GIS at the North Carolina Department of Transportation
Bridging Spatial and Nonspatial Data
The North Carolina Department of Transportation (NCDOT) directs, plans, constructs, maintains, and operates one of the largest state-maintained transportation systems in the nation. With a staff of more than 50 permanent employees, its GIS Unit in Raleigh provides GIS, mapping, and road inventory services to NCDOT. Among the many challenges faced by NCDOT is modeling its road network in an integrated yet efficient manner while preserving the lineage of the network in the face of constant change.
Which Road Segment Is Which?
Consider a 4-mile stretch of road. As it advances, attributes change by speed limit, surface material, road condition, and other factors. From a particular start point, the road may have a speed limit of 35 miles per hour for a distance of 1 mile, then decrease to 25 miles per hour for the remaining 3 miles. The road surface may be asphalt for the first 2 miles, then gravel for the final 2 miles. The pavement condition may start as fair, continue for 3 miles, then change to good for the fourth mile.
Note that none of these virtual road segments "lines up" with any other in the sense of sharing identical portions along the underlying road. Representing this variability with separate line datasets would be inefficient in terms of storage, processing, and effort to maintain. The challenge is to capture attributes that vary by length without duplicating geometry.
Linear referencing efficiently meets the representational challenge. It allows virtual networks to be constructed upon a single, stable base network. Attributes are attached (referenced) to linear features along virtual networks using relative (known locations and offset distances) rather than absolute (x and y coordinates) measures. Linear referencing simplifies spatial data management by eliminating the need to maintain separate coordinate geometries for each portion of a road where attribute values may vary independently.
Several linear referencing methods exist, and NCDOT supports four. One commonly used by NCDOT is Route and Mileposta conceptual measurement system where locations and distances are measured from the start of a road to a specific distance or distance range. For example, a stretch of road with a speed limit of 35 miles per hour may be defined as beginning 1.25 miles from where a route enters a county and ending 2.75 miles further along.
"We've taken a customercentric approach," says Frank Winn, director, NCDOT's Engineering Transportation System. "By supporting four linear referencing methods, internal customers may integrate their data with our road network, choosing whichever method works best for them."
The NCDOT Linear Referencing System (LRS) serves as a central repository for the state-maintained transportation network. Several objectives and requirements guided its construction. The system had to
ArcInfo helped meet these requirements by providing a bridge between the LRS (nonspatial data) and linework geometry (spatial data).
Conceptually, the NCDOT LRS is composed of two main components: a road linework geodatabase and the LRS Database Core Module. The geodatabase stores all explicitly spatial data for the state-maintained road network. At LRS version 1.0, this component has been implemented utilizing the personal geodatabase format within ArcInfo.
The LRS Database Core Module represents the road network in a nonspatial, tabular format utilizing an Oracle9i database running on Red Hat Linux servers. It has a set of related subcomponents that model the transportation network, the NCDOT route number system, and change histories for both.
Integration of spatial data with LRS will enable customers to focus on maintaining data unique to their units rather than core transportation network data. By accessing a centralized road network, customers can be confident they are working with the most up-to-date and accurate information available. Centralization puts everyone on the same map, providing a common understanding of the road network.
NCDOT units may bring data together that was previously stored in "silos" across the enterprise and perform new queries and analyses. Prior to NCDOT LRS, such work was impossible to perform or could only be undertaken with great difficulty.
NCDOT intends to enhance the LRS by migrating the Road Linework Geodatabase from a personal geodatabase to Esri's multiuser geodatabase, incorporating non-NCDOT roads as part of its core dataset and creating a unified edit/management tool.
"The LRS is a long-term investment for NCDOT," says Jun Wu, GIS database administrator for the system. "It is built for the long run and provides a framework for integrating fragmented applications and database systems. This system will help us reduce maintenance costs and improve data quality."