An Incremental Geographic Update System (IGUS) for a Large Geographic Database in New York City

Over the last decade a number of states and cities in America have successfully established spatial data infrastructures that incorporate core spatial data such as the physical infrastructure layers, topography, road and rail networks and critical administrative databases like land parcel and administrative boundaries. New York City built its digital photogrammetric base map (NYCMap) from 1996 1:4,500 aerial photography and successfully integrated it with the City’s existing land parcel file (COGIS) and street centerline file (LION) in the year 2000. NYCMap is composed of over two-dozen planimetric features (e.g. buildings, street centerline, towers, subway lines, railroad, hydrography and topography, etc.) and a 0.25-meter resolution digital orthophotos. It is one of the core spatial data sets widely used by City Government Agencies, decision makers, planners, lawyers and utility companies. A key issue facing the City of New York was what was the best methodology for updating NYCMap and how could it be done in such a way as to maintain the integrity of the database and track changes over time. This paper discusses a solution to the update of NYCMap with the creation of an Incremental Geographic Update System (IGUS). IGUS establishes the techniques for the incremental maintenance of large spatial information system in three fundamental ways. First, because objected oriented GIS technology represents the geographic entities as objects instead of layers or geometries, changes can be tracked for each entity over time rather than changes of layers over time, which doesn’t make sense from a temporal standpoint. Second, the approach combines the quality control and data integration process into one step and only operates on the changed features. This increases efficiency of GIS database maintenance and ensures data integrity. Third, incremental changes can be propagated to other database environments through the use of difference streams between database alternatives created on different dates. These changes can then be propagated out to other users in a generic data transfer format.