Across the parched borderlands of northern Kenya and southern Ethiopia, the stakes of a seasonal forecast are measured in sacks of maize, surviving calves and meals not skipped. Supercomputers in Nairobi generate increasingly sophisticated predictions, while elders still read the stars and the behavior of cattle for signs of rain. Yet between those two worlds lies a quieter vulnerability: when the rains come late, or fall too hard, or not at all, there is rarely a clear path for farmers and pastoralists to say the forecast failed them
Mzee Guyo Sora has been reading the sky since he was a boy. His father taught him. His father’s father taught him.
For generations, the men in his family have watched the stars and advised their community along the Kenya-Ethiopia border on when the rains would come. The knowledge is carried in memory and ritual, in the observation of animal behavior and the dissection of entrails. It has long shaped decisions about when to plant, when to move livestock and when to brace for hardship.
This year, Mr. Guyo says, the rains will start on April 3. There will be short rains in March and more toward the end of April. Some areas along the border may receive below-normal rainfall, but most places will manage. He is advising farmers to plant maize, green grams and beans. He warns that flies could spread disease among animals, but says no livestock will die from starvation.
He reads these signs from the behavior of his cattle and the condition of the sky. When cows hesitate to enter their enclosure in the evening, he explains, it signals that rain is coming and the ground will soon turn muddy.
Then he pauses.
“The evening star has come earlier than expected,” he says quietly. “I am confused. All our expectations have been turned upside down. People used to come to me for advice. But now I tell them: I am also wondering what is happening.”
Mr. Guyo is a traditional forecaster whose knowledge runs deep. But climate change is moving faster than inherited patterns can reliably track. The signs that once aligned no longer do. The stars, he says, have become liars.
More than 600 kilometers away in Nairobi, supercomputers are grappling with the same uncertainty. They process data from satellites, ocean buoys and atmospheric sensors. At the IGAD Climate Prediction and Applications Centre, known as ICPAC, scientists produce some of the region’s most advanced seasonal forecasts.
Between Mr. Guyo’s sky-reading and Nairobi’s models lies a stubborn divide. It is not simply a gap between tradition and science. It is a gap between information and understanding, between prediction and action, between those who generate climate data and those whose harvests depend on it.
In February 2026, an unusual meeting took place in Moyale, the town that straddles the Kenya-Ethiopia border. Forecasters from both traditions gathered for a climate outlook forum focused on the Marsabit–Borana cross-border region.
Oliver Kipkogei, a regional agro-meteorologist with ICPAC, acknowledged a central problem.
“We need to design advisories that help local farmers understand both the benefits and the risks of each season’s forecast,” he told Impact AI News.
The science, he explained, is strong. But it often falters in the final stretch between forecast producers and users.
Farmers along the border had just endured a disappointing October-to-December season. Crops germinated poorly. Yields were low. Warnings of below-normal rainfall had been issued, yet they did not translate into effective action.
The March-to-May rains are critical across the Greater Horn of Africa, providing up to 60 percent of annual rainfall in some areas. For millions of smallholder farmers and pastoralists, the season determines whether families eat or go hungry.
For decades, seasonal forecasts were communicated in blunt categories: below-normal, normal or above-normal rainfall. By the time such forecasts filtered down to rural communities, they often arrived stripped of nuance. Farmers heard probabilities for regions that did not match their farms. The message felt distant and abstract.
In 2019, ICPAC shifted its approach, becoming the first regional climate center in Africa to adopt objective, traceable and reproducible forecast methods. Instead of relying on group consensus, it began combining outputs from multiple climate models using statistical techniques to determine the most reliable signal. Through projects such as CONFER, the center developed high-resolution simulations that zoom in to 10-kilometer grids.
The science has improved markedly. But improved science does not automatically produce improved outcomes.
At the 5th Sub-Regional Climate Outlook Forum, held from February 10 to 12, 2026, in Ethiopia, the focus was not just on models but on dialogue. Around one table sat representatives from national meteorological services, ministries of agriculture and health, the Red Cross, local radio journalists, farmer group leaders and traditional forecasters like Mr. Guyo.
The process, known as the Participatory Scenario Process, unfolds in stages. Meteorologists present the scientific outlook, detailing probabilities and expected onset dates. Traditional forecasters share observations of stars, trees and birds. Sector experts assess implications for crops, livestock and disease. Farmers articulate what information they actually need to make decisions.
“This approach bridges the gap between climate forecasters and sector-specific expertise,” Mr. Kipkogei said. “Local stakeholders can give information on the status of their livelihood based on the past season, and give their input on critical steps to be taken in the next season.”
The result is a set of advisories translated into local languages and tailored to specific locations. Farmers who once doubted official forecasts now incorporate them alongside indigenous knowledge.
But such workshops reach dozens of people at a time. The region is home to millions.
The last-mile challenge remains formidable. How does a forecast generated in Nairobi reach a pastoralist in northern Kenya in a form that is understandable and actionable?
Mobile technology is one avenue. In 2022, ICPAC launched HUSIKA, a platform that sends early warnings via SMS and mobile app. Users receive alerts about weather hazards and can reply with reports from the ground.
Radio remains essential. In Taita Taveta County, Albert Mwanyasi of Sifa FM delivers daily weather information. “After receiving the information from the weatherman, we have to interpret it so we can communicate it using language that our listeners can actually understand,” he said.
Training from ICPAC and NORCAP has helped broadcasters strip away jargon and connect climate information to food security, health and education. When farmers call with questions, Mr. Mwanyasi relays them to county officials.
Even with these efforts, structural weaknesses persist. Analyses comparing regional forecasts against models such as the UK Met Office’s GloSea5 have identified systematic biases, including a tendency to overestimate the likelihood of near-normal rainfall. Forecasts often cluster in the middle, even when outcomes turn out more extreme.
Yet accuracy is only part of the story. When forecasts fail, there is little formal mechanism for farmers to report that failure. The system is designed primarily to send information outward, not to absorb feedback.
Research from Burkina Faso found that farmers who attended workshops were more likely to understand and use forecasts. But one-time engagement is insufficient. Participation must be continuous, and inclusion intentional. An ICPAC report noted that “Involvement of women in the process required a conscious and intentional effort in all the sites,” with female representation in some cases below five percent.
Accountability in climate services is not only about statistical verification. It is about whose voices shape advisories and whose experiences are counted as evidence.
For Mr. Guyo, the tension is deeply personal. Indigenous forecasting systems are rooted in culture and trust, built on centuries of observation. But climate change has disrupted the ecological signals on which they depend. Trees flower, but the rains do not follow.
The path forward is not to discard one system for another, but to bring them into conversation. Traditional forecasters sit with meteorologists, comparing observations and seeking alignment. Scientific forecasts gain local legitimacy. Indigenous knowledge gains analytical support.
Dr. Guyo Malicha of IGAD’s livestock unit, ICPALD, says forecasts can also help communities manage pasture and avoid conflict over scarce resources. When expectations are shared, planning becomes possible.
John Nguyo of the Kenya Meteorological Department recalls the October-to-December 2025 forecast, which predicted below-normal rains in northern Kenya. Scientific and traditional forecasts aligned, strengthening confidence in the warning.
The 72nd Greater Horn of Africa Climate Outlook Forum has now issued its outlook for March to May 2026. Above-normal rainfall is expected across parts of western Kenya, Uganda, Rwanda, Burundi and Tanzania. Near-normal conditions are forecast for Somalia, eastern Kenya and coastal areas. Flood and drought risks are highlighted, and sector-specific advisories attached.
The models are more sophisticated than ever. Investments are growing. The commitment to co-production is evident.
But the forecast remains incomplete without a functioning loop of accountability.
It will be complete only when farmers who act on advisories have structured ways to report back on their usefulness. Only when verification reports are shared in the same communities that receive the predictions. Only when climate services are embedded in agricultural extension and disaster planning at the district level.
Journalists, too, play a role. Reporting the forecast is not enough. Its performance must also be scrutinized.
As the sun sets along the border, Mr. Guyo looks again at the evening sky. The star still unsettles him. But last season, he sat beside meteorologists for the first time, sharing observations and listening in return.
“The scientists also have knowledge,” he says slowly. “Maybe together, we can find the truth.”
The coming forecast is expressed in probabilities and thresholds. Its credibility will not be measured in charts, but in fields that yield crops and families that avoid hunger.
The stars and the supercomputers are beginning to speak to one another. The larger question is whether the system that connects them will learn to listen.
Get the latest news and insights that are shaping the world. Subscribe to Impact Newswire to stay informed and be part of the global conversation.
Got a story to share? Pitch it to us at info@impactnews-wire.com and reach the right audience worldwide
Discover more from Impact Newswire
Subscribe to get the latest posts sent to your email.
