In the spring of 1997 the Red River valley was affected by one of the worst floods on record. Each spring when temperatures rise and snow melts there is a possibility that flooding of the Red River valley will occur. The application of remotely sensed data and geographic information systems to predict and manage future floods to minimize impact is currently being investigated. The Red River is an international river that originates in Lake Traverse and flows north into Lake Winnipeg. Canadian and American agencies are working together to reduce the impact of future flooding in the Red River basin in part through the efforts of the International Joint Commission. The Canada Centre for Remote Sensing and the EROS Data Center are currently integrating a number of Canadian and U.S. spatial data sets to create a seamless international GIS database for the Red River. ARC/INFO is being used as a tool to integrate the data that will serve as the first level geospatial database required for the development of a system to monitor and predict floods within the Red River watershed. The system will have a requirement to integrate raster, vector, and point data at regional and local scales. This paper identifies challenges in integrating the best available data acquired in different formats to create a seamless geospatial database for this large international watershed. Issues related to the development of a flood monitoring system that utilizes remotely sensed and GIS data in a decision support system (DSS) are also discussed.