Feature Analysis
Ablation studies, feature importance, and data characteristics
Ablation Study — Feature Sets
RMSE comparison across 7 feature configurations (lower is better)
R² by Feature Set
R² comparison across 7 feature configurations (higher is better)
Key Finding: Lagged Yield Dominance
The operational model relies heavily on lagged yield history rather than weather features alone. Removing lagged yield increases RMSE from 0.658 to 0.889 (R² drops from 0.730 to 0.507). This suggests the model memorizes past yields rather than learning weather-yield relationships — a critical limitation for true early-warning capability.
Soil Attributes (10 features × 2 depths)
Soil properties at topsoil (0–30cm) and subsoil (30–100cm) depths
Available Water Capacity
Bulk Density
Clay Content
Sand Content
Silt Content
Soil Organic Carbon
Total Nitrogen
Total Phosphorus
pH (Calcium Chloride)
Exchangeable Cation Sum
Data Coverage
Regional Climate Summary
| Region | Rainfall | Avg Temp | Evaporation | Solar |
|---|---|---|---|---|
| NSW | 520mm | 21.3°C | 4.8mm | 22.1MJ/m² |
| QLD | 610mm | 24.7°C | 5.6mm | 24.3MJ/m² |
| SA | 310mm | 22.8°C | 6.2mm | 23.5MJ/m² |
| TAS | 980mm | 15.6°C | 3.4mm | 18.7MJ/m² |
| VIC | 580mm | 17.4°C | 4.1mm | 20.2MJ/m² |
| WA | 420mm | 23.1°C | 5.8mm | 23.8MJ/m² |
Forecast Window Details
Each window aggregates daily weather from May through the end month, capturing conditions during critical growth stages.
May–Jun
Earliest forecast — 12 weeks before harvest
~12 weeks before harvest
May–Jul
Early season outlook
~16 weeks before harvest
May–Aug
Mid-season prediction
~20 weeks before harvest
May–Sep
Late season forecast
~24 weeks before harvest
May–Oct
Pre-harvest — most accurate
~28 weeks before harvest