📊 Forecast Evaluation
Transparency and accountability matter! This page tracks the accuracy of our weekend ski forecasts by comparing predictions against actual observed conditions.
Data is collected from SNOTEL stations, NWS observations, and resort reports. Evaluation helps improve future forecasts and builds trust with readers.
Season Statistics (2024-2025)
Weekly forecasts issued
Forecasts within predicted range
Average snowfall error
Statistics will be updated as we collect more forecast evaluation data throughout the season.
Recent Forecast Evaluation
| Forecast Date | Ski Area | Forecast | Actual | Error | Accuracy |
|---|---|---|---|---|---|
|
Evaluation data will appear here after the first forecast weekend. Run the evaluation scripts after collecting observation data. |
|||||
By Ski Area Performance
Mt. Baker
Forecasts Verified: --
Mean Absolute Error: -- inches
Within Range: --/-- (--%)
Stevens Pass
Forecasts Verified: --
Mean Absolute Error: -- inches
Within Range: --/-- (--%)
Crystal Mountain
Forecasts Verified: --
Mean Absolute Error: -- inches
Within Range: --/-- (--%)
Other Areas
Evaluation data for Snoqualmie, White Pass, and Blewett Pass will be added as we collect observations.
Methodology
Data Sources
- SNOTEL Stations: Automated snow and weather stations from NRCS
- NWS Observations: National Weather Service station data
- Resort Reports: Official snow reports from ski areas
Evaluation Process
- Forecast data is saved at time of posting (stored in data/forecasts/)
- After the forecast period, actual observations are collected
- Automated scripts compare forecast vs. actual values
- Evaluation reports are generated and posted here
Accuracy Metrics
- Within Range: Did actual fall within the forecast range?
- Absolute Error: |Forecast - Actual|
- Percentage Error: (|Forecast - Actual| / Actual) × 100
- Mean Absolute Error (MAE): Average error across all forecasts
Why Evaluation Matters
Forecast evaluation serves multiple purposes:
- Accountability: Being transparent about forecast accuracy builds trust
- Improvement: Identifying patterns in errors helps refine future forecasts
- Learning: Understanding which weather patterns are harder to forecast
- Comparison: See how different ski areas verify relative to each other
No forecast is perfect, but consistent evaluation helps us get better over time!