Vector Datasets and Formats

Much vector data is stored in conventional GIS formats, such as those of ESRI SHP or MapInfo formats.  However, there are a number of vector data formats which are specific to particular data types and US agencies.  These are important because they are the format in which a large amount of public-domain data is distributed.

Global Vector Datasets

• VMAP.  The US NGA has a global vector dataset called VMAP.  Read more about it below and on the Data Openness page.
• Natural Earth is a public domain map dataset in very low resolutions.
  • It contains mostly cartographic features like political boundaries which aren't much use for realistic terrain visualization.
• There are many free Shapefiles for global political boundaries, but surprisingly they are all quite deeply flawed in many respects, as detailed in the very impressive 4-part blog entry Free Shapefile of Countries of the World

USGS DLG-O

• "Digital Line Graph - Optional format"
• the USGS's older vector data format, which they are trying to phase out in favor of the (very bad) SDTS
• 1:100,000 DLGs are available free online from the USGS for most of the USA, higher-resolution DLGs can be purchased from the USGS
• DLG-O is much simpler to read than SDTS, with a file naming scheme that's also much easier to use
• the format has some documentation here
• the format does support specification of a Datum, although this is not mentioned in the USGS spec!
  • beware: pretty much all of the USGS data that says NAD27, can actually mean OHD, NAD27 Alaska, NAD27 Aleutians, or NAD27 CONUS, depending on where the data actually is.
• DLG layers are missing a lot of important information, including: road widths, names of streets and rivers, bridges, and traffic direction
• common DLG layers:
  • HYF - Hydrography
  • MTF - "Miscellaneous Transportation Features" includes airfields and major power lines
  • RRF - Railroads
  • RDF - "Roads and Trails"
• Road layers
  • It's possible for a road to have no attribute codes, or many.  The data is highly variable.  Most (like 99%) have only the "road class" attribute, which doesn't tell you much.  However, i have encountered a single road with all of the following:
    • 170 0 Photorevised feature
      170 201 Primary route, class 1, symbol undivided
      170 216 Arbitrary extension of line (join or closure)
      170 607 Underpassing
      170 654 Road width 86-95 feet, 0.045 inches at 1:24,000
      171 4 Number of lanes = 4
      174 17 State route number = 17
• source code for reading DLG files is part of the vtdata library

TIGER^®

• TIGER/Line ("Topologically Integrated Geographic Encoding and Referencing") files contain geographic features: roads, railroads, rivers, lakes, political boundaries, census statistical boundaries, etc. covering the entire United States
• produced by the US Census Bureau, distributed online or on 8 CDROMs at $260 each
• each county is represented by a directory of files with a common basename, and different extensions - for instance, county 9 in state 23 (Maine) consists of the following files:

  TGR23009.RT1 TGR23009.RT2 TGR23009.RT3 TGR23009.RT4 TGR23009.RT5 TGR23009.RT6
  TGR23009.RT7 TGR23009.RT8 TGR23009.RT9 TGR23009.RTA TGR23009.RTC TGR23009.RTH
  TGR23009.RTI TGR23009.RTP TGR23009.RTR TGR23009.RTS TGR23009.RTZ

• there is free online software that reads TIGER files: LandView II, a very primitive DOS program
• bitmaps utilizing TIGER data can be generated by the Tiger Map Server, TMS Version 2.5
• software libraries: OGR supports TIGER

Other Vector Formats

• BTS (Bureau of Transportation Statistics)
  • Part of the U.S. Department of Transportation
  • Provides free geospatial data for highways, railroads, and all other kinds of transportation, for all of North America.  Historically, this was in their own ASCII file format.
• SDTS
  • SDTS is used for vector data as well as other data types

U.S. Military Formats

• VMAP
  • (supposedly) unclassified military data from the NIMA Terrain Resource Repository (TRR)
  • previously known as DCW (Digital Chart of the World)
  • 3 levels of detail: VMAP0, VMAP1, and VMAP2
  • all  based on the VPF format (Military Standard Vector Product Format, MIL-STD-2407)
    • according to an experienced developer: "Don't try to read VPF unless absolutely necessary. It's a dog of a format."
  • distributed on CDROM, VMAP0 takes 4 CDs to cover the world
  • Although VMAP0 is available (see Locations - International) is it extremely low detail.  VMAP1 should be available, but only an early version of a few tiles has actually been released (see Data Openness, and mapAbility.com VMAP1).  Making the full (public domain) VMAP1 actually available will apparently require a very painful, expensive legal process involving the FOIA (Freedom Of Information Act), which (as of 2006) nobody has yet done.
• DFAD
  • a military vector format, originally developed for radar applications
  • defines terrain features, both natural and cultural
  • based on 18bit words, requires a particular ordering of features, and uses feature codes
  • reportedly DFAD is not well designed to support 3D visualization
  • was DFAD phased out in favor of VPF?  yes, and DFAD is basically unavailable unless you are working for the DOD
  • it is/was read by some high-end commercial Vis-Sim software (e.g.. MultiGen, Coryphaeus, EaSIEST)