Frost Risk Intelligence — Argamasilla de Alba
Case Study: How to predict frost events in orchards.
Frosts are critical events that can be devastating for the production of some crops. We present below an automated solution based on local sensor data collection and global ensemble weather forecast models to predict the risk of the next frost event. Our system is accurate up to 5 nights ahead. The honest forecast error (RMSE) stays between 1.4 and 1.8 °C, with fewer than 10 false alarms per year. To test our approach, we present this case study based on historical forecasts and sensor data in Argamasilla de Alba.
How we provide the forecast service:
- On-site sensor installation — local high-quality data that we use to build local models and calibrate weather forecasts.
- Calibrated forecast system — at least three global ensemble models, bias- and spread-corrected for your site and blended based on evidence.
- Real-time monitoring — a dashboard that can look like this one, refreshed every forecast cycle.
- Alert system — WhatsApp/SMS alerting you when it is advisable to take action.
Case study site: Argamasilla de Alba (SIAR CR07) — grapevine , critical temperature -1.5 °C.
Verified nights
Observed frost nights
Next-night error (°C)
Frost discrimination (AUC)
The model reproduces the observed values within its confidence band
A perfect point forecast is impossible, so we forecast the honest range instead. The blue band below is the model's own 90% confidence interval, verified against two seasons of data. Frost alerts are computed from probabilities, not from a model that can silently miss.
Reliable up to 5 nights before a frost comes
Forecast error grows from 1.4 °C the night before to 1.8 °C five nights out (CRPS 0.78 → 1.00 °C), and frost discrimination stays at AUC 0.96–0.99 across all five leads.
Beyond five nights only a single frost night is scoreable in the archive (the horizon limit) — the practical service window is 1–5 nights ahead.
Check how we can help you
Sensors installed in your orchard, forecasts built and operated through every frost season, alerts on your phone up to 5 days ahead — a service that answers for its numbers, orchard by orchard.
Explore the service →Every frost night, caught
Raw weather forecast models fail to predict local conditions. Our calibrated system instead issues alerts on all analysed frost nights:
- Performance — every reliability metric, measured on ~200 nights.
- The model — how the calibration and model blending work.
- The service — what we install, operate and answer for, orchard by orchard.