A Watershed Year: Anatomy of the Iowa Floods of 2008

6. Forecasting a Record Flood

6 Forecasting a Record Flood A. Allen Bradley, Jr. In Iowa, nearly everyone has used weather forecasts to make everyday decisions —such as how to dress for the day, or whether it is time to plant crops. We know that weather forecasts are not perfect. But from our everyday experience, we also know that forecasts do not have to be perfect to be useful. In contrast, the public looks at river forecasts only when a river is expected to overflow its banks. When it comes to using a flood forecast to make a decision, the stakes are very high—such as how to place sandbags to keep a building dry, or whether it is time to evacuate homes and businesses. TheNationalWeatherService(NWS)isresponsibleforissuingriverforecasts and flood warnings to the public (Mason and Weiger 1995). River forecasts predict the river stage (the future height of the water surface above a base elevation , such as the streambed) and the river flow or discharge (the rate at which the water is moving). Flood warnings predict the severity of an anticipated flood—minor, moderate, or major flooding. And when anticipated flood stages are higher than any recorded in historic times, the NWS will issue a warning for a record flood. During the Iowa floods of 2008, the NWS Forecast Offices in the Quad Cities and Des Moines issued hundreds of river forecasts and flood warnings for locations along the Cedar and Iowa Rivers. 54 rising rivers, spreading waters From the forecaster’s perspective, a record flood event is probably the most difficult river-forecasting situation imaginable. A record flood is uncharted territory; past flood experience is no longer a reliable guide as to what might happen. The forecasters are also under the intense scrutiny of emergency managers and the public, many of whom are looking at river forecasts for the first time. Everyone wants the forecasts to be perfect. Some even make decisions as if they were. But forecasting a flood that is larger than anyone has ever experienced is fraught with uncertainty. Imagine what it would take to make a perfect flood forecast. First, you would need a perfect weather forecast to know exactly how much rain would fall and where. Then, you would need to track how much of the rainwater flowed across the landscape, and its path. Once the water entered a river, you would need to calculate how quickly it was moving and the obstacles it would encounter— levees, sandbags, bridges, even buildings—as it spread across the floodplain. Such precise accounting of water exceeds our scientific ability, making a perfect forecast a practical impossibility. Given these inherent limitations, just how does the NWS make river forecasts ? The NWS continuously monitors river and weather conditions, and combines this information with imperfect weather forecasts, to make approximate numerical predictions of future river flows and stages. The forecasters refer to these predictions as numerical guidance, because judgment and experience also play a role in issuing river stage forecasts and flood warnings. River forecasting is a data-intensive process. Perhaps the most important data are recent observations of river stages and flow rates. These tell the forecaster how much water is already in the river and making its way downstream. For the most part, river data used in forecasting come from U.S. Geological Survey (USGS) streamgages (see plate 1), where river stage and flow rates are continuously monitored. Precipitation observations are also crucial to river forecasting. These data are obtained from both rain gauges (cylindrical buckets designed to record rainfall accumulation) and NEXRAD weather radar estimates. These tell the forecaster how much water has recently fallen—critical information in estimating how much water will run off and soon reach the river. These data also let the forecasters check the validity of recent precipitation forecasts. River and precipitation data are transmitted by satellite, landline, and computer network to one of the NWS’s 13 River Forecast Centers. Hydrologists thereassembleallavailableobservations,combinethemwithweatherforecasts, and feed them into computer models that simulate watershed processes and forecasting a record flood 55 river hydraulics (Braatz et al. 1997). The computer model subdivides a river basin into smaller drainage areas known as forecast segments. Precipitation observations and forecasts are used to predict the amount of runoff. Runoff from each forecast segment enters the river, where it combines with water already in the river from upstream. The river hydraulics model then computes how quickly that water volume will move downstream...