General Ensemble Theory for Streamflow Forecasts
NWS River Forecast Centers across the U.S. utilize a system called the National Weather Service River Forecast System (NWSRFS) to generate short term flood forecast guidance at nearly 4000 locations. NWSRFS is a collection of data processing, hydrologic simulation models, and associated interfaces. NWSRFS has three subsystems (1) calibration, (2) operations, and (3) ensemble. These three subsystems share architectural and information resources.
NWS hydrologists use the Calibration System to select appropriate models, develop and adjust parameters and fit the resulting simulated flow to that historically observed. This also includes integrating those parameters and supporting realtime data into the Operations System.
The Operations System is used daily by NWS river forecasters to produce short term river discharge forecasts and guidance. Real time data are used to update model states (soil moisture, snow water, etc.) Forecasts of precipitation and air temperature are included in the simulation of the “future” to develop the streamflow forecast. Simulations are deterministic (single value) and extend 3 to 10 days into the future. The CNRFC uses a standard 5 day horizon for operational forecasts. You can view these forecasts at:
The Ensemble System utilizes the same models and states as the Operational System, but in a slightly different fashion. In this case, the Operations System is run, starting with today’s current conditions, once for every year of historical calibration data. Most watersheds in the CNRFC area are calibrated with between 30 and 40 years of data. Thus, the Ensemble System will produce 30 to 40 possibilities. These possibilities, called “traces”, can extend for up to one year into the future.
This figure shows the “traces” for the West Fork Carson River at Woodfords starting on June 15th, 2005. There are 49 traces, one for each year starting with 1961 and continuing through 1999. Each trace has a rather simple interpretation best described by an example. Starting with the model states (soil moisture, snowpack, etc.) on June 14th, 2005, the trace for year 1961 represents the model simulation using the precipitation and temperature that occurred from June 15th through September 1st of 1961. In practice, the June 15th, 2005 would be replaced by “today”. Note that the simulation is based on conditions as of June 14th, 2005.
Once traces have been generated, the true power and flexibility of the Ensemble System becomes evident. You can already see that the magnitude and variability of the simulated streamflow is greatest in the first half of the time period. For any period within the duration of the traces (June 15 – September 1 above), you can develop a probability distribution. Using this distribution you can answer risk-based questions. For example, what 5-day volume during the first week of July will be exceeded 9 times out of 10? If you’re concerned about low flows, you might ask what flow rate has 90% chance of being exceeded throughout this period. The possibilities are endless. This is why the CNRFC supports a user interface that allows for the development of custom products.
For assistance in selecting options and interpreting the products available through the AHPS / ESP Trace Analysis web page, please see the ESP Trace Analysis Interface (Making Selections and Interpreting Results).