River Forecast Center Overview and Climate Requirements

Kevin Werner   [ Listen to the speech ]

2nd NCPO-NWS Dialogue, Silver Spring, 4 Jan 2006

    RFC Overview

  • NWS Hydrology Mission:

      - Provide river and flood forecasts and warnings for the protection of lives and property.

      - Provide basic hydrologic forecast information for the nation’s environmental and economic well being.

  • River Forecast Center Functions:

    •   - Continuous hydrometeorological data assimilation, river basin modeling, and hydrologic forecast preparation.

      - Technical support and interaction with supported and supporting NWS offices.

      - Technical support and interaction with outside water management agencies and users.
      - Applied research, development, and technological implementation to facilitate and support the     above functions.

    Water Supply Forecasts

  • Water Supply Forecasts issued jointly by RFCs and NRCS for western US basins.

  • Water Supply Forecasts may directly apply climate forecasts
  • Existing procedure for incorporating climate forecasts into ESP was developed largely by OHD
  • Little effort by NOAA climate community to apply climate forecasts to water supply forecasts

Historical Synopsis of Climate In Water Resources Forecasting

(credit: Dave Brandon)

1917–Church,J.E. The first documented application to forecasting using correlation of snowpack to water level in Lake Tahoe
1943-USWB. Created bi-monthly 30-day weather Internal outlooks, and went public in 1953.
1947-USWB/SCS. Started publishing seasonal water supply outlooks.
1955-CBIAC Report. Evaluated use of 30-day outlooks in forecasting Columbia Streamflow…potential could be great but there was little/no skill.
1958-USWB. Created seasonal 90-day internal outlooks and went public in 1974 (temp) and 1978 (precip).

1964-CBIAC. Follow-up Report Forecast skill improving but hard to apply broad forecasts to specific basins.
1974-NWS. Seasonal 90-day temperature forecasts release to the public followed by precipitation forecasts in 1978.
1976-Marron(NRCS). Began using SOI in forecasts for Lake Tahoe
1977-Schaake, J. (NWS). Used 30-day precipitation outlook to remove a series of anti-analogs in ESP.
1987-Croley/Hartmann. Used climate outlooks subjectively to alter ESP traces in forecasting Great Lake Levels.
1995-Rundquist, L. Developed ESP post weighting scheme.
1988-Perkins, T.(NRCS). Began using SOI as predictor in lower Colorado.
1989-Cayan/Peterson. Investigated El Nino and western streamflow
1994-Hartman, (NWS) Investigated using SOIs at CBRFC
1995-CPC. Begins issuing new climate format, with tercile probability anomalies for 13 overlapping months.
1997-Mantua et al. Development of PDO
1997/1998-El Nino spurred variety of research
1998-Brandon, D. (NWS). Began using SOIs in preliminary Outlooks issued in the fall.
2000-Perica, S. (NWS). Developed CPC pre-adjustment technique to be used in NWSRFC ESP.

  • Two major tools:

      - Statistical Water Supply (SWS)

    • Regression based approach

    • Based on snow observations, antecedent conditions, climate conditions (e.g. ENSO), etc.

      - Ensemble Streamflow Prediction (ESP)

    • Based on continuous calibrated hydrologic model

    • Calibration requires long record of streamflow, precipitation, and temperature

    Climate Requirements

  • Increased skill in medium to long range precipitation forecasts - Castuccio et al (1981) estimated that a 6 percent improvement in forecasting accuracy would produce $36.5 million in annual benefits to agriculture and hydropower in the west

  • New methods for incorporating climate forecasts

      -  Must be based on data that is available when forecasts are made

  • New methods for incorporating and assimilating data into hydrologic forecasts.

  • Improving forecast verification to validate impact of climate forecasts

  • Expanding and improving use of water supply forecasts (including ensemble based decision support systems)

  • Longer term forecasts (e.g. 2 year and beyond) – RFCs are increasingly being asked to produce water supply outlooks for 2 years, sometimes even longer.

  • Data – High quality, long duration, high frequency records of precipitation, streamflow, and temperature are necessary for calibration and real-time operation.
     

   A (partially) Successful Collaboration

        2002-2004 CBRFC/CIRES/CDC partnership to improve medium range forecasts

  • Incorporated probabilistic meteorological model output into ESP

  • Documented improvements in medium range streamflow forecasts (Werner et al, 2005)

  • Implemented in CBRFC operations

  • No direct national implementation although concepts have helped drive national software development

      Elements of Successful Collaboration:

  • Focus on mutually interesting problem:

      - Directly address operational requirement(s).
      -     Also address relevant research question(s).

  • Early and active NWS involvement: Must include both RFC or WFO and OHD as active participants.

  • Proposed solutions must be possible in real-time NWS forecasting environment.

Contact Kevin Werner