How satellites are supporting farmers across Africa | Catherine Nakalembe

0:07You're listening to TED Talks Daily, where we bring you new ideas to spark your curiosity every day. I'm your host, Elise Hu. Today's talk is part of our new 2025 TED Fellows Films, adapted for podcasts just for our TED Talks daily listeners. We'll be releasing these special episodes showcasing our amazing fellows on certain Fridays throughout the rest of 2025 and into the new year. So stay tuned. The TED Fellows program supports a network of global innovators, and we're so excited to share their work with you. Today, we'd like you to meet the satellite food security specialist, Catherine Nakalembe.

0:41It might be hard to imagine how satellites orbiting in space can help farmers tending to the soil on Earth. But for Catherine, who uses satellite technology and machine learning to monitor smallholder farming, this connection is essential. She shares how satellites can play a vital role in helping to feed the world, and why, even with her bird's eye view, a grounded, bottom-up approach to supporting farmers is just as key. As she puts it, true innovation isn't about high-tech systems, but about making technology fit the problem. After we hear from Catherine, stick around for her conversation with TED Fellows Program Director, Lily James Olds.

1:25If farming was your primary source of income, and if you can't grow anything, there's not much you can do. It's very demoralizing. It's very overwhelming. A lot of the countries where I work, farmers face fires, pests, diseases, droughts, and floods. If crops fail, no food is available for a lot of people who depend on what they produce. It translates into, you know, sometimes an entire generation being undermined. My name is Catherine Nakalembe, and I'm a satellite food security specialist. I use satellite data to map and monitor crops and then work to make sure that information can be used by decision makers and organizations that support farmers.

2:15I primarily focus on Africa, Uganda, Kenya, Tanzania, Zambia, Mali, Senegal. Last time I checked, there were over 8 ,000 satellites observing Earth. They take pictures every day. You can use those images to map what crops are growing where, how much it's going to rain, where the system might be coming from, where it might be impacted and how badly it might be impacted. I can just sit on my computer and tell you anywhere in the world, rainfall, drought, floods, name it. I can tell you where it is. We have tons of information, tons of data about what's happening. We're living in a fantastic age.

2:57We have huge advances with AI. We're able to process a lot more data and information. The problem is when you look deeply at any place, you start to see problems with a lot of the existing products. They're not tailored to the underground contexts. Sometimes it's simply that the data is just wrong. A lot of the models are trained very well to predict for European or U.S. agriculture. In Europe, most of the farms have like a single crop. They're really big, and it's very easy to model. In Kenya, in Uganda, in Rwanda, however, the fields are so tiny. They have so many different crops in them, and farmers do things so differently.

3:39It's like a tapestry. With those images, fields are misrepresented, so there are places where there are no crops but are labeled as crops. There are places where there should be crops and people that are completely missing. When you go and try to assess something for the ministry, and you use that as input. They're basically feeding them garbage. You want to make sure that what you're feeding in is really good and that actually requires work. To train a model to understand that complexity, you need a lot of examples. You have to go on the ground. What we did is we used GoPros. You wear a GoPro as you're driving on a motorcycle or you can do it in the car.

4:20And as you drive, we take pictures. Basically, Google Street View, not for streets, but for crops. So the camera is actually facing towards the field and then adapt basically what would be face detection. But instead of detecting faces, cats, dogs, et cetera, we modified it to detect maize, beans, cassava. We covered all of Western Kenya in two weeks with just two teams, collected over 5 million images, a lot of them with volunteers, everyday motor taxi drivers, students. This would allow us to build a more complex model that can learn from all these different examples from all the different contexts.

5:00There was a flood in Kenya in 2024. It happened really, really rapidly and you know the entire country was affected pretty much. I got an email from the Ministry of Agriculture asking to do an assessment using satellite data to look at where floods happened, where were crops and give an estimate of what the total area of problem that's been affected was. And then what the ministry does with information is they make their response programs, which is where do we need to go to provide seeds so people can replant as an example of an action that was taken. It's really powerful to be able to do that.

5:38True innovation is not about high-tech systems, but about making the technology fit the problem. We have to provide really good information to the people who can do something with it. If we do this correctly, we can save time, we can save money, we can save livelihoods.

6:01And now a special conversation between TED Fellow Catherine Nakalembe and TED Fellows Program Director Lily James Olds. That's coming up right after a quick break.

6:17Hi, Catherine. It's so good to have you here in the TED offices, a rare treat. I also just want to say to our listeners that if you hear some people or sounds, it's because we are surrounded by our colleagues here. I'm so grateful to be here. I'm really excited to be in the TED office, but also for the opportunity to sit down and chat with you during Climate Week. So many different things to get into with you, but I would love to just start by knowing what first sparked your passion in using satellite data to tackle food insecurity? I used to play badminton. I was going to do sports science. That's like literally I was called Catherine Badminton when I was in high school.

7:00That was it. But by chance, I ended up getting a scholarship to do an environmental science undergraduate. And at the end, I was able to do this, you can call it like a capstone, a research project that for the first time I left my parents house and went to the west of the country. And I did field mapping with a GPS, a backpack, and a forest ranger. And I remember I have this photo, like in this, it rained completely. I was completely covered in dirt, but I was so happy. And I thought that I wanted to do more of that. So I applied to do a master's, and I ended up going to the Johns Hopkins University where I did geography and environment engineering.

7:46It's a long story, but I didn't get the opportunity to sort of do what I'd hoped that I would do, getting a degree and then being able to apply it to a problem back home. Then I discovered the department at the University of Maryland, and I had a chat with my PhD advisor, and he was like, how would you like to go do fieldwork in Uganda? I was like, what? So he worked on agriculture. That was his main domain. I never actually considered like crop science, et cetera. I was always like forests because of my previous experience. But then on, I was like, all right, this is it. Fast forward, I had the opportunity to not only do field work back home.

8:25It got me to be able to learn how things actually work because I spent a lot of time in the field. Yeah, I love that. You talk a lot in the conversations that we've had about the disconnect between the massive amounts of data and knowledge that we have and the, as you call it, failure of translation, you know, where the data doesn't actually reach farmers in real time or in a useful format. From your perspective, what's the biggest barrier standing in the way of turning the data you have into action? Data is the new oil. You know, there's a lot of investment in data infrastructure. a lot of investment in methods, a lot of publications.

9:05It seems really exciting. It seems like, you know, we're breaking a lot of boundaries, a lot of barriers. But in reality, if you were to visit my sister who has a farm and is trying to grow maize, none of what I do has anything to do with what she has to do. I could do my very best analysis for her, but it's disconnected from the resources that she has access to. So it does not tell her where the fertilizer is, where the seeds are, when to really plant in reality. So that's one part of it. Sometimes we think about what satellite data and AI, et cetera, can do in agriculture. I think a lot of people think about it the same way as how ChatGPT enables you to write an email.

9:52You can immediately check that the text is wrong. You can correct it, right? but it's gotten information from a lot of text that's been digitized. Agriculture has not been digitized. There's not a lot of text from which my model will learn from to give me the right sentence. So in the context for my sister who has some coffee here, some maize here, I think there was cacao too. Her field is not represented in existing data. So I could try to map it really well. I could get all the information directly on the ground. But I'd still do a pretty bad job because the model has learned about other things that have nothing to do with our reality.

10:33I'm curious, is it about where the technology is in development, which we'll get into later, obviously, and you talk a lot about, or is it the way that that information is then delivered to the farmer or the person on the ground that can make use of it? So there's one side of it. One side of it is in order to get useful information from satellite data, we build models. You have to train the model using some existing example. And then the model goes through all the satellite data and tries to find that thing that you were looking for. And that's how we would create like a crop type map. Today, most models do very well for farmland in the United States.

11:15The fields are really big. They're homogenous. So there's like one crop over a really large area. There's been a lot of investment in collecting data to have those examples. So when I build a model and train to predict what is growing in Kansas, it's easy. But for me to produce a map of where maize is in Kenya, a whole different story because I don't have those examples. It's much harder to collect the data because there's no default data collection for this purpose. The other side of it is that the satellites that we have access to, where we have open data that allows you to scale for all of Kenya, the resolution doesn't fit the small Kenyan fields.

12:01And so if I use the same amount of labels but trained on the same data set for Kenya and the U.S., it wouldn't work because this field is so much smaller and much more complex. So the products we have are usually not relevant at the farmer's scale. They're relevant at a larger scale, like I can tell. Yes, there is a drought in western Kenya. It has affected approximately 20 ,000 plus or minus 5 ,000 acres. A week ago, there was a huge hailstorm in Kenya, and there was hail everywhere, literally. So we can actually see, we can show the damage, we can give an approximation. However, it is not the same as me, a farmer, Jane, and my field in Hoima, for example.

12:53This is in Uganda. What exactly happened to me? Was my field completely destroyed? Maybe. Was it not? There's like that kind of like mismatch. But then the other thing is, even with what we have, there's so much that we can do. So knowing a drought is coming is so powerful because you would know whether, well, is it worth me planting this season? I don't have an irrigation system, so maybe I shouldn't do anything. It's going to be terrible anyways. So you don't waste your energy. You don't buy or pay for labor. You don't buy seeds. You don't put them in the ground for nothing. So that's one side of it.

13:28But it could also tell you that maybe in this area generally there's been droughts continuously. We should really invest in some irrigation infrastructure. And so being able to come from what we measure and predict and learn over time and connecting it to we should have irrigation infrastructure here and that irrigation infrastructure arrives, whole different story. Yeah, I mean, it's interesting. It almost feels like part of what you're describing is that last mile of translation, as you're saying. It's like there's the specificity on the ground and there's the specificity, you know, from building the data sets, but not from the right places.

14:06So as you said, where does that connect? And I think one of the things that I was that I think is so kind of fascinating about where you are and that some of our conversations have revolved around is you wear so many hats. You sit in so many different kind of institutional and otherwise silos, right? It's like academia, you connect to the government, you connect to policy, you connect to smartholder farmers. And so much of your work to create change is about connecting those dots. And how do I sort of bridge through the messy middle to get the information to the right people? So I guess I would love to hear first, not just from a data perspective, but from a, I guess, almost like project management perspective.

14:53What are those challenges you encounter in connecting those dots? And what is your kind of process or system in approaching these problems? This is a fascinating question. I'm in different places that give me different contexts. So it helps me understand a little bit better, like, what the perception is on one side to make it relevant on the other side. Everything is connected in some sense. And so I want to kind of give an example of why these different contexts matter so much. So when we think about Madagascar, one of my favorite places nowadays, I talk about it a lot. We usually read about biodiversity loss.

15:42We see maps on deforestation. We see all these challenges, right? So from an outside perspective, this is the perspective you will gain. and you would be worried and you would, you know, put on your boots and you want to go do something about it. However, if you were in Madagascar and you're an everyday person and you were actually trying to fend for your family and you don't have options, there's not much else anybody could do. If the only thing that you have is the wood from that tree to make a meal for your child, that is the only option, right? So depending on what room you're in, how you hear it will inform what you think about people's actions.

16:24So when I was in Madagascar, what I saw is everyday people trying to have a meaningful life and doing their very best with what they have. If you just observe the process of what they do on a daily basis, I can figure out where my tools can be useful. I can create a really good map of cacao, a really good map of where rice is. I can show to the farmers the contribution of cacao, but also I can utilize what I've learned in this process to communicate to the outside. Like this is the reality of what a cacao farmer goes through. They harvest every morning. They have to take the cacao out of the pod.

17:14They have to put it in the bins to start fermenting. They have to then dry it, and then they have to weigh it. And there's a whole records-keeping process. And in that process, not only will you interact with the farmer, the person who does the deshelling, the one who will harvest the cacao, you will talk to a company that works for them, buys cacao from them, the person who fixes the tractors that help them carry things. you will experience the fact that when it rains you literally have to take off your shoes and walk for like a whole kilometer to get to a place where you could just drive and then you experience the fact that because the weather is so unpredictable the drying process of cacao is so painful but you don't want it to get wet because that ruins the value and so it's like all of these like things that I wouldn't otherwise have access to if all I did was map tree loss, you know, think about it as deforestation.

18:14I wouldn't be able to see that. Yeah. And as you said, just the understanding of obviously the context for how this is relevant. My question is back to you as this individual that sits in all of these different sectors, spaces, academia, policy, government in the farm otherwise. It sounds like what you're talking about is having to experience those things in context to make the data relevant. How do we scale that from where you sit? Do you think we need more of you? We need more humans in that kind of connective role doing this translation to make that last mile connection happen between the data and it being useful and relevant?

19:07This is a really good question. For me, what I think is, it's more about the why we're doing it. What is valued in academia versus what will be valued on the ground, sometimes completely different things, right? It's okay to develop models, to publish papers, to create platforms, to measure everything the best possible way. That is okay. It is not okay to say that we are very good at measuring this and this is how it's going to help 10 million farmers somewhere on the continent of Africa or Asia. Very different because it's like any model that you can imagine. It is so far, far away from the reality of an everyday person.

19:55All this obsession with, you know, oh my God, I wrote this, or this model works better than this, and everybody's LinkedIn posts are getting more polished and longer, etc., like that kind of thing. But when you go to this scenario I was describing in Madagascar, there was not a single person who worked at a farm who had a smartphone. and so they do not know chat GPT. Chat GPT has not done anything in their life and what that means is that all of these things literally do not change anything for them. So if I'm like, we're going to figure out how to make this farmer find what we're doing relevant.

20:35We don't want to waste their time so we need to think about their context and then we can build from their context. figuring out where the things that are accessible in their context, if you can receive information via a text message, or if you have to call somebody, or if there's some kind of village structure where people meet. And my coolest idea that I had recently was using my knowledge, tools, etc., and getting somebody who runs a radio program to help translate it. Not translate it in a language, but explain it the way they would explain it to the farmer. They have listeners who would get that information.

21:19They'd say, well, we know there's a drought coming. Information shows all of this area will be affected. These are some of the recommendations, some of the things you can do. That would be so much better than my next bulletin on drought in Somalia, for example. I mean, so it sounds like humility is important. It sounds like deep listening, being very intentional with your observations. It's like what we're missing is the connective tissue of understanding. But creating that knowledge by working with local people is such a long process that doesn't fit a regular timeline. My work, if you can believe it, involves a lot of WhatsApp groups of trying to figure out things on the ground.

22:13I'm going to tell you about a project that I have that was literally blood and sweat, love and care. So for my long work looking at drought analysis, et cetera, one of the biggest problems we have and there lies the solution is in order to predict a drought, you have to know how much water is available to the plant. And so when you look at the moisture in the soil, it gives you an idea of what the plant might look like two weeks, three weeks, four weeks in advance. And so there's this cool satellite that was launched in July. And while it was going to be launched, some of the preparation that is done is you have to have these calibration centers or stations, ground calibration.

22:59So in order for us to know what the satellite is measuring and what the true value is, we need sensors on the ground. And in the U.S. and in Europe and in India, there are huge calibration networks that are part of this mission planning, et cetera. So when the satellite launched, when the data start coming in, they can calibrate and know what the true value of what is at the sensor is for what is on the ground. But guess what? There's not a single station in Africa. No. Yes. There's not a single reporting live station in Africa. There was not a plan for like a calibration infrastructure for this upcoming mission that is amazing for that whole continent, where the biggest burden of drought is, I should say.

23:49And so, because I knew about the mission, et cetera, and I worked on drought, I have a PhD student who's working on drought. The plan was, we're going to make sure we have some sensors in the ground before the satellite launches. No funding. So I buy four stations with my, what do you call it, leftover money. So I bought one when I was in Nairobi. We have a station in Uganda, in Karamoja, actually, installed by my favorite, oldest friend, who took me everywhere when I was with my PhD, with a farmer there. We have one in Tanzania with my friend that I worked with at Tsokwena University. We have one in Kenya at my friend's friend's farm.

24:36We installed them, and we have now our WhatsApp group of groups. We followed the requirements for calibration. They fall within the grid of the sensor. They fall within these requirements for it to be actually a good data set on the ground. But the only way this project would be magical is if we can install like 2 ,000, 5 ,000, or 20, I don't know how many stations considering the size and the scope and the scale of drought on the continent. Wow. And so, again, that's only happening because you're using your extra money, your friends. My friends. And carrying the stuff yourself. Yes. So, like, I think you asked, how do you scale me?

25:19From what I've experienced is there are so many really brilliant people who, like, a lot of people reach out to me on LinkedIn. and like I'm inspired by your work because I want to do what you do. They want to do something meaningful. But the opportunity to do that something meaningful literally is like, sometimes I think about like breaking rules, being defined, figuring out things that otherwise would be like, why would you do that? But it's like in the end when it comes to life, it's so meaningful. It's magical. Now we have all these dashboards with all these lines where there was zero lines.

25:55And the other thing that I kind of think about is going back to my PhD, I had the opportunity to go figure it out. And I had the space to learn to figure it out. And so like that opportunity to learn, to grow, to be able to think and create your own ideas and link them back to the things that matter. It's like, I think we need more of that. Like it's so much effort. It's so much energy. And it's like, but in the end, it will be worth it. And learning how to ask good questions. Yeah. Not just ones that are in a classroom or academia, but actually that apply to the people that are living them. Yeah.

26:31So what is something that is scaring you right now? And what is something that's giving you hope and excitement? What gives me hope is the fact that there are so many young people eager and willing to learn and contribute. The other thing that gives me hope is my ninjas. I call them ninjas sometimes, my sons, because they keep me on my toes. They ask me questions about stuff, like I have to explain black holes. I have to do all of this, like, things that I wouldn't normally do if I didn't have them. But the other thing is, so I just did my talk yesterday, and I practiced on them. What I practiced on them was the fact that, you know, we can predict drought, but crops fail, and people don't have food.

27:24And they were like, oh, but can't they, like, have an irrigation system and pump water? And I was like, very good point. That's what everybody says. and they see it, but like in reality, doing it is very different. And so like being able to like have these experiences with them, it gives me a lot of hope and I can't wait for them to be my field buddies for sure. But what worries me I think right now is there are huge challenges for earth science per se, which is being able to do my work the way I think it should be done through connections, like ground-based innovation that then links to the most sophisticated workflows is harder and harder.

28:17Funding to do the probes in the ground, I call it root sense, it's like we sense what's at the root to figure out what will happen in the weeks to come. is something that for me, in my perspective, is like amazing. It should just happen. But unfortunately, it's so complicated to do it. And so trying to figure out how to continue to do my work in the current context is very hard. The pathways I used to be able to do capacity building have basically shut down. One of the things I do, I call them crop monitor champions. I do these learning exchanges, have people from different ministries. We meet in one country.

29:00The country's the priority. We focus on their monitoring, help them kind of get up to speed, and then others learn and share from their experiences. And I was doing them as rotational, and they're amazing. It is not possible to do that anymore. Yes, we can do things on Zoom, but the connection and the problem-solving and the generosity that happens when we're in rooms with people is very different, that everybody walks away empowered. And what I love to do in these particular sessions is I get the countries that have done something to present it from their perspective and what they were able to do.

29:39Because it's easy for me to say, you can do X, Y, Z. Very different from how that works when the real person who needs to do it does it. And so when they explain it from their perspective, but it arms you to be like, I can do it too. And so I'm overwhelmed by the fact that those pathways are closing. That's kind of like the biggest, saddest thing. Yeah. Yeah. I mean, I guess back to where we started the conversation, it's in that I have to figure it out. Like, what's the pathway through when it seems there is not one present? Yeah. I keep going. There's so many people willing to share and give and collaborate.

30:21there is so much hope and so much perseverance, so much resilience. In the storm I mentioned that happened in Kenya, the hail storm, people sent photos from there, from where it's happening. And one of the agents said, we really need to educate and prepare our farmers for what's coming because this is absolutely devastating. So imagine hail that covers the ground completely white in Kenya. What month are we? September. So maybe it's a little later and maybe people have already harvested, but if they didn't, it's completely ruined. And so like him saying that, right, is like, it is really important that we prepare our farmers to know that these types of things are coming and they need to know before they happen.

31:14Or we need to have an immediate solution that can help address their loss. Imagine the scale at which it's a banana plantation. It literally kills it. And so whoever's field is in that area is completely ruined. They had no warning, nothing, and it's completely ruined. And hail can actually be predicted using microwave radar because some storm has to form somewhere and we can actually track it with radar. But sometimes it happens really quickly. But so when it happens, then what? That's why I criticize that whole, I can write you a nice report about what happened with the event, but then what?

32:04So I find hope. I get discouraged. I find, but there's actual things that need to be done that can make a huge difference. And if we did more of that, we would be in a much better place. I mean, I feel like I could just talk to you about this forever. And thank you for being here today and having this conversation. It's really been a treat. Thank you.

32:33That was Catherine Nakalembe, a 2025 TED Fellow. To learn more about the TED Fellows program and watch all the TED Fellows films, go to fellows.ted.com. And that's it for today. This episode was produced by Lucy Little, edited by Alejandra Salazar, and fact-checked by Eva Dasher. The audio you heard at the top comes from the short film made by Divya Gdenghi and Owen McLean. Story edited by Corey Hajim and produced by Ian Lowe. Video production manager is Searing Dolma. Additional support from Lily James Olds, Leonie Horster, and Allegra Pearl. TED Talks Daily is part of the TED Audio Collective.

33:11Our team includes Martha Estefanos, Oliver Friedman, Brian Green, Lucy Little, and Tonsika Sungmarnivong. Additional support from Emma Taubner and Daniela Ballarezzo. I'm Elise Hu. I'll be back tomorrow with a fresh idea for your feed. Thanks for listening.

33:36Thank you.