Flooding recurrence intervals in a given watershed are strongly affected by long-term changes in the spatial and temporal character of precipitation events. These characteristics include the intensity and size of precipitation events (e.g. convective to regional scale) as well as the manner in which the precipitation is delivered (e.g. short vs. long duration). In this study, these characteristics are identified for a large sample of precipitation events collected during the period 1950-1996. To carry this out, daily precipitation totals from the cooperative observer and hourly precipitation data (HPD) networks are spatially interpolated onto a fine scale (10 by 10 km) grid over the eastern two-thirds of the United States. An automated algorithm is developed to identify precipitation events for two-day periods across a range of spatial scales (1000 km2 to 500,000 km2). Assessments are made of the spatial and temporal character of each event. Long-term trends in these characteristics are identified across different regions of the eastern U.S. Gridded synoptic (i.e. NCEP Reanalysis) data are then used to assess long-term trends in the frequency of synoptic scenarios (e.g. superpositioning of high tropospheric moisture and upper level divergence) that contribute to the development of heavy precipitation events. Associations are made between the changes in the character of precipitation events and changes in the frequency of "wet" synoptic scenarios.