AI for Climate Global Forum - New technologies for ecosystem restoration
![AI for Climate Global Forum - New technologies for ecosystem restoration AI for Climate Global Forum - New technologies for ecosystem restoration](/pic/ai_for_climate_global_forum_-_new_technologies_for_ecosystem_restoration/QmJ0NjZiRjUwaW8_.jpeg)
welcome everyone to the panel new technologies for ecosystem restoration this panel is part of the ai4 biodiversity track of the 2021 a.a for climate global forum my name is lucia trojes i am the global forum director and it's a pleasure to be here with you in this incredible conversation i would like to welcome our special guests for today clara rowe katie raider and jorge ricards a group of experts that will discuss the role of ai systems and new technologies in accelerating restoration efforts worldwide i would like to tell you more about them rowe is the ceo of restore a science-based open data platform to support and connect the global restoration movement she is a sustainability leader with a decade of experience in natural resource management international development and responsible commodity supply chains clara has expertise in bridging the gap between global sustainable solutions and underground challenges to ensure interventions and impactful and strategic katie radar is a senior research associate at the world resources institute supporting the world the forest program's global restoration initiative she leads wris technical work on forest and landscape restoration monitoring at the global scale she specializes in gis analysis remote sensing and spatial data management and last but not least our incredible interviewer for today's session jorge ricard is the general director and representative in mexico for the world wildlife fund where he also served as ceo and director of conservation since 2008. he has received several awards including the medal on honorary merit recognition awarded by the department of state of the united states of america the recognition of excellence for his server to usaid and for service to the mexican fund for the conservation of nature clara katie and jorge thank you for joining us now i would like to pass the microphone to jorge ricardo who will lead this conversation jorge please go ahead thank you very much good day good evening to everybody it's a pleasure to be here uh it's a real pleasure to share this panel with uh such level of experts with claro with katie with lucia and we're very happy from uh the ai for climate platform and group and in general to have the time to have this discussion about restoration and technology artificial intelligence there are so many things happening in the world as we as we begin the decade for restoration that the ua in the united nations uh established or declared there are so many challenges and we have recognized that it's not only about maintaining what we still have in the world in terms of biodiversity and ecosystems but it's about bringing back about restoring so many areas that we have degraded and it's a moment in which the voices of young people the voices of local communities and the voices of women and other genders are being heard it's an amazing moment of change and opportunity but it's also a tremendous moment of challenge we have a decade a few years to set the basis for the next number of years and as the conference of the parties for climate change in glasgow happens we all have the expectation of what our leaders are going to agree on but we also have the opportunity and the responsibility to act from our own from our own places from our own sites organizations to contribute so it's very very uh i'm very happy to be here i i welcome you and uh i would uh put the first question on the table and i would ask uh clara if you would answer this but of course uh please katie feel free to contribute and and uh as an open dialogue and i think that the first question would be uh as a start is what is the potential of these tools that already exist in terms of technology artificial intelligence like for example gis to assess the needs and potential for restoration in the world as as as uh with its open applicability in different ecosystems in different sites what what are we talking about here what what does it bring to the table that can add to accelerate the solutions that we're looking for big question jorge thank you it's nice to be here with all of you and i'll put some ideas out there and and katie do jump in as well because i know that wri is huge in the space um you know so ai can play several important roles in assessing the potential of an area for restoration one is just around understanding land clever class land cover classification over time right so ai is now performing better than traditional remote sensing approaches a new map came out in 2020 sorry 2021 esri land cover map land cover maps from wris land and carbon watch are coming out and it's this really exciting time for being able to at scale determine what's happening now and therefore what could happen um there's also the possibility to determine you know potentially suitable plants for restoration trees shrubs herbs both today and under future climate scenarios using ai and so you know species distribution models are working to make this easier and more scalable across the globe and and another category of things is prioritizing areas for restoration right so we can think about criteria that we might use potential carbon sequestration and soil or in biomass um and there are lots of scientific um you know data sets out there and restore where i work houses many of those that help us make decision based on potential so those are just a few of the ways that we can think about ai contributing in this space and i think it's really important um as we dive into this discussion to say okay land cover carbon potential you know species potential are certainly not enough it's a part of it but not enough in order to make responsible restoration decisions so it's really key that we think about what the original ecosystem type was are we planting trees where there shouldn't be trees um and we need to ask questions about land tenure and local community needs and their priorities so it's something where ai is a piece and science is a piece but there is always i think the undercurrent in this conversation has to be and what else and who else right right absolutely uh katie would you like to to add sure yeah i think clara covered it very well um yeah i would just i would just add there that um yeah like like clara said there's so much data available now that is that is you know has the potential to help us understand ecosystems better what's happening on the ground the change that's happening and ai is really helping us to get so much more finer scale with those estimates as well i mean just a few years ago in one of the first projects i worked on at wri we were mapping restoration opportunity at a one kilometer scale which seemed amazing at the time but now we're down to 30 meters and and actually and even transitioning out to 10 meters um using sentinel satellites and and my colleagues at wi just released last week actually a new data set called trees and mosaic landscapes which is looking at a 10 meter scale trees outside of forests so the trees that were traditionally invisible on most land cover maps the ones that are in croplands and pasture that are so critical for um you know for restoration stabilizing soils improving um yeah soil fertility and those sorts of things but they were so hard to to find on on that macro scale until more recently and so we're pretty excited about that data set and think it'll be kind of a game changer and us estimating you know where where the trees currently are and where there could be more and doing these kind of finer scale assessments which can then help inform i guess clara was saying more on the ground and getting getting you know you know understanding what's there but then also working with people more locally to um to understand what really the potential is on the ground making sure it's appropriate given the ecosystem and the community thanks i i agree and and i have found that uh one of the challenges that we that we face uh is having the technology having it advance at at a very quick rate because we're talking about a few years in which as you say we had some level of resolution and then suddenly we have much higher resolution uh we also now have other other tools no that are more popular uh that are more accessible for people but it in my my experience it has also been a challenge to put this these technologies uh accessible for local communities so what what what what would you suggest to to sort of like bridge that gap between uh scientific organizations or uh ngos or uh other sort of institutions and local communities that many times still unfortunately do not have the capacities within their own you know people uh to adopt these tools and make the decisions um what's your opinion about that considering that in many places of the world uh local communities are the owners of some of the best conserved areas or some of the areas that urgently need restoration so the question i think there's this larger question about consultation and involvement and um without diving into that large framework i'll try to speak more specifically to your question about technology and not link from a technology and scientific community and then local community perspective and i i can't say that restore is all of the solution at restore we hope to start to be some of the solution because we're making it really easy to access data although right now it's mostly easy on the computer to be honest you could open us up on a phone but um it's not optimized and we really need to move into the app space and i think that's true across the board anything that can be used on a mobile phone is something that is scalable and is something that will reach communities in rural places and it's not perfect i've worked in many many places where there's no cell service and where people don't have smartphones yet but it's a start and it's spreading um i think there's really neat applications around you know supporting indigenous communities to be able to detect deforestation through bioacoustics around chainsaws through the bio acoustics um of you know gunshots trying to prevent illegal poaching and so i think those are some nice examples where the the tech world has been able to create a really practical tool that enables reporting at a local level and ultimately puts agency in the hands of people who um both you know are responsible for management of a place and also have the most stake in the success of that so i think there's a lot of exciting things there but it is it's it's about accessibility all right absolutely i absolutely agree with you katie would you like to to add to to this point yeah sure yeah i'd absolutely agree with that as well and um yeah i know one experience we've had with the global forest watch platform which has been around for seven or eight years now um you know it has these weekly deforestation alerts on it which have been really critical and interesting for people to see um online you know you can choose an area of interest and then actually get subscriptions and emails saying when deforestation is happening um the recent studies that the team has done and with partners have shown that just having it online and on the websites not enough really having a mobile device where you can get that information as well and having it offline as well has been really critical so with the the forest watcher app which is takes those alerts from global force watch and kind of makes it into a more field-based tool they recently did a case study in the peruvian amazon where they found that deforestation actually when they when they're able to work with local communities and indigenous peoples in the peruvian amazon um and train them on how to use the device and to get alerts and to work with the community on and and leaving it you know to them on on how that could help them improve their own monitoring that they're already doing they found that in the first year the deforestation was reduced by over 50 percent um so having access to the technology is really critical um and so yeah and so just you know having having that that available and accessible is really important so that it can actually make change on the ground yes and i i i fully agree with with uh one thing that has worked for for for us and has been working with local communities is uh you know gathering the information uh and then presenting it you know as as it is and then going to the field to verify whatever the imagery is telling us because in uh at least with the communities or some of the communities that we have worked with when they see the imagery uh sometimes they're rather shocked with what they're seeing sometimes positively because they they are seeing more forests or than they imagined in their territories but sometimes negatively because they are surprised that they are having some sort of degradation so so um field verification uh still seems to be an important part of the of the process um but of course uh uh starting from a platform you know like restore that that exists that that sets the basis now following this this line uh what kind of data uh can these technologies provide for decision makers communities governments etc to be able to detect areas that have the potential for restoration what what what is the information that can now be available for decision makers that want to do restoration oh katie would you like to start with that one sure um yeah i mean there's there's so many options here i feel like in kind of the grand scheme of things there's a lot of data available now um i mean more than ever before and at a finer scale than ever before as we were saying earlier um but it now it's it's up to everyone what what do you do with it and how do you get into the people who need it um i would say you know in the in the realm like the or one of the gaps i guess in the realm of available data now is um is in just investing in long-term monitoring um and also documenting um not just changes in land cover and tree cover but but also you know documenting changes and improvements or or even failures um you know and and what you can do differently in adaptive management things and in the social aspects of um of restoration and then those impacts um so yeah there's there's you know a lot that's being done on the biophysical side but i still i think there's a lot more that can be done on the social side to make sure that you know the restoration that's happening is having the improvements that um that you know are anticipated or you're striving for um but yeah i mean i think the first the first goal of course for us is getting that that you know kind of biophysical mapping done that that's been kind of our goal for a while to get that um kind of ecosystem change and the land cover change mapping done well and then moving on from there and doing more monitoring of kind of the impacts of that and and supporting and also getting more local with our our um work and trying to work more with local communities to to really see the impact and not just like doing it from you know from your behind your desk but also working more with local partners to get there um yeah to to get you know them involved and um yeah supporting their needs yeah so absolutely absolutely tara would like to comment yeah i'm really glad katie that you brought us into the monitoring space a bit because i think that there is the how do you decide where to restore and then there is the how do you tell whether restoration is working and there are many different goals of restoration of course and there's you know a continuum from you know restoration in a more agricultural context all the way towards conservation of existing ecosystems and then everything in between but you know katie mentioned how there's been a lot of focus on being able to map the biophysical right can we see are there trees or not trees is there vegetation cover or not vegetation cover but then from a social perspective and also from a biodiversity perspective i think that there's a lot of room to grow in terms of the ability for scalable monitoring and ai has a key role to play there right there's lots of focus on measuring carbon right now above ground biomass carbon so i'm a little bit less worried um about investing in that innovative space but i do think that the people and the biodiversity is a place where we can really use solutions um so we talked about bio acoustics i think that's an incredibly exciting space that ai will speed along in terms of being able to process bioacoustic data and evaluate species presence or be able to estimate biodiversity you know in a restoration site compared to an intact landscape site our partners at the the crowther lab are working on some really exciting research looking taking maxar imagery and basically coming up with a spectral diversity measure which can be a proxy for biodiversity and so trying to use really high resolution satellite imagery to start getting a sense of biodiversity and that way you know we would love to on restore for example being be able to give a little flag of is this a monoculture or not a monoculture and that being an important part of that measurement space katie you mentioned the the mosaic landscape tree mapping and i think that's such a cool example of this um tree crown delineation is kind of a holy grail that many people are going after now and the ability to be able to monitor an nai is a key piece of that um i i do not yet know of ai as a way of helping us measure social indicators so maybe we ignore that for the time being um but it well i should i shouldn't say ignore ignore that from an ai perspective um but in the biodiversity world i think there's a lot that can really happen and should happen um in terms of being able to monitor at scale and and what about predictability can we talk we've talked about about the monitoring and you're absolutely right it's key to see i mean is this really working are we making a difference um what about predictability can technologies allow us to predict to the point in which we can for example convince uh investors to uh uh channel resources to a restoration initiative yeah with enough certainty that that that will there will be success is is do we have the tools to do that it's all about i mean katie jump in here with me but it's certainty how much certainty do you want is a big question and then the scale at which you're operating is another big question if you're operating at large scales we can reduce certainty uncertainty quite a bit in terms of our our ability to predict what is the carbon potential here what is the biodiversity potential here if we're working at much smaller scales you know katie's mentioning the one kilometer to 30 kilometer to 10 sorry to 30 meter to 10 meter progression in terms of remote sensing um those are still important numbers to take into account as we're thinking about the ability to predict at scale the impact of something and again i still think that most investors are paying attention to carbon which of course is important but is not the only piece of the picture and so again as we're thinking about those two different things those are two different challenges that we have to get to in terms of a predictability um perspective what do you think uh katie about about this yeah i yeah i think clara summarized it well um i would say predictability is really helpful in agenda setting um like you know estimating how much of the world's land has the potential for restoration i mean that was really critical for some of the studies that have been done through the crowder lab and also wri um had had done these in the past on you know how much land could be restored or could you know have trees planted and then that was helpful for estimating you know when countries were making pledges to the bond challenge or the new york declaration force and you know making these commitments that was really helpful um and that was a little bit of a predictive or opportunity setting to help you know make make these goals um and and make them that were actually achievable um but but yeah i i would echo what clara said in terms of yeah it's kind of like how how much error there's always some error there's always many variables that you can't quite predict especially when it comes to restoration there's there's you know that social element that's really hard to predict on whether something will be a success or not um and so yeah i would i would say ai and and gis and the predictability side is you know really good for getting that big picture and you know supporting the agenda but um yeah i would say with monitoring and on the on the ground work yeah there's there's a lot of variables there you have to consider and um ai can support that yeah with like with the mapping or even machine learning say with looking at policies and comparing policies of different countries um but but yeah the um yeah it's it's kind of a complicated thing sometimes when you get closer to the ground so um yeah ai definitely has its place but um yeah that's it's uh it can get a little bit more difficult as you get closer to the ground with that but it's definitely been been critical for the agenda setting part yeah and especially when we move from the concepts of of reforestation well the affordable station that that was so popular in past decades to to the real concept of what is restoration what are we really wanting to achieve what what is the definition and then how does that definition apply to different conditions biological conditions social conditions because you know maybe what we believe is the ideal habitat might not be for the local needs of the local people so there needs to be this i guess this definition of what restoration means that can be or needs to be flexible uh and adaptable to each one of the different areas that one is going to working but but uh following on on what we were talking about uh finance and investment uh what what would your uh idea be of the potential of finance for restoration particularly based on advancing technologies for restoration advancing artificial intelligence etc uh what is the potential for finance or uh to further advance this this this approach uh uh clara would you like to comment on that sure um so i i'm going to say something a little risky which is that we sort of know how to deploy finance to large-scale restoration efforts i really caveat that sort of because there are a lot of things that we still need to figure out um but given that i i do think that the big challenge we face is deploying to small-scale efforts um around the globe which ultimately add up to this huge impact and i would say that there are twin challenges here right there's assessing restoration efforts at a relevant scale which we were just talking about how do you do that at a few hectares or maybe even smaller at a time and then how do you deploy funding to remote and disconnected areas you know we've started to have conversations with groups in the brazilian amazon who want to be able to compensate farmers who are at the edge of the amazon rainforest for restoration work that they're doing but they're areas without electricity you know let alone bank accounts and so you you have to think about solar panels to charge phones to be able to accept mobile money and those are the challenges that we're talking about again when we're talking about this small-scale dispersed challenge of being able to compensate for restoration so that's an area that i think is incredibly exciting um and again whether it's mobile money whether you're thinking about cryptocurrency i'm a bit agnostic there but it is the you know those twin challenges of measuring small-scale progress and deploying money to those rural places katie what do you think yeah i mean i i think something promising that that i've noticed recently um is that a lot of technology companies have become more active in the environmental space um i mean between google i know that's involved in restore and um like amazon and and facebook who are supporting um non-profits to work in on restoration and deforestation prevention it's been really heartening to see that kind of commitment and support on the technology side um so that's yeah that's been really really good and really valuable to have that kind of um um support but yeah as clara was saying the kind of the the real where the real finance um is needed um is at the local scale and actually making restoration happen at that local scale and so i think there's yeah there's there's a lot of creative ways i think to make that happen that that need to be explored more um and and yeah i mean one just to give one example um through the afr 100 initiative um just last week or a few weeks ago we were accepting um proposals for small-scale restoration through organizations that are local and so it's kind of like a small grants fund to get local restoration practitioners either non-profit or for-profit entrepreneurs involved um so so kind of taking that macro scale financing and getting it um to into the hands of local people who are you know doing the restoration in their own communities as i think is really critical for actually making the change on the ground and um kind of getting that connection between the top level of the the big funding and then the small implementation you know yes and and i was thinking as as i hear you i fully agree with i was thinking about for example where where the finance should be should be directed to and um in terms of the priorities like uh highly endangered ecosystems but also ecosystems with high carbon uh sequestration potential like mangroves perhaps for example that also have the adaptation uh angle or importance to them uh and that happened to be are beginning to be in my opinion very attractive to to finance mechanisms and donors and multilateral funders because adaptation has is becoming very visible as unfortunately we are dealing with so many natural disasters around the world so so being able to use technologies to inform those plans and set priorities in terms of adaptation particularly for local communities that are more vulnerable i i think is is is a priority and hopefully more financing can be channeled uh to those uh more vulnerable countries populations uh you know all over all over the world and following that in a way when we talk about countries adaptation uh challenges to climate change and how technologies uh can can be useful what's your opinion about how um artificial intelligence and technologies can help inform policymakers or build policies uh policy tools um how how are these being used or are they being used how can we better use these technologies to build uh uh a local policies national policies for example in adaptation mitigation what's your opinion about the the policy angle uh clara would you like to start i always get to start poor katie it's all right i am kicking off so well you're making it so easy sorry for that you really want to start creating no i'm i'm happy to i have i have a few ideas to add here and and i'm sure katie will fill in much more detail um i i think you know echoing some of the things we've already talked about so many options from a policy perspective in terms of using you know ai derived data in order to make decisions about restoration i think that um you know we can think specifically about how to inform payment for ecosystem services schemes in terms of assessing suitability of applicants in terms of monitoring compliance over time we can think about enforcing no deforestation commitments right both at the government level and also at the corporate level so we haven't talked at all really about sustainab about supply chains and about agricultural supply chains and the impact that they have um both in terms of you know habitat loss and and we know the figures around agriculture and habitat loss but also the potential for supply chains agricultural supply chains as um carbon sinks right what does it mean to transition from traditional monoculture to an agroforestry scheme so on both that again government and corporate level i think there's a lot that data can do to help enforce and inform commitments in the no deforestation and in the restoration space and then you you hinted at this jorge but you know protected areas will need to expand and or shift to the ecological realities of a changing planet and so as we think about where it is that species are going to move and how quickly are they going to be able to move and what it means to be able to not react to that but predict that and respond to that i think is something that um is incredibly important for technological insights to be able to contribute to at a policy level no i i agree we're going to be uh closing our our panel uh very soon but i would like to uh katie uh ask you one final question more related to the ethical part the ethical angle of of the use of information especially uh when as we have talked already uh earlier when we are in this moment in which the world is recognizing more clearly the rights of local communities of of uh original peoples um and when uh many of these these communities have not have access to information about their own about their own lands and that is changing so so what what is your your thought of in terms of from that ethical angle to use the information to advance the information in a way in which we are meeting those ethical requirements uh uh that that are now so important and so clear in the way forward yeah that's a good question and a really important one um yeah i mean there's kind of a trade-off between having open data and and being very um you know transparent on on you know and providing that kind of fine scale information that can help influence change but then there's on the flip side there's you know of course there's gonna be privacy issues and you know people not necessarily wanting certain types of data being shared um for instance yeah like in some work i've been doing on on landmark which is a global platform of indigenous and community lands it's it's kind of a global map of where indigenous and community lands are located in actual boundaries and so um it's been kind of a yeah you know whether that information wants to be shared is by the communities themselves is really important um you know on on one hand it can help show you know investors and you know companies who are looking to purchase land like they would have this information to know that that land's not available at the same time you know if there's local controversy or something about the boundaries you know that could be uh detrimental to the community themselves and so they might not want that information shared and so there's always this trade-off yeah on what's what information you know should be out there and i guess what the benefit first the drawback is and that really gets down to the local level on what you know what that accepts acceptable level of risk is for you know a certain community in that instance um yeah so there's right so as you know satellites and and this all the mapping we do gets more advanced yeah there is this kind of yeah the issue of you know how much information is too much and what should be shared um so that's i don't have a great answer here but i just say it's kind of a case-by-case basis um you know on and doing that kind of risk assessment on on what's too much and or you know or should be shared and also getting you know especially you know if if a certain community is is going to be effective getting there getting their buying or additional thoughts on that what do you think clara it's a really hard question and and katie i think you've laid out a lot of the different considerations um it feels a little bit too easy to say that it comes down to local choice because it's not that simple we're taking pictures all the time you know we as in whoever is managing those satellites out there and so imagery is there and that's a really hard thing to take away um now that it is there but i do think that platforms like restore like global forest watch who are sharing that data and then sharing it in a way that has context and that allows people to do something with it have the responsibility of giving local communities choice so i don't think that we know how to do this perfectly at restore yet but one of the places that we have come to is when communities want to share information about the work that they're doing or local projects want to share information about the work that they're doing they choose how much and they choose is it public or is it private is this something that i just use from a management perspective or is this something that i think is worth sharing more broadly and i hope over time we're going to get better at making that even more adaptable um so that you know literally piece of data by piece of data it is up to the person sharing that um and again i i don't think that it's a simple solution to say choice because there are many many different pieces to that but i do think it has to be at the heart of it is respect and choice and we need to grapple all together you know about what that actually means in practice thank you i i agree and and it's also about um who uses that information i think that strengthening local governance is always key because information can also be used for bad purposes not for illegal trade of species to detect habitats where there is potential for illegal logging so so um the ownership of the information in terms of the communities in terms of the institutions is so important and and uh thank you this ethical part is is i think was very important to touch on because it's as you said it has many angles to it but it's certainly very very central so thank you thank you so much uh uh i'm going to be uh wrapping up very very quickly but this has been a fantastic a fantastic conversation i think we have very clearly established there is enormous potential that there is enormous potential to use technologies artificial intelligence for restoration we are at that point in which we have reached a level of detail in which we can even talk about species we can talk about for example plant species that can be used in different habitats uh we have also talked about how to use that information in relation to the local communities it's it and who owns that information who manages that information to make the right choices we also have i think established that there is potential for finance to advance technologies and and also the need to set priorities in terms of where do we use these technologies in the context of climate change adaptation um um in a in a ever-changing ever-changing and complex complex world but certainly it's very clear that that all this information is allowing us to make better better informed choices from a governmentals uh perspective to a more local perspective to a very micro local perspective it certainly is something that is allowing us to make to make a big change we have a much finer scale now as well compared to a few years back and now we have these public tools not like restore like earthwatch that are allowing us to very easily see with very high definition what is happening in the field and i think you also mentioned something very important all the information that is up there is uh in accordance with the landowners so the information that is there is because the landowners want to share that so so hopefully more people can use these tools and begin to be accustomed to using them to make better choices and then we can talk about connectivity for example uh where do we need to invest where do we attract funding to restore under a logic of not only isolated plots but connecting you mentioned protected areas for example and and the shift of protected areas in the light of climate change where do we establish the biological corridors that we will need to provide services uh to people agriculture food etc so so it is very clearly something that still has enormous potential but already can provide us with extremely valuable tools for for advancing advancing restoration globally so thank you thank you so much clara thank you so much katie this has been a fantastic fantastic conversation and and i hope to continue the conversation further lucia so much for having us yeah thank you thanks so much for facilitating thank you so much jorge katie and clara for those uh for that conversation uh we feel so honored to have been part of it and i i live with a very specific thought regarding the accessibility part um i think that that that content right on how we can make that um all those efforts that are regarding technology more accessible to communities and indigenous uh communities is gonna definitely add up i think one of the things that that really i really liked of the conversation is that all the small efforts are gonna be adding up to to that big result right um so again thank you so much for being part of this global forum it was really great to learn from your experience and to those watching join us at the next presentation of the ai for climate global forum [Applause]
2021-12-03 21:34