The 1993 adverse weather and floods in the midwestern United States caused enormous damage. Apart from the impacts to urban areas, most the ponding and flood damage in the Upper Midwest occurred on farmland with significant effects on agricultural yields and production.

Although public officials and policy makers knew that the agricultural damage was extensive during the summer of 1993, information was imprecise. Because the setting of policy parameters, such as those relating to disaster assistance, emergency wetlands reserve, and the emergency conservation program, depended directly on expectations of harvested yields, intraseason quantification of weather-induced impacts was required. Given this need, the Food and Agricultural Policy Research Institute (FAPRI) of Iowa State University was asked to estimate the extent of the flood damage in Iowa, detailing the impacts on acreage, yields, prices, and farm income (Smith et al. 1993). This experience of 1993 induced FAPRI to examine alternative procedures for estimating yields throughout the growing season. One of the more promising alternatives, in terms of parsimony and data availability, is presented in this paper.