National comparisonWFO breakdown →

The forecaster’s edge

Human vs. the Machine

Where does the human forecast add value? Follow the NWS high and low temperatures, then measure NDFD against the National Blend and other models at the same stations.

NDFDNWS forecaster-edited gridLoading the station network…
Temperature
Score window

Loading paired station verification…

Where NDFD adds value
Subtracts value0°F Adds value

Model error − NDFD error, in °F. Same station-days and lead. Select a station to see its temperature ribbon and the forecasts behind the score.

Daily minimum → maximum · °F

NDFD low → high Model low / high◆ Observed low / high
Forecast values and issue timing
The temperature heat mapSelect any cell to inspect that station-day

The tiles above show all paired stations for the selected model. These filters apply to the map and heat map.

The score columns use the selected window. Heat-map cells show the most recent 21 verified calendar dates plus 7 future dates; pending forecasts have an outline. A dash means no available forecast or accepted observation. Purple / green are reserved for value; temperatures use Ryan’s original Fahrenheit colors.

How the comparison works