Re-graded on November 4, 2026 under the new grading rule (changelog 2026.11.4.2). The original graded a prediction issued after the creek had already reached its crest.
Classification: partial Confidence: Confidence remains moderate; the underprediction highlights a gap in the model's representation of runoff efficiency under dry conditions, which needs to be addressed through coefficient tuning. The empirical model's accurate threshold-based forecast is a positive indicator for operational reliability.
The model correctly predicted a rise but underestimated the peak magnitude by 27%, likely due to insufficient runoff coefficients for the dry but intense rainfall event.
| Metric | Predicted | Actual | Error |
|---|---|---|---|
| Peak height | 2.06 ft | 2.81 ft | -0.75 ft |
| Total rise | — | 2.22 ft | — |
| Band | Zone | Precip | Predicted Rise | Intensity | Moisture |
|---|---|---|---|---|---|
| 1 | ? | 1.18" | 0.11 ft | MODERATE | VERY DRY |
| 2 | ? | 1.81" | 0.97 ft | INTENSE | VERY DRY |
Headline: No rise indicated by recent rainfall Forecast graded: the last forecast before the creek reached its crest, issued 11:18 at 0.98 ft (24 forecasts today) Outcome: LOW (3 ft) was not reached within 12 h; the gauge's high after this forecast was 2.81 ft today, so this call held. Notes: The empirical model correctly indicated no rise to the 'low' threshold (3.0 ft), as the actual peak (2.81 ft) stayed below this level, consistent with the model's 1.92% probability of exceeding it.
The event occurred on November 3, 2026, with a 2.22 ft rise from a very dry antecedent condition. The hydrograph shape was broad, indicating a combined response from both bands, with Band 2 (Falling Water) contributing significantly more rainfall (1.81") and higher intensity than Band 1 (1.18"). The model predicted a peak of 2.06 ft, while the actual peak was 2.81 ft, resulting in a -26.7% error. This underprediction suggests that the current runoff coefficients may be too conservative for 'VERY DRY' moisture conditions, especially when rainfall intensity is moderate to heavy. The timing was well-captured, with only a 0.1-hour error, indicating that the lag and dispersion parameters are appropriate for this event.
Predictions issued today (archive): 23 active issuance(s) from 2026-11-03T07:20:03.109149+00:00 to 2026-11-04T05:20:02.125120+00:00; issued peaks ranged 0.57 to 2.86 ft. The creek reached its crest at 2026-11-03T12:15:00-06:00. The prediction graded below is the HIGHEST peak issued before then (issued 2026-11-03T17:20:02.248811+00:00, 0.9 h ahead). 12 later issuance(s) don't count: the creek was already there.
| Band | Change | Reason |
|---|---|---|
| 1 | +5% | Underprediction of peak contribution suggests a slight increase in response coefficient is needed to capture the runoff from moderate-intensity rainfall on dry ground. |
| 2 | +10% | Band 2 contributed more rainfall and intensity, yet the overall peak was underpredicted, indicating its response coefficient should be increased more significantly to reflect the actual runoff. |
Confidence remains moderate; the underprediction highlights a gap in the model's representation of runoff efficiency under dry conditions, which needs to be addressed through coefficient tuning. The empirical model's accurate threshold-based forecast is a positive indicator for operational reliability.