No Radar Tower Required: Reinventing Weather Forecasting in the Age of AI

Ignitia is a deep-tech company committed to democratizing weather and climate intelligence around the world. It provides AI-powered weather insights for adaptation and resilience designed to keep you ahead of the storm.

This is the third in a three-part sponsored series on Ignitia. You can listen to our special podcast episode with Ignitia CEO Andrew Lala HERE, and read our introductory article on the company HERE.

Interested in being a Climate Proof sponsor? Get in touch HERE.

Emmnauel Sani Bello had a fraught 2025. A hydrologist with the International Rescue Committee (IRC) in Nigeria, Bello’s year was consumed by a series of brutal flooding events that swept through the country from April onwards, claiming over 500 lives and causing billions of naira in economic damages.

Many of these were flash floods, occurring over the course of mere hours amidst sudden, heavy rainfalls. As a key member of the IRC’s early warnings team, Bello’s job is to monitor these events and push out alerts to at-risk populations so they have time either to evacuate or prepare flood defenses. 

This is no easy task. With flash flooding, what matters is where and when a cloudburst occurs, and how much water it releases. Forecasting accuracy is therefore essential to ensuring the right warnings reach the right people at the right time. Older tools available to the IRC had been found wanting. But last year, Bello and his team were able to use a new tool: AI-enhanced flood prediction data from Ignitia, a climate intelligence tech company specializing in high-precision weather forecasts.   

“One of the major problems within the humanitarian sector is there is a common belief that flash floods cannot be detected,” says Bello. “But I was able to see first hand that Ignitia was able to provide [an] accurate and reliable solution that was able to help us to detect flash flood … It played a great role in ensuring that our set of models were effective — and at a finer scale.”

The company's data and tooling strengthened the IRC’s early warning program, extending its flood alerts from two weeks out to three — an extra week for communities to prepare. “There are areas where we saw this risk of flood and it translated into reality like magic. In fact, one of our own offices was almost about 60% submerged in one part of Borno state, and this was something we saw three weeks before it unfolded,” says Bello.

Bello is one of an emerging cohort of climate resilience practitioners who hoover up extreme weather forecasts, evaluate the risks to people, businesses, and infrastructure, and issue timely alerts so they can prepare. It’s a role becoming more important as heavy storms — supercharged by the changing climate — unleash havoc at an ever greater frequency. 

While these early warnings experts have most visibly been deployed by humanitarian organizations like the IRC, they are increasingly in-demand across infrastructure operators and private businesses. Rising climate risks have also shone a harsh spotlight on the quality, credibility, and timeliness of legacy forecasting systems — and opened an opportunity for savvier, more tech-forward providers to muscle in. 

Ignitia is among their number. The climate and weather intelligence company started off some 15 years ago by finetuning local-scale rainfall modeling for communities in West Africa. Back in 2018, it started embracing AI and machine learning to augment its predictions, and now delivers faster, cheaper, and more accurate forecasts to more than three million users across 15 countries in Africa, Latin America, and Asia — without the need for costly physical infrastructure.

“With machine learning and AI techniques, we have been able to completely get off that vicious cycle of expanding geography equaling expanding costs, to something that can have a near global coverage without adding any additional cost — other than essentially running one server. This has allowed us, as a small Swedish company with a social mission, to actually expand beyond West Africa,” says Andreas Vallgren, Co-Founder and Chief Scientist at Ignitia. An important recent innovation is the company’s ‘virtual radar’, which is used to issue nowcast alerts and inform a suite of data services, from daily weather forecasts to climate-smart agriculture advisories.

Some of the company's newer users work at Porto do Açu in Brazil — one of the largest private industrial port complexes in Latin America. The facility accounts for roughly 30 to 40% of the country's oil exports by volume and houses a major iron ore terminal. André Ribeiro, the Risk, Emergencies, Safety and Environment Manager at the port, says around 20 members of the Emergency Response and Operations Center (CORE) have been using Ignitia's platform since February 2025. CORE’s role is to issue alerts and communicate potential risks to users and clients across the complex — communications that Ribeiro estimates could reach some 7,000 people circulating within the port on any given day.

“Maritime maneuvers, cargo handling, vessel movements, and emergency preparedness all rely on reliable short-term weather information,” says Ribeiro. “For us, having access to precise and timely forecasts is fundamental to support our risk management processes. It enables us to anticipate potential impacts, take preventive measures, adjust operational planning, and protect people, assets, and the environment,” he adds.

An additional draw is the system’s integration with WhatsApp — a widely used communication platform in Brazil, with 150 million users. This allows weather alerts and updates to be shared rapidly across teams via a medium they are familiar with, improving response times and coordination.

This user-focused approach makes Ignitia a good fit for the UN’s Early Warnings for All (EW4A) initiative. Launched in 2022 by UN Secretary-General António Guterres, the ambitious plan aims to cover everyone on Earth with extreme weather and climate disaster early warning systems by the end of 2027.

Turning this vision into reality requires new capabilities, including high-quality forecasting for remote regions in the world — without the imposition of costly new hardware. It’s a challenge Ignitia is well-positioned to meet. Last year, the company was selected as a top four finalist for the EW4A innovation challenge and invited to present at the AI for Good Global Summit in July 2025. It is now engaged in a pilot scheme of its ‘virtual radar’ in Uganda.

“Ignitia really encapsulated the spirit of the Early Warning for All Initiative,” says Newton Campbell, a consultant supporting the International Telecommunication Union’s work on the AI for Early Warnings for All Group. “Their models are there to really deliver these hyper-local storm alerts. On top of this, they developed what is essentially a virtual radar combining satellite imagery, ensemble forecasts, and deep learning — so you get radar-like performance without installing a single radar tower,” he explains.

Because the EW4A initiative exists to deliver potentially life-saving alerts to climate-vulnerable populations, standards for tech partners are high and the margin for error low. While Ignitia has implemented its own safeguards to ensure its AI-powered forecasts are rooted in observation data sets and won’t ‘go rogue’, the EW4A initiative has augmented this with Human-in-the-Loop and Human-on-the-Loop processes to improve and strengthen these defenses.

“Deploying doesn’t just mean: ‘hey, here’s our software, use it for alerts, and it’ll be 100% correct all the time’,” says Campbell. “It means we’re going to come up with a concept of operations where there’s a Human-on-the-Loop that can actually do some checking, do some validation, before, during, and after”.

For his part, Vallgren says the team is obsessive when it comes to measuring the accuracy of its early warnings. “We look at those metrics that are important to end users: false alarms, misses, those sort of things that build confidence in a forecast,” he says. “We also look at what is the economic consequence of doing the wrong action, given a certain rate of false alarms and misses, in order to try to optimize the value for an end user.”

While Ignitia’s early warning capabilities are already proving their worth with a wide range of organizations, users are always pressing for more – and understandably so, given the escalating frequency and severity of climate-related shocks. 

The company has proven reactive to user needs. At IRC, Bello says he asked Ignitia for more “local realities” to be factored into its flood data, arguing that earlier iterations missed the hydrological parameters governing water movement across the Earth's surface. Ignitia responded in July 2025, updating its modeled forecast to incorporate local hydrological factors — including a six-week outlook on flash flood risk and dry spell risk.

What Bello’s request illustrates, though, is the sheer complexity of forecasting in a world where past experience no longer informs future predictions as reliably as before, and the importance of diverse data sources to building a holistic picture of flood risk. It's a challenge Vallgren is eager to overcome.

“By having developed such skillful forecasts and past weather datasets in atmospheric environments that are really hard to get right — and where both global models and new AI models struggle — we are able to focus more on the weather impact predictions. There is simply so much more one can do to support users when the underlying data is good. We can more directly help drive actions and answer the real questions posed by the users through use-case modeling instead of just passing raw forecast data,” he explains.

At the EW4A initiative, Campbell is watching the Uganda pilot closely — as is the country’s government. The latter is particularly eager to understand how the company’s flood warnings will reach rural farming communities and fast-growing urban centers exposed to sudden deluges.

If Ignitia’s forecasts prove effective and mesh with existing communications infrastructure, the next step would be to expand the technology to other jurisdictions, with Cambodia and Ecuador potential targets. The success of these trials, in turn, could determine whether the 2027 EW4A target is within reach or wishful thinking.

Campbell says a great deal depends on country readiness and financing — not to mention overcoming practical challenges integrating solutions with existing systems. “The match-making with countries is important. But just as important is having the UN there to nurture and guide — because one; a pilot is not going to be executed in a way we can predict, two; there’s a much larger network at play here, and three; because we’re using artificial intelligence, what are both the restrictions and constraints of a particular area, and what happens when things fail in this area?” he asks.

Thanks for reading!

Louie Woodall
Editor