Forecast Accuracy Analysis
Track how well our predictions match actual weather conditions
Forecast Performance by Parameter
| Weather Parameter | Dayboro.au | BOM | Inigo Jones | Best |
|---|
Model Performance Comparison
90-Day Accuracy Trend
Each point is a 14-day rolling composite score so single-day noise from rare events does not distort the line. A static snapshot can be cherry-picked. A trend cannot.
Dayboro Model and BOM blend temperature, rainfall, wind, UV and humidity. Inigo Jones forecasts only temperature and rainfall, so its line uses just those parameters.
How We Calculate Accuracy
+Temperature Accuracy
Graduated scoring with realistic thresholds. A 5°C error is considered significant.
- MAE: Average forecast error magnitude
- Bias: Positive = forecast too high
- Accuracy %: Based on MAE vs threshold
Precipitation Accuracy
Categorical and binary accuracy with proper rain/no-rain detection.
- Rain event threshold: 0.5mm
- Categories: None, Light, Moderate, Heavy
- POD: Probability of Detection
- FAR: False Alarm Ratio
Composite Score
Weighted average of all metrics for overall accuracy.
- Temperature: Weight 3.0
- Precipitation: Weight 2.5
- Wind: Weight 1.5
- UV Index & Humidity: Weight 1.0 each
Data Sources
Daily comparison of forecasts against recorded weather station data.
- Dayboro.au: Local Dayboro Model
- BOM: Bureau of Meteorology
- Inigo Jones: Temperature & rainfall only
Last updated: --
Accuracy metrics are calculated by comparing forecasts with actual recorded conditions
The data is more accurate — now get it delivered
Members get 10 email reports covering daily weather, solar output, soil conditions, and market prices. You pick which ones land in your inbox.
$3.95/month or $39/year • Cancel anytime