The New and Improved Base Framework For National Atlas Data Rupert Brooks Canada Centre for Remote Sensing, Geomatics Canada Brooks@NRCan.gc.ca Abstract Reliable framework data is essential for the practical use of geospatial data. It is one of the objectives of Geo-Connections (formerly the Canadian GeoSpatial Data Infrastructure or CGDI) to provide reliable framework datasets for Canada. The National Atlas of Canada is addressing this need by developing integrated and reliable framework datasets based on the Vector Map Level 0 (VMAP0) product. The VMAP0 data product is a global digital dataset produced and maintained by the United States National Imagery and Mapping Agency (NIMA) at a nominal scale of 1 : 1,000,000. The National Atlas, in wanting to choose an international global digital map base, has adapted the VMAP0 for use as a base from which other, smaller map scales can be created and to which thematic layers can be added. The 1:1M data had to be processed and quality controlled to allow it to be used. Since the National Atlas represents data at a national level, a variety of derived scales are being produced from the VMAP0 data. This is being done using automatic model generalisation techniques, as developed by Richardson [1996]. To prepare the data for these techniques, however, existing topology had to be corrected, and even more rigorous topology had to be defined. Having overcome these shortcomings, the National Atlas of Canada VMAP0 data for Canada is of the highest quality. In addition, geostatistical atomic units of area such as Census Subdivisions have been fit to an enhanced VMAP0 base at the scale of 1 : 1,000,000, which should make the dataset even more useful. Throughout these processes, rigorous quality control has been applied to ensure the best possible result. This paper discusses both the techniques used to achieve these results, and the quality control processes used to yield products of the highest quality. Introduction In order to better relate its datasets to global initatives, and to better meet the needs of base data users at Atlas scales, the National Atlas of Canada is reconstructing all its bases at a nominal scale of 1:1,000,000. (The word nominal is used because the concept of scale has changed somewhat with the advent of digital maps.) To date, the hydrology and boundary layers have received the most work. These layers have been derived primarily from the VMAP (Vector Map) Level 0 revision 4 (hereafter, VMAP0) data. It is anticipated that other layers will come from a variety of sources. The VMAP0 data was first released as the DCW (Digital Chart of the World) dataset. The DCW has been widely used and has proved to be very useful because it was the first global multitheme dataset at such a scale in the public domain. Despite its success, the DCW dataset is widely known to have had some deficiencies. The revision 4 data for North America was much better in quality, but still had to be improved before it would be suitable for use by the National Atlas of Canada. Two kinds of improvements were performed on the dataset. First, improvements were performed where the data was occasionally in error, out of date, or incomplete. In particular, some parts of the James Bay power project and the Nunavut territory were not properly represented. International marine boundaries were not part of the dataset, and had to be added. Certain rivers which were connected in reality were not