Get Your Weekly Digest

Receive the best podcast summaries and insights delivered to your inbox every week.

We respect your privacy. Unsubscribe anytime.

TED Talks Daily

How satellites are supporting farmers across Africa | Catherine Nakalembe

with Catherine Nakalembe

Published October 24, 2025
View Show Notes

About This Episode

Satellite food security specialist Catherine Nakalembe explains how she uses satellite imagery and machine learning to map and monitor crops across African countries, and why many existing models fail when applied to smallholder farms. In a follow-up conversation with TED Fellows Program Director Lily James-Olds, she describes the gap between powerful data systems and farmers' realities, the importance of ground-based data and local context, and her efforts to build practical, human-centered ways to turn drought and flood information into action. She also shares a grassroots project to establish soil moisture calibration stations in Africa and reflects on the institutional and financial barriers, as well as the sources of hope that keep her pursuing this work.

Topics Covered

Satellite mapping Climate change and agriculture Smallholder farmers Food systems Ecological data Data-driven decision-making Natural disasters Drought Sustainability Resilience

Disclaimer: We provide independent summaries of podcasts and are not affiliated with or endorsed in any way by any podcast or creator. All podcast names and content are the property of their respective owners. The views and opinions expressed within the podcasts belong solely to the original hosts and guests and do not reflect the views or positions of Summapod.

Quick Takeaways

  • Existing satellite and AI models for agriculture often work well in Europe and the U.S. but fail to capture the complexity of small, mixed-crop fields in countries like Kenya, Uganda, and Rwanda.
  • To make satellite-based crop monitoring accurate and useful, Catherine emphasizes extensive ground truthing, including creative data collection such as GoPro-based "Google Street View for crops."
  • Many African farmers remain untouched by sophisticated AI tools because these tools are not connected to the resources, communication channels, and constraints of their daily lives.
  • Catherine argues that true innovation is not about high-tech for its own sake, but about fitting technology to real problems and local contexts, often through low-tech channels like radio programs.
  • A major blind spot in global earth observation is the lack of soil moisture calibration stations in Africa, despite the continent bearing a large burden of drought.
  • Catherine personally helped install a small network of soil moisture stations in several African countries using leftover funds and informal collaborations, highlighting both ingenuity and systemic underfunding.
  • Ground-based experience with farmers, such as cacao producers in Madagascar, deeply changes how satellite data is interpreted and how environmental challenges are framed.
  • She stresses that predictive information about droughts and extreme weather is only meaningful if it leads to concrete support, like irrigation investments or rapid replanting assistance.
  • In-person regional learning exchanges among government staff and practitioners have been powerful for building capacity and confidence in using crop monitoring tools, but many of these pathways have recently closed.
  • Despite systemic challenges, Catherine draws hope from eager young people, her inquisitive children, and local collaborators who are determined to prepare farmers for increasingly extreme climate events.

Podcast Notes

Host introduction and TED Fellows Films context

Explanation of TED Talks Daily and special Fellows series

Elise Hu introduces TED Talks Daily as a show bringing new ideas every day[2:48]
Announcement of new 2025 TED Fellows Films adapted for podcasts for TED Talks Daily listeners[3:03]
These special episodes will be released on certain Fridays through the rest of 2025 and into the new year[3:11]
TED Fellows program supports a network of global innovators whose work will be showcased[3:15]

Introduction of guest and topic

Elise introduces satellite food security specialist Catherine Nakalembe[3:22]
She notes it can be hard to imagine how satellites help farmers working the soil on Earth[3:30]
Catherine uses satellite technology and machine learning to monitor smallholder farming[3:33]
The talk will show how satellites can play a vital role in helping to feed the world[3:39]
Elise highlights that a grounded, bottom-up approach is also key despite the "bird's eye view" of satellites[3:46]
Catherine emphasizes that true innovation is about making technology fit the problem, not just building high-tech systems[3:51]
Listeners are told there will be a conversation afterward between Catherine and TED Fellows Program Director Lily James-Olds[4:00]

Film segment: Catherine's work and the stakes of crop failure

Emotional and economic impact of failed harvests

Catherine asks listeners to imagine if farming were their primary source of income and they couldn't grow anything[4:03]
• She describes this situation as demoralizing and overwhelming for farmers
In many countries where she works, farmers face biases, pests, diseases, droughts, and floods[4:24]
• If crops fail, there is no food for many people who depend on what those farmers produce
Crop failure can undermine an entire generation[4:36]

Catherine's role as a satellite food security specialist

Catherine identifies herself as a satellite food security specialist[4:42]
She uses satellite data to map and monitor crops[4:44]
• She then works to ensure that this information can be used by decision makers and organizations supporting farmers
Her primary geographic focus is on African countries: Uganda, Kenya, Tanzania, Zambia, Mali, and Senegal[5:00]

What satellites can measure and why we live in a "fantastic age"

She notes that last time she checked, there were over 8,000 satellites observing Earth[5:06]
• These satellites take pictures every day
Satellite images can be used to map what crops are growing where, forecast rainfall, and track weather systems[5:11]
• They can show where a system might be coming from, where it might be impacted, and how badly
Catherine says she can sit at her computer and tell you rainfall, drought, or floods anywhere in the world[5:28]
She describes this as a fantastic age with huge advances in AI that let us process more data and information[5:42]

Limitations and misfits in current satellite products

Looking deeply at any place reveals problems with many existing products[5:49]
• They are often not tailored to on-the-ground contexts
Sometimes the data is simply wrong[5:56]
Many models are trained very well to predict for European or U.S. agriculture[5:59]
• In Europe, most farms have a single crop and are large and homogeneous, which makes them easier to model
In countries like Kenya, Uganda, and Rwanda, fields are tiny, contain many different crops, and farmers do things very differently[6:19]
• She likens the landscape to a tapestry, emphasizing its complexity
With current images and models, fields are misrepresented: some non-crop areas are labeled as crops and some crop areas and people are missing[6:31]
• Using such flawed inputs for ministry assessments means "feeding them garbage"
Ensuring high-quality input data requires significant work[6:46]

Ground-truthing via GoPros and machine learning adaptation

To train models to understand complex smallholder fields, many examples from the ground are needed[6:51]
Her team used GoPros worn while driving on motorcycles or in cars, taking pictures of fields along the way[6:57]
• She describes this as similar to Google Street View, but for crops rather than streets
They adapted face detection algorithms to detect crops like maize, beans, and cassava instead of faces, cats, or dogs[7:16]
They covered all of Western Kenya in two weeks with just two teams[7:24]
• They collected over five million images, many with the help of volunteers such as motorcycle taxi drivers and students
This rich dataset allowed them to build a more complex model that can learn from different examples and contexts[7:39]

Using satellite analysis for rapid flood response in Kenya

She recalls a rapid, extensive flood in Kenya in 2024 that affected much of the country[7:48]
The Ministry of Agriculture emailed her to request an assessment using satellite data[7:54]
• They asked for maps of where floods happened, where crops were, and estimates of the total area of cropland affected
The ministry uses this information to design response programs, such as deciding where to provide seeds so people can replant[8:11]
Catherine describes this capacity to inform concrete action as really powerful[8:27]

Philosophy of true innovation

She reiterates that true innovation is not about high-tech systems[8:22]
• Instead, it is about making technology fit the problem
Good information must be provided to people who can act on it[8:27]
If done correctly, this approach can save time, money, and livelihoods[8:35]

Conversation segment: Catherine's path into satellite-based food security

Setting and timing of the conversation

Lily welcomes Catherine to the TED offices and notes there may be background sounds from colleagues[11:24]
Catherine expresses gratitude for being at the TED office and for the chance to chat during Climate Week[11:47]

Catherine's early academic and career trajectory

Lily asks what first sparked Catherine's passion for using satellite data to tackle food insecurity[11:54]
Catherine recalls that she used to play badminton and planned to do sports science[11:58]
• In high school, she was literally called "Catherine Badminton"
By chance, she received a scholarship to do an environmental science undergraduate degree instead[12:09]
For her capstone research project, she left her parents' house for the first time and went to the west of the country[12:22]
• She did field mapping with a GPS, a backpack, and a forest ranger
• She remembers a photo of herself completely covered in dirt after rain, yet feeling very happy and wanting to do more of that kind of work
She applied for a master's and went to Johns Hopkins University for geography and environmental engineering[12:50]
She did not initially get the opportunity to apply her degree directly to a problem back home[13:00]
She later discovered a department at the University of Maryland and spoke with her eventual PhD advisor[13:03]
• He asked her how she would like to do fieldwork in Uganda, which excited her
Her advisor's main domain was agriculture, which she had not previously considered as a focus[13:16]
• Her earlier inclination had been toward forests because of her prior field experience
She ultimately had the opportunity to do fieldwork back home, learning how things actually work by spending a lot of time in the field[13:31]

Gap between data abundance and on-the-ground usefulness

Failure of translation from data to farmer action

Lily notes Catherine often talks about a disconnect between massive data and a failure of translation to farmers[13:51]
She asks what the biggest barrier is to turning data into action[14:01]
Catherine references the phrase "data is the new oil" and mentions heavy investment in data infrastructure, methods, and publications[14:07]
From a distance, it looks like boundaries are being broken, but in reality, if you visit her sister's farm, there's no connection[14:19]
• Her sister grows maize and relies on agriculture, yet none of Catherine's work directly informs what her sister must do
Even Catherine's best analysis would be disconnected from her sister's access to resources[14:31]
• Her analyses do not tell her sister where fertilizer or seeds are, or when to plant in reality
Catherine compares common expectations for AI in agriculture to how ChatGPT helps write an email[14:57]
• With text, users can immediately check if the AI-generated text is wrong and correct it, because models learn from vast digitized corpora
Agriculture, by contrast, has not been digitized, so there is not much text for models to learn from[15:07]
Her sister's mixed field with coffee, maize, and possibly cacao is not represented in existing data[15:26]
• Even with detailed ground information, models trained on other realities would still do a bad job

Model training and resolution challenges for African smallholder farms

Lily asks whether the main issue is the state of the technology or how information is delivered to farmers[15:57]
Catherine explains that to get useful information from satellite data, models must be trained on existing examples[16:09]
• The model then scans satellite data to find the patterns associated with what you seek, such as specific crop types
Today, most models perform very well for farmland in the United States[16:21]
• U.S. fields are large and homogeneous and have benefited from substantial investment in data collection for training examples
Predicting what is growing in Kansas is comparatively easy[16:37]
Producing a map of maize in Kenya is very different because examples are scarce and data collection is harder[16:52]
• There is no default data collection in Kenya for this mapping purpose
Another issue is that open-access satellite data suitable for scaling across Kenya lacks the spatial resolution to match small Kenyan fields[17:07]
• Training on the same dataset for Kenya and the U.S. would fail for Kenya because the fields are smaller and more complex
Current products are usually not relevant at a farmer's scale[17:25]
• They are more suited to larger-scale assessments, such as indicating a drought in western Kenya affecting a certain number of acres
She gives an example of a recent hailstorm in Kenya where satellites could approximate damage but not individual farm outcomes[17:41]
• For a farmer named Jane with a field in Hoima, Uganda, current systems cannot precisely state whether her specific field was destroyed or not

Potential and limits of existing drought information

Despite limitations, she says there is still a lot that can be done with current data[17:15]
Knowing a drought is coming is powerful information[18:13]
• A farmer without irrigation might decide not to plant in a given season, saving energy, labor costs, and seed investment
Repeated drought information could signal that an area needs investment in irrigation infrastructure[18:37]
• When such infrastructure is built based on long-term measurements and predictions, the situation for farmers becomes a "whole different story"

Bridging institutional silos and learning from ground contexts

Catherine's multi-hat role across sectors

Lily observes that Catherine sits in many silos-academia, government, policy, and connections to smallholder farmers[19:33]
Much of her work is about connecting dots and navigating the "messy middle" to get information to the right people[19:49]
Lily asks about project management challenges of connecting these dots and Catherine's approach[20:03]
Catherine says being in different places gives her multiple contexts and helps her understand perceptions on each side[20:15]
• This broader perspective helps her make information relevant across boundaries

Madagascar example: biodiversity versus daily survival

She uses Madagascar as an example of why context matters[20:35]
From the outside, people read about biodiversity loss and deforestation and want to intervene[20:46]
If you are an everyday person in Madagascar, you may have no options other than using wood from a tree to cook a meal for your child[20:05]
• In that scenario, cutting wood is the only option, despite external concerns about deforestation
Depending on which room you are in, you will hear different framings of people's actions[21:25]
On the ground in Madagascar, she saw everyday people striving to have meaningful lives and doing their best with what they have[21:35]
By observing their daily processes, she can see where her tools might be useful[21:41]
• For example, she can create good maps of cacao and rice and show farmers the contribution of cacao
She can also use what she learns to communicate to outside audiences the realities of a cacao farmer's work[20:53]
• Daily tasks include harvesting cacao, removing beans from pods, fermenting them in bins, drying, weighing, and record keeping
• She interacts with farmers, workers who shell cacao, companies that buy cacao, mechanics who repair tractors, and witnesses difficult logistics in rainy conditions
• Unpredictable weather makes drying cacao very challenging because getting it wet ruins its value
She notes that if she only mapped tree loss and thought in terms of deforestation, she would miss this lived context[21:19]

Scaling, humility, and locally appropriate communication

Need for more connectors versus changing incentives

Lily asks how to scale Catherine's kind of contextual, connective work and whether more people like her are needed[22:09]
Catherine thinks it is more about why the work is done and what is valued in academia vs. on the ground[22:43]
• In academia, developing models, publishing papers, building platforms, and measuring precisely are valued
She says it is fine to do such work, but not fine to claim it directly helps millions of farmers when it is far from their reality[22:43]
• She criticizes obsessions with model performance and polished online self-promotion when they don't change farmers' lives
In Madagascar, none of the farm workers she met had smartphones[22:13]
• They do not know GPT, and GPT has done nothing in their lives, underscoring the disconnect
To make farmers find her work relevant, she insists on starting from their context and not wasting their time[23:27]

Translating information through low-tech channels like radio

She suggests figuring out which communication modes are accessible: text messages, phone calls, or village meeting structures[23:27]
Her recent idea is to combine her knowledge and tools with a radio program host[23:43]
• The radio host would not just translate language but explain information in ways farmers understand
Listeners could receive warnings about droughts, which areas will be affected, and recommendations about what to do[24:01]
She believes such local radio communication would be more impactful than her next technical bulletin on drought in Somalia[24:13]
Lily observes that humility, deep listening, and intentional observation are important for this connective work[24:03]
Catherine notes that creating knowledge with local people is a long process that does not fit regular timelines[25:10]
She says her work involves many WhatsApp groups to figure out on-the-ground situations[25:07]

Soil moisture calibration project and systemic gaps

Importance of soil moisture for drought prediction

She describes a project she calls "literally blood and sweat, love and care"[25:29]
To predict drought, one must know how much water is available to plants[25:20]
• Soil moisture gives an idea of what plants will look like 2-4 weeks in advance
A new satellite was launched in July, and preparation required ground calibration stations[25:29]
• Sensors on the ground provide true values to compare with satellite measurements
In the U.S., Europe, and India, there are huge calibration networks as part of mission planning[25:46]
She reveals there is not a single live reporting station in Africa for this mission[25:16]
• There was no planned calibration infrastructure in Africa for the upcoming mission, despite the continent bearing a large drought burden

Grassroots installation of calibration stations

Knowing about the mission and working on drought, she and her PhD student planned to install sensors before launch[26:26]
There was no funding, so she bought four stations with leftover money[26:29]
• One station was brought to Nairobi, one installed in Karamoja, Uganda, one in Tanzania, and one in Kenya
• Installations involved old friends, colleagues at Sokoina University, and a friend's friend's farm
They formed WhatsApp groups to coordinate and maintain the stations[27:14]
They followed mission requirements so the data would be suitable for calibration[26:18]
Catherine notes the project would be truly effective only if they could install thousands of stations across the continent[27:20]
• She mentions numbers like 2,000, 5,000, or 20,000 stations relative to Africa's size and drought scope
Lily underscores that this is happening only because Catherine uses extra money, friends, and carries equipment herself[27:44]
Catherine says many brilliant people want to do meaningful work like this but lack opportunities[27:49]
• She describes sometimes needing to break rules and be defiant to pursue meaningful projects
Now, where there were zero lines on dashboards, they see data lines from these stations, which she finds magical[28:24]
She values the space she had in her PhD to learn, figure things out, and tie ideas back to what matters[28:26]

Hope, worries, and the future of capacity building

Asking good questions and generational engagement

Lily highlights the importance of learning to ask good questions that apply to people's lived realities[28:35]
She asks Catherine what is scaring her now and what gives her hope and excitement[28:48]
Catherine says she is hopeful because many young people are eager to learn and contribute[28:58]
She affectionately calls her sons her "ninjas" because they keep her on her toes[28:26]
• They ask her about topics like black holes, forcing her to explain complex ideas
She practiced her TED talk on them, which involved explaining that even though we can predict drought, crops still fail and people lack food[29:44]
Her children suggested irrigation systems and water pumping as a solution, echoing common external suggestions[29:44]
• She points out that while that sounds straightforward, actually implementing it in reality is very different
She looks forward to her sons becoming her field buddies[30:02]

Concerns about funding and shrinking pathways

She worries about challenges in earth science work, particularly doing it through ground-based innovation linked to sophisticated workflows[30:32]
Funding for ground probes and what she calls "root census" is hard to secure[29:57]
• By "root census" she means sensing what's at plant roots to infer what will happen in coming weeks
She finds it amazing and obvious that this should happen, yet it is complicated to implement[30:13]
Pathways she used for capacity building have largely shut down[30:54]
She used to run "crop monitor champions" learning exchanges with ministry staff from different countries[31:28]
• They met in one priority country, focused on that country's monitoring, and others learned and shared experiences
• These rotational meetings were powerful, but she says it is no longer possible to do them
While Zoom is possible, she feels in-room connection, problem-solving, and generosity are different[31:33]
She likes having participating countries present their own work, as it is more empowering than her telling them what they can do[31:56]
She feels overwhelmed that these in-person pathways are closing[32:02]

Balancing discouragement with concrete needs and hope

Returning to the earlier theme, she notes her work often involves figuring out pathways where none seem to exist[33:06]
She emphasizes that many people are willing to share, give, and collaborate, which provides hope and resilience[31:45]
She describes the devastating hailstorm in Kenya where hail covered the ground completely white[33:13]
• In some places, banana plantations were completely destroyed and fields ruined without warning
One agent in a WhatsApp group said they need to educate and prepare farmers for what's coming[32:41]
She notes hail can be predicted with microwave radar, but sometimes events happen very quickly[33:31]
She criticizes the idea of only writing a report after the event, asking, "then what?"[33:18]
Despite discouragement, she insists there are concrete things that can be done which would make a huge difference[33:11]
Lily closes by saying she could talk to Catherine forever and thanks her for the conversation[33:40]
Catherine thanks Lily and reiterates her gratitude[33:52]

Outro and credits

Closing information about TED Fellows and production credits

Listeners are informed that Catherine is a 2025 TED Fellow and directed to fellows.ted.com to learn more and watch films[37:43]
Production credits are given for the episode's producer, editor, fact-checker, film creators, story editor, and production staff[38:05]
TED Talks Daily is noted as part of the TED Audio Collective, and Elise Hu signs off, promising a fresh idea the next day[38:14]

Lessons Learned

Actionable insights and wisdom you can apply to your business, career, and personal life.

1

Technology only creates real impact when it is shaped around the lived realities, constraints, and communication channels of the people it is meant to serve, not around what is easiest to model or publish.

Reflection Questions:

  • • Whose real-world constraints and daily routines should you better understand before designing a solution or making a decision?
  • • How might you adapt one of your current tools or processes so it fits more naturally into the way your customers, colleagues, or community members actually live and work?
  • • What is one concrete step you can take this month to test your ideas with the people most affected, and incorporate their feedback into your approach?
2

High-quality inputs and ground truth are essential for trustworthy insights; investing in accurate, context-specific data upfront prevents "garbage in, garbage out" decisions later.

Reflection Questions:

  • • Where in your work or life are you relying on assumptions or generic data that might be misrepresenting the true situation?
  • • How could you build a simple form of "ground truthing"-like direct observation, interviews, or small pilots-into your current projects?
  • • What is one decision you're facing now where slowing down to gather better input data could significantly improve the outcome?
3

Bridging silos requires people who can translate between worlds-technical, policy, and local-and who are willing to do unglamorous connective work like relationship-building and low-tech communication.

Reflection Questions:

  • • In which "worlds" do you currently operate, and where are there gaps in understanding between those worlds?
  • • How might you take on more of a translator role-explaining complex ideas in the language and channels that different stakeholders actually use?
  • • What is one relationship or cross-group collaboration you could initiate or deepen to reduce misunderstandings and increase shared impact?
4

Small, improvised actions-like repurposing existing funds or informal networks-can create critical proof-of-concept infrastructure that highlights systemic gaps and opens the door for larger change.

Reflection Questions:

  • • Where could a small, scrappy initiative in your context demonstrate what's possible and draw attention to an overlooked need?
  • • How can you better leverage your existing relationships and "leftover" resources to start progress instead of waiting for perfect conditions?
  • • What is one modest experiment you can launch in the next 30 days that would generate real data or stories to influence bigger decisions?
5

Sustainable progress in complex problems depends on long-term capacity building-helping others learn, own, and adapt tools themselves-rather than centralizing expertise in a few specialists.

Reflection Questions:

  • • Who around you could become a "champion" if given the chance to learn, practice, and teach others in a particular area?
  • • How might you redesign a current project or meeting so that participants present and problem-solve from their own perspective instead of relying on you as the expert?
  • • What recurring forum, training, or exchange could you create or revive to help people share practical experiences and build confidence together?
6

Forecasts and reports only matter if they are paired with pathways to action-preparedness plans, support mechanisms, and rapid responses that translate knowledge into tangible protection.

Reflection Questions:

  • • What important risks or trends in your environment are currently only being monitored or reported, without clear contingency plans attached?
  • • How could you design a simple "if this happens, then we do that" set of responses for the most critical scenarios you face?
  • • What is one early-warning signal you already have access to, and what concrete action will you commit to taking when that signal appears?

Episode Summary - Notes by Reese

How satellites are supporting farmers across Africa | Catherine Nakalembe
0:00 0:00