← 2026-06-16  |  2026-06-20 →

Cossatot River — Daily Analysis: 2026-06-19

Classification: correct Confidence: Physics model confidence remains stable. Accurate magnitude prediction on a headwaters-dominated event validates current Band 5 coefficients. Empirical model confidence should be reviewed due to overprediction of tier.

Event Summary

The physics model accurately predicted the peak stage (2.0% error) driven by headwaters rainfall, though it overestimated the rise speed by 3.8 hours.

Prediction vs Actual

Metric Predicted Actual Error
Peak height 3.09 ft 3.03 ft +0.06 ft
Total rise 0.33 ft

Band Contributions

Band Precip Predicted Rise Intensity Moisture
5 0.73" 0.31 ft MODERATE WET

Empirical Forecast

Headline: Watch — Recent rainfall is in the lower-bound range preceding rises to MEDIUM Settled outcome: None (reached None) LLM said headline was correct: False Notes: Empirical model predicted MEDIUM (>=3.4ft), but actual peak was 3.03ft (LOW). The headline was not borne out; the rise was smaller than anticipated.

Analysis

Today's event was driven by widespread, low-intensity rainfall across all bands, with the gauge recording 0" precipitation. The model correctly identified that the rise would be dominated by Band 5 (headwaters) contributions, predicting a peak of 3.09 ft against an actual peak of 3.03 ft. The magnitude prediction was excellent, falling well within the ±15-20% tolerance for a 'correct' classification.

The timing error of 3.8 hours (predicted 04:50 next day vs actual 20:00 same day) suggests the lag for Band 5 is currently too long for these specific moderate-intensity, wet-condition inputs. However, given the 'very_broad' hydrograph shape and the conservative principle of small adjustments, we will not adjust coefficients today. The model's ability to capture the magnitude of a multi-band, headwaters-dominated event with no gauge rain is a strong validation of the current Band 5 coefficient.

The empirical forecast predicted a 'MEDIUM' tier rise (threshold 3.4 ft), but the river only reached 3.03 ft ('LOW' tier). This indicates the empirical model may be overestimating the potential for rises in these moderate rainfall scenarios, or the percentile bands need refinement. No calibration changes are recommended for the physics model at this time.

Coefficient Adjustments

No changes made.

Notes

Physics model confidence remains stable. Accurate magnitude prediction on a headwaters-dominated event validates current Band 5 coefficients. Empirical model confidence should be reviewed due to overprediction of tier.

← 2026-06-16  |  2026-06-20 →