 Hi everyone and welcome to this session for World Rain First Day on innovating tech for conservation posted by XPRIZE. My name is Kat and I'm the Program Activation Specialist for Climate and Conservation at XPRIZE. So I support communications for XPRIZE Rainforest. It's a five year competition to innovate the autonomous technologies needed for biodiversity assessment and ecosystem conservation. We'll learn a little more about the solutions being developed for that prize in this session, along with many more working and prototyping conservation tech solutions from our four incredible panelists who have firsthand experience in this field. So let's jump right in and meet those panelists. Hi everyone. Thanks for joining us on World Rain Forest Day. My name is Shaw Selby. I am a conservation technologist. I came from the world of engineering so I spent the first decade of my career working on satellites and then learned about the challenges we were facing broadly in conservation across the world, and started doing research in the area of this intersection between technology and conservation. That was around 15 years ago now when I first started doing that. I've since started a couple of non-profits. The first being Conservify and kind of the focus around Conservify was, even though it's a non-profit, can we design it as a kind of a tech development lab, something similar that you would see in places like the Silicon Valley, full of all this prototyping equipment, with one caveat being that every single thing we work on is and will be completely open source. So once we learn how to design and code these things, we put that information out for anyone to replicate. And the second thing is we only work on wildlife conservation environmental protection sort of projects. And the second nonprofit I started is called FieldKit. If you're interested, you can go to fieldkit.org and learn more about it, but that's really kind of looking at how do we change the fundamental issues that we see in environmental monitoring. And it's like the high cost of sensors, these complex and really hard to use systems, how a lot of the stuff out there is incompatible and proprietary. And so our thoughts were like, how do we redesign this from the ground up using a lot of the principles from UI and UX, user experience sort of stuff that you see with smartphones and bring that into understanding how our environment is changing, how the air quality, weather systems, things like that. And so that's kind of what we're focused on there. I'm also an explorer with the National Geographic Society and I was a fellow there for for three years leading up to the pandemic, helping them think through what does conservation technology mean and the work that that Nat Geo does as well. And this background behind me that you see here is actually one of our field sites in Cameroon, where we're working with the Congo Basin Institute to deploy a radio network throughout the rainforest that's going to allow you to bring back data from any sort of sensor or other tool that you're doing there so we've already put up some of it. And we're returning later on this year to finish it up. So that's going to be completely open. Anyone can use it, you know, local scientists, scientists from all over the world can come in and try and understand what's happening in the Congo Basin, and bring that data back live for anyone to see on the internet. So that's, that's my intro I'm going to, I think I pass it on to to Kev next to talk a little bit about his background. My name is Kevin Marriott. I'm the tech lead for X Prize Rainforest. I'm responsible for the technical operations of the competition as well as its execution. The X Prize Rainforest is a $10 million five year competition to enhance our understanding of the rainforest ecosystem. Prior to joining X Prize, I was, I worked with in wildlife conservation all throughout Africa, mainly dealing in technical capability development, so radio systems, technical surveillance systems, radio trackers, camera systems, that kind of thing. And, you know, I really kind of got my experience and forged everything I kind of put into wildlife conservation through serving in the military for 14 years and eight years of that was with special operations doing the same kind of work really with radio systems in a tactical environment and technical surveillance systems. So, you know, I kind of come from a military background, and then I transitioned quickly after leaving the military to work within wildlife conservation and you know I started studying and you know just recently finished my master's degree in wildlife biology and conservation so I kind of blend the two kind of backgrounds really with conservation and technical as well so kind of working as a technical conservationist technologist, you know all in one really. So yeah that's me and I'll pass over to Sarah. Yeah, so I'm a researcher at the intersection of computer vision, AI and biodiversity monitoring and environmental monitoring. I just finished my PhD from Caltech in computer vision about three weeks ago, and I'll be, I'll be starting up an academic lab at the intersection of computing in the environment. Probably fall 2023. And a lot of the work that I've done has been really focused around, you know, not only how do you train accurate machine learning models that are able to monitor biodiversity you know recognize species recognize individual animals, recognize environmental covariates for different types of data, things like images or audio or sonar or even, you know, GPS tracks. But also, you know, beyond just training these models how do we think about deploying what makes a machine learning model robust and usable and actually useful for conservation technology. When can things be used off the shelf. How do you figure out you know, good systems for quality control over time, because we know machine learning models tend to to struggle with things that they haven't seen before so how do you handle anomalies how do you handle rare species or endangered species and how do you make models adapt really efficiently to to new parts of the world and new species that maybe they haven't seen. Yeah, and, you know, I've been trying to actually work on not just you know training the models actually deploying them and working directly with a lot of stakeholders so I work with wildlife insights on their AI team I have for about three years trying to figure out how we scale up species identification and camera trap images. I was one of the initial builders of the Microsoft mega detector, which is now actually been incorporated within wildlife insights but you know last year we processed 100 million camera trap images from through that model from 60 different NGOs worldwide. And lately I've been doing a lot of work field work actually I got back from Kenya last week with the Mara Elephant Project and Impala Research Center looking at how do we actually bring these tools and build these systems that combine the best of AI with the best of human expertise to get really usable monitoring for biodiversity. Yeah, Eric I think that's that's you. Yeah. So, well, first of all Sarah congratulations on your PhD. Thank you. Yeah, it's one form of torture with it. Nice to have behind you. Hi everybody, I'm Eric Dinterstein. I've been a gosh a wildlife biologist and conservation biologist since about 1975 when I went over to the kingdom of Nepal and plucked out of undergrad to with the assignment of sentencing the tiger population and low in Nepal and a newly created tiger reserve which I didn't have a clue how to do and we didn't have any technologies at the time. From there, I developed a strong interest in the tropics and in graduate school ended up doing my PhD in Costa Rica and Panama, studying tropical fruit bats and working on mapping rainforests and Barocara Island the first time it was done. And then post that I joined the Smithsonian for five years and went back to Nepal to lead a research project on studying tigers and rhinos in the field using technology. And then after that spent 25 years as the chief scientist of World Wildlife Fund which took me to lots of rainforest spectacular ones from New Guinea and New Caledonia to Peru and Panama and Hawaii and everywhere. It was a great appreciation for rainforest conservation. I left WWF about eight years ago to create a new program called biodiversity and wildlife solutions at resolve a minnow of an NGO compared to some of the others that swim in the ocean, but very focused on bringing new technologies to bear. And much of what Shah said and Kevin and Sarah to find ways to make the most advanced technology really affordable and to really scale it. And so, I'll be talking about that as well and in the next hour of the innovations that we've created that from my early days of trying to detect tiger footprints in the sand with rulers and and sketch paper to state of the art AI cameras with embedded AI in the cameras that can detect tigers and send images back in 30 seconds to the end user. I was asked to start this out and I'll go back to my early days in tropical biology, which probably a number of people, you know, listening to this call are familiar with the organization for tropical studies which is probably trained more tropical biologists than any other program in based in Costa Rica now in many countries. And one of the things that they like to do to get people to become more aware of what's in the rainforest and how does it work is you go out with your field notebook and a pair of binoculars and hopefully a keen mind. You had to have too many beers the night before and your goal was to write down 20 questions, and often those questions should begin with why or they could be as how many or whatever but it's like, those questions should reflect how you try to get to a deeper understanding of the rainforest and uncover its mysteries of how it all fits together. And that was the way that biology worked for so long tropical biology is you just went out with a notebook and you tried to come up with an original idea. I remember listening to a lecture during my OTS course Dan Janssen probably the most preeminent tropical biologists in the world, and I was listening I was sitting next to my friend, and my friend whispered to me said Eric, you know, our goal is to try to come up with like one original idea in our lifetime, and like this guy comes up with one like every week. And that's what you need is that great keen mind that signed that innovation and then from that innovation, then the technology comes in and like okay how do we do this. How do we do it better. How do we do it faster. How do we do it over a wider area. What's exciting about what we're going to talk about today is we now have all these tools at our disposal that we didn't have 50 years ago of how we can monitor and understand rainforest in a much deeper way. In real time, in many places where those biologists with their field notebooks if you multiply them 500,000 times. That's what we're trying to achieve with the technology. But we're, we're entering this new frontier of understanding tropical forest, our big challenges can we conserve it before it's too late, because this is a very time sensitive problem that we're addressing. I'll stop there and hand it over to the next person. Thanks Eric I think that's a that's a really great look at how things are and you know I think this this whole intersection of technology and conservation a lot of us call it conservation technologies is just in such a such a unique space right now. It's just because the ability for us to think of an idea, come up with a solution around it, get circuit boards made or you know write some software around it, and get that stuff out in the field and testing it's just the speed at which we could do that is unlike any, any time in history right and that's, that's kind of the real exciting thing about where we sit right now. I think, traditionally what I've seen in terms of how people bring technology into the field, a lot of times. It's through a couple different ways right maybe that technology is something that's used in other areas so in the military or an agriculture and with just a few small tweaks we can use that we can bring that over to help in field science and conservation work. What I find is really exciting now is our ability to start from scratch and design these technologies to be used with like the conservationist as that core user how do we create something that is easy for them to understand that makes sense culturally and they kind of region that it's being used, it's like you're not putting a band-aid on something you're actually bringing something to the field that was designed explicitly for that and, you know that the cool thing about that approach is it also falls back on to kind of the more traditional ecology and field science sort of studies as well where a lot of those tools were created. In a university or by some small firm and they're not really optimized for that sort of thing but once like things become cheap and easy, then folks look at like, you know, one example would be an acoustic monitoring right, you know there's these big expensive acoustic monitors that a lot of folks have to listen to soundscapes in the environment. And then, you know, you have a small tool like audio moth that comes out and that really changes the way that people think about how they're doing that research because, you know, when you for the price of one of these things you can buy 150 of something else, then that changed the types of questions you can ask which is exactly what you're just talking about here. Yeah, and I like the way he mentioned about the, you know, bringing in, you know, existing solutions, especially from the military like my background, you know, that's how I got involved in conservation was, you know, I was supplying military-esque capability to anti-terrorism organizations around East Africa and mainly Somalia and, you know, the anti-poaching operations and the law enforcement on, you know, in protected areas, they needed the same capabilities. But, you know, it's almost as if there's a real need for military-style surveillance systems, but, you know, the budget isn't there. And that's always the struggle within conservation, I feel, is that, you know, we're trying to deliver a similar kind of surveillance systems, especially within the law enforcement side, but with a conservation budget. Yeah, I think actually one of the really interesting things from just from the perspective of like, what are we actually trying to ask and how do we answer these questions is, you know, I think there's a lot of potential given the amount of data that we're able to collect. So, you know, Shah talked about, okay, you can place 100 audio moths, but I think that one of the things that's then pretty challenging when it comes to this is people who get excited about the technology go put out 100 audio moths, get terabytes and terabytes of data back, and then don't know what to do with it. You know, and the actual human power that's needed to go through and process and extract the scientific insight from that data is huge. We are seeing really great advances on the AI side when it comes to helping to go through and process a lot of that data but unfortunately, at this point it's still to really be able to use machine learning effectively. To be able to, you know, get stuff that works well for your data usually it means you still need to have a data scientist or an expert in house. There are very few examples and mega detector is one of them where, for most users that model does tend to work pretty well for their camera trap data. But it comes at a cost right in order to build a model that would generalize that would work off the shelf. We had to step back from you know our original goal which was identifying species and go back to just detecting animals humans and vehicles. It means you still need humans in the loop. Now I've talked to a lot of different stakeholders using that model it means that they're able to reduce their processing time from, you know, five years to a few months they're now you know able to filter through that data but it's still this interaction. It's still an interaction between machine learning results and humans and there still is this verification and correction stuff. And actually this is something you know Eric you mentioned that you have your cameras that are detecting tigers on the edge and I think that's amazing and also it's something that you know detecting tigers is a very specific and well constrained problem. And one that actually I think state of the art machine learning is kind of ready to tackle with the right training data so you know I could see intuitively how that would be a pretty robust and pretty reliable model but particularly when you're talking about tiger populations, where there's you know it's a very at risk species and also one that's very politically charged mistakes can be significant and so I'm curious Eric like for when you have these edge based cameras where you're processing detecting tigers in real time and sending those results. How do you think about quality control and risk. When it comes to the data that's coming in. Well, let me just respond quickly to some of Kev's great points to, you know, I, and, and early on is your point, Sarah about the terabytes of data being generated is the way I look we cut through that is, and avoid what we used to call in field fishing expeditions is you start with a really powerful question like what's most important, like what's most useful. So in this case like how many tigers out there I can't tell you how many heads of wildlife departments in Asia have asked that question and that's what all the newspaper reporters want to know and that's what the donors want to know because has our intervention made a difference so it's often that the hardest number to come up with is how many are there but that's what everybody wants to know. It relates to what Kevin said his military background, because so much of what we're doing. I mean the military has all of this right. Basically and Kevin would probably test that it's just that there's lots of zeros attached to if you were, if you even had access to get it. We've got a lot for it and so that's and shots made that point is we've got to figure out how to move this vital technology into the hands of, you know, citizen scientists, academic scientists that this can be useful so we can save rainforest and wildlife before it's too late. I think where a lot of this comes down to is a very basic question is, does the data that we collect, does it need to be transmitted in real time, or it can just be collected, you know, occasionally or annually or whatever it is. Unfortunately for biodiversity, it's not like, like CO2, like a single to a single sensor on a volcano in Hawaii. Because CO2 mixes so well tells us what we need to know about CO2 levels in in the atmosphere every year. That's not going to even as rich as the biodiversity of Hawaii is that's not going to work for us for biodiversity is far too complex. So that's our problem. So this, but then we if we disaggregate what do we need to know in real time and what do we need to know once a month once a year. As to Sarah's point about moving to identifying species getting that information like tigers, it really is time sensitive we need to know where they are. So this could be killing livestock and involved in human wildlife conflict. We want to know if a corridor we've created for tigers is actually be using being used by sub adult male tigers that in their biology they have to disperse to another area. So this is the status of the population like in 2010 when we created the global tiger initiative the goal was T times two to double the tiger population in 12 years. 2010 was the year of the tiger 2022 is this year is the next the year the tiger, and I'm happy to say in about five of the 13 range states, we have doubled the number of tigers, and we've documented that by using camera trap technology. But as Sarah said, all of that data was collected it had to be someone had to go get the cameras, pull out the SD cards, the old fashioned way, and then go through them frame by frame and look, do we have a tiger and then gosh is Tiger a the same as Tiger B. Now, that's all changed because we have this, which is our new camera that trail guard AI, and in here is a computer chip from Intel, the Movidius Maria to chip that runs AI on the edge. And we can just change the models whether we want to detect tigers or other species which by changing the SD card. So this will be new to you as we figured out how to move your great mega detector program to the edge. So we can now do animals vehicles people or just animals by essentially changing the probability threshold which you can talk about to a much lower level so you're basically getting all four legged creatures that walk in front of the camera at the time, trigger and then through a communication system we can talk about later, send that to headquarters to the tiger authority in under 30 seconds, where you have cellular connectivity and where you don't we have a different system so this is the revolution we're stepping into. And everybody has a big role to play and like Sarah saying that we couldn't have done this without working with a company severe that is one of the leaders in creating AI algorithms from synthetic data. But yeah, so it's, it's a really exciting time to be a biologist and a technology person. Yeah, I think you know that's a great example is your, your, your camera trap, just because it shows the potential that we have in designing these solutions specifically for the task hand you know, because you know the way they did it before was they bought a bunch of camera traps and all that data sat on SD cards and then all those SD cards had to be manually collected right. We worked on a project in the Republic of Congo around gorilla monitoring and basically the team out there spent 100% of their time changing batteries and SD cards and camera traps that's all they did, you know, so when it came to the data data came back and just sat on hard drive somewhere you know it's really amazing to be able to create these tools very specifically to try and answer these sorts of questions and I think that's something that we could do now quite easily. We started talking about challenges so I'm really interested to hear what what other challenges, maybe the rest of you see in this space around technology one thing that that I've seen over the years is there's tends to be like this you know excitement curve that comes up around a technology and a lot of people want to implement that technology. And, and that's just, you know, you can't, you can't fault the conservationist because they need better tools to do their job right and so they get very excited about these technologies but sometimes, you know in this space we didn't put enough thought into it before we take a piece of tech out into the field. If that's actually answering the question that we want to answer, or if the data that's coming out of that text going to actually help to solve the problem that we're working to solve there. I've worked in this space for a long time that drones are a perfect example of that right drones were thought to be the silver bullet for everything. And there's plenty of field programs that went out bought a drone used it and found nothing of consequence to come out of the use of that drone. And if like, people would have stepped back and thought about this tool. What's the monitoring program look like how do we understand what the limitations are to that technology and make sure that we're, you know, not over promising and all this sort of stuff is really important to think about, about before you even take that first piece of tech out into the wild. Yeah, I think that this is huge, especially considering that we all know that, you know, conservation is not an area with unlimited resources, far from it, right. So if someone's going out and hyping a new technology, and people buy in and there's this excitement curve and then they invest a lot of resources that really are hard one and those resources are coming out of then other programs. And then the technology is a failure. It's not just like a hype bubble bursting it's actually, you know, actively and negatively impacting conservation efforts. And so I think that that's our responsibility as as technologists and as conservationists to stop over promising right and be very, very careful about the tests that we undertake. And I think that can be hard, you know, especially given like the I think one of the things I've kind of come across as I've worked with different conservation organizations is like with with the ways that funding happens in conservation, being able to talk about like wow we're using this amazing technology can bring in funding for organizations, you know, it can be it can be used to fundraise against. But then I think, you know that has consequences as well right. There's something I've been thinking about a lot and especially you know when I start coming in, building out the sustainability plans and the relationships I have with stakeholders you know when we're working on bringing technology to the field to do beta testing and being very very clear about, you know, where the responsibility to maintain the system lies and who's going to be paying for it and how we're going to pay for it and you know, if it doesn't work, because it's research and it might not work, you know, what does that mean. I think all of those things are really important and definitely can be very challenging to navigate. And when that comes down to, you know, developing technology as a nonprofit. Yeah, there's no, you know, you've got to try and you've got to try and make people realize the impact that it's going to have by showcasing the technology maybe in its in its stages, when it's not quite ready to be deployed but you have to show that it's going to work and you know people get excited by that. And yeah it's just kind of managing the expectations that this technology and the algorithms and things take a long time to develop just trying to manage those expectations is the biologists and protected area managers that really have an urgent operational requirement for it and they want it now and yeah it's difficult. And I think when that comes down to just purely the, you know, our mechanism of fundraising for people to, you know, fund the development of these things as a nonprofit. And I'll pick up on that is about 10 years ago I was, I'll keep the names out of this but I was involved in a major project, funded by a big technology company to a conservation NGO. And relating to what sauce said, you could pretty much define the project as the answer is a drone what's the question. And it was trying to shoehorn a technology that might not have been appropriate for the use case or for what the needs were, but that was what was decided upon by people who weren't the technologists or the biologists of like, it was decided at a PR level. And it broke my heart to see $5 million wasted in that way, and not going to doing things that could have funded really important things. And to what I think all the three other panelists are saying is I was at a conference about eight years ago at Google, and a number of the major conservation groups were there and technologists, and they invited people from Google X to to attend some projects and this one guy who probably didn't have a conservation bone in his body, you know, or, but it was a great engineer and just listened to everything and came up with the most astute insight of the whole three days, which he said that to me, everything you're trying to do it all boils down to to battery life and connectivity. And I took that to heart when I heard that because, yes, that's the problem that we have we have camera traps that people used to put out there that would, if they did transmit data they would send everything and exhaust their batteries in two months, or they would take the image out in the first place, because they hadn't solved the connectivity problem. And this is where I think bringing in some of the advances from and combining technologies can make a major difference. Let me give you one concrete example that plant you have behind Sarah I'm waiting for you to pick it up and throw it at me when I'm about to say this but what one of the, the most important reasons and advances of adding AI embedded into camera to our system is saving on battery life. It's not the brilliance just of, you know, being able to detect a tiger from a leopard from a sloth bear in India, but that by filtering out all of the noise. So like everybody knows who's camera trapped when you put your camera out there you're going to get 75 to 95% of the images the camera takes are going to be of no value. And a big part of mega detector was filtering all that noise out, but what if you could do that on the edge, so that you're saving all that battery life such that this single 3.7 volt battery I have in my hand here, just transmitted 2758 transmissions over the cell network on a single charge with probably another 2000 false positives. I mean you multiply that out. And if you set the cameras up right. That could be if you're doing this for anti poaching. This could be like five years of use in the field without changing batteries so back to Shows point of people going out to change batteries. Hopefully, you know that's a thing of the past and then to Sarah's point the more accurate the algorithms, the higher we can raise the probability threshold, the more stuff we can flush out right at the edge and never transmit. And then on the connectivity side by engaging with, you know, companies that are already doing this but using our advantages conservationists we can get great discounts on the critical piece of all this which is airtime costs. So, all the, all the technology we have if you couldn't transmit the data in near real time at a low cost, then it's back to the drawing board because you know you can't get the data out so fortunately, their companies like in Mars that that make this incredible. This is a satellite modem. I call it for those old enough to remember this is the host is Twinkie size we have one that's more like a Belgian waffle that's a little bit bigger, but you put this out there. So as long as you and kev has used our system is as long as you have this to the right angle to the satellite this 36,000 kilometers above the equator, and the right azimuth. Like for me I have to point mine to the Galapagos here in Washington DC, you'll have 100% satellite connectivity and you can send data very quickly to the satellite to your system to your headquarters to your research lab, you know, in under a couple of years. So, that all makes that that gets back to this question of, we've got to solve the battery life and the connectivity issue and a lot of different technologies, impinge on making those solutions possible. I really learned that point, actually, I was just going to say like around with AI saving battery life but I think this has been something. It comes back to that question I had about risk earlier this isn't something I've been thinking about a lot which is basically learning models we know that there's model drift we know that over time, you know, model accuracy on one data set, it might start to degrade. And one of the things that I've been thinking about is how do we make decisions about when and what data to send, so that we can both try to capture, you know, as many of the animals as we can, we're minimizing the number of false positives but we also have some protocols in place that are sending data periodically as just quality control, just as a way to try to understand and have this verification over time as the real world changes that our model is not, you know, deviating from reality and starting to detect fewer and fewer animals, for example. I think that that's actually a really important research question we have to solve like, how do we do real time consistent verification of AI on the edge. I just, you know, I wanted to add on that battery life and connectivity thing that the great benefit is that you know the this problem is one that the entire technological industry is trying to solve right. So, so us as conservationists are benefiting off of the development that's going into that that sort of work but it is that is the thing that that's the limiting factor on on a lot of these sorts of deployments particularly since, you know, some of us work in some very remote parts of this, the world I mean the the project that I talked about here you know we have to hike with everything on our back for, for 30 kilometers into the middle of the rainforest and there's nothing around. And we're already in a pretty remote part of the country and so this is what a lot of these products, you know the solutions end up looking like how do we create something that's going to be reliable. It doesn't last for a long time, you know a big part of it is, you know if the if it fails. How can we make sure it fails in a way that's that's more graceful so you know, don't destroy all the data on the device don't fail in a catastrophic way that's going to mess up the ecosystem even further, you know like there's a lot there's a lot that goes into the design of these services. I'm particularly excited about with field kit is like we work a lot with indigenous scientists and folks in those communities so you know conservation decades ago was really a lot of scientists from Western universities flying into high biodiversity areas collecting a lot of data and then and going back to, you know the places where they came from, and the interaction with folks on the ground just wasn't quite there in the way that I think it is, it's getting there now and and so, you know a lot of the stuff we when we're thinking about what a low, you know, how can we build a low cost sensor that can help X, Y and Z, like we're working with those communities and deploying them I'm actually going in a week and a half to the Amazon and Peru to work with a lot of our partners down there that are out there, collecting water quality data and water level data and trying to understand how they're, you know their little corner of the Amazon is changing, changing over time and I think that has potential both on the, on the hardware side, even though, you know now's very hard time to build hardware because of the supply chains, but also on the data side like how do we make sure that the data comes out as accessible to these communities, easy to understand and explore and you know there's a lot of really cool opportunities there. I could just maybe just jump ahead here just to say, working, we published a paper, two papers in the last few years and science advances, one called the global deal for nature and the other global safety net that maps out where 50% of the earth needs to be conserved to reverse biodiversity loss and stabilize the climate. And one of the most important findings of the paper was that of the 50% that we need to protect to achieve those those goals and avoid these existential threats to humanity that indigenous areas areas that are occupied or maintained or indigenous groups have title to, we're about 35% of the global safety net. In other words, the single most important thing we could do to protect life on earth is to empower and finance indigenous peoples to be able to better protect their lands. We've just been involved in our first project to protect an indigenous reserve in Brazil using our technology, and really even beyond just helping to monitor like how many jaguars are there out there or tapers. It's everything in the Amazon moves by boat so the boats are the roads of that system. And there are a lot of intruders in indigenous reserves that shouldn't be there and they're doing things that they shouldn't be doing whether it's gold mining illegally or hunting, or cutting trees or whatever the activity, and the indigenous peoples they have no early warning system of how to do that so now we have, we have river guard that can detect boats coming up rivers which you could add an acoustic element to as well to pick up the sound of the motors, so that they can have an early warning of intruders in their reserves and give them the heads up. So that's also vitally important for protecting nature in the wild and supporting indigenous people at the same time. Other indigenous First Nations groups have approached us as well in Canada, asking for similar technology so they can monitor incursions into their homelands. I mean I just like to say as well you know the challenges of working in the rainforest environment, how that plays on the integrity of the equipment and the solution to withstand being used within a rainforest. And in an area it's probably the most challenging, demanding environment, you know, on land anyway, you know, especially for communications and getting that communications link. You can choose a more difficult environment really than a rainforest. So, you know, it's one of the things that during the rainforest XPRIZE competition is one of the biggest challenges really is getting that connectivity out. And also making sure that the systems are going to last and work and I know Eric you've done a fair amount of work in making sure that your systems will withstand being used in those environments. So we get to kind of if you can elaborate a little bit on those as well that'd be great. Sure, maybe I'll let somebody else people talk to since I just jumped the line here. Yeah, yeah, I'll I mean it is it's a very challenging environment to work in some of these places just get almost unbelievable amounts of rainfall, you know it's a rainforest and so what the some of the stuff that we've really struggled with is just kind of the topography of those lands a lot of them are very hilly and so yeah communications gets quite difficult. Also, you know we like to do things solar power just because that extends the life on them and solar panels are dirt cheap. However, that doesn't work super well when you're underneath the canopy. So there's a lot of consideration that has to go into into those sort of things as well. So particularly if you're looking at things like, you know, like water quality and stuff in those regions, then you have to deal with biofouling and you know it's just the list of things that you have to you have to worry about grows grows more and more. We've had a lot of stuff eaten by insects and and so it's quite a complicated thing it's a tough problem to solve for. But once you once you get something that's working quite good it's it's it's pretty rewarding slide easier than putting it you know out in my backyard in Los Angeles or something like that right. Yeah I mean even even you know here really can help sometimes you know flat is also a problem you know if it's too flat, you've got no elevation. So there's no where you know for kind of getting a line of site radio connectivity and things like that. I mean that's a great point I mean this this region of Cameroon in this picture behind me, part of the reason why we picked it for this sort of technology that we're implementing is because they have these giant insulbergs that jet out of the rain force that make these beautiful radio tower you know effective radio towers if you put things at the top of that thing then the view you have around it is quite phenomenal right so that's you I think that's a really great point as much as you can use that environment to your advantage like that is the way that you can make these things successful. Yeah I mean rainforest, pretty much the only way you're getting communications is up yeah, you know you're trying to try to punch through all of that, you know, foliage and trees and everything, especially when it's when it's misty and you know then you just dealing with even more humidity it's almost impossible so you know with working radio communications in the rainforest, you're pretty much limited to just work trying to punch straight out straight up through the canopy, and then using all kinds of different techniques to try and use signals especially with HF and using it to basically refract off the atmosphere and back down, you know almost like like an umbrella type coverage, but we're you know we're firing out straight out through the canopy trying to work line of sight on the forest floor is almost unworkable. Which is why I think going to I think where this has to go are sat modems like this this is about $350. It all depends on how valuable is that data to get out you could have multiple sensors talk to the gateway that we built which is this little guy right here that can transmit over sell long range radio or satellite. It's the same as the transmitter and as the receiver you just change the software and job we should talk about this later on, but this is now being manufactured and to me as long as we can get cheap airtime rates and cheap sat modems. You can be anywhere in a rainforest I mean so kev like where we were working with you and in the DRC where you've got the worst situation where you don't have insulbergs but you just have hilly rainforest, how to get stuff out. I think this is really about the only way I've come up with, you know that can reliably transmit data when you need it. So one of the other things so obviously you know I think the data transmission is a universal challenge. One of the things that I think is super interesting. Also about the express is you know we've talked a lot about static sensors static cameras but I think there's a lot of power and the potential of these multimodal data collection systems you know how do we think about you know going beyond just a network of static sensors but like actually active data collection in the environment through autonomous systems I think is really exciting. And some of these other newer types of technology like EDNA like meta bar coding ways that we're trying to to collect variable types of data I don't think that images alone are going to give us the whole bad diversity picture audio alone will give us the whole bad diversity picture and so I'm really excited to see what comes out of the express in terms of these like complicated systems of static and mobile sensors collecting all these different modalities of data. I think that something that if I was to give my vision of where this is going to go in the next couple of years is building on what Sarah said is the most important innovation that I see on the horizon is when we get spectrometers in space. So Greg Asner who I consider the world's leading environmental scientists is head of a NASA mission carbon mapper that's going to send up two spectrometers that basically every 18 days are going to give us a new vegetation map of the earth that can distinguish oil palm plantations from intact rainforest things that are really hard to do from satellite right now. And basically I almost see this as global ecology before spectronomics and global ecology afterwards because the data that we're going to have to ask these questions are going to be phenomenal will have a map of the world's canopy every 18 days. Our challenge everybody on this panel is like how do we figure out if that diversity is still there under the canopy that the spectrometers can't see but if we can put the two of those together. That's the vision for how we monitor life on earth going forward. Yeah, yeah I mean that Greg Asner's work is going to be a game changer once it's once it's put into orbit it's going to be make a huge difference. One of the other things is your kind of collaboration on on producing these solutions you know we all work for nonprofits. And that makes it a little bit easier for your collaborating between ourselves and you know just trying to help everybody out and I think that there's been a great shift. You know over the last kind of five years I've noticed where people are tending to work together a lot more and not be as protective over the solutions and help each other out because we're all trying to achieve the same thing which is you know conserve what's left on you know in these protected areas. So yeah I think it's you know collaborations also pretty strong within the conservation technology industry I suppose. Yeah I mean that you know I'll echo some of the stuff that was already spoken about I think that you know this Rainforest X prize is going to be really a game changer and pushing some of these technologies forward. And then you know like the stuff we're doing the field kid or trail guard or you know a lot of the thought around making these sort of systems smaller cheaper more accessible then you can get that ground true thing that can go with like the space based stuff and we can get a really you know interesting understanding of how these places are changing over time. You know conservation in general like we're not doing so good globally. We have small successes here and there but I think you know with climate change we see the the general trends and the faster we can get these better tools out to people I think the the better we can understand how those changes are happening and potentially try and step in and make some mitigations or or you know help to reduce some of the pressures on some of these ecosystem and and I. You know I think that's the key thing right is like how do we make sure that the work that we're doing contributes to this larger goal and this this final outcome of making sure that the rainforest are there for our children our children's children and everything you know into the future so. Well we're all out of time but I would like to thank our panelists so much for joining and teaching us so much about these new tech solutions that are coming up to help conserve biodiversity and rainforest ecosystems I feel like I learned a lot and hope you did too. If you would like to learn more about XPRIZE and the XPRIZE rainforest, please follow us on social media at XPRIZE and you can find us at XPRIZE.org. Thank you so much for joining and happy World Rainforest Day.