How Runoff Risk Works - NOAA Runoff Risk
Runoff Risk guidance is produced from a hydrologic model ran in real-time at the North Central River Forecast Center (NCRFC). The model is ran continuously on a 1-hour time-step across 4km x 4km grid domains for each of the four active states: MI, MN, OH, and WI. There are currently two forecast runs per day each going out 10 days into the future. To enable the model to be ran across large sections of the country in real-time the required forcings are limited to only observed and forecast temperatures and precipitation. The model continuously simulates soil moisture, soil temperature (frozen ground), runoff, and snowpack for each grid cell on the hourly basis.
To maintain accurate model states a fresh carryover run using only observed forcings is conducted each night. This run starts back 10 days and stops at 00Z (00 UTC) of the current day. Observed temperature forcings used are Real-time Mesoscale Analysis (RTMA) or Unrestricted Mesoscale Analysis (URMA) grids. Observed precipitation is derived from stage-IV quantitative precipitation estimate (QPE) created by the National Weather Service (NWS) River Forecast Centers (RFCs).
Starting from the most recent carryover model states, the forecast runs use observed forcings until the model T-zero hours where they transition to forecast forcings. Currently there are two morning forecast runs per day. The first is based on the 06Z forecast cycle and the second is a bit later based on the 12Z forecast cycle. Each forecast run relies on forecast temperatures from the NWS National Blend of Models (NBM) model suite. For forecast precipitation, or QPF (quantitative precipitation forecast), Runoff Risk still relies on 7 days of QPF from the NWS Weather Prediction Center (WPC). Note that this QPF guidance is still in 6-hr accumulations and is evenly disaggreated to hourly values for the hydrologic model.