Forecasting Deep Dives
Technical deep dives into how physics-based solar forecasting works — from weather models to transposition to accuracy analysis.
← Back to blog- 1
Physics-Based vs ML-Based Solar Forecasting — Why the Approach Matters
Two fundamentally different approaches to predicting solar production. One models the physics, the other learns from data. Here's why the distinction matters for your forecast.
- 2
Where Weather Forecasts Come From: NWP Models Explained for PV Owners
Your solar forecast is only as good as its weather input. Here's how numerical weather prediction models work, what they get right, and where they fall short for solar energy.
- 3
Why Cloud Cover Is the Hardest Variable in Solar Prediction
Temperature is predictable. Sun position is calculable. But clouds? Clouds are chaos theory in action — and the single biggest source of solar forecast error.
- 4
The Irradiance Journey: From Space to Your Solar Panel in 8 Steps
Follow a beam of sunlight from the edge of Earth's atmosphere through scattering, absorption, clouds, and reflection until it reaches your tilted panel and becomes electricity.
- 5
Transposition Models: How Volcast Calculates What Hits Your Tilted Panel
Weather data tells us what reaches a horizontal surface. Your panels are tilted. The mathematical bridge between the two is called transposition — and the choice of model matters.
- 6
Nowcasting vs Day-Ahead vs Weekly — Accuracy at Every Time Horizon
A 15-minute forecast and a 7-day forecast are fundamentally different problems. Here's what each time horizon can realistically deliver and where accuracy breaks down.
- 7
Why New Installations Don't Need Training Data With Physics-Based Models
Most forecast tools need months of history before they work. Physics-based forecasting works from day one. Here's why — and what that means for new PV owners.