realize that there's something wrong and then you go and realize that one of them is full of leaves you know. Sometimes you have to have that duplication. Yeah, yeah. The duplication is across the tree scape in that sense in that for example with the spectrometer for windows blocked you'll see a big drop in transmission in just one sensor whereas if it's cloudy days you drop in transmission and all of the or no light levels in all of them so we can. We can unpick a lot of that based on where one sensor is a very clear anomaly for a short period something rather than a long term trend. Are you using Laura one or 5G or how are you getting the data out of the woods. It's worth using comes, it's right a SIM card so it's not even 5G I think it's just a GSM network. I use Laura to talk to the hubs. Yes. So that's the biggest issue is when a hub goes down, because you lose all 25 sensors transmitting the data. Then you get a gap. You can retrieve the data because each individual sensor has memory inside the sensor but it means a manual download, which means going out and on screen. It's quite time consuming so that that has been an issue with the hubs but as soon as you replace one it's it's it's absolutely fine you do the manual download. Barbara if I could turn to you I mean we've heard some very interesting anecdotes about the relationship formed with the tree wardens and the. We've heard some photos of the training sessions you've done and we've heard about these batteries being changed but I'm just interested really because the the role of people in in the in this project has been really quite key it seems and maybe if you could just talk us through that a little bit more and you know what what what the interfaces between people and technology and a project like this and what what can we learn learn from that. Yeah, as I said as we heard an integral part of this project is working with citizen scientists in this case the tree wardens. And I think it's been quite a steep learning curve for us at the beginning so while volunteers is obviously one state called a group there is naturally a lot of diversity in terms of the individual backgrounds, which meant that some were actually very comfortable with different aspects of the technology so for example we heard this multiple times now and you've just mentioned it as well. Initially the team was planning to do all the battery changes themselves but actually the volunteers were more than happy to help with this so not learning is that sort of discussions like that always benefit from just being a discussion directly with the volunteers. And yeah so at the back of this as well then there was an update to the app, which you've seen in Paul's presentation which made it possible for the volunteers to monitor the battery status themselves and similarly. Same similar elements which the tree warden started doing their own spreadsheets to record which trees that they've looked at. And again, that was something that was then integrated into the app and made it easier for them to coordinate their trips. And yes, all the volunteers were just really interested in what the technology could do and could offer and what the outputs looked like. Initially we or at least I wasn't sure about exactly why the volunteers might have wanted to join this particular project and whether they might want to use the data outputs for themselves it actually became clear that in general there wasn't as much interest in the detail that integrated off the data but rather that it delivered something exciting new science that could help contribute to keeping our woodlands healthy. So there was definitely a lot of interest in finding out about how this process is going, whether the data is delivering what we thought it would deliver or what we wanted it to deliver. And yeah so the technology gave the project an interesting new aspect that was worthwhile contributing to. And there's also clear desire right from the start to find out about the legacy of the project and how the outputs would be used. And I think that's also highlighted by just a continued interest in just maintaining the tree talker network beyond the project and one tree warden even purchasing and deploying their own tree talkers. So, it was, it's all really good stuff. In terms of the general public there was wasn't as much engagement and there was never any plan to. So the project team team did put up signs along with a tree talkers just with a little bit of background and an email contact email to get in touch with. And it has encouraged or at least contributed to probably one dog walker actually wanting to find out more signing up as a tree warden and joining the project team and I think as well whenever the team was doing maintenance or anything similar down there they got quite a few questions from interested passes by as well. Yeah, so you did have some dog walkers commenting on the, on the, on the. And as a question I think perhaps for for you Bob and indeed Rachel. What one on the sort of the people on the technology side I'm just intrigued by the, the level of engagement that you had you clearly had a lot of people Barbara who are very interested these wardens are very intrigued by what you're doing and interested to take part. So what would you say if to yourself if you are starting again other particular learning points about how you approach these, these, these colleagues and how you integrated them into the project what are the learning points of how you've dealt with them. I think. Yeah, please and I'm right so afterwards that. I think it's one of it was just keeping them informed because it's as that before they weren't necessarily to bother about the exact details but they were really interested in finding out sort of what the project was achieving. And I think at times we probably have left it quite long just because we were setting up the project we were trying to deal with different things, but it's just making sure that we just kept them in or keep them in a loop and that they sort of, they know that things are moving things are ongoing and that what they are doing is contributing to the project until the outputs. Rachel sort of a similar question really just in terms of the placing of the technologies and the sensors and so on. You know what are there any particular learning points you would you would reflect on over the over the project. The tree talk as we always knew we'd have a challenge and there's only so many treat all because we could deploy in and demonstrate a project for the first time. But coverage of therefore of the different species within the woodland and the health status is perhaps meant there's not as much replication as we might like across all of those groups. I'm not sure necessarily there's an ideal solution to that other than a, you know, a larger project with a bigger scope and a bigger, bigger budget but I do think there's an issue around, you know, how do we choose to deploy these where do we put them and part of the project was to try to come up the answers to that so we knew we didn't have them up front. I think the other thing perhaps is because of the interaction with the tree ones and because we were aligned on them for batch changing and so on that did somewhat limit where we could put the tree talkers so we did actively avoid for obvious reasons trees that were extremely unhealthy and we're at that far end of the health status scale because we couldn't have tree ones interacting with trees we actually thought were likely to be a health hazard to them. I think that's limited the range of data to some extent. I'm not sure if it's necessarily means I do anything differently in that respect we obviously wouldn't pick really dangerous trees to get the tree ones look at now either. And but I think maybe we think about some alternatives to try to capture some of those more extremely unhealthy trees within the woodland. I think otherwise, and some of that relate perhaps having a site a little closer to to fairer where they could go more readily to to change batteries of a small group of trees themselves or teller. I think the other thing was probably around the drone acquisitions I mean the sort of challenges there in terms of having to do them within a limited time limited field campaigns that means sometimes it was sub optimal lighting conditions and so on but in the most part that's that sort of worked all right I think. So a bigger network more tree talkers ball species captured or focusing in on a single species and just looking at that for the demonstrator in the first place perhaps. And one of the, one of the aspects of these demonstrator projects John, to come to you is the sort of translational aspect of the of the work and I'm actually looking at thank you Laura Kelly for posting the question about you mentioned data from the survey of ash dieback disease to parameterize models and to what extent you think signs of poor health from a specific threat such as ash dieback disease would be predicted predictive of general good or poor health and trees, or my trees display their unhealthiness in different ways and I think maybe I could just ask, ask you that and in answering, of course the differ and another key policy makers in the country who are trying to address some of these challenges. How can the work that you've done inform inform that that sort of translational piece as well. If I could just before I answer that one Stephen and I will if I could just one of the key factors in relation to the previous question. One of the you asked about learnings one of the key learnings for me of the whole project was the ability of the project team. And there's a there's a really interesting thing of when hard science hits the real world. If you had if we'd gone with the original plan it probably wouldn't have developed the amount of data and the resources that it did. And it was because the project team adapted and flexed and did things that weren't pure science but as Paul said in his introduction, we had a walk in the woods in September just to get to keep the community engaged. And that ability to flex and adapt was for me the main learning that came out of it, because without that we wouldn't have got half the engagement half the data half the results that happened so I think that thing that we should all bear in mind when we're doing these things the technology is brilliant, but it only works if you get the people to to engage with it. And then flipping the flipping that round into then how do we use the data that we've gathered to influence policy. Well obviously ashtabak is one of one of DEFRA's main concerns at the moment we tree counselor heavily involved in those conversations at policy level, and and helped guide local authorities to developing their management plans which is why the team wanted to have the conversation. So we've all the way through this project we've been connecting the policy, though the sorry the science to the policymakers and using the information to have the conversations to keep the to keep the conversations going. And, and that is critical if you're there is that element of real science is fantastic for and of itself. But if it's to have a policy impact you've got to be able to build the interface to build that. So the project teams need to have that capacity to interact with the policymakers so that. So the output of the work actually reaches people. I'm afraid it's, it's one of the academic world's biggest problems is that we do all this work when we produce some very nice dusty paper in some very nice dusty journal and nobody ever sees it. So, so one of the things that we need to do is to ensure that we translate all of the good science for the policymakers so that they get what they need to be able to use it in a way that that actually means it has impact. I, we haven't got to the point in this project where we've got the real data the X is coming because that's part of the next part of that, but the analysis of the data will be fed into all the policy conversations, because that that was always the point. But you have to have the channels and you have to have the opportunities and every project that you've got in the bundle of projects that you've done just needs to find those, those ways to get their information. So that's what policymakers but it won't be through always through academic publication. It needs to have webinars like this it needs to have sessions and Barbara's got Barbara's got policy workshops planned when we get to the, you know, to the to the data. So all of that will help to use this in a very positive way. Yes. So, Barbara, I mean, I'm just interested in how you've gone about engaging with all the different stakeholders that you've that are relevant for for this area, what, how are you going about that? I guess, sorry, I was still with the john's policy question because, as john said we've got a workshop planned sort of at the end of the project we were hoping to do that a bit sooner. And as I was already mentioned that there's obviously while policy teams have shown a lot of interest already. They've started asking questions and understandably so quite targeted questions. So, yeah, because this project is very much a scoping project it made more sense after discussions once we've got a firm idea about the shape and implications of the data, and just to add as well that for that policy workshop we're also hoping to invite other state called the national rail. So it's big stakeholder groups with an interest in tree health as well so yeah, watch this space. In terms of the other activities again, these things have already been mentioned in addition to the ongoing project work with with our tree warden group. That's the ongoing tree health assessments the technology maintenance sharing their experiences. We also had a local workshop in Norfolk which split into two parts. So one was for the wider tree warden groups on networks to share how the project's going and also to explore what the level of interest might be from these other groups and there was actually lots of positive engagement and people actively thinking about what this could mean to their individual communities and what barriers that might we might need to overcome. And yeah so that was later followed up then as well by a wide even wider survey on the motivations for citizen science. And the second part of this particular workshop was a walk through the woods and on site showcasing off the technology. And again this was very well received, especially by the county council other representatives working with parks highways, education. engagement there. Something else that came through the county council was a request from the museum service, I think, which led to the tree talk or a tree talk of being displayed in there from access access I can pronounce this to a corns exhibition. And it's also being used as an example on Christian thorns BBC radio Norfolk show tomorrow so just a little bit of advertising there for that as well. It's an element of this project and reached out got got publicity got people interested got people talking. And it's because it was vested in the community the community. So Norfolk radio probably doesn't care about the internet of things but it does care about the fact that the community have helped with the internet of things. It's, it's that element of community buying. And I don't know, all I know we've, we've got a little bit of video I don't know if this is the point where we could just show how the community truly engaged with this because I think hearing it. We've talked about citizen scientists but to hear it from their, from their own words. Hello, my name is Anna Rodriguez. I've been a tree warden for just over a year now. I have no background or previous experience dealing with trees. And soon after I started as a tree warden, the Sentinel project came up. And it's been a really good way of meeting a lot of the other tree wardens in my network. So much on the Sentinel project from experienced tree wardens, especially about how to assess the health of trees, but also basics like tree and plant identification and overall woodland health. It's been very exciting to be involved with this Sentinel project. It's a really big project, gathering lots of data. And we're now seeing how that data is actually being used to obtain new information on how to assess tree health. It's also been really great spending time in the woods and deliberately looking at the trees and spending time with other like-minded people. It's been easy to have you found that the interface, John, between the science and policy, easy to find or is it always an ongoing challenge? There is a huge amount of interest in DEFRA in basing policy on science. They want science. They want policy to be science-led. So any information that's available to them that is packaged and presented in a way that is useful and helpful and gives them what they need is to be desired. And there are many sister projects that are paralleling this work that is joining up the capacities from this project to share its resources. So there is plenty of opportunities to make the links. And as Barbara said, the fact that people want to come to workshops to hear what's been done and want to hear about how we're going to engage it, there won't be problems in getting the information into the policymakers. So I know DEFRA are very keen on evidence-based policymaking. So here is one excellent source of evidence that could be fed into that process. Exactly. So just looking back over the project, Paul, constructing a digital environment, have you constructed a digital environment? What strides have you made in that area? What are the challenges, I suppose, that remain in conducting this sort of research? Yeah, I think we've been successful in demonstrating a digital environment, using environmental sensors wirelessly, and then also integrating that with citizen sciences observations. I think it's been very successful in that point of view. There are challenges remaining. There's network connectivity challenges that the noise in the data is a bit of a challenge, how we overcome that. There has been new hubs released by Nature4, new sensors, so that might help with that a little bit, going back to that technology changing and developing all the time. Another challenge, as I mentioned before, is the expense of data, expense of remote sensing data, so extrapolating these sensor networks, these observations to satellite observations, and then using these central treescapes as almost like weather stations. You have one extrapolate between with remote sensing, have another one. But it's the expense of that data and the availability of that data. Another thing is the huge amount of data that you record. These sensors are recording, they've started on an hourly basis, multiple, multiple data streams an hour for 18 months. We've knocked that down to three hourly just because of the battery life. But it still is a huge amount of data. And how do we bring those data streams together to create a robust baseline and remove that short term noise. And so, are we measuring the things we can measure digitally? Is it correct? Are we measuring the right things? We have sensors given to us saying they measure this, this, and this, but is it the right tech? Are the things that we're missing and bringing in the multidisciplinary aspects of project teams is key to that, to get it right. And it's a challenge of this project. We may be missing an area in the plant physiology, the actual tree physiology. We've got a little bit of knowledge missing in that multidisciplinary aspect that going back to lessons learned, we would have incorporated that in the beginning. Yeah, I was going to come back to the, you know, we've talked about the technology, the policy links, the stakeholders, the tree ward and so on. But how about you guys as a team that the addressing these things, as you say, is a multidisciplinary interdisciplinary challenge and how have you all worked as a team? How did you set out to address this complex? I mean, you've mentioned that maybe there's some skills that could have been added to the mix. How's it worked out? Well, it's been great. I mean, we're not new to working with each other, the main core team you see on the call here. We have links with the tree council from DEFRA funded projects, myself and Rachel have worked together for a number of years on different projects in the tree skate. And we have bought in new collaborators, Michael from the sensory ecology and hydrology, bringing in that ecological knowledge and learning that sits in science. And, you know, he's been in one of these webinars before and you can see his knowledge and sits in science so that has been extremely beneficial. Yeah, there's been some really great, great discussions on that. I mean, I think there's time for one, one last question, which is something that's intriguing me and maybe Rachel this is one for you and there's been a lot of discussion in the demonstrated projects and in the I'm going to provide a CD program about digital twins. And what, you know, what the role of a digital twin is in trying to help manage and understand look at different scenarios for the future. Do you think you're, you know, this this project is moving towards a digital twin other other things that should be done to what's your take on on digital twins having having done what you've done. So, you know, a digital twin is going beyond measurements and into the ability to model and simulate and test out what future scenarios would look like based on that digital data so I think within the project we have some rich data sources that could integrate digital twins we're starting to look at how we could integrate those predict tree health but I think in terms of a true forest conditions sort of digital twin one that can let you experiment with changes in future climate and what effect that will have for example, you really need to go beyond the measurements and start to see how can we integrate some of these remote sensing observations into that of things observations into forest models that can tell us how the trees are going to respond to that and those models exist in many cases along history of forest models of course and many of them are physiologically based in some form, so have the capability to take these data streams but there's a sometimes disconnect between the types of parameters those models use and those that lend themselves to for example being derived from remote sensing so you know a lot of traditional forest models would be based on the diameter of the trunk whereas remote sensing there's things like the height and the ground size more readily so I think there's work to do there I think one challenge is bringing these things together the digital data side with the the modeling community to really build models that work. It's been great work in that direction already I think there's lots of technologies we're maybe not using in this project but that can build forest models lidar and grab my side scanning for example to reconstruct the sort of forest structure and so on. I think if those, you know, they're being brought together increasingly I think we're not that far off in this context from being able to produce these kind of digital twins that could help us to explore those interactions between different stresses so I think at the sort of siloing a bit you know we've got ash by back we've got drought stress but actually how are those interacting what would we expect the impact on tree health to be when we have both drought stress and a new pest and disease hitting at the same time for example. The challenges though I think remain the speed of change I mean not least in terms of new pests and diseases arriving for example and new stresses to the treescape arising. But also in the technology pace of change so even since we started this project there's new tree talkers available you know we can we can upgrade the system already. Which is quite a challenge I think to keep up with that if you like and the other big challenge the funding streams we need long term funding covering multiple sites which lets us deploy these kind of technologies to build up the data that's needed to produce a true digital twin of these kind of environments. To support those projects of course one needs the demonstrators first to show you the value and. Well that's been fantastic I think I think that's with that last comment on digital twins will, if we may just bring things to a close now looking at the time. Thank you so much for that it's really fascinating to hear about that project, and also it's wonderful to sort of hear about this project in the, in the context of the other demonstrated projects that are running and many of the challenges that that you're facing are actually ones that are. are clearly ones that other other folks have grappled with as well. So, thank you very much to all the panelists today and thank you very much for the audience has been a very interesting discussion all about the way that digital environment practices and principles are put into place for solving environmental challenges. It's been a good good series that we've had in this webinar, the video for this as with all our other ones will be available on our YouTube channel, and there's a link in the in the chat coming up on that. YouTube.com at the digital environment, you can have a look at look at this video which will be shortly available, and the other ones and we now look forward to the next series of webinars in the construction of digital environment, where we're going to focus on a bit more on the the science around artificial intelligence and machine learning in in environmental science. And we'll start that seventh webinar series on 31st of March Friday the 31st of March, three weeks time and we'll receive a keynote for that from Professor David topping from the University of Manchester. I'll be speaking about AI and environmental sciences from research developments to underlying infrastructure and policy implementation and I'm sure, like me or you'll be looking forward to hearing the professor topping speaking about that. So once more thank you so much to the panelists to the audience, and I wish you a good rest of the day. Thank you very much.