Conclusions and research opportunities have been summarized after each main section. Overall, the quadrennium 1991--1994 has been marked by considerable advances in the understanding and modeling of fundamental elements of flood forecast and warning systems. But probably the most distinguishing trait of the period has been efforts expanded on integration, or coupling, of elements into systems. I venture to predict that this is just the beginning of a fruitful research path toward improving our forecasting capabilities, through optimal integration of all predictive information, and increasing societal benefits of hydro-meteorologic science and services, through optimal coupling of flood forecasts with warning decisions and emergency response plans.
Acknowledgments. Thanks are extended to Margaret A. Heritage and Patricia J. Reiman for assistance in retrieving references. This article is based upon work supported by the National Science Foundation under Award No. BCS-9016979, ``Models of Warning Systems for Natural Hazards,'' and by the COMET Program of the University Corporation for Atmospheric Research, under a cooperative agreement with the National Oceanic and Atmospheric Administration, Subaward No. UCAR-S94-43833, ``Probabilistic River Stage Forecasting.''