Investing in Regenerative Agriculture and Food
Investing in Regenerative Agriculture and Food
299 Cameron Frayling - Forget biodiversity credits (for now). Regen ag farm land funds and regulation are driving the biodiversity sector
A check in conversation with Cameron Frayling, CEO of Pivotal Earth, about biodiversity, one of the most important sets of things we should track and measure, and yet it is super difficult and mostly hasn’t been done until now at scale at all. The data is simply not there, so what do we do? With Cameron we check in with one of the leading companies trying to bring technology to this space and make biodiversity measured at scale and cost-effective.
We learn a lot about the current tracking devices and new hardware Cameron would love to see developed, how little most biodiversity experts actually know and not many are able to identify the right insects, etc. What data to trust and how to build trustworthy data, plus the most active customer of the company, not biodiversity credit developers, but regen farm land forestry developers that want to report to their investors about biodiversity gains because the investors are asking for it or regulation is forcing them.
---------------------------------------------------
Join our Gumroad community, discover the tiers and benefits on www.gumroad.com/investinginregenag.
Support our work:
- Share it
- Give a 5-star rating
- Buy us a coffee… or a meal! www.Ko-fi.com/regenerativeagriculture
----------------------------------------------------
More about this episode on https://investinginregenerativeagriculture.com/cameron-frayling-2.
Find our video course on https://investinginregenerativeagriculture.com/course.
----------------------------------------------------
The above references an opinion and is for information and educational purposes only. It is not intended to be investment advice. Seek a duly licensed professional for investment advice.
Thoughts? Ideas? Questions? Send us a message!
https://foodhub.nl/en/opleidingen/your-path-forward-in-regenerative-food-and-agriculture/
https://www.landalive.co.uk/
Find out more about our Generation-Re investment syndicate:
https://gen-re.land/
https://www.freshventures.eu/
https://investinginregenerativeagriculture.com/2023/02/21/bart-van-der-zande-2/
https://investinginregenerativeagriculture.com/2024/03/22/chris-bloomfield-daniel-reisman/
Feedback, ideas, suggestions?
- Twitter @KoenvanSeijen
- Get in touch www.investinginregenerativeagriculture.com
Join our newsletter on www.eepurl.com/cxU33P!
Support the show
Thanks for listening and sharing!
Biodiversity Maybe the most important set of things we should track and measure, and yet it's super, super difficult and hasn't been done until now at scale at all. The data is simply not there. So what do we do? Today we check in with one of the leading companies trying to bring technology but software and hardware to this space and make biodiversity measured at scale cost effective. And, wow, we learn a lot. But current tracking devices, new hardware our guests would love to see developed, and how little some biodiversity experts actually know and how some are not able to identify the right insects. So what data to trust and how do you build trustworthy data? Plus, the most active customer set of the company, not biodiversity credit developers, but region, farmland and forestry developers that want to report to their investors about, hopefully, biodiversity gain, not loss, because their investors are asking for it and or regulation is starting to force them.
Speaker 1:This is the Investing in Regenerative Agriculture and Food podcast Investing as if the planet mattered, where we talk to the pioneers in the regenerative food and agriculture space to learn more on how to put our money to work to regenerate soil, people, local communities and ecosystems, while making an appropriate and fair return. Why my focus on soil and regeneration, because so many of the pressing issues we face today have their roots in how we treat our land and our sea, grow our food, what we eat, wear and consume, and it's time that we as investors big and small and consumers, start paying much more attention to the dirt slash soil underneath our feet. To make it easy for fans to support our work, we launched our membership community and so many of you have joined us as a member. Thank you. If our work created value for you and if you have the means and only if you have the means consider joining us. Find out more on gumroadcom slash investing in RegenAg that is, gumroadcom slash investing in RegenAg or find the link below. Welcome to another conversation. Today, we tackle nature. Nature is invaluable and we're talking with someone who's building a company. They're back on the show and they're building a future where nature and business are in balance by incentivizing the regeneration of biodiversity at scale. Welcome back, cameron. Well, thank you very much, and it's been.
Speaker 1:I was checking in the preparation. I was like, oh, it's been a year, maybe a year and a half, no, two years ago, may 2022. And the company was still very fresh, I would say, and very new, I think, less than a year old Super exciting. Fresh, I would say, and very new, I think, less than a year old, super exciting. We were talking about a number of your pilots you were doing, but they were still literally pilots because you're flying, but they were still very, very small, I think.
Speaker 1:Two years in, a lot has happened. A lot has happened in the biodiversity space. At the same time, more needs to happen. So I'm very excited to have you back on the show to check in to see what you've learned surprises and non-surprises and all of that. But for the people that didn't listen to the previous conversation which of course we'll put in the show notes a brief intro to what Pivotal is now. I don't think it has changed too much, honestly, but we're talking in April 2024. So if you because you're talking to investors and to others if you have to give you your slightly larger elevator pitch, what would that be? If you have to introduce yourself, I don't know, at a dinner table or something?
Speaker 2:It is hard to believe it's been two years. It doesn't feel like that much time has gone by, but my goodness. So as a background at Piml, I guess the fundamental thing to understand is that we are a data and analytics company. We're based in the UK, but we're operating globally and we're focused on providing the data people in companies need to invest in outcomes for nature and biodiversity, and that means reducing the cost such that people can actually track change over time so you can make investments in outcomes rather than invest in projects. It's a very different way of looking at how you can make investments in outcomes rather than invest in projects. It's a very different way of looking at how you might make investments in nature, but we think it's a very powerful one and scales to the global economy. How you look at supply chains, how you look at real estate investments lots of different things and in those two years, biodiversity I mean.
Speaker 1:Obviously I'm not deep in the biodiversity space. I remember discussing it last time. You say there's a lot of hype about carbon and around carbon. I think it's sort of slowing down maybe a bit. Carbon sequestration, I mean regeneration, et cetera. And we discussed like there are ways to do that which are not great for biodiversity Monoculture, anything monoculture, honestly and monoculture plantations, tree plantations, etc. But if you do biodiversity well, if you have biodiversity gain, you almost always have carbon gain as well. And is that notion of? Those two are so interlinked? If you start from the biodiversity piece, has that, let's say, landed more over the last two years? Has have people, instead of just throwing the word biodiversity oh it's horrible what's happening with biodiversity but, as let's say, landed more over the last two years? Have people, instead of just throwing the word biodiversity, oh, it's horrible what's happening with biodiversity. But, as let's say, the underlying knowledge, has that shifted, changed? Are we becoming a bit more mainstream or not yet?
Speaker 2:I think it's very interesting. I think, if I go back something like two years ago, there was the very beginnings of an idea that you could have carbon projects and you could also, of course, measure the biodiversity co-benefit from that, and then perhaps, by being able to evidence it, you could get a premium, because you could actually evidence to your purchasers that you were having a real positive impact on the ground. You were having a real positive impact on the ground. The reality of what happened is that there was an interest in doing that and then the purchasers never got on board, factually requiring that data, and so, of course, the producers, the developers of these projects, never do it, and so the data around carbon projects is more or less the same as it always has been. So I think I have seen, in the past year and a half, 0% of those projects turned into anything commercial. People just do the same as they've always done, and I think that's not amazing in terms of carbon projects and the actual impact they have and evidencing what goes out of the ground.
Speaker 2:I think what I've seen is a real understanding that nature and biodiversity are incredibly important to our economies and then a real will to try and figure out how to now look at some of these things, but it is so complicated people struggle to know where to start. So that includes new regulation coming in, things like the EU deforestation regulation, which doesn't just speak to trees and lumber, it's all kinds of things, and this includes soy. There's the CSRD, there's reporting requirements, there's voluntary frameworks which are kicking off this year, so there's the TNFD, for example, there's the GRI, and all of these help frame and produce the kind of framework for looking at reporting around biodiversity disclosures, your impacts. But there's still this enormous question is how you do some of these things at scale. And so, as requirements come in, as stakeholders push for, more typically, companies come looking for us to try and figure out how they might measure on the ground some of these things, like if they have a supply chain, how would they actually do the sampling so that you've got something that's statistically robust? How would you do it so it's affordable? And these are sometimes really complicated questions.
Speaker 2:I can give you lots of interesting examples of problems people have come to us with. One of them, for example, would be regenerative farming backed by funds. I would say the majority of our clients are now asset managers who really want this data. They want it, they're going to pay for it, they're going to get it. And then they come looking for us, and some of the things that they're looking at might be 5,000 hectares of shake-grown coffee in, let's say, south America or somewhere in Indonesia, and they're looking at not just that individual plot of land but also the landscape in which they're operating. And they want metrics. Exactly, they want metrics. They want to understand how close are they to other areas of interest for nature and biodiversity?
Speaker 1:And why do you think they have that interest? Is that is the investors? Did they promise this data to the investors when they invested and they now have to deliver? And thus is it important to the investors? Is it because it's costly, even though it might be getting cheaper and at scale, et cetera, but still this is out-of-pocket costs. This is not something like what drives that interest from asset managers.
Speaker 2:So one would be. Of course it's regulation, so there's a month they have to do some of these things, and of course, it's a staged requirement, so starting this year and ramping up over time. But that means that there are these reporting requirements that people somehow have to solve. People will try and solve them in different ways, in cost-effective ways, but still you're looking at sometimes vast areas of land, really complicated analysis. You need someone that can work just on these problems and help solve them. You need someone that can work just on these problems and help solve them. But what we don't see is the companies themselves being the driver, but the investors behind the companies, and they want this data. I think there's a lot of people who want to make ethical, moral investments, investments that are quite long term sometimes, and they're creating funds specifically to make these kinds of investments. So we're seeing that with, for example, there's Swiss banks, there's funds in the US, there's funds in the UK all kinds of different asset managers.
Speaker 1:So, interestingly enough, it's not the carbon developers that we were discussing two years ago that maybe wanted to add on, not add on, but also show that they were doing amazing things on biodiversity, but it's the more on the agriculture and timber asset managers that manage large pieces of land and actually we see, honestly, if we we survey the space in terms of who are, we have a list of funds that we track of, who are putting money to work in in regeneration and that scale, by and large, it's the group like the largest amount of money flows into asset managers that hold land and that invest in land, because simply that's where most money gets locked in, obviously because you need to buy a lot of land and it's been the most developed because it's sort of real estate. We partly understand it. To a certain extent it's. It's a really um quote, unquote safe investment if you're really good at it and and of course, there's a long-term impact there and a long-term trajectory. So it's interesting for pension funds and all of that and I expect that to grow pretty exponentially over the next years because we've seen a lot of these funds started small and started to pick up over the last year and a half, two years like a big, a big inflow of money, and then that means also there's a management fee. There's means other costs are like some money is available to do other things and that maybe these pension funds require from you.
Speaker 1:And thus they end up at Pivotal to ask, okay, how we're going to measure biodiversity in a shade grown coffee plantation of 5,000 hectares? So how do you do that? Or how do you do that or how do you? I mean depends what you want to know, of course. But when somebody comes, like knocking on the door, and says, uh, cameron, cameron, we have a problem, um, or we have, we have a question and we don't know how to do this and we cannot send a huge team of of ecologists into into the field because that's too costly, as we discussed last time. Um, how do we do this over time? How do we even use it for management decisions, because that would be great for coffee production, quality of coffee, etc. And what's the first questions you ask? How do you guide an asset manager through a process to even start thinking about this?
Speaker 2:so I'm going to step back a little bit. So last year we spent a long time working on the methodology for the first bite of your two certificates with a certifier called Plan Vivo. Now, plan Vivo is a lovely, lovely group to work with, incredibly sincere in what they do. They turn down most projects, which is why they have relatively few, but the ones that they do have are incredibly high quality, and they asked us if we would write the methodology for how you track change over time for projects that might be anywhere in the world Tanzania, uganda, new Zealand, south Africa, the USA, europe, you name it.
Speaker 1:Very different biodiversity, so how do you cover that in one framework?
Speaker 2:Yeah, I'm not sure I can talk that through.
Speaker 1:No, no, no, I'm not saying here, but it's for people to understand the complexity when somebody asks a question like that.
Speaker 2:But perhaps in the show notes we could pop the link to the methodology, which is available online. It was a lot of work to put through to put together and then we set it for expert review. It went for public consultation and it was published December last year. It's, I think, a really nice piece of work. If nothing else, you could read the first few pages pages two to five where we talk through the design choices that were made. Nothing will ever be perfect. So what compromises did we deliberately make?
Speaker 2:In thinking through the balance between the amount of information that we're collecting, the data we're collecting, the cost implications, the scalability, all of these things, and because we had gone through that process of having to put all this down, work it through, be able to explain it to a broad audience, we were in a really good position to then work on other frameworks.
Speaker 2:So when these funds have come to us, they usually have some part of a reporting framework they want to look at, and what they're missing is the niche part around nature and biodiversity data which, when you start looking at it, actually encompasses quite a lot of their framework.
Speaker 2:And so the process for now a few different asset managers has been there's an existing framework. They have an existing need. We work with them on the reporting on the framework, on developing some of these things, and then we're now increasingly involved with the due diligence process as well, because, let's say, you're a fund and you're going to back let's call it 20 different projects around the world. These may be quite large, they could be thousands of hectares each and they could be in different countries. You're, of course, going to assess the returns that you might get on any different property, on any different set of investment, and we're now increasingly part of that, helping to judge the kind of the restoration potential for a given site um, it's in the process of selecting, basically way earlier than oh, we bought this, we're working on it.
Speaker 1:Please follow it over time this is, we have three options, which one should we choose?
Speaker 2:that's. That's something that's evolving organically, just as, um, people are trying to make these decisions. It turns out actually we can be quite helpful there. And then once they do make that decision, then we can handle the ground-based. You know collecting that primary data and doing all the analytics Beyond. You know figuring out how to do the analytics.
Speaker 2:You, of course, also have to figure out how to do the data collection, create the sampling plans. The sampling plans can be very complex because you might have to deal with time constraints. People are, there's seasonality, there's migrations, there's crop planting cycles, there's forestry cycles, like whatever there might be. There's access requirements. So you have to create a sampling plan that works with time, it works with the different habitats, and then you also have to be able to work with local communities to be able to collect that data.
Speaker 2:The most efficient way for data to be collected is that it's always the people who are there locally doing the data collection. For that to be done, well, of course we also have to provide training, and so we've spent quite a bit of time now putting together training materials and testing those training materials, such that people can, within a short period of time, be shown how to use different bits of technology, adapt it to their environment that they're in, because of course Ecuador is not going to be the same as Tanzania. There's going to be different requirements, different issues. The local communities can help solve those. But a huge part of collecting that data is also evidencing that data, because if you're a financial institution, of course you care about auditability, and that is what we built in from the very beginning with Pivotal is this concept of. Not only do we want to track change over time, we want to be able to evidence it. And without that evidence, how can we expect people to put real money?
Speaker 1:And what does it mean in practice, Like how do you evidence data?
Speaker 2:Yeah, sure. So in reality, that means digital data collection, not manual data collection. So what you'll see, historically for the last hundred years or so, people would go up clipboards. They would be experts in their field, they would note down what they see, but you can't prove any of it. It doesn't mean they're wrong, doesn't mean they're right.
Speaker 2:And if you're a fund, what you care about is being able to have at least some evidence that a process was followed, that there's some understanding of how likely it is that any given result is correct. And you can do that if you do digital data collection meaning cameras, acoustic sensors, drones, anything that you can evidence, your satellites, lots of different things. And then you have data pipelines. You've got databases, you've got auditability. You also need to provide field apps so that people can be out in the field, navigate to a site where they're meant to be collecting data, have checklists of things for them to work through, be able to evidence things with additional photographs, gps locations, timestamps all of this, and without that, it's hard to be able to have the metadata that you need to be able to evidence that the data was collected at the right place, at the right time in the right way.
Speaker 1:And that was, until relatively recently, impossible because you could fill out your clipboard in your hotel and nobody would ever know, because we didn't know where it was filled out. How it was filled out um, maybe the color of the pen said something, but not really. Um. And now, of course, with with this, this process becomes uh, not impossible to fraud, but way more difficult. Like you have to go through a great length and a lot of photo editing and Photoshop and sounds, etc. To do that, which makes it much more trustworthy for financial institutions to do something with, or at least to build up on top.
Speaker 2:Yeah, that's right. That means you can use all of the modern tools that are available machine learning, all the different sort of data science methods that you can throw at this kind of thing which gives you much better data, much better data analysis and ultimately the results can be just far more meaningful and more relevant to how people invest money and to the global economies.
Speaker 1:And that's really that combination of the two things the data piece and the database piece and the software around it, but also the hardware. I mean, drones weren't around 20 years ago I mean a few, but not the ones we need now. Cameras weren't that cheap or that good, microphones, acoustics, the same. So that combination of the two, plus the need of the huge biodiversity loss we've had, and makes it a very interesting, uh, interesting moment in time to to unleash that. And do you like what? What's like of those different pieces, what's the least developed or where would you love to see more development? Do you need better drones? Do you need more more interest from from potential customers or clients, or is the software piece still, or is it all sort of developing at the same time? And actually it's a good flow. What's holding you back? I think?
Speaker 2:so if we focus just on hardware for a moment, I think, yeah, there are now tools that are really good. So, for example, there are acoustic sensors you can buy that are inexpensive, but you could probably do much better with some relatively straightforward engineering. So we actually hired an engineering firm in Sweden to help develop a prototype of an improved sensor. And instead of having one microphone, it has four. Instead of being able to set the frequencies and time in during which you're recording, it has a, you know, and have only very few choices as to how you do that. We have a much more complicated um sort of way of setting different frequencies for different times of day to be able to enable to, to help preserve your memory, your battery, all kinds of things. You can do things like beamforming. If you have multiple microphones, you can have background noise subtraction. It means that the range of the sensors can be improved, the sound quality can be improved, and that's just one example.
Speaker 2:I would love to extend that to things like ultra-low sounds and being able to do a rumble detection for different things. There's lots of stuff you can do with acoustics. It is generally fantastic value for money and we love to use it and, of course it feeds really well into machine learning, machine learning models, but you I'll come back to that later. On the other side, I would say the drones are fantastically capable. You can program flights and with the press of a button you can fly them almost without looking. They're really remarkable pieces of kit. What I would love would be improvements in autonomous navigation through some of the very complex environments in which we operate.
Speaker 1:Forest, anybody forest yeah.
Speaker 2:Exactly so. I know there are people working on this. There's a amazing professor in switzerland working on sort of how you have autonomous drones navigating obstacle courses I think he now has it so they can outperform human pilots in these course of race courses. But in a forest you don't have a clearly marked target or path. You have twigs that are sub-millimeter. You hit some of these, you're going to crash.
Speaker 2:I have flown myself drones in these sorts of scenarios and, yes, it's very difficult to not crash. But if you can, then the kinds of data you could collect would be really so much better. So I would love to see small, capable autonomous drones that have really high-quality cameras. That would be fantastic. But the second you start adding improvements to some of these drones, like better cameras, they become bigger because they become heavier, and so there's always these playoffs. But that would be another step forward, I think.
Speaker 2:For marine data collection, I think there's tons of stuff to do. Right now. You have to hack together relatively simple things called BRUVs, and we would love to have arrayed acoustic sensors that we can use to detect different species that make sound underwater and be able to identify how far away they are. Machine learning, identify what they are lots of different things like that, so I think there's lots of scope. I think the number one thing for me that would make a big difference would be better insect technologies. You would be surprised at how uninspiring some of the current technologies are in terms of their capability, and I think when you say technology, what do you mean to?
Speaker 1:to just for people that are not used to collect intact samples? What should? What does the technology do now and what should it do?
Speaker 2:there are. Let me give you two examples. One example and I'm not trying to be too dismissive of these technologies they do what they're meant to do. One would be you have a colored background, say something yellow, and it's flat, and you have it perpendicular to the ground. You set up a camera facing this colored sheet whatever it is felt or paper and then insects land on it, and when they landed it you take a photo of them and that's fine. But you could imagine that that only works during the day. It only attracts things that like yellow, and it's relatively expensive and clunky to put out.
Speaker 2:When you think about other ways of doing it, you could have light based attractors. Those tend to be a little more capable and of course, you set them out at night. You can do things like adjust for the amount of ambient light around you and you can attract all kinds of nighttime insects, but also, believe it or not, a lot of insects that are normally present during the day. Typically, what they do is they attract them into kill jars and you then get a big jar full of dead insects that you can then pick through, identify or send for DNA sequencing. The DNA sequencing side is also relatively inaccurate. But what I would love to see as just a first attempt would be a combination of a light-based detractor instead of a kill jar, actually cameras, cameras that are taking good, high quality, up close photographs of these insects as they come in. Things like that, simple things that would actually move us way forward in terms of our ability to do detection of different insect types and calculations abundance in relatively cheap ways.
Speaker 1:Because the photos you can then feed to software to recognize. And with dna that's way more more complex. And are people working on that in in like? Is that a not saying a hot, topic in the hardware scene? But, um, are people starting to take, at least, let's say, biodiversity sampling? And because a lot of the drone pieces that come from other spaces I mean the professor developing, uh, this kind of software is mostly, I think, coming or looking at the racing side, like if you haven't ever seen a drone race, definitely not now, but go and Google that, search for it on YouTube or Vimeo et cetera because it's fascinating to see how capable we are with training to navigate an obstacle course. But it's from another sector, sort of. Do you see the biodiversity and hardware theme becoming more and more? Do you see more companies starting to build specifically for biodiversity piece, or is it usually coming from another space and then sort of adjusted?
Speaker 2:I think I don't see. I do definitely see some technology development, some in acoustics, some in drones, I think, some in insects, but most of that seems really early stage. Some are commercial but they're specific for agricultural applications. They use scent attractors, for example, to look at specific kinds of insects like hornets or whatever it might be. And I think what I'm talking about is something that is not necessarily going to pick up every kind of insect, but it's going to be consistent over time. It's going to be relatively easy to deploy leverages all the machine learning you can do to identify insects and is also relatively inexpensive. Now, at that I have not seen. That's quite a wish. We could build one.
Speaker 1:Yeah, no, of course, but you want to at the same time.
Speaker 1:As a startup, you want to not reinvent too many things at the same time. As a startup, you want to not reinvent too many things at the same time, because then it just gets very complex, even though you might have to at some point do a few things because it's a crucial piece. But, as you're saying, a lot of these are good enough until now. You would love them to do more, but it hasn't been holding you back too much. Let's say, and if you look forward a bit in terms of where do you see if we do this call again in two years? Where do you see? Is the oceans piece that's starting to take off? Is it? What do you hope? And what do you see actually in terms of, okay, the asset manager piece is starting to get serious credits, maybe not yet. Frameworks are starting to come with Plan Vivo, etc. Where do you see the most potential and most development in, let's say, near short term?
Speaker 2:So near short term, I would say, as in the next year. I would say it's asset managers backing regenerative farming projects or forestry projects around the world. One of our clients is in Japan, for example, and there's a project that we're going to be doing with them this summer and then hopefully expanding to lots of other areas that they have around Southeast Asia. I would say just those two groups, so regenerative farming, forestry and those could cover an absolutely enormous number of hectares.
Speaker 2:Yeah, absolutely, and it's also incredibly impactful because people are doing it partly because they need to report some of these things, but mostly it's the investors who are incredibly annoyed with ESG markings and how little they actually mean, now wanting real data to back these things up, and that's what I would say. Those are our customers, as the people who are angry with how little ESG marks actually mean, and then, of course, that's a great sector of people that we would like to work with. I think there's also going to be things around insurance, because I think fundamentally and I've felt this for a long time when you think of biodiversity metrics, what do they really mean? If it goes up, you've got a higher set of biodiversity indices. What does that really mean?
Speaker 2:Well, it really means that the ecosystem is more resilient, and if you think of resilience and how that impacts economics, the two really go very well together, and his view is that some of the indices that we're looking at. In regards to mainstream farming, you know major supply chains, commodity crops, there are different ways that you can grow and, of course, if you produce things in a very intensive way, maybe you will produce much more, but if you have a dry year or you have a very wet year, you may be less resilient to those impacts versus We've seen that actually, yeah, in research, yeah, exactly Difficult years, resilience is key and you might not harvest anything.
Speaker 2:Exactly, and so this is where, actually, you may produce less in normal years, but you might produce more in years where the weather is produce less in normal years, but you might produce more in years where the weather is actually less predictable. It feels right to all the farmers that I speak with. It feels like there is definitely something there, and I think that makes absolute sense that if you were a commodity purchaser, you would start to pay attention to biodiversity metrics as a way in which you actually source your supply, and so I think things like that are going to start to happen, I think starting this year in terms of producing that data in those contexts, and that I think working on supply chains is also something that hopefully, we're going to start working on this year.
Speaker 1:Specifically, how do you produce supply chain data that, if you are a, a broker or purchase or a company and you just need this data, you could purchase it and be able to then, you know, look at your reporting um and have a cost a fraction of what it would do if you try to collect this data yourself um and have you seen surprising like analytics or surprising results until now that maybe people thought they were doing way better um than they actually were, or or worse um, and they were pleasantly surprised or not like, in terms of working specifically with the region ag um crew and the farmers you, you with, for maybe with the asset owners Would have been some surprising stories you can share of biodiversity that was there or wasn't there where you thought or they thought it was going to be.
Speaker 2:I think there's a few things actually. So the first thing I was absolutely shocked by was the lack of multi-year, multi-species data sets. So it turns out if you went and looked let's say your job was to find these data sets and you went out and look for them you would find five in the world. Wow, there are private data sets, but in terms of ones that you or I could just go and access and then draw some kind of conclusions from, there are five, and even there and there it's not like they're looking at 100 species. They may be looking at three, that is.
Speaker 1:It comes back to the point we made like in the other, like the, the, the little we know, like how many hectares we've actually surveyed and then how we surveyed them until now, for good reasons, because it's super expensive and we didn't have to date. But, like just for people to understand, when we talk about biodiversity, most of it we don't know, like we have very little view on very few hectares or acres, and that's an issue.
Speaker 2:Yeah. The other thing that I found surprising is we've now done multiple years for our first sites, cause we going for a little over two years now for our first sites. We're going for a little over two years now and I guess we've got our third year's data set just coming up on the horizon. And if I look at change over time, the first thing is that the metrics that we see, because I've been to most of these sites, almost all these sites, I've been there myself.
Speaker 1:If I look at the metrics that we produce, I am very surprised at just how well those metrics correlate to how it feels on the ground yeah, I was going to ask you because you've been there and you've used your sensors, like your eyes, your nose, your ears and that's, and you're not an expert in the, you're not a bird expert and it's exactly, etc.
Speaker 2:That's interesting yeah, and so how it feels on the ground is actually really well reflected in the sort of the heat maps, the different bits of analytics that we can put together, and I found that surprising, the degree to it which is that that's true, and I've seen that in our clients as well. So when you present the data to your client they're looking at these reports, they're looking at graphs, they're looking at maps to your client they're looking at these reports, they're looking at graphs, they're looking at maps and of course, they're wondering, okay, well, how much can I trust this data? And what always happens is they find something that they can relate to, that they know. So they see it in the data and they know why that might be true. They make a connection and then after that they realize, oh, actually, this is actually reflecting really accurately the reality on the ground, and you see that happen again and again and again. And that's the moment when they've done that themselves, that I know that they'll start to trust in this data Fascinating. So that's the other thing that I found really interesting and really, you know, gives me a lot of hope for how much we're doing now, but also how much better we can do going forward.
Speaker 2:I think the other thing that I found absolutely shocking this is less positive is the degree to which major institutions will come out with a framework and a set of metrics around nature and biodiversity and proposed that as the way that things should be done, and as part of the work we did with Plan Vivo. We would take all of those things, all these different frameworks and metrics, and we would test them. We have an analytics team of PhD ecologists and data scientists and we would produce data. We would create synthetic data and we would test to see how they performed under different scenarios, and what we found was that they performed in sometimes very strange ways. We would then go and ask these institutions have you actually tested these frameworks that you're proposing? And 100% of the time the answer was no. They had never actually tested the frameworks they were proposing the world used as a way of measuring nature and biodiversity. I found that very shocking, but it did mean that we became the experts at testing methodologies and frameworks.
Speaker 1:So no, just letting that sink in, that probably the most important thing to measure is is life around us and that biodiversity. And we're starting to attempt to do that, and with frameworks that have never been tested um, which cannot go well. Let's say, like there's a, there's, there's a. I mean, it might be, we might get lucky and these frameworks work brilliantly, but, knowing history, they usually don't um, unless you test and and refine, etc. Etc. Especially with something as complex as biodiversity.
Speaker 2:Um, in what other scenario would that be okay? None, yeah, like it's, it's that's never okay. Is that one of the reasons why we haven't?
Speaker 1:seen credits so much yet like we haven't seen, like this huge uptake, because it is still in the fake space and not really okay. How do we then actually cost effectively on large scale? Measure, track over time changes, because that's what we need, otherwise what are we going to sell?
Speaker 2:I think the way I think what companies and the potential purchasers of credits are thinking at the moment is that it's involuntary. For one thing, it's unclear what high quality and low quality is. Because of that particular problem, we actually put together a document which I'd be happy to share, so if anyone would like to read it, please just email me at info at pivotalearth, and what we've put together is a series of questions that you can ask. So if you're looking at biodiversity and nature data, you could ask these questions and very quickly you should get a sense of whether it's high quality or low quality and whether you should be basing your actions on this data. But going back to companies and purchasers, I think there's just a lot of uncertainty. They don't have to.
Speaker 2:It's hard to know what's good and what's bad. You need to be an expert. There's no consensus as to what is required. There are a lot of strong opinions. There's a lack of data on which you base these things, and so, to some extent, you know our role is to provide data. Our role is not to judge what is right or what is wrong, but because we do spend a lot of time testing how different things behave, it does give us the ability to have a view, sometimes on what doesn't work. Why?
Speaker 1:something really might not be a good metric, and it seems like the voluntary market is not really going to lead this. It seems more the regulation piece, like what's the biggest one that people should really know about or not the biggest, the most impactful one? Is it deforestation? Is it the disclosure requirements? Like, what do you see most? What do you see that make companies actually move? We've had a discussion on deforestation I don't remember with whom actually, or it was not on.
Speaker 1:Anyway, the deforestation law in Europe for a number of commodities start to be very relevant, because you have to prove a lot of things and show a lot of things, and most companies have no data, let alone where their coffee comes from or their cacao, let alone which farm, let alone which farmer, and if he or she is making a living wage or not, or if there was deforestation. So it creates this huge because there's a significant fee based on your revenue in europe, and it seems to be. Let's see, of course, that this has teeth, in the sense that it starts to actually matter if you don't have, uh, have this. So we see this whole scrambling around. Okay, what are my supply chains actually, etc. Etc. But what do you see in your world biodiversity, which one seems to have most teeth scared, scared companies knocking on your door I see the same thing.
Speaker 2:I think tnfd was um was well done in lots of ways, as imperfect as it might be. It's a great attempt to synthesize some of the reporting that you might want to do as a company, but it's voluntary, so from that point of view it has a carbon disclosure.
Speaker 1:Was it the carbon? I just put links in the um I'll, I'll send you a link after.
Speaker 2:Voluntary, so it's not going to bite, it doesn't have teeth. Lots of people will say they'll do it and they're just not. And also, even if they do do it, if they do it badly, there are no repercussions. So for me, the drivers really are regulation and stakeholder pressure there are a lot of people who really do want this kind of data to be part of their investment decisions and for their companies to report it and stakeholder pressure there are a lot of people who really do want this kind of data to be part of their investment decisions and for their companies to report it. I think the requirements from regulation do drive companies to start acting and to start trying to figure out how to do these things.
Speaker 2:I think a staged approach definitely makes sense. So it's staged between this year and 2026 in terms of a ramp up for disclosures. People need to start to figure things out. What I always expect with some of these things is that there'll be pushback from industry, that some things will be delayed, it'll be watered down a little bit, but by then there's still an impact. There is still something that's happened where a company has had to start to figure this out, and I think you then have a spectrum of behaviors from companies, but inevitably you will be moving in that direction of requiring this kind of reporting. It'll be easier and easier. People will then push back less. So even if it's slightly delayed, that's more than enough for us as a startup in terms of business. It's a huge number of hectares and we would expect it to continue to ramp after that as more and more companies sort of get on board.
Speaker 1:Anyway, and actually coming back to surprises, have you seen negative surprises as well in your, of course, without mentioning people, places et cetera where somebody thought to have done very impactful work, let's say, over years, maybe decades, even on biodiversity? Or coming to your point where you said you feel it when you walk the field, you feel it when you're there, so you're maybe not surprised because you know these certain forests or certain fields are just not that biodiverse at the moment. Have you seen the surprises on the negative side?
Speaker 2:So we wouldn't necessarily see that, because typically people have bought the land or carrying out some actions to do some kind of restoration or it's conservation, and so in both those cases we're evidencing it's already rich, there's no loss, or it's already terrible and it's being improved. So we wouldn't necessarily see that and I would say I can't think of anyone who we work with who I think isn't sincere about what they're trying to do. I think I mean, maybe perhaps as we scale that will happen, but I have not seen that.
Speaker 1:There are different parts of your portfolio where you hoped there would be more and actually it's going maybe slower, or it's just less, but less under the surface, less active or less lively.
Speaker 2:There are definitely things that are negative, but I think it's more around the ability of people who are very loud to sometimes be heard far more than they should be. And without naming any names, like in the credit space, you could imagine there's a whole range.
Speaker 1:It was an interesting year in the credit space last year. Let's say yeah, exactly.
Speaker 2:And there are people who, like they're managing to get into things now, they're making a lot of noise but they're not going to get very far ultimately. But in the meantime, some people, not having enough knowledge, will be taken in by it and invest in something that is absolutely meaningless and that we see everywhere.
Speaker 1:Yeah, no, but that's, I mean the, the region washing I don't know if it's even a term, but that that we see as well. You slap a few sdgs on your um, on your presentation. Uh, some asset managers should be, should be listening now, um, and and you put okay, we're also measuring our water uh usage and now we're a regen fund and people will fall into that because it's hyped. People don't have the knowledge. I don't think we can stop that. I don't think there's a. People will get disappointed. There will be a disillusion.
Speaker 1:I hope they keep searching and Googling and going down the rabbit hole and will realize that there are more questions to ask than a few SDGs photos that you actually are not even supposed to use on your slides and say you're measuring your diesel and water uses and then you're fine, um, but yeah, it's sort of um, we cannot really avoid it. And I think a biodiversity, as it's going to hype, as we're going to understand more and more how important it is and how much we lost and how much we need to restore and how much can be restored. There will be a lot of yeah, there will be a lot of. Yeah, there will be a lot of hot air.
Speaker 2:Let's say and there's other questions around marine. So we're seeing a lot of interest in marine from the kind of philanthropic side of things. People want back marine protected areas and so they should absolutely, and there's some really amazing projects that are coming down the pipeline. One's in very lovely tropical places like Belize or coral reef off the coast of tanda, tanzania.
Speaker 1:I see someone who wants to go there and feel the biodiversity live, etc. Yeah, he's on the call here and it's not me. I mean. Also, I would like to do it, but, um, there are not bad places to go, that's for sure. Yeah, a lot of potential in restoration. I'm not making the joke of the trip, of course, but a lot of potential restoration, because we're going to do an ocean series or a region aquaculture series, which I want to talk to you about in a second. But the, the speed of, okay, first of all, the amount of knowledge we have on the oceans, like dwarves, compared to what we have on land, like we. We know more about the moon, probably, than about the ocean, so we need a lot of research there. The amount of speed and potential of restoration, especially in the tropics, where things grow fast is, is just astonishing. And and the amount of destruction we've done to the oceans and the amount of heat we stored and and, like the, it is ground zero probably of regeneration. Sorry, I will get off my pedestal I completely agree.
Speaker 2:I think there's some really complex questions to answer that I'm not sure we'll be able to provide any kind of meaningful resolution to. Such as, if you have bleaching of coral reefs and it's a protected area and you have less bleaching versus somewhere else because of actions you've taken, should that result in some kind of credit? Have you carried out some kind of conservation? What should you think of in terms of your change over time? Some of these are going to be incredibly hard things to answer, but do you get approached by the?
Speaker 1:farming side of things, the aquaculture side as well. As you see, the most interest for now, pivotal on land is asset managers that manage the regenerative side and the regenerative farming side of things. Aquaculture is a massive industry mostly under the radar A lot of scandals, a lot of issues. We're repeating what we did on land Huge monocultures, a lot of inputs. We're going to get into that in the series. But do you get the first questions of multi-trophics like ocean plays that are wanting to look okay, how does my kelp influence the, the muscles and and the oysters underneath, etc. Etc. Is that still wishful thinking that they reach out?
Speaker 2:yeah, no, not at all. I think we um on the, the, the different approaches that we've had. They're all on the um protected areas. They're on things like improving kelp, carbon sequestration, these sorts of things, but never from industry. I think that will come and I think some of these mass bleaching events that are happening now they are going to hit fishing. Of course they are, it's just a matter of time. And when that does happen in some meaningful way, people will then need to figure out is it worthwhile starting to restore and put real money behind some of these restoration efforts? Can that even be done? Who knows?
Speaker 1:And how to even start researching and start tracking. I think that's potentially where the most hardware is also needed, because you said, some people hack some things together, but that's pretty much it. There's not a lot.
Speaker 2:Yeah, and I mean we would love to collaborate with people developing hardware, people who want to write grants with us. By all means reach out. I think there's a huge amount of opportunity, as you said, to do research into how you collect this data. What metrics are the most appropriate? How do you report analytics?
Speaker 1:How does it relate to mainstream economics, phishing, et cetera, all of these things, and do you see any interest or maybe regulation will come as well to, like the salmon industry and and some of these highly in many cases, not all cases highly polluting industries, uh in in the aquaculture space that they may be, start being pushed for reporting as well, and biodiversity and things like that, which will be an interesting uh case to to to, to tackle that do you will regulation? I mean, it's impossible to predict, but do you see a future there for regulation to step in and then these.
Speaker 2:I hope so. I haven't seen anything yet, but I do hope so and certainly something that, yeah, we would love to to to be involved. I think, generally, marine is something that's, you know, very dear to my heart, so I hope to do more and more of that, but I have not yet seen any commercial interest whatsoever.
Speaker 1:Yeah, oceans are absolutely fascinating and greatly under. They're greatly neglected. I think we're coming to terms with it slowly now as a I mean an ocean board species. Let's be very, very clear there as well, and we've not been treating that nearly as well. Even we've done horribly with land. In many cases, the oceans is even worse, and so, as a… Well, let me tell you about a few more things I would love to tell you about.
Speaker 2:So one of the things that has been very surprising is the degree to which the existing kind of ML models that you can just use off the shelf to identify different species are amazing but wrong, and large language models for people that are like what is he talking about?
Speaker 1:Okay, the AI hype, because that didn't exist two years ago at all. I mean, it exists in some corners, but not like this. And so there's image recognition. There's a lot of things you can do with video and audio and you're saying in many cases they hallucinate.
Speaker 2:It's really interesting. So they are incredibly good and very accurate at identifying common species. They are very poor at identifying more rare species. That's kind of what you'd expect, but what does happen with these models is that they they will take a very rare species and just assume it's one of the common ones. That's something that we've seen again and again.
Speaker 1:That's not good for your papers and it's not good for your database. It's not good for your papers.
Speaker 2:The thing is is that for the use cases for which they're developed, that could be fine. But for our use case, where you're looking at species diversity indices, you've got distributions of species. Sometimes you may have something that's very rare, that fulfills a very important niche within an ecosystem. You want to know if it's there. You want an accurate identification, and so we've had to modify the whole process of how we think about using machine learning towards something that we're calling calm clustering and local models. So instead of looking at an image or a set of audio files and say, please identify these things, what we first do is say, please cluster all the things into things that are similar. And, of course, language models get better and better and better at doing that, especially when you have input from experts who can help identify these things, and then it gives us a much better way of finding rare events. The math bind is a little hard, but I'll talk that through later at some point.
Speaker 2:Finding rare events, but also being able to evidence that they're there, because the sampling process also changes. It's no longer a black box. You can show the process from start to finish, or how everything was identified, how everything was sampled. You can use statistics for sampling different clusters of different sizes, and it also means you can start operating from scratch anywhere in the world. So if you are asked to go somewhere for which there are no language model or no ML models at all, you can do that. You start off doing lots of annotation, but whether it's frogs of Ecuador or bats of Kenya, you could handle it, and so this is something that allows us to operate globally even in the absence of existing models, but to be much more transparent about how we're getting to our results.
Speaker 2:And this came by the realization that the ML models were quite inaccurate in lots of different ways, and also that the way we were working with annotators also needed a huge amount of improvement. So over the last, I would say, year and a half, we have run many, many, many, many experiments as to how to ask the right questions of these amazing experts who could be anywhere in the world and able to have these almost superhuman powers to identify different species. But if you ask the question the wrong way, you get biased results, and so working through that process of asking the question, yeah like with large language models, if you, exactly if you make the mistake with the prompt um, you're not going to get anywhere.
Speaker 1:Let's say yeah, and it's the same with experts.
Speaker 2:Interesting exactly, and so I mean. One example I heard of this outside of our work was actually from this amazing person who has worked as a tracker in Africa for most of his life, and we call it the honey badger problem. So there were tracks in the ground, there were scratchings on a log, and he was there with three expert trackers and they all said, oh, it's this honey badger. They all agreed with each other. And there was one more junior tracker who looked a bit more closely and was like I don't think it is, I think it's this. At which point they then went back and were like, oh, he's right.
Speaker 2:And of course, that's just human nature. If you have someone saying this is a honey badger, the next person is more likely than not to say, ah, I agree, that is also a honey badger, especially when they're incredibly expert at what they do. Say, ah, I agree, that is also a honey badger, especially when they're incredibly expert at what they do. And you have to really go out of your way to disprove that which this younger tracker did. Um. So, with that in mind, changing how you ask the question, of course, is incredibly important, but so is having the right process for quality control, so you need multiple stages of quality control, and this is not something that we're seeing any of our um, our competitors, even starting to look at yeah, because I mean with everything, bad data in, bad data out, basically, um, you need to to to track that like asking the right question.
Speaker 1:So how is it asking? The question changed over time with experts on butterflies or whatever you were looking at like what was a just to give an example was it was a less optimal way of asking a question to an absolute expert with almost superhuman capabilities well, when you start out, you have a smaller network of people that you can ask.
Speaker 2:We now have more than 80 people who are actually contracted to work with us. So the first thing is we said, well, we have an expert in this particular region. We have these machine learning results. Let's create a sampling process and see how accurate we actually were. And that person that goes through and identifies things and says was the machine learning right or wrong? You can see that's clearly flawed. It assumes, makes all kinds of assumptions. It's not saying what do I think it is. It's saying do I think the machine learning was right or wrong? So you can change that and then you can add a second set of annotators. So now you've got more than one annotator looking at the same set of data and you can change the questions that you ask. And then you realize that, well, actually coordinating how you ask the questions between two different sets of annotators is also really important.
Speaker 2:But on top of that, we then started to realize that we would have people being recommended to us as annotators to identify these different species, and they may have a PhD, they may have decades in the field doing this. But you were getting a high level of disagreement between two annotators as to what a set of things was, and we realized that, well, wait a minute. How do we know whether they're accurate or not? How do we know how good they are really? They may have a PhD and be highly recommended and have decades in the space, but are they actually accurate at identifying these species groups? And so we started testing.
Speaker 2:Figuring out how to test also took a really long time, because for every new part of the world and for every target species, you have to create a set of tests. But we figured out how to do this and then we started testing people, and what we found is that a lot of people were failing these tests, and when you had people who were untested, sometimes you would get almost 0% overlap in terms of agreement between the different species lists. When you tested them. That would go up to 92%, things like that. So a huge difference. And so quite a while ago, we moved to the regime where only people who have passed these tests can actually work for us at all, and that was a huge change. But it just shows how far you know you have to push some of these different models and these different assumptions before you get to something that is even remotely accurate.
Speaker 1:Well, it's a good chunk and of course you don't know, because you base it on past experience. Phds, a lot of paperwork you would imagine they would be really good, but if you then put them to the test, a number of them aren't, and that also questions a lot of the papers they've written and produced in the past. A lot of this biodiversity data might be flawed, to say the least, but the problem is we don't know which ones. You start to know, of course, because you start to see which ones actually are able to identify, but it starts with that. If you're not able to identify species, then what are you going to base your research on? And a lot is just very difficult.
Speaker 2:It becomes a bootstrapping exercise. You've got people who are they're validating other annotators. They're all working blind, they don't necessarily know that they're doing that, but then you can gradually figure out who actually is more accurate. You can work with them more. You who actually is more accurate, you can work with them more. You can incentivize them in different ways. I mean, another example is we did this test where we took insects, we send them for DNA sequencing and we set the same set of insects to be identified by entomologists via the images that we took of these insects and we found there was some overlap in the species list, but it was not very good. So the question was was the annotator, the person identifying the insects, wrong, or the lab? Or the lab? I mean, it's not going to get the sequences wrong, but who identified the insect and identified specific sequences with that insect in the first place? So the data in the database, how accurate is that? And so we have no way of knowing who is right in this case and so how do you approach that then?
Speaker 1:like what was your conclusion? I don't trust edna results, probably potentially because of the database there, because they might be based on the same expert that also weren't able to identify. Yeah, that's such a tricky one.
Speaker 2:Yeah, so all you can do is this gradual bootstrapping of improving the quality and standards around data collection annotation.
Speaker 1:Yeah, Fascinating, fascinating. I want to be conscious of your time as well, but thank you so much for a deep dive into the world of biodiversity, data collection, the mist and the, let's say, the unknown. Unknowns there and the unknown knowns that we absolutely know, we don't know, and how to collect, how to scale this, how to build it up, and also, what are the most interesting commercial parties you're working with now. What are the driving forces there? Good to hear that investors are part of the driving force and regulation. When those two come together, it's a strong force.
Speaker 2:There's one other thing I would like to add, and I find this very interesting.
Speaker 2:Sometimes we're asked to look at really large areas, by which I mean hundreds of thousands of hectares, and they may be in countries like Tanzania, they may be elsewhere, and it turns out that no one has carried out any kind of sampling at that scale where you're looking at multiple species over periods of time.
Speaker 2:For obvious reasons, that would cost a lot of money, especially with using historic methods, and so last year we spent four months or so just looking at how would you look at something that's hundreds of thousands of hectares and create a sampling plan for it, and this is something that I'm hopeful will kick off this year. But basically we're going to be taking a section of that. So let's say, instead of 100,000 or 500,000 hectares, taking 30,000 hectares something that we could handle and we're going to oversample that 30,000 hectares and then from there we're going to do data analytics on that site and figure out where we can sample less and then expand that to the much larger area. So then optimizing for, you know, cost of the data and the integrity of the result and that really gets to a landscape based approach.
Speaker 1:like what does a landscape with multiple functions maybe some high intensive farming, maybe some more wilder areas, maybe a national park, maybe there are many different functions, of course, potentially some cities, villages, infrastructure how does that interplay? And the answer is we don't know, because nobody has done that research.
Speaker 2:No, no one's ever done anything at that scale with this kind of granularity of data, so it's really interesting.
Speaker 2:I think the other thing that's interesting so those are contiguous areas. No one's ever done anything at that scale with this kind of granularity of data, so it's really interesting. I think the other thing that's interesting so those are contiguous areas, very large, obviously, well away from normal large cities. The other thing I think that's going to be really interesting to figure out and we've had this presented to us a few times now are areas that may be, say, 5,000 hectares, but it's 5,001 hectare plots of land and they may be distributed over a landscape, and so it's figuring out well, one, how do you sample for that? And two, what conclusions can you draw? What can you actually say If someone's doing an amazing job on their plot, what does that mean in terms of impact on the surrounding area? And so figuring out not just how to sample, which I think we can do, but the conclusions that you can draw from that data and what you're really reporting against in that context, is also going to be really interesting.
Speaker 1:And that will be very interesting for investors as well as you're usually touching a small percentage of the hectares, let's say in a watershed, in a region, and you want to, first of all, be able to select the ones where you want to intervene. Then also, within your portfolio, select, okay, where does it make sense to plant what and why? Where does it make sense to do certain things or not? And if you want to have landscape effect without owning the whole landscape, which in many cases, luckily, isn't possible. But those kind of questions have been not even asked until now.
Speaker 1:It's usually I do my thing on my one hectare, 10 hectares, 100, etc. And and sort of as an investor, which is even odd, because it really is important what your neighbors are doing. It really is important what your neighbors, neighbors are doing for the effect you want to see on your land. And if that isn't even taken into consideration, you might be a little island trying to row against the tide and everybody else is going in another direction, which might be quite detrimental for your investment, your land, your future.
Speaker 2:Yeah, absolutely so. Then you look at restoration potential, and then you also look at the incentives that you could build into that kind of planning. Who knows if people will do that, but it makes complete sense in certain circumstances and that could be very exciting as well. So lots to look forward to over the next year or two.
Speaker 1:Absolutely so. Thank you so much for coming here and sharing an update, and I hope it's not going to be two years before we do another one, because it's always good to do a deep dive Literally. Maybe it will be in Oceans 1 at some point when you start working there as well. So thank you so much, cameron, for coming here and sharing with us.
Speaker 2:It's my pleasure. Thank you, it's my pleasure.
Speaker 1:Thank you. Thank you so much for listening all the way to the end. For the show notes and links we discussed in this episode, check out our website investinginregenerativeagriculturecom. Forward slash posts. If you liked this episode, why not share it with a friend? Or give us a rating on Apple Podcasts? That really helps. Thanks again and see you next time.
