A good amount. The right place.
Rain Bullseye
Was the rainfall forecast on target, or did the rain land nearby?
Reading the native rainfall verification…
Every radius tells a different story.
| Footprint | Forecast | Analysis | Bias | RMSE | Forecast rain area | Analysis rain area | Matched area |
|---|
Each circle is a cumulative neighborhood, not an annulus. Station values are analysis-grid cells, not rain-gauge measurements. Regional mean bias can cancel wet and dry errors; RMSE retains their magnitude.
Amounts, placement, and verification
Scores come from native NDFD forecasts against URMA or Stage IV, with fractional boundary-cell areas and the exact same valid footprint. Measurable rain means at least 0.01 inch. Forecasts must have been available before the accumulation starts.
The placement diagnostic compares rainfall-weighted centers of paired, displayed analysis-grid cells whose centers fall inside the selected radius. It uses spherical cell areas and excludes missing or sub-threshold rainfall. Unlike the scoring table, this sampled diagnostic does not use fractional circular boundary cells. It requires at least 95% paired sampled area and rain in both fields. An offset is not proof of one displaced storm, especially with multiple rain areas.
The headline uses a 0.10-inch mean-bias threshold and, when available, a 5-mile center offset. These are interpretation guides, not a new accuracy grade. Probabilities of precipitation are not scored.