Delayed rainfall is disrupting planting across Rwanda, leaving some farmers counting losses and others racing to sow crops weeks into the September-to-December growing season, as climate scientists explore artificial intelligence (AI) to improve weather forecasts. The concerns coincide with an AI summer school on weather and climate modelling in Kigali from October 15 to 23, bringing together meteorological experts and young researchers to strengthen forecasting capabilities across Africa. “Climate has changed. We used to have rains from the beginning of September, but some farmers received rainfall in mid-October. Maize seeds planted in September struggled to germinate, resulting in losses for maize farmers,” said Evariste Tugirinshuti, president of the Rwanda Maize Farmer Cooperatives Federation. ALSO READ: Many parts of Rwanda to get minimal rains He said the uncertainty had complicated planting decisions, with some areas receiving rain only once every two weeks and other farmers beginning to plant much later than usual. “Last year, during this season, we received rain earlier, but it stopped earlier than expected, affecting farmers while their maize was still growing,” he said. “There is still hope for the coming months if the rains do not stop early.” Rwanda produced about 600,000 tonnes of maize in 2025 and aims to raise annual output to approximately 900,000 tonnes by 2029. Farmers fear increasingly unpredictable rainfall could make that target harder to achieve without stronger climate-resilience measures. The Rwanda Meteorology Agency had forecast staggered rainfall onset across the country, including the start of rains in Kirehe and remaining parts of Ngoma and Kayonza between October 6 and 13. Potato growers in the normally cool volcanic region have also felt the effects of prolonged dry spells. “Some potato seeds have not yet germinated due to prolonged dry spells,” said Jean Providence Bigirimana, a farmer in Burera District. He said selected Irish potato varieties were selling for Rwf1,100 per kilogramme in Musanze, a price he described as unusually high. ALSO READ: How climate change effects are leaving farmers vulnerable Another Burera farmer, Eric Muyizere, called for an expansion of irrigation subsidies to help growers cope with unreliable rainfall. Government urges planting as rains return Florence Uwamahoro, acting director general in charge of agriculture development at the Rwanda Agriculture and Animal Resources Development Board (RAB), acknowledged that extended dry conditions had affected production in some areas. “We experienced prolonged dry spells from May to October this year,” she said, noting that farmers in Nyabihu, Musanze and Burera, who traditionally grow crops during Season C in July, had faced unusually dry weather. With rains beginning in some locations, she said farmers were moving to plant potatoes and other crops, but planting elsewhere remained slow. “Farmers in areas where it has rained should already be planting,” Uwamahoro said. She also urged those with emerging crops or access to irrigation to remain vigilant against pests and diseases. Expanding irrigation is another priority. National mapping has identified 610,595 hectares of potentially irrigable land, while more than 417,000 hectares with irrigation potential remain undeveloped. ALSO READ: Inside Kayonza's irrigation bid to mitigate climate change impacts Currently, irrigation infrastructure covers 37,273 hectares of marshland, 10,159 hectares of hillsides and 27,791 hectares under small-scale irrigation technology. Rwanda estimates it will need at least 693.7 million cubic metres of water storage by 2034 to meet projected irrigation demand. The Food and Agriculture Organization says agrometeorological advisories can help farmers manage water and fertiliser use, control pests and diseases, and plan planting and harvesting. Such information is important for smallholders whose livelihoods depend on rainfall and who have limited alternatives when seasons shift. AI offers promise, but data gaps persist Against this backdrop, experts meeting in Kigali say AI and machine learning could complement conventional weather and climate models by processing large datasets, improving local forecasts, correcting systematic errors and supporting earlier warnings of extreme weather. For farmers, timely and location-specific information could guide decisions on when to plant, what crops to grow and how to prepare for droughts or heavy rainfall. But experts caution that advanced tools cannot compensate for missing observations. “AI is an important opportunity for Africa, but we cannot make forecasts without data. We are not doing divination,” said Dr Ousmane Ndiaye, director general of the African Centre of Meteorological Applications for Development (ACMAD). Ndiaye warned that gaps in readings from weather stations across Africa could undermine forecast reliability, limiting the ability of farmers and authorities to prepare for damaging conditions. ACMAD is seeking to strengthen the skills of national meteorological and hydrological services and young climate researchers through an annual AI training programme aimed at improving locally developed forecasting tools and climate information services. Athanase Hafashimana, a climate data scientist at the AIMS Research Innovation Centre and Jomo Kenyatta University of Agriculture and Technology, said AI could help analyse historical and current climate data more quickly. Scientists generally use at least 30 years of observations to establish a climatological baseline and assess how current conditions compare with long-term patterns, he explained. “AI is a big opportunity because what we have traditionally done using physical and statistical methods is no longer sufficient,” Hafashimana said. He stressed, however, that better forecasts matter only if they reach users early enough and are relevant to local conditions. Seasonal outlooks could help farmers choose planting dates and crops, while alerts about heavy rain could give communities time to prepare for flooding. For farmers facing uncertain rains, the test will be whether advances in forecasting translate into practical guidance before the next planting decision has to be made.
