%2520(1).png&w=3840&q=75)
The energy sector runs on decisions made in narrow windows – when to store, when to sell, when to curtail. Those decisions depend on one thing above all else: knowing how much solar energy is actually arriving at the panel, right now and over the past hours. OpenWeather's Solar Irradiance and Energy Prediction service has just received a significant upgrade to the data engine behind it, and for existing users the benefits land immediately with nothing to change on their end.
What has changed
The core update is the integration of satellite observation data into the irradiance model. Previously, the service drew on numerical weather modelling and reanalysis datasets. Now, satellite-derived measurements are fused directly into the data pipeline, giving the model a real-world observational anchor that atmospheric modelling alone cannot provide. Cloud cover, which is the dominant source of irradiance error in operational forecasting, is now constrained by what satellites actually see, not just what models estimate.
The impact is measurable. Hourly Global Horizontal Irradiance (GHI) accuracy has improved significantly, with root mean square deviation (RMSD) falling by 10–20% depending on location. For solar asset operators and energy traders, this translates directly into tighter generation estimates, fewer dispatch errors, and more confident trading positions.
Real-time data has also become substantially more current. The update window for live irradiance data has been reduced considerably, meaning the gap between what is happening at a site and what the API returns is now much shorter than before. For grid balancing and intraday trading applications, where stale data erodes the value of any model output, this matters.
The update also significantly improves the underlying spatial resolution of the dataset. Geospatial resolution now ranges between 3–8 km depending on satellite viewing geometry.
Performance has also improved at the infrastructure level, with Solar Irradiance API responses now up to 2 times faster.
No integration changes required
For teams already using the Solar Irradiance API, Solar Irradiance History Bulk, or Solar Panel Energy Prediction endpoints: nothing changes. API endpoints, call logic, authentication, and response structure are all identical. The improved data flows through the existing integration transparently. There is no migration, no versioning, no re-certification of pipelines. Users who query the API today will simply receive more accurate data than they did yesterday.
Coverage
The updated model currently covers Europe, Central Asia, and Africa - regions where solar deployment is accelerating and where accurate irradiance data is increasingly operationally critical. Coverage is being extended: Asia and the United States are planned additions for later this year, bringing the satellite-enhanced accuracy to two of the world's largest solar markets.
The upgraded satellite-enhanced dataset is currently available for the 2004–2026 period.
Why it matters
The solar industry is at the point where the difference between good irradiance data and precisely accurate irradiance data is a commercial question, not just a technical one. A 10–20% improvement in GHI RMSD is not just a marginal gain. At scale, across a portfolio of assets or across an energy trading book, it compounds into meaningfully better outcomes.
OpenWeather's Solar Irradiance service already covers historical data going back to January 1979, 15-minute and hourly resolution, DNI, DHI and GHI parameters for both clear-sky and cloudy-sky models, and direct solar panel energy output prediction. This update makes all of that more accurate, more timely, and immediately available, without any work required from the teams already relying on it.