← 2026-11-03

Richland Creek — Daily Analysis: 2026-11-04

Classification: correct Confidence: Confidence remains stable; the model correctly handled a zero-rainfall day, reinforcing the robustness of the 'no rise' logic. The recent upward adjustment on 2026-11-03 suggests the model is well-tuned for active events, and today's performance confirms it does not over-predict in quiet conditions.

Event Summary

The system correctly identified a quiet day with zero rainfall, though the gauge showed a minor 0.59 ft residual rise not attributed to precipitation.

Prediction vs Actual

Metric Predicted Actual Error
Peak height 1.43 ft 2.02 ft -0.59 ft
Total rise — 0.59 ft —

Band Contributions

Band Zone Precip Predicted Rise Intensity Moisture

Empirical Forecast

Headline: No rise indicated by recent rainfall Forecast graded: the day's strongest forecast, issued 00:18 at 2.06 ft (24 forecasts today) Outcome: LOW (3 ft) was not reached within 12 h; the gauge's high after this forecast was 1.99 ft today, so this call held. Notes: The empirical forecast correctly indicated no rise to the low tier (3.0 ft) within 12 hours, consistent with the gauge staying below that threshold despite the minor residual rise.

Analysis

Today's event was classified as 'correct' because the predictor issued an explicit 'no rise' call, which aligns with the zero-rainfall observation (0.00" in both Band 1 and Band 2) and the 'NORMAL' moisture tier. The gauge registered a peak of 2.02 ft, representing a 0.59 ft rise above the daily minimum of 1.43 ft. However, the hydrograph shape is described as 'sharp' with a duration of 0.0 hours, suggesting this was likely a residual effect from the previous event on 2026-11-03 (which had significant rainfall) or a non-rain-related fluctuation, rather than a new storm response. Given the lack of rainfall input, any rise is outside the scope of the rainfall-runoff model's predictive capability for new events.

The recent history shows a trend of the model correctly handling low-rainfall days (2026-10-11, 2026-10-31, 2026-11-01) and recently adjusting upward for a high-rise event (2026-11-03) where it underpredicted. Today's 'correct' grade reinforces that the model is appropriately suppressing rise predictions when QPE is zero. Since no rainfall occurred, there is no physical basis to adjust the band response coefficients, as the error is not a mismatch between rain and runoff, but rather the presence of runoff without rain. Therefore, no calibration changes are warranted.

Graded Prediction

Predictions issued today: NONE - the predictor never issued an active rise prediction today. It is graded as an explicit 'no rise' call at the day's minimum level; a real rise today therefore counts as a false_negative.

Coefficient Adjustments

Grade-only mode: no adjustments suggested; calibration is fit offline.

Notes

Confidence remains stable; the model correctly handled a zero-rainfall day, reinforcing the robustness of the 'no rise' logic. The recent upward adjustment on 2026-11-03 suggests the model is well-tuned for active events, and today's performance confirms it does not over-predict in quiet conditions.

← 2026-11-03