Periodic Reporting for period 2 - SKYFORA (Next-generation weather intelligence for more accurate decision making throughout the economy)
Reporting period: 2024-02-01 to 2025-07-31
Accurate weather forecasting is vital for various industries, including utilities, agriculture, transport, construction, and disaster risk management. The World Bank estimates that these weather-sensitive industries could potentially benefit by approximately €160 billion annually from improved forecasting. Weather's impact is increasing due to climate change, creating demand for advanced weather monitoring solutions from public and private sectors.
Private weather companies face challenges due to the high cost of weather observation infrastructure, relying heavily on public agencies' data. Quality and coverage vary by region, hindering comprehensive forecasting. However, the growing global network of 5G base stations offers an untapped opportunity. Equipped with dual-band GNSS receivers, they capture navigation signal delays caused by atmospheric interference, which can be processed for precise weather forecasting.
Unlike traditional methods, GNSS signals can be captured from multiple directions, enabling volumetric estimation of atmospheric conditions. During the project, Skyfora advanced its WeatherCTScan software to process GNSS signal delays and integrate them with external weather datasets, producing 3D atmospheric tomography models. These models were validated against real-world and synthetic data, demonstrating significantly improved accuracy and resolution compared to baseline approaches.
The system was able to extract humidity, temperature, pressure, and wind fields, providing more granular forecasts for pilot users. The results confirmed measurable forecast improvements, showing how GNSS-based tomography can enhance both public safety applications and industry-specific planning.
These validated outcomes provide a foundation for future integration into public warning systems and tailored forecasting services for weather-sensitive industries, supporting both societal resilience and commercial adoption.
Private weather companies face challenges due to the high cost of weather observation infrastructure, relying heavily on public agencies' data. Quality and coverage vary by region, hindering comprehensive forecasting. However, the growing global network of 5G base stations offers an untapped opportunity. Equipped with dual-band GNSS receivers, they capture navigation signal delays caused by atmospheric interference, which can be processed for precise weather forecasting.
Unlike traditional methods, GNSS signals can be captured from multiple directions, enabling volumetric estimation of atmospheric conditions. During the project, Skyfora advanced its WeatherCTScan software to process GNSS signal delays and integrate them with external weather datasets, producing 3D atmospheric tomography models. These models were validated against real-world and synthetic data, demonstrating significantly improved accuracy and resolution compared to baseline approaches.
The system was able to extract humidity, temperature, pressure, and wind fields, providing more granular forecasts for pilot users. The results confirmed measurable forecast improvements, showing how GNSS-based tomography can enhance both public safety applications and industry-specific planning.
These validated outcomes provide a foundation for future integration into public warning systems and tailored forecasting services for weather-sensitive industries, supporting both societal resilience and commercial adoption.
The project carried out several pilot deployments across different sectors, demonstrating clear benefits in forecast accuracy and operational value. These pilots showed how GNSS-based weather intelligence can improve decision-making in real-world environments and created strong references for further adoption. Engagements with telecom and infrastructure partners expanded significantly, establishing a foundation for large-scale deployment. Production capabilities were built with a European manufacturing partner, including quality control and testing processes. Intellectual property protection was strengthened through new patent filings and extension of existing protection to additional key markets. Dissemination activities included participation in major international events, media publications, and industry awards, which raised visibility and helped attract new partners.
The hardware component of the system was advanced to a production-ready stage, with antenna design, electronics, and enclosure optimized for scalable manufacturing. Connectivity options were enhanced, and diagnostic features were built in to ensure reliability during large-scale deployment. Production testing processes were established to support efficient quality control.
On the software side, significant progress was made in developing atmospheric tomography and AI-based forecasting. A new tomography approach was created that combines GNSS signal delays with external weather data to reconstruct 3D atmospheric fields such as humidity, temperature, pressure, and wind. This represents a step change in the level of detail and accuracy compared to traditional methods.
In parallel, advanced AI forecasting algorithms were designed and benchmarked against the latest international models. These algorithms were tested with multimodal data sources and validated in real-world pilot settings, confirming their ability to deliver more accurate and timely forecasts. During the project, the rapid evolution of AI weather models led us to strategically refocus from building a fully proprietary architecture to adapting and enhancing leading open models with our unique GNSS-based data. This ensured scientific relevance, competitiveness, and faster progress.
Together, these hardware and software achievements — reinforced by their successful use in pilots — demonstrate that the integrated system has reached a high level of technical maturity. The results provide a strong foundation for continued development and commercial scaling after the project, both in Europe and globally.
The hardware component of the system was advanced to a production-ready stage, with antenna design, electronics, and enclosure optimized for scalable manufacturing. Connectivity options were enhanced, and diagnostic features were built in to ensure reliability during large-scale deployment. Production testing processes were established to support efficient quality control.
On the software side, significant progress was made in developing atmospheric tomography and AI-based forecasting. A new tomography approach was created that combines GNSS signal delays with external weather data to reconstruct 3D atmospheric fields such as humidity, temperature, pressure, and wind. This represents a step change in the level of detail and accuracy compared to traditional methods.
In parallel, advanced AI forecasting algorithms were designed and benchmarked against the latest international models. These algorithms were tested with multimodal data sources and validated in real-world pilot settings, confirming their ability to deliver more accurate and timely forecasts. During the project, the rapid evolution of AI weather models led us to strategically refocus from building a fully proprietary architecture to adapting and enhancing leading open models with our unique GNSS-based data. This ensured scientific relevance, competitiveness, and faster progress.
Together, these hardware and software achievements — reinforced by their successful use in pilots — demonstrate that the integrated system has reached a high level of technical maturity. The results provide a strong foundation for continued development and commercial scaling after the project, both in Europe and globally.
The project delivered results that go well beyond the current state of the art in weather observation and forecasting. Unlike conventional approaches that rely mainly on sparse radiosonde launches and weather radars, Skyfora demonstrated how GNSS signals from existing telecom infrastructure can be transformed into a dense and continuous global weather observation network. This breakthrough enables the creation of 3D atmospheric data at higher resolution, lower cost, and broader coverage than has previously been possible.
Needs for further uptake:
- Continued demonstration and scaling of pilots to prove reliability in diverse climates and geographies.
- Access to markets and finance to support global deployment of sensor hardware and cloud-based data services.
- Ongoing research and model refinement to maintain leadership in rapidly evolving AI-based forecasting.
- Commercialization partnerships with telecom operators, energy providers, and public agencies to ensure integration into existing infrastructure.
- IPR support and internationalization to safeguard innovations and enable competitive global expansion.
Needs for further uptake:
- Continued demonstration and scaling of pilots to prove reliability in diverse climates and geographies.
- Access to markets and finance to support global deployment of sensor hardware and cloud-based data services.
- Ongoing research and model refinement to maintain leadership in rapidly evolving AI-based forecasting.
- Commercialization partnerships with telecom operators, energy providers, and public agencies to ensure integration into existing infrastructure.
- IPR support and internationalization to safeguard innovations and enable competitive global expansion.