The Monsoon Data Visualizer is provided to allow ready access to physical and meteorological data that can support open-ended student investigations. The data sets include both station (STA) and assimilated data (DAO) for 107 sites (defined as 2 degree x 2 degree regions). The station data represents the monthly mean of a parameter measured locally at a site and the visualizer covers all years of data made available by the station sources (the data record at some stations span several centuries). The assimilated data represents a hybrid data product which merges station data, remotely sensed data and computer models to provide a best-fit to global weather patterns. This assimilated data has been processed for the years 1980-1993 and so does not provide the historical coverage of the station data but allows computation of important but difficult-to-measure parameters that help improve our understanding of weather phenomena.

The data made available through the data visualizer comes from two main sources:

• The Global Historical Climatology Network V1, V2 data sets (website): The authors wish to thank the Carbon Dioxide Information Analysis Center at Oak Ridge National Laboratory and National Climatic Data Center for producing and distributing the data. GHCN is a comprehensive global surface baseline climate data set designed to be used to monitor and detect climate change.
• The Data Assimilation Office (DAO) Monthly Mean Subset Data (website): The authors wish to thank the Data Assimilation Office at the Goddard Space Flight Center, Greenbelt, MD, 20771, for producing these data, and Goddard's Distributed Active Archive Center for distributing the data. These activities are sponsored as part of NASA's Mission to Planet Earth program.

The sites were selected based on:

• Geographic Distribution: the globe was divided into equal area regions and a fixed number of sites were selected from each region. There were 18 divisions of longitude and 4 divisions of latitude.
• Quality of Data: all GHCN station sites were compared for time span, available parameters, and completeness of coverage during the years 1980-1993. A simple algorithm that combined all of these measures of quality into one metric was used to select among sites within each equal area region of the globe.