Graeme Forward

So as it says up there, Mark's introduced me, I'm Graeme, Digital Insights manager at KaarbonTech.
I manage a team of data analysts of which a couple are here today as well, James, Lucy, and the job of what we do really is to take all that data that you gather out in the field and make sense of it and try to help make better decisions. So we're gonna talk for a little bit now about a couple of things. One that Kieran was talking about earlier about the, the tree data standard. And actually, are you collecting the right information out in the field to make the decisions you need to? And secondly, how we can help use what you've got our insights, other available data sets to make better operational decisions, save time and money and all those kind of good things that you want to deliver. So we've seen this slide a couple of times already. Why are we interested in data data is at the core of everything? And I'm not just talking about that car tech. Is it in your pockets? It's, it's everything you deal with it day to day. So we're just surrounded by it, but we make software, we've got the surveying, we've got the advisory, obviously the dr the digital insight scene, the vast majority of the work we do is in that advisory section. And those are some of the kind of key things that we look at all the survey at the bottom. But on top of that, as well as matching the insights team, I also head up our operational survey division and that puts it in a real key advantage because I get to see first hand the information that's coming out from the field and what it means for us in the in the data team. So the key part of that is that I can start to think about, well, actually, if we were getting this data, then we would understand this and we could get that data out and when we can make those decisions. So we're seeing that flow and that's key to making sure that we're asking the right questions at the right time in the right way to get what we know from it. Because that is where we start to make these strategic long term decisions rather than being very reactive and having to deal with things time to time. Mark's talked about all the different types of surveying we can do. So the use of A I, whether it's warts ridden driven, liar, which I'm going to talk about in a second as well. And traditionally, lots of the surveys would be camping the streets, walking, miles and miles of highway, not necessarily the most effective or the most cost efficient way of doing things. So we can use technology, we can bring all those things together. And that purpose of that is we want to save you time, we want to save you money. We want to help reduce co two emissions. What is big data compared to just standard data? There are five vs that people talk about and they talk about big data and the key one being the value, that's what we can get out of it. That's what we can do stuff with, that's made up of a number of different things. So initially, you've got the value and that means the amount of stuff you've got over a million trees being handled inside context systems at the moment side tree smart, but there are roughly 3 billion trees in the country and a need to increase that tree population by about 50% over the next 2030 years. We wanna meet our climate targets. So we're talking about 4.5 billion trees, the rate of collection, the rates of understanding these is gonna be astronomical and exponential over the next 20 years, you've then got the variety of data. So that's what information are we collecting or different stuff do we have? So there's lots of different things you can collect about the tree itself, the height, the breast type, the age, the life's age, all those kind of things, we also use some supplementary data sets. So we've got ordinance survey topographic data. We can look at at heights and things above sea level, all the mapping or where it relates to where it sits in the world and iot data. So, rainfall sensors and those other things that actually help us understand the situation where the tree is a bit more tree height and diameter, breast height are absolutely crucial when you want to quantify and work out CO2 storage and sequestration. So they're important things to understand the age, the life stage, the the structural condition of a tree, what defects and what history has had help you understand the risk of failure. So you need to be really thinking about what sort of data am I collecting and why and what do I want to do with it? Where is it taking me and where, where we're going with this? You know, what's the outcome? The final element is the velocity. So that's how quickly am I collecting data? I've looked at the volume, I've looked how much but actually how quickly is it coming in because the rapidity of that helps you understand things over time and how things change. So we're having approximately 2000 inspections or work requests or orders coming through our systems every day, we're bringing in the rainfall center to the polls every 15 minutes, every other A I thing. So there's a huge amount of data that kind of all pools together to help you make the decisions. So those three things together help make up the value. And there's one other aspect veracity. So the veracity is the accuracy, the trustworthiness of the information you're getting in. If all the information in the garage you're getting in is garbage, then the decisions you're gonna make at the other end are gonna be rubbish too. What does veracity mean for us? And there's three sort of core aspects we need to look at is the information, accurate and consistent. Have you got one surveyor who's measuring everything in feet and inches and another one measuring everything in metric because you won't be able to make any decisions. We often see people who come to us and say, well, we like your system. We don't want to give you what we had before we want to throw it away because we don't trust it. It doesn't seem like it's gonna fit for us. Well, was it a complete waste of time in the first place collecting it is the data relevant? If you've got a tiny little sapling that you measured at a meter high 20 years ago, can you make a decision based on that? If you haven't been back in that time, it's probably not going to be the same as it was there. So, are you using information that allows you to make the decisions you need to? And finally, am I capturing the correct information quite often? We speak to people and they'll say, well, we've got to go through all these trees and we've got two surveys and this budget. So we're only gonna go and capture the height and the species, which is understandable because no one has the time money to everything. You can't make an informed decision unless you do that. And what I'm going to be looking at and showing you is actually we probably can have everything if we work smarter and we work with technology and we work other ways. How do we turn data into insights? Everybody is collecting data in some sense, everyone's a different maturity with their data, but we all have some even if it's a rudimentary level. So you've got co collect. This is an example of a tree planting scheme for an authority, three different wards over seven months. How many they could they'd put in at that time. That's a captain of, well, we planted 831 in ward 2, 377 in, in ward three that tells us what has happened and it tells us what happened in the past. A lot of people do some analysis on that as well. Oh, we don't just know how many it was. We know what species we planted. Oh, fantastic. Ok. So wards one and two, the majority they planted the sycamore. Ward three, a different makeup that helps you understand a little bit about the areas possibly and the needs. So there's some other information going on, but it doesn't actually give you much insight. There are different things you can do. There's visualization which we'll move on to a second where you can actually look at area and area, what we've learned what we've done here. And the key thing is that, well, actually, we had this scheme, we needed to plant these trees. We're a bit behind, we're gonna get there by this point in the future. So it's naturally it's an operational judgment, an operational decision rather than just a count of how many things you've done. And the key thing is if you look at the bottom here where we want to get to is moving from a very reactive, oh, well, we know what we did to setting ourselves up for a proactive approach and a strategic thing for the future. One of the ways that we're doing this is helping set up the Tree Management Agreement. So this is set up in collaboration with, with Kew Gardens, the idea to work alongside with local authorities to improve that kind of understanding that that collection. So James is saying, oh, wouldn't it be great if we were all working the same way. This is the kind of first steps idea that we can collaborate, share in form, work together different local authorities with experts and researchers from Kew Gardens, sharing their knowledge site to build consistency with that, that data standard so that we can improve the management and widen the sc a lot of what we do. What James and Lucy do. I kind of look at it afterwards is data visualization and there's a real key thing to this. So I can show you a map here and I can say all these black points are trees and you look at it and say so what, then we could do some stuff behind the scenes. We've done some analysis, looked at the risk and actually visualized it and immediately it's name. I don't need to say anything. You've already made a judgment about that about. Well, where would I start if I was to look at something where's good and where's bad. So, visualization bringing things to life is key because it makes things easier to digest. It helps us bring that data to life. It makes it easier to identify patterns, the next decisions really quick, direct, they'll walk you through real quick. A study of a very recent thing we did with the customer. So they they had a true management system. They had true smart for a number of years being used by the, the highways team and the highways team had built a beautiful inventory, knew about how they were using it. We were managing it to, to cyclically go and go and tell them their trees and they said, but we've shown this to the property guys and they love it as well. And we've got all these parcels of land we can bring in the property polygons, but we don't know where any of the trees are. You think? Well, that's a huge amount of area that could say if you do traditional methods that will take years and years and thousands of pounds to go and serve. She said, well, why don't we use liar? So not mentioned light up before for anyone who doesn't know it's a laser, you bounce it off something and it tells you how far away it is. So the planes fly over and that can very accurately tell you the height of the trees. This is freely available. This is something we can harness and we use. And we did use it in this situation when it mapped the heights and it can be polygons. So what we've done here, you can see yet again, visualization, red is good down to green. You've got red is your highest and green are your your lowest trees because we know that an impact zone of a tall tree is gonna be bigger than a small one. If it falls, the radius that it can fall is gonna be a much bigger distance. So we plotted it. The only thing we know at the moment is the height we plotted them, we identified them, the small green areas, kind of the hedgerows and things we brought in as polygons, the tall trees, as individual trees instantly create an inventory. And then, because at the moment, we've only got the height, we validate that for a survey, but we can target that and say, well, look, this has got a load of really tall individual trees right next to a school. Probably start there rather than this little hedgerow. That's that's just outside. you know, someone's house. So key benefits, it adds value to the existing approach. It's complementing. We need to go the survey. We've got to do that, but we can't survey everything. It's gonna take forever removes the risk of walking roads without trees, cos we know where they are creates that inventory makes targeting the sites much simpler. Again, it comes back to this thing we want to do every single time, save time, save money, make more efficiencies, reduce emissions, how we use data to solve your real world problems. So again, James mentioned that he wasn't very big on the voodoo and things behind that. And that's kind of my role. I think this is what to bring to life that actually what is the data? What can we do? How does that kind of all fit? So I've got a few examples of of issues that we've kind of been working through with people of how we've shown the skills we've got answered some questions that actually possibly we didn't even know we were asking beforehand, but it builds into our knowledge and shows how we can support to do other things. So where should I plant trees? I mean, this is a kind of a key start for 10. We know that we need to plant 1.5 billion trees. And I'm pretty sure that involved everyone involved in this room will have planting as an element that they need to work on. Woodland trust recommends that all local authorities have a tree and woodland strategy. Just a quick show of hands in this room if you know that the way you work has a work strategy or some kind of risk management plan or an approach, which is ok. Fantastic. So this is gonna be something that we're gonna talk about in the in the wrap table just after and hands up again if you are directly involved in either coming up with that plan, designing that plan or implementing it. Fantastic stuff. Ok. So how do you go about identifying suitable areas for trees? Sure. There are loads and loads of different approaches, different ideas, but this is just a quick one that we come up with. So UK government target to achieve 17% tree canopy coverage does that mean as a whole in certain parts of the lab, how do you start? Do you do that as a as a counting as a ward? What we can do is take that information, we've got that data and plot it ward by ward and say, well, this one here got 41%. These down here are very, very low. Maybe that might be somewhere to start. Obviously, there are so many other factors involved. You know, the the suitability of the land could well be that there are trees here have then since then. It's all industrialized. They've been planted over, it might have been the best place for a tree anyway, there's no runoff. So certainly different factors. But it gives you a starting point to say, let's focus here. Let's work where our task is, let's try and make things quicker and more efficient. It's a way of moving from being reactive. It's proactive, the isa isa tree risk assessment. So the isa for anyone who doesn't know you do is the International Society of our Horticulture. They use a system called track tree risk assessment qualification. In order to go and survey tree. It's a very, very long form. I'm not sure everyone doesn't use every single bit, but there's an a section about aspects of the site to consider. So I'd use this and thought, well, actually, this would be quite a nice thing to work through. So patterns of previous three fair general and specific site specific drainage lab disturbances, flooding grave drains, hydrology, general characteristics of the tree. We've kind of touched on already land use history, its rain wind patterns and sort of characteristics. So I kind of mentioned very, very briefly in passing there that you know, when you're looking at black a tree, the lack of history, the changes of something that kind of a good effect where suits go. So then we looked at planting trees by risk exposure. So as I said before, we know that no one has the resource or budget to allow for every single tree to be surveyed in the traditional way. But actually, can we help identify which of the high risk areas where there are your problems where actually you should focus time and effort. We don't need every tree to be on the same cycle. I visit every single tree three or four, every three or four years. That's fine. But actually, if you've got trees with defects, could you potentially go to them once a year and not the others? Not for every four or five, manage your risk. Your your time litigation is a key driver for tree maintenance to understanding where your risk is is gonna be absolutely crucial to avoiding that because anything that you do is gonna take time and money. So here is a map of a load of trees. You just, it's got a bit fuzzy, but in the top corner there, risk scores high, medium and low, again, red and green. And what we've done is we've like we've drawn the trees by the impact zone. So if a tree is 10 m tall and it falls over, it could be 10 m, that's the radius of the tree. So the bigger the circle, the tall of the tree, we then looked at some ofthe criteria involved. So species defect history, pathogens, the life stage of the tree, remaining contribution, whether it was in a flood zone and all these other kind of bits that we could bring in to identify and score the risk by the tree. So you've then got a score right down from grid that really simply it just starts to pop out and show you. So you've got here, 35 trees here, all small but grow too green. Actually, this doesn't seem like much risk compared to this area in a recreation ground, lots of red trees, history of failure, two big tall trees, actually, there's much more exposure there and likelihood that it's gonna hurt someone. Ok. And what that does is it allows us to target the inspections. We can schedule centrical visits, we can prioritize our work, we can start doing more with let's identifying trees at risk from flooding. So that tree risk assessment mentioned al hydrology, flood disturbances and things as a factor for planting. When you're, when you're surveying, managing planting trees, climate change is obviously a key issue and seems to be able to be going one way. So flooding is gonna be more and more of an issue for us over the years. And certain trees are more susceptible than others and less tolerant of being in flooded areas. So we could look at those, all those correlations bring them together and say, well, these are potentially more likely to be defective than others. So his little trees in wing chosen working because of the data that they capture. And this again comes back to, are you capturing the right stuff? It's crucial for us to understand the tree species. We also quite like having the the height diameter, breast height can be spread and those kind of things. So we've got working data set and knowing the species, we can say these trees are more sensitive, these trees are more tolerant. We then have the environment agency flood zone data. So this is the fluvial flood zones. So rivers and streams, we also have the surface water flood zones, so much more reck full and we can put them together. And we've got six here, six risk factors that could affect trees. If we overlay those trees back up, we can then identify the sensitive trees that are at most risk of flooding. If I get rid of the noise for you, you're left with those trees. Obviously, the darker blue areas are in more flood zones. So there's about 1500 trees on that map. But instantly I've reduced the ones that you key risk areas. The things that you really, really need to get to from 1500 to 100 or two. Actually, that's where I need to concentrate my time and effort. We took it a step further and have a look at oak trees in the area because something we know is that oak session moth doesn't really like living on wet water locked trees if he can avoid it. So actually, these could possibly less be less susceptible to opf than this area here. This is an example of like how we're using these insights to target key areas, target survey areas, only surveying the areas you need to frees up every resource allows you to do more of those things and ensuring that key information can be captured at each visit. You tell you doing fewer, fewer surveys, which means that you've got more time to capture more and more information. Going back to our risk assessment, we've looked at the patterns of previous tree failure and that risk scoring and the general characteristics of the tree. And then we looked through flooding at general and specific sub drainage and flooding and hydrographic. Is it possible to select suitable trees for planting based on the soil type? If we know it, we know that trees are my preferred soil type and we know what trees are better in certain landscapes, certain ecosystems. So we can map those and help you with the planning. So we've got this, this information, local knowledge is vital. We can't get away from. Actually, you're the experts out in the field. You know, your areas, we can bring this up to life, but you still have that insight. I'll talk to a guy a week ago, happened to be in London, but it, it's, it could be an issue all over the country about some very big trees, big thirsty trees in the dry air dries kind of season last year. Salt as much water out as they could. You've got clay areas and then compacts dries out the clay potential, struggles with subsidence and other issues. So actually knowing what you've got around you where, where buildings are other factors are key, but we can look at the site. So here is a flower that shows you the soil escape soil types. You've got some loamy and clay floodplain soils and loamy soils, naturally wet soils, all those kind of things, which is absolutely fantastic. But wouldn't it be easier? Much simpler if we could just say what, what trees are better in those areas? So we've got that data set as well. So in the yellow areas, Beach Oak, some load of birch and curb areas. So actually, if we use this as an overlay with that data and analyze what you've currently got what the defects would be potentially where you could would look for. Actually, this is what we want to do next. We need to plug here. What sort of things should I be doing? All kind of just comes to love again, this is immediately kind of applicable to those strategies that we can plant all those kind of things that we were looking at. Topography also affects tree success. So we can do some 3d modeling left to identify wind weather, other factors, trees that are on slopes exposed where roots may be coming out where they may be less stable distances above the water table. So and suitability would allow us to do it and see it. So if we bring this up, we can look at the slope of the gradient no longer in woking because there weren't enough hills to make it look pretty. So there is a bit of cumbria and you can see that you've got lower gradients here up to steep. So you already know from that where you imagine that Theresa more likely to survive. We've also got the source scape information for that. And the aspect and the aspect is quite interesting because you've got north, south east west, you can see where all these things are facing and that allows us to then start to put in the prevailing winds, the rainfall, the sunlight, all other factors that affect how we plotted. If we bring all of those together, then we can plot the trees that are in there. Then that's when you see, actually this is a tree that's exposed, doesn't have access to much water in a soil type it doesn't like and understand why it's typically having issues. So that's covered all those areas of the isa pre assessment. No talk on dates for trees to be complete if I didn't cover some carton as well. So comics, the origin sequestration is key. It's true. It's something that you discussed and you're aware of, but I don't want to teach you. Ok. I'm just gonna run through this bit quickly. You've got your above, above ground biomass and your below ground biomass. You've got your trees and your roots, you then got the stuff. But when it dies, your litter, you dead wood, the leaves coming off the trees dying and the soil, ideally, we wanna get all that carbon into the soil, but we can't just put it there. So that's why we need the trees. Interestingly, globally, 55% of our carbon is stored in soil. But actually the sequestration, what would bring it in and how we do that the vast majority is through troops. What I was talking earlier about it's vital that you're capturing the right fields in order to, to capture the sequestration and understand the carbon. We need the diameter breast height and the height of the tread and the species that helps us calculate the weight. We know that 50% of a tree is carbon. So we can calculate the weight of the carbon. No, the weight of the carbon. We know what CO2 is made up from and the, the electorate made up of that. So we can calculate the CO2 and knowing that over that, if we know the apes tree, then we can work out also how much it takes in per year doing that. We can make some pretty bubbles. So we've got here, we're showing the KL carb sequestration in kilograms per year. The bigger the more there is. So you've got this big common line right in the middle, 100 and 80 kg per year. A smaller ravinia, smaller circle 68 kg a year and then a willow at the bottom 17. Now that's kind of interesting in itself at a really some micro level. But actually the application for it is probably gonna be more a bigger wider area. So we take all those together, add them up and put them on a map. And you can see here, huge amount of sequestration. This area is and almost nothing 18 kg a year compared to 4500. So there's a real lack of trees there. We're talking about not so much risk here but also benefit. So when when K and Pinson talk about actually the the good things that trees bring. If there are no trees there, then potentially it's not a healthy place to live, there's more pollution, it's less aesthetic, got noise pollution, things that it can, it can start to bring in so really crucial for your, for your tree planting. And we could do the same thing with the carbon storage. So now we know where all the soil, where the tree, where carbon is stored in those trees. And in addition, we take that same data, the same information about the we use to calculate the sequestration to other pollutants. I think it was Kenton who showed a slide about air pollution and there was a 14% in the corner about in, in London, some of the pollutants they removed. I was reading a study which had a very similar thing based in Strasbourg. We say around yeah, 10 to 15% was removed city as a whole. But actually, they've done some speed and some studies in smaller areas and in some of those pockets of localized areas, it was reducing the pollution by 55 to 60%. They also understood that hairier leaves are what kind of captures it all it all gets sucked into those. So the understanding what varieties and species of trees to uh to capture plant is kind of key and using all that data and understanding where the council land is, understanding the hydrology and the flooding. We can start modeling those perfect trees and those perfect ways to actually start to reduce some of that. So here it is, this is what it looks like as a whole, is a water open, but possibly this is better at a micro level. So these are the little council parcels of land there. And actually, if you start to plant those uh those bushes, those trees, those barriers, the right depth, then actually maybe we could reduce the pollution that is in our public spaces in our schools. So that's a bit of a whirlwind kind of a, you know, rush through some of the things we'd be doing with data and some of the kind of areas we're going and the way that actually we can kind of reduce the risk. But it's kind of a big cycle really to how this works. First of all, it's capturing the useful data. So you need to make sure you're getting the right stuff in. Once we do that, then we can do the analytics, we can make the inside, we can identify those patterns. Once we identify some patterns, we can actually develop a strategy that's achievable where we've got some real Practicals and goals and prioritize our work based on what actual resource we've got because it's unlikely you're gonna get more time or more money. So we need to be working smarter or if you understood that you put a plan in place and it moves around and round and round and yet again, a bit of a theme for the day what's at the center of that is data? thats me done, Thank you very much.