Classification: no_prediction Confidence: Model confidence remains stable. The previous event (May 23) validated the SATURATED tier coefficients and Band 1-5 responses. Today's lack of prediction aligns with the lack of rainfall input.
No prediction was generated for today due to zero QPE rainfall in all bands, despite an observed 0.79 ft rise attributed to residual flow from the previous day's significant event.
| Metric | Predicted | Actual | Error |
|---|---|---|---|
| Peak height | N/A | 4.11 ft | N/A |
| Total rise | — | 0.79 ft | — |
| Band | Precip | Predicted Rise | Intensity | Moisture |
|---|---|---|---|---|
The physics-based prediction engine generated no output because the QPE data showed 0.00" of rainfall across all bands for the entire 24-hour period. Without rainfall input, the model correctly produced no new peak prediction. However, the gauge recorded a daily high of 4.11 ft with a calculated rise of 0.79 ft from the daily minimum. Given that the previous day (May 23) was a 'correct' classification with a peak of 4.14 ft and widespread rainfall on saturated ground, today's observed 'rise' is almost certainly the tail end of the previous hydrograph rather than a new independent event.
The gauge minimum (3.32 ft) was reached late in the day (23:00), indicating the river was still high and receding from the previous peak. The 'Rise duration: 0.0 hours' and 'Hydrograph shape: sharp' are artifacts of how the daily max/min were calculated across the 24-hour window overlapping two events. Since the model did not predict a new rise (because there was no rain), it did not make an error in prediction logic. Classifying this as a false negative would be incorrect because the driver (rainfall) was absent. Therefore, the system functioned as designed by producing no prediction.
Because the prediction engine did not generate a peak to compare against the observed value, no coefficient adjustments can be derived from this specific 24-hour window. The previous event (May 23) already provided a successful calibration point for SATURATED conditions. This day serves as a confirmation that the recession limb was tracked, but offers no new data to tune rainfall-response coefficients.
No changes made.
Model confidence remains stable. The previous event (May 23) validated the SATURATED tier coefficients and Band 1-5 responses. Today's lack of prediction aligns with the lack of rainfall input.