About this site

History of the NBI

The National Bridge Inventory is a response to a series of bridge disasters. In December 1967, the Silver Bridge at Point Pleasant, West Virginia, collapsed, killing 46 people. Hoping to avoid a repeat catastrophe, the following year Congress created the National Bridge Inspection Standards to require uniform bridge inspections across the whole country. Bridges over 20 feet long located on federal-aid highways were required to be inspected at least once every two years. That requirement has since been expanded to all bridges on public roads. All of the inspection and inventory data is compiled at the federal level into the National Bridge Inventory.

Two more bridge disasters in the 1980s led to more stringent inspection requirements. A section of the Mianus River Bridge on Interstate 95 in Connecticut collapsed in 1983, killing three people. An investigation into the cause of the collapse revealed the hazards of "fracture critical" designs in which a failure of one part of the bridge (in this case, a pin-and-hangar assembly) could cause a failure of the entire bridge. Fracture critical bridges now receive more attention.

In 1987, the Schoharie Creek Bridge on the New York State Thruway collapsed during a major flood, killing 10 people. It was determined that the bridge suffered from scour, or an undermining of the piers caused by the force of water over time. The flood was the final straw that caused one of the piers to give away, causing the rest of the bridge to fail. As a result, vulnerable bridges now receive underwater inspections to check against scour problems.

Following the failure of the Minneapolis I-35W Bridge, killing 13 people, the National Bridge Inventory re-entered the spotlight. Inspectors fanned out to check for problems on similar bridges (deck trusses), but errors in the NBI data made it difficult to locate all deck truss bridges.

Problems with the NBI

While there has been much talk about improving the quality of the NBI since the Minneapolis disaster, progress has been slow so far. Some problems I've encountered while building this site:

• Some bridges are listed in the wrong county or city.
• GPS coordinates are often way off. The Bureau of Tranportation Statistics has tried to improve the accuracy by using modern GIS tools, producing an alternate version of the NBI, but they can only do so much.
• Bridges maintained by federal agencies are included in the NBI, but the data is often terrible, with many missing fields. In some cases, especially involving U.S. Army Corps of Engineers bridges, the data is so spotty that it's completely unuseable.
• Some states include records for things (ferries and overhead signs) that aren't bridges or tunnels at all.
• On the other hand, a handful of bridges that should be included are not.
• In a few states, the bridge ID numbering systems has changed over the years, making it difficult to compare inspection reports for the same bridge over time. I've tried to compensate, but I haven't been able to match all records.
• In some states, the "route carried" field refers to an internal route number or name that is not shown on signs, nor widely known to the public. For example, good luck trying to find a Tennessee "NFA" or "FAU" route on a map or street sign.
• The year built is often wrong for older bridges. A date of "1900" typically means "Don't know" rather than an actual date of 1900. In many states, years ending in "0" or "5" might just be guesses.
• When a bridge is completely replaced, sometimes the year built field isn't changed, making it seem like the bridge is much older. The NBI is littered with prestressed concrete bridges that are listed with construction dates at least 50 years before prestressed technology was even invented!

Building this site

To deal with some of the problems above, here are the steps I followed to build this website:

1. Took the raw ASCII files from the latest available version of the NBI (2009) and processed them using the OVERPASS program I developed earlier.
2. Updated GPS coordinates from the BTS version of the NBI were incorporated where possible.
3. Eliminated GPS coordinates that are clearly wrong (in the wrong hemisphere, for example)
4. Filtered out records that were missing key information, such as the county. These are almost always listings from federal agencies.
5. Filtered out records that were obviously not bridges (such as overhead signs).
6. Matched records from earlier versions of the NBI (1992 and 2000) based on bridge ID to show how the inspection reports have changed over time
7. In states where the bridge IDs have changed between years, tried to match records using other methods
8. Grouped together bridges by city, if given by the NBI.
9. Fixed other annoying problems from particular files (such as South Carolina 1992 using a different numbering system for counties)
10. Connected NBI listings with bridges already listed on Bridgehunter.com