The advances subject to this review are those published by U.S. scientists during, or just prior to, the quadrennium 1991-1994, to the extent I have been able to trace them. With respect to flood warning systems, the review intends to be all encompassing. With respect to flood forecast systems, the delineation of scope required a decision. A narrow perspective might include only those studies in which forecasts were produced operationally, or at least were simulated under field-like conditions, and were verified against actual observations. A broad perspective might include all science of hydrometeorology that contributes the knowledge used in forecasting. This review is selective: it covers hydrologic forecasting models plus two inputs which are critical to improving flood forecasts and which have been advanced to the point of operational utility: these are radar-based estimates and predictions of rainfall, and quantitative precipitation forecasts.
The review is organized into three sections: forecasting flash floods (events that occur within 6 hours of the causative rainfall), forecasting main-stem floods (primarily in headwater basins), and flood warning systems.