In the U.S. flooding typically causes more damage than any other weather-related phenomena. In the 1990s, flooding caused over $60 billion in damage. Although flood damage in the United States fluctuated greatly from year to year, there was a generally increasing trend over the past century. Policies related to flood damage will be more effective if decision makers know why flood damage is increasing. Increased damage due to climate may suggest different policy actions than would flood damage increases due to failed flood policies. Flood damage data are essential for any study that seeks to understand the role that climate, population growth and development, and policy play in determining trends in damaging floods. Scientists have access to reliable climatic (e.g., precipitation and temperature), hydrological (e.g., streamflow), and societal (e.g., population and wealth) data. But little attention has been paid to historical flood damage data needed for policy evaluation, scientific analysis, and disaster mitigation planning. There are no uniform guidelines for estimating flood losses, and there is no central clearinghouse to collect, evaluate, and report flood losses. In order to improve the existing knowledge of flood damage data for the United States, we have re-analyzed historical National Weather Service data and as a result developed three data sets containing flood loss estimates that range spatially from river basin to national scale. We have also evaluated the quality and identified the limitations of these data sets. At the conclusion of this project, we are making these data sets available to scientists, decision-makers, and members of the public.