Quantitative Precipitation Forecasting

What is it? ... Why is it important?

Introduction | | The Necessity of QPF | A Primer on Areal Precipitation |Point vs. Mean Areal Precipitation | BIBLIOGRAPHY

1. Introduction

I. The Role of Hydrometeorology in National Weather Service Operations Accurate forecasting of river conditions and timely issuance of flood watches and warnings requires the integration of meteorology and hydrology. Operational hydrology must respond to the dynamic nature of river systems. Because of the lag time which usually exists between a rainfall event and a basin response Figure 1 , forecasting river stages and flows using only observed data allows for a limited period of forecast reliability. To extend the lead time of river forecasts, information pertaining to anticipated weather is required. Consequently, meteorological information becomes a valuable tool in river forecasting.

The linkage between the scientific disciplines of meteorology, hydrology, and hydrometeorology is illustrated in Figure 2 . Hydrometeorology is the discipline which bridges meteorology and hydrology. Within the National Weather Service, hydrometeorology facilitates evaluation and conveyance of significant meteorological information to enhance hydrologic modeling efforts. The goal of this training tool is to provide an introduction to Quantitative Precipitation Forecasting (QPF) and illustrate the importance of precipitation forecasts in hydrologic operations.


2. The Necessity of QPF

Use of QPF in river forecasts allows for longer lead time and increased forecast reliability and usefulness. QPFs and the products generated using QPFs are important in flood and non-flood conditions and are used by flood control agencies, water managemnt districts and emergency management agencies ( Opitz et al. 1995, Krzystofowicz et al. 1993, CA DISASTER SURVEY 1995) In particular, forecasting flash floods and rapidly evolving floods which develop within 24 hours should be improved by the incorporation of QPF. In a 1994 study (Zevin 1994), it was determined that 75% of flash flood warnings are issued with less than 1 hour lead time and 50% are issued when a flood has already been observed. Not only does incorporation of QPF enable more reliable river forecasts to be issued, it also reduces the need for forecast revisions during a precipitation event. The goal of QPF is to facilitate the timely dissemination of potentially life- saving information to emergency management agencies and the public (Figure 3) . The ability to extend the forecast period resulting from QPF allows for better, more timely responses to river forecasts.


3. A Primer on Areal Precipitation
The goal of this section is to introduce the concept of mean areal precipitation which is important in the determination of the basin average QPF values needed for input into river forecast models. Whereas point precipitation is the depth of rainfall at a discrete location, areal precipitation describes the depth of rainfall over an area. Usually, in the National Weather Service, precipitation is forecast in 6 hour time periods, out to 24 hours. The impact of using a point precipitation values rather than a mean areal precipitation value is illustrated in example 1.


4. Point vs. Mean Areal Precipitation

Example 1: Consider the Sample River basin show in Figure 4 . An average of 1 inch of rainfall in 6 hours over the basin will generate 0.5 inch of runoff and cause the increase in flow. An average of 0.5 inch of rainfall in 6 hours over the basin will generate 0.25 inch of runoff and also produce a hydrologic response. Understandably, the responses will be different since the QPF differs by a factor of 2, as shown in Figure 5 .

Consider the precipitation forecast for Location 1 and Location 2
Forecast 6 Hour Precipitation


1.0 inch


0.5 inch

If the maximum precipitation forecast for the basin (1 inch) were used as input into the hydrologic model, the peak flow forecasted would be 4050 cubic_feet/second. If a simple arthimetic average of precipitation at the two forecast locations were used, the mean areal precipiation computed would be 0.75 inch. The 0.375 inch of runoff that would be generated would cause peak flow of 3037 cubic_feet/second to be forecast for the Sample River, nearly 25% less than the peak flow forecast from a QPF of 1 inch. Current National Weather Service River Forecast System hydrologic models are designed to calculate flows using mean areal precipitation values. Consequently, use of a maximum precipitation value rather than a mean areal value can lead to overestimation of flow and stage in river forecasts.


5.BIBLIOGRAPHY *Recommended Reading
*Arkell, R.E. and R.E. LaPlante, 1991: Verification of River Stage and Quantitative Precipitation Forecasts. Eastern Region Technical Attachment, No. 91-5A, National Weather Service, NOAA, U.S. Department of Commerce, 7 pp.

Amburn, S.A., 1994: Quantitative Precipitation Forecast Training and Methodology at NWSO Tulsa, Oklahoma, Tulsa National Weather Service Office, National Weather Service, NOAA, U.S. Department of Commerce, 19pp.

Chow, V.T., D.R. Maidment, and L.W. Mays, 1988: Applied Hydrology. McGraw-Hill, 572pp.

Cooperative Program for Operational Meteorology, Education, and Training, 1995: COMET Hydrometeorology Course Notes.

Fenbers, M.J., D.F. Innes, and M.J. Van Tress, 1995: WinQPF User's Guide. NOAA Eastern Region Computer Programs NWS ERCP - 25 MC, National Weather Service, NOAA, U.S. Department of Commerce, 39pp.

Hart, J.A., and J. Korotky, 1991: The SHARP Workstation version 1.50: a Skew-T/Hodograph Analysis and Research Program for the IBM and Compatible PC. National Weather Service, NOAA, U.S. Department of Commerce, 30pp.

Holton, J.R., 1979: An Introduction to Dynamic Meteorology. Academic Press, 391pp.

*Junker, N.W., 1992: Heavy Rain Forecasting Manual, National Weather Service Training Center, National Weather Service, NOAA, U.S. Department of Commerce, 91pp.

Krzysztofowicz. R., 1993: Probabilistic Hydrometeorological Forecasting System: A Conceptual Design in Postprint Volume, Third National Heavy Precipitation Workshop, 16-20 November 1992. NOAA Technical Memorandum NWS ER-87, National Oceanic and Atmospheric Administration, U. S. Department of Commerce, 388pp.

*Krzysztofowicz. R., 1992: Bayesian Correlation Score: A Utilitarian Measure of Forecast Skill, Mon. Wea. Rev., 120, 208-219.

*Krzysztofowicz, R., W. J. Drzal, T.R. Drake, J.C. Weyman, and L.A. Giordano, 1993: Probabilistic Quantitative Precipitation Forecasts for River Basins, Wea. Forecasting, 8, 424-439.

Opitz, H.H., S.G. Summer, D.A. Wert, W.R. Snyder, R.J. Kane, R.H. Brady, P.M. Stokols, S.C. Kuhl, and G.M. Carter, 1995: The Challenge of Forecasting Heavy Rain and Flooding Throughout the Eastern Region of the National Weather Service. Part II: Forecast Techniques and Applications, Wea. Forecasting, 10, 91-104.

Scofield, R., 1995: Hydrometeorologic Workshop - Satellite Applications, in COMET Hydrometeorology Course Notes, Cooperative Program for Operational Meteorology, Education, and Training, January.

*Scofield, R. and J. Robinson, 1990: The "Water Vapor Imagery/Theta-E Connection" with Heavy Convective Rainfall, Satellite Applications Information Note 90/7. NOAA/NESDIS/Satellite Applications Laboratory, Washington, D.C., 7pp.

Shinko, J.A., 1995: An Introductory Guide for the New Quantitative Precipitation Forecasters. Eastern Region Training and Applications Notes, ER TAN-8, National Weather Service, NOAA, U.S. Department of Commerce, 12pp.

*Zevin, S.F., 1994: Steps toward an integrated approach to hydrometeorological forecasting services, Bull. Amer. Meteor. Soc., 75, 1267-1275.