← 2026-10-02  |  2026-10-15 →

Hailstone (Upper Buffalo) — Daily Analysis: 2026-10-14

Classification: partial Confidence: The physics model's overprediction in dry conditions continues to be a concern, but the conservative adjustment helps mitigate risk. The empirical model's accuracy in this scenario reinforces its value as a cross-check for the physics model.

Event Summary

Physics model overpredicted peak magnitude by 82% and misjudged timing by 8.9 hours, while the empirical model correctly indicated no significant rise despite a modest 2.8 CFS increase.

Prediction vs Actual

Metric Predicted Actual Error
Peak CFS 7 CFS 4 CFS +3 CFS
Total rise — 2.8 CFS —

Empirical Forecast

Headline: No rise indicated by recent rainfall LLM said headline was correct: True Notes: The empirical model correctly predicted no rise above the low threshold, which was borne out by the actual peak of 3.77 CFS.

Band Contributions

Band Zone Precip Predicted Rise Intensity
1 ? 1.68" 3 CFS HEAVY

Analysis

The physics model predicted a peak of 6.9 CFS, which is an 82.2% overprediction compared to the actual 3.77 CFS peak. The timing error of 8.9 hours indicates a significant lag mismatch, likely due to the model assuming a longer response time than the steep, small Boxley basin actually exhibits. The empirical forecast model correctly identified that the rainfall was insufficient to cause a meaningful rise above the low threshold (500 CFS), which aligns with the observed peak remaining well below operational thresholds.

The antecedent moisture condition was 'VERY DRY', which should have suppressed runoff response. The model's overprediction suggests that the response coefficient or the handling of dry conditions is still too generous. The empirical model's accuracy in this low-flow, dry-condition scenario highlights its strength in filtering out false positives, a critical feature for this gauge given its small catchment and high sensitivity to false alerts.

Graded Prediction

Predictions issued today (archive): 19 active issuance(s) from 2026-10-14T10:21:02.130249+00:00 to 2026-10-15T04:21:02.472295+00:00; issued peaks ranged 0.9 to 6.9 CFS. The prediction graded below is the HIGHEST issued peak of the day (issued 2026-10-14T20:21:02.110534+00:00), not the last card of the day.

Coefficient Adjustments (SUGGESTED by the analyzer, NOT applied - grade-only mode)

Band Change Reason
1 -10% Conservative reduction to address the persistent overprediction trend in dry conditions, while maintaining a buffer for potential future wet events.
← 2026-10-02  |  2026-10-15 →