Richard C. Smith

Implementing GIS-based Highway Safety Analyses: Bridging the Gap

Abstract

The GIS Safety Analysis Tools v1.0 CD-ROM represents a recent example of the work the Federal Highway Administration (FHWA) Office of Safety Research and Development has done to promote the capabilities and potential of Geographic Information Systems (GIS) in highway safety analyses.

One of the purposes of distributing the Safety Tools CD is to encourage users to explore the capabilities of the Tools and adapt the ideas and applications to fit their particular needs. However, due to the variety of implementations of GIS by the States Department of Transportation (DOT), developing capabilities in highway safety analyses at the State level requires special considerations and an understanding of the requirements of GIS by all persons within an agency who wish to participate in this GIS application.

To help with a better understanding of State DOT capabilities, FHWA recently conducted a survey to determine the status of GIS for the States that participate in FHWA’s Highway Safety Information System (HSIS) program. The results of the HSIS States GIS Survey provided insight into the States’ capabilities in using GIS for highway safety analyses.

From the results of the GIS Survey and through our experiences working with the participating HSIS States, we have recognized there are barriers in implementing GIS-T applications. This paper discusses the results of our survey and addresses the considerations in bridging the gap between the desire to implement highway safety analyses within an organization and the development of a GIS-T infrastructure that will support that effort.


Background

The FHWA operates and maintains the HSIS, a highway safety database that uses data already collected by HSIS States for the management of the highway system for the study of highway safety. The HSIS is used in support of the FHWA safety research program and as input to program and policy decisions. The HSIS is also available to analysts conducting research under the National Highway Research Program, university researchers, and others involved in the study of highway safety [see FHWA-RD-00-033]. States participating in FHWA HSIS are California, Illinois, Maine, Michigan, Minnesota, North Carolina, Utah and Washington State.

Under various initiatives the FHWA HSIS and their contractors have been developing GIS capabilities and tools for highway safety analysis and GIS-based safety analysis tools are available to traffic and safety engineers to aid in their work. Early efforts to provide workstation capabilities for GIS-based safety analysis were developed in 1996. These GIS-based tools, which range from hazardous spot location to corridor analysis software, were initially developed for FHWA and the North Carolina Department of Transportation (NCDOT), as the NCDOT Crash Referencing and Analysis System application, by the North Carolina Center for Geographic Information and Analysis under a separate FHWA contract.

The GIS Safety Analysis Tools v1.0 represent a recent example of FHWA to place GIS-based safety analysis tools in the desktop environment. These GIS-based tools were adapted from the NCDOT tools, a UNIX-based application, to provide the same safety analyses capabilities in a more user-friendly environment, for FHWA Office of Highway Safety Research and Development and the University of North Carolina Highway Safety Research Center by GIS/Trans, Ltd. The Safety Analysis Tools were developed in Windows NT as a GIS software upgrade, using ArcView GIS, the Avenue scripting language and the Arc Macro Language (AML) for use with Arc/Info and Arc/Info Network Extension.

One of the purposes of distributing the GIS Safety Analysis Tools is to encourage the safety engineers, State DOTs, and Metropolitan Planning Organizations (MPO) personnel to explore the capabilities of the GIS-based highway safety analysis tools and adapt those ideas and applications to fit their particular needs. However, due to the variety of implementations of GIS by the State DOTs, developing capabilities in highway safety analysis at the State level requires an understanding of the requirements of GIS, Linear Referencing Systems (LRS), and GIS-based highway safety analysis applications.

Purpose of Current Effort

The primary goal of this current effort is to discuss GIS/Safety integration in terms that can be understood by both safety engineers and GIS specialists, and to describe issues and solutions involved in the integration of GIS into safety-related analysis efforts. It is important to present this discussion in language that can be understood by both safety engineers and GIS specialists -- to initiate a "common dialogue" in a manner understood by all.

FHWA has undertaken to provide this technical dialog in various forums with plans to publish and present various aspects of a common theme. To accomplish this, a narrative has been adopted to provide a foundation for understanding GIS linear referenced highway and safety data as it is applied in safety analyses. This narrative includes:

  1. Provide a discussion showing relevance to highway safety analysis using illustrations and examples.
  2. Provide general background information on Linear Location Referencing System (LLRS).
  3. Provide a general understanding of how GIS manages routes and integrates linear referenced data.
  4. Provide background on the benefits GIS technology offers in general analyses.
  5. Discuss the use of safety-related data, accident data, and roadway inventory data in GIS.
  6. Discuss the use of GIS-based safety applications tools.
  7. Discuss the considerations in implementing GIS for safety analysis.

Addressing the Issues

Through the experiences of working with selected HSIS States, it has been recognized there are barriers in implementing GIS Transportation (GIS-T) applications, specifically GIS-based highway safety analyses applications. These barriers include:

This paper specifically looks at overcoming these barriers in implementing GIS for safety analysis and will address the following questions:

This paper addresses the considerations in bridging the gap between the desire to implement highway safety analysis within an organization and the development of a GIS-T infrastructure that will support that effort.

What GIS Analysis Potentially Offers Safety Analysis

GIS is a graphical information system that supports the display and analysis of spatial data. GIS has its strength in providing capabilities to model the physical proximity of spatial features. One powerful aspect of GIS is the flexibility in modeling spatial objects to suit particular user needs or application requirements. These capabilities have developed as the technology has matured. The present-day benefits of GIS are well established -- GIS provides capability to store and maintain large data sets of spatial and tabular information and support complex system architectural design with modern network connections for large enterprise applications that serve inter-departmental needs.

Over the past 10 years GIS has adapted to accommodate linear referenced data common to DOTs. Most transportation data, such as accident and roadway inventory data, are "located" or referenced on the roadway using linear referencing methods (LRMs). GIS has been extended to offer capabilities to define route systems in coordinate space and to manage linear event data using available GIS tools, so that accident and roadway inventory data can now be brought into GIS for display and analysis. It is the application of LRMs in the GIS that make possible the integration of linear referenced data with spatial data.

The integration of linear referenced data in the GIS provides new capabilities of data comparison and analysis that were not before available in non-GIS LRS. These capabilities are used in GIS to bring together linear referenced data from different route systems and spatial data from other sources into a single environment for data conversion and data enhancement, such as integrating one route system with another, or joining spatial data attributes to the linear referenced data.

Such capabilities in GIS can be used to improve the quality of the linear reference data. For example, in the accident report data entry process, the recorded "posted speed limit" of an incident report could be compared with the legal speed limit attributes in the GIS roadway database to provide quality control and data integrity checks. The GIS LRS can also be used to provide a historical location reference to improve the long-term quality of linear referenced data. For example, event locations in the LRS would be misplaced subsequent to realignment of the roadway. In such case, using the spatial coordinate from the GIS roadway network to attribute events such as crash locations, as an example, would provide one solution to historic location referencing of accident data.

GIS offers the safety engineer specific analytical and functional capabilities for understanding the spatial relationship of data not found in the DOT LRS. As an example, these capabilities would allow the safety engineer to analyze the relationships of transportation data to other non-DOT data, such as streams and wetland data, rainfall data, etc. The safety engineer could display culverts that were inventoried with roadway data to see how they are located relative to the streams and rivers, or use rainfall data to determine accidents that might be related to storm events, as two examples. The safety engineer might compare pedestrian accidents along a roadway segment with land use and zoning data, or population and other demographic data to gain a better understand of the relationship of crash incidents to community. Or, if construction zone data is available, that data could be integrated with accident records to provide a true picture of work zone related crash incident, as examples.

For the safety engineer, GIS is used to assist in locating accidents. This is made possible in various ways. With GIS, the safety engineer can analyze traffic data using spatial tools that perform overlay analysis, proximity analysis, cluster analysis, corridor analysis, and the like. The GIS display and mapping capabilities permit visualization of high accident areas as clusters, such as might exist around roadway intersections or fixed objects. Using classification and symbolization of accident attributes, areas of interest can be portrayed for thematic mapping, such as for crash type and severity. Proximity analysis can be applied with GIS, so accidents involving school aged pedestrians can be "found" that are within one kilometer of schools. And, GIS network analysis tools can be used for corridor analysis to pinpoint truck related accidents that are further than 3 mi (4.8 km) from a designated truck corridor.

These examples illustrate the potential of GIS for safety analyses. The rapid application development available in the GIS desktop environment makes it possible for the development and application of specialized highway safety analyses, such as is suggested above. The GIS Safety Analysis Tools is one such application that is currently available to the highway safety community.

The GIS Safety Analysis Tools

The GIS Safety Analysis Tools v1.0 use GIS capabilities to select, display, edit, analyze, and report crash locations. The Safety Analysis Tools was developed using ArcView GIS Avenue script and Arc/Info AML programming; however, these capabilities are not vendor specific and can be developed in a variety of vendor environments. One tool in the NCDOT application is used to assist in locating accidents based on officer description, for location that cannot be automatically located through dynamic segmentation. The other tools are analytical tools designed to assist in analyses to identify areas with specific accident characteristics or densities. For discussion, capabilities available on CD-ROM are presented below categorized by type into three broad types: (1) location coding tools, (2) site location analysis tools, and (3) network analysis tools. Five analytical tools fall into the last two categories and are referred to as the "safety analysis tools". The functionality, example of use, and requirements of these safety analysis tools are presented below.

The tools demonstrate safety analysis with county data from Wake County, NC. The sample data set that is included on the CD-ROM is a NC State route system LRS and crash records data set. The Safety Tools will need to be adapted to meet the data requirements of each DOT, and can be adapted to the ArcView LRS model with some programming. For a State DOT, MPO or others to adapt the applications to fit their particular data needs, the Tools must be modified to accommodate file names, field names, and data type conventions used in the Tools, as they were defined for the NC sample data set and are considered "hardwired" for that data set. Data requirements for the tools are provided below and a full description of data requirements will be available on the GIS Safety Analysis Tools second release in July 2000.

Figure 1 represents one view from the GIS Safety Analysis Tools showing the edit window displaying the NC Route System, the locator window and analysis selection capabilities that are available in the application.

 

A view of the GIS Safety Analysis Tools

Figure 1. A view of the GIS Safety Analysis Tools.

 

Location Coding Tools

Location Coding Tools provide capabilities to assist in locating accidents based on the description of a crash location provided by the officer in the accident report. The operator can use a hardcopy of the accident report to assign x,y coordinates to the crash location. The tools also provide the capability to display a scanned copy of the accident report to see all the information pertinent to a crash location that was recorded at the scene, but the use of this capability is not required. Using that information made available, the user can select a precise location in the GIS using the road network data, digital aerial photography or other imagery, and other source materials at hand, to code the accident record with a coordinate location.

Site Location Analysis Tools

Site location analysis tools provide capabilities for analyzing crash trends at a specified point along the roadway. The GIS Safety Analysis Tools CD-ROM contains three site location analysis applications:

Each site location analysis tool is described below with examples and data requirements for their use.

Spot/Intersection Analysis

The Spot/Intersection Analysis routine is used to evaluate crashes at a user designated point or intersection for a given search radius. For selecting crashes on-route and reference-route are used as criteria for selecting routes and miles (converted to kilometers for the NC State plane coordinate system) are used in determining the distance from a point.

Example: Using the mouse, the safety engineer would position the cursor over the spot to analyze. If the point falls on at least one route, a pop-up menu permits the safety engineer to select crash data attributes, on-route, and reference-route. The Spot/Intersection menu allows a range for crash data date selection and search radius to be entered.

The end result of this analysis is a report which lists the number of crashes, fatalities, injuries, costs, etc., and a graphic which can be output as a hardcopy map depicting the spot, search radius, selected crashes, and roadway. Figure 2 represents graphical output from Spot/Intersection Analysis.

Requirements: The Spot/Intersection analysis tool requires Route Milepost linear referenced data and crash data, located by geographic coordinates. ArcView GIS application software is required.

 

Output from Spot/Intersection Analysis

Figure 2. Output from Spot/Intersection Analysis.

 

Strip Analysis

The Strip Analysis routine is used to study crashes along a length of roadway rather than a finite location, spot, or intersection. Crossroads can be included in selecting crashes within the Strip Analysis.

Example: Using the mouse, the safety engineer would position the cursor over the route to analyze and "click" the points indicating the start and end of the strip. A pop-up menu permits the safety engineer to select crash data attributes, on-route, and reference-route. The Strip menu allows a range for crash data date selection and search distance to be entered.

The end result of this analysis is a report which lists the number of crashes, fatalities, injuries, costs, etc., and a graphic which can be output as a hardcopy map depicting the buffer which makes up the strip, selected crashes, and roadway. Figure 3 is a graphical output from Strip Analysis.

Requirements: The Strip Analysis tool requires route data and crash data located by geographic coordinates. ArcView GIS application software and Arc/Info application software using a Remote Procedure Call (RPC) connection are required.

 

Output from Strip Analysis

Figure 3. Output from Strip Analysis.

 

Cluster Analysis

The Cluster Analysis routine is used to study crashes clustered around a given roadway feature. Crashes are selected within a given distance of a feature from a selected theme, on a selected route. The analysis will cluster around any theme or selected set of features. A threshold is used to select the number of clustered crashes.

Example: The safety engineer would select the appropriate route or group of routes for the analysis from the route system theme using the graphic selection tools or the table selection tools. The Cluster menu allows data range for crash data date selection and weed value, and selection distance parameters for crash data. The type of feature associated with crash clusters is then specified, e.g., intersection, bridge, etc.

The end result of this analysis is a report that lists various summary statistics based on the valid features and graphics, which can be output as a hardcopy map depicting high crash locations.

Requirements: The Cluster Analysis tool requires route data, crash data located by geographic coordinates, and, feature data such as intersection features, bridge features, etc. ArcView GIS application software is required.

Network Analysis Tools

Network analysis is a special class of analyses performed along linear features, such as along a route system or the road network feature elements (as opposed to a spatial analysis for proximity in any direction). The safety engineer can use network analyses tools to traverse a network to perform analyses and find accidents for a specified distance along a route.

The GIS Safety Analysis Tools has two tools that perform network analysis:

Each network analysis routine is described below with examples and data requirements for their use.

Sliding Scale Analysis

The Sliding Scale Analysis routine is used to run a sliding scale analysis familiar to safety engineers, where a section of the route is specified as a window to search for accidents. Each route section is analyzed one at a time to build a strip composed of similar route sections that meet threshold parameters specified by the safety engineer. The end result is a table that lists each valid strip, which was identified using the parameters that were entered.

Example: A safety engineer is attempting to locate all 0.25 mile sections of roadway which have experienced more than X crashes in the past three years. Various parameters are entered to calculate the validity of a strip, such as start length, extension length, crash rate, etc. The initial strip starts at the beginning route milepost or other specified location and traverses along the route in increments equal to the Extension Length parameter that was entered. The calculations are made on crash rate using ADT, and then overlaid with Interchange file to determine strip validity in order to add the extension or begin another strip. An overlay is done to check the Functional System Classification of the strip to validate for change in number of lanes. Additional calculations are performed on critical crash rate, and the process repeats, completing the strip according to all parameters provided. Figure 4 shows the tabular and graphical results of Corridor Analysis.

Requirements: The Sliding Scale Analysis tool requires route data containing measures, intersection data, the North Carolina Universe File, which contains linear event data of system designations, physical characteristics, ADT, and the like, for NC state maintained roadway segments. ArcView GIS application software is required.

 

Results of Corridor Analysis

Figure 4. Results of Corridor Analysis.

 

Truck Corridor Analysis

The Truck Corridor Analysis routine provides a visual means to examine truck-related accidents that occur outside STAA truck routes.

A methodology is used to incorporate Sliding Scale Analysis to first select strips of high concentrations of truck crashes along each STAA truck route. Second, the 3-Mile Drivable Intersection Summary was run to select road segments comprising a 3 mi (4.8 km) drive from each intersection along each STAA truck route.

To determine the 3 mi drive from the designated STAA truck route, Truck Corridor Analysis traces the road network data. Truck related crashes beyond the road network designated 3 mi (4.8 km) drive would be shown outside of the normal driving area for trucks. Figure 5 represents graphical output of Truck Corridor Analysis.

Requirements: The Truck Corridor Analysis tool requires Arc/Info Network Extension to trace the road network. ArcView GIS application software and Arc/Info application software using an RPC connection are required.

 

Output of Truck Corridor Analysis

Figure 5. Output of Truck Corridor Analysis.

 

Current Use of GIS for State Safety Analyses

The use of GIS for display and analysis is well documented. Recent publication by Miller [2000], Breyer, [2000] and others provide a review of GIS capabilities that are available to the traffic safety community. These application reviewed have identified prototype systems and tools having capabilities useful for highway safety analyses. In the synthesis of literature review, Miller points out that "except for reporting a picture of crash locations, the majority of states are not using many of these GIS capabilities."

Although GIS is use in transportation planning, for such applications as mapping landslide hazard zones, modeling disaster response planning, routing of overweight and oversized vehicles, flood prediction, to study the health impact of air emissions, one-way traffic network planning, and the like, GIS is less used for highway safety analyses. This author would like to suggest possible reasons why GIS is not use for safety analyses by the majority of the states:

  1. Not having a champion within an organization for the use of GIS in safety analyses.
  2. Having a need for progressive communications between the GIS leader, other departments, and the GIS users, such as the safety engineer.
  3. Having a lack of knowledge of how to start and proceed.
  4. Lacking a commitment of resources for the high cost of GIS data development.
  5. Lacking the funding for capital investment in GIS hardware and software.
  6. Not having in-house technical skills in GIS and IT to develop or customize available safety tools to meet the needs of the DOT.
  7. Currently not using GIS to managing the DOT LRS.
  8. Having a closed system for existing DOT data, such as is attributed to legacy systems or standalone computing environments.

However, this paper is meant to address some of the technical issues raised here.

Considerations in Implementing GIS for Safety Analyses

Linking highway safety data to GIS will provide challenges for the State DOTs and MPOs. A good understanding of how any LRM has been implemented by that agency is necessary to develop an appropriate GIS to avoid linkage-related issues. Many potential data related issues could also inhibit implementation of GIS-based Safety Analysis Tools. The GIS Safety Analysis Tools depend upon crash data and road network data being available, properly formatted, and fully developed as a route system. Also, other data may be required for specific applications, like a properly identified truck corridor for the Truck Corridor analysis.

This section presents consideration, listed below, in implementing GIS for safety analysis through a discussion of types of problems related to linkage of the GIS and the LRS. This discussion will conclude with two case studies: the Maine Case Study and the Washington State Case Study, which are used to illustrate the different linkage-related problems encountered using the State data, and to illustrate the successes in dealing with types of linkage related problems.

Types of Linkage Problems

LRS linkage problems will be the most evident and pervasive for the safety engineer who desires to implement the GIS for safety analysis, but through explanation and understanding provided below these linkage-related issues anticipated or avoided. The types of linkage problems presented here are:

Standardization Issues

Data standardization is a fundamental consideration in developing GIS for integration with existing databases. All working groups depend on standards being established within an organization for inter-departments and external organization cooperative effort.

Standards for hardware platforms, operating systems, network environments, database systems, and applications software are not the issue. Interoperability evidenced by the success of the Internet has shown this. Standards for data definitions are much more important to provide reliability and portability in developing and maintaining systems and applications. GIS software standards can also add to complexity, as not all GIS share a common route system that is easily transferred from one vendor specific application to another.

Standards should be established for the LRS and the GIS. This includes simple integration and understanding of data in GIS. DOTs have established LRS built on standards that should include the linear referenced data modeling, data file naming convention, attribute coding, and design of relational database tables. For placing accidents on the map, spatial data standards are less an issue, however the GIS route system standards should include the spatial data modeling, considering scale, accuracy, resolution, and generalization, as well as establishing a mapping datum and projection.

Attribute Coding Issues

Key field names are required to establish a database linkage for the crash and roadway inventory database. Planning is required to model existing database design in order to integrate newer technology to older, well established computing environments. Small mistakes, such as improperly defining field name, data width or data type, could add unnecessary hurdles and delays in GIS development and linkage to linear referenced data. The key fields and items for linking LRS to the GIS route system are Route-ID and Route Measures, for point and linear event data, such as crash data and roadway inventory data. Other LRS items could be key depending on the scale and level of generalization of the LRS, such as with the implementation of multiple route systems.

For some states, a fully functional LRS linkage depends upon additional key fields. A County-Route-Milepoint (CRM) is one example of an LRM common to some state DOTs, where an additional LRS attribute would have a "functional dependency" in the GIS LRS data model. In the CRM LRS, county or jurisdiction is a key attribute because the route beginning mileage measure is reinitiated for each jurisdiction the route passes through. For CRM as an example, where the GIS was modeled using only Route-ID and Milepoint, an incorrect linkage would likely occur. This is because without the use of the County attribute (where milepoint is measured independently for each county) the CRM data model would be more like a Route Mile Point (RMP) data model (where a single route measurement is established once across the entire state). In such a case, several accidents occurring in different counties but having the same route and mileage attributes would likely be improperly mapped to the same point location, making any case likely as an improper mapping to the wrong county for any given route and milepoint combination.

Resolution and Generalization

Modeling the road network as a spatial or graphical layer in the GIS is a planning exercise that needs to be compatible with, and reflect the needs and requirements of the DOT, as it might support the daily operations of the organization. Mainline routes, secondary routes, collectors, and interchange features can be represented in the GIS at various levels of detail and generalization. Large-scale mapping of state maintained roads could be developed to depict mainline roads, ramps, and collectors, as illustrated in figure 9A. At this mapping scale, lanes of travel, i.e., lanes for same direction of travel, would be generalized to a single pavement centerline and provide for most LRS application needs. Whereas, small-scale mapping of the same roadway could be represented as a simple roadway or right-of-way centerline, as illustrated in figure 9B. This depiction would show the intersection of mainline roads as a point, but does not show interchange features, such as ramps.

 

Same roadway interchange represented in GIS at two levels of generalization and detail

Figure 6. Same roadway interchange represented in GIS at two levels of generalization and detail.

 

At a moderately large-scale mapping, such as 1:24,000-scale, roadway features could be resolved at a spatial accuracy of ±40 ft (±12.2 m)[1] to fully depict the road network for all directions of travel, showing ramps and collectors. To graphically depict edge of pavement or actual lane designations, such as required in urban GIS initiatives, much larger scales of 1:400-scale, or similar scale, would be needed.

In general, scale can be debated, however, we find that much site location analysis, to find areas that may be studied further offline, can be performed with nearly any scale mapping. Thus, to exploit the LRS in GIS, establish an LRS linkage to the GIS route system, and make fully functional the data integration, the GIS spatial data model is dependent on the linear referenced data model. Route-ID and Route Measures are key items in the LRM needed to fix the data to the graphic of the roadway. However, other data fields, such as route direction, may have a functional dependency in the LRS, (this example of functional dependency is discussed below in the case study for Washington State).

Scale, Accuracy, and Precision

The accuracy of linear referenced data is relative to, and thus mostly dependent on, the calibration of measures along the road network, and less dependent on the accuracy of the road network data in the GIS. However, the accuracy of spatial data will come into play in two ways: (1) when the road network data is overlaid with other spatial data, and (2) as the LRS is linked with the GIS route system through dynamic segmentation; the latter as being the degree to which calibration needs to be performed to improve placement of accidents given the spatial resolution of the road network data set.

In GIS data development, precision in the road network database is determined by two factors: (1) the scale of the source data, such as hardcopy maps to be digitized, and (2) the skill and abilities of the person digitizing the road network. Small-scale source offers a more generalized representation of features than does large-scale source. The use of scanning, character recognition, and raster to vector conversion technologies have aided in the task of converting hardcopy to digital data and to mitigate operator introduced errors.

For GIS-LRS integration, precision in computer data storage is not an issue. Since a GIS can produce maps at any scale, scale primarily affects accuracy and precision of a spatial layer. In today’s computing environment, the precision to which accidents are mapped to the GIS route system is generally dependent on the linear referenced data collection method, not the precision in the GIS. The following example is given to show calculation of precision of LRS conversion to decimal degrees in a GIS route system.

Calculation of Precision of LRS conversion to Decimal Degrees

A GIS using double precision will generate point features to six significant figures. Given a roads network defined in geographic coordinates of decimal degrees and a route system with an attached system of measure defined in miles, dynamic segmentation of accidents will place a crash in coordinate space to 0.000001 degree of latitude and longitude. At the equator one degree of latitude is equal to approximately 68 mi. One-millionth of a degree is equivalent to 0.359 ft [68 mi/° x 5280 ft/mi x 0.000001]. For the State of Washington, as an example of being at higher latitudes, this distance is reduced to approximately 0.248 ft in the east to west direction for double precision storage of the spatial data!

It may be desirable to plan the GIS for future data locational accuracy (i.e., when GPS is more readily in use and made available for collection of crash locations and roadway inventory) rather than design the road network resolution to accommodate existing data. There are tradeoffs to each approach. Cost/benefit must be weighed in choosing the best scale, and a decision made whether to support all LRMs that may exist for various data sets.

Multiple LRMs

Route systems can be modeled in various ways, and have been done so in legacy systems long before the advent of GIS for Transportation. Many different LRS have been defined and implemented by DOTs and MPOs, each using variation of LRMs having various designations and naming conventions. The LRS is developed from the LRM, and at times the reference to the types of LRMs, LRSs, and route systems are used interchangeably; however, LRM should not be confused with LRS when discussing route systems. As a reference for discussion of LRM, two definitions are provided:

Linear Location Referencing System is the total set of procedures for determining and retaining a record of specific points along a linear feature. The system includes the location reference method(s) together with the procedures for storing, maintaining, and retrieving location information about points and segments on the highways. (NCHRP Synthesis 21, 1974).

Linear Referencing Method is a set of procedures used in the field to identify the address of any point and provides a means for designating and recording the geographic position of specific locations on a highway. The LRM designations are used as a key to store information about the locations. (TRB Highway Location Reference Methods, 1974)

The BTS Resource Guide on the Implementation of LRS in GIS provides a resource to define LRMs and a common basis in the discussion of LRM types. The BTS Resource Guide groups LRMs into four general categories:

Some DOTs have over time developed multiple LRMs, which are used for various business operations. In the 1999 FHWA HSIS States GIS Survey, the HSIS States reported the use of various LRS, being variation on Route-Mileage, Route-Reference Post, and/or LN. In general, all HSIS States do not reference accidents for non-state maintained roads, such as in urban area streets, which are not state roadways. The response by the HSIS states to "LRS in use", and "whether more than one LRS is in use for accidents and inventories" are provided in table 2. This table shows the complexity and variations of LRM implementation by the HSIS states.

Table 1. HSIS States’ use of LRS.

HSIS State DOTs

LRS in Use

More than One LRS for Accidents and Inventories

California

Route-Mileage

No

Illinois

Link-Node and Route-Mileage

Key Route and Marked Route mileages are directly tied to Link-Node base.

No, however inventory primary reference is Key Route, accidents are Marked Route and mapped to state routes only.

Maine

Link-Node

Also maintain Route/Cumulative Mileage based on Ordered Links.

No

Michigan

Control Section /Mile Point

Currently updating the network and adding a new LRS.

Yes

Minnesota

Route-Mileage

Moving to Anchor-Point/Anchor-Segment in separate Location Data Model project.

No

North Carolina

County-Route-Milepost and Link-Node

A statewide unique ID system is under development.

Yes, accidents use route-milepost and pavement use link-node.

Utah

Route-Mileage and Node-Offset

Yes

Washington

Route-Mileage

No

Support of each LRM will vary by GIS vendor, requiring adaptation of some commercial-off-the-shelf products to satisfy specific LRM requirements in support of the functional specifications of a particular DOT LRS implementation. This particularly may be the case for RRPO or LN Models. The GIS Safety Analysis Tools have been developed to anticipate the use of a RMP LRS in an Arc/Info Coverage data format with an attached route system.

LRS versus Coordinates

Dynamic segmentation is used to develop spatial coordinates of accidents and other linear referenced data. GPS technology is beginning to be used to assist the crash data collection task by providing x,y coordinates of the crash site location, and will be an improvement far superior to most current collection methods for crash locations. However, it would be expected that when overlaid with linear referenced crash data that the LRS data would not align well with the GPS data due to the difference in the datum, or the set of parameters and control points used to accurately define horizontal or vertical measurements. Data derived from different sources can be resolved for accurate display and meaningful analyses if the datum is known. It is suggested that metadata be available for all coordinate data and include projection, datum, and unit of measure information.

The case will be that as GPS data is more widely used the precision of the road network layer will be in question relative to the precision of the GPS data. The solution for the road network data to spatially "fit" other data from a higher spatial precision, such as GPS data, is to "conflate" the data to the more precise dataset, i.e., rectify the spatial accuracy through rubber sheeting techniques.

Another technique is to adjust the GPS data positional accuracy to the linear datum or snap the GPS data to the linear features in the GIS route system. Coordinate-based crash data derived from GPS or other sources such as a different road network will require adjustment to "snap" to the roadway as depicted in the GIS roadway network, such will be expected for site location analysis mapped against the road network data. For states having GIS located crash data, the buffering distance along routes, available in the GIS Safety Analysis Tools, will have to be considered to allow for the spatial margin of error in accident x,y placement relative to the GIS defined roadway feature.

Historical Linear Reference

Over time roads change. New highways are built, roads are realigned, roads are abandoned, routes renamed, and roadway inventory continually changes. Route identifiers and road measures may change in the process, and the system that maintains this linear information would be updated accordingly. As changes in the LRS occur, changes in the spatial representation of linear features in the road network layer need to be updated also. Often this synchronization of databases requires an inter-departmental cooperative effort.

As early as 1985, the HSIS States have provided accident and other related data to HSIS. Each year’s dataset represents an annual snapshot of the linear representation and events of the state’s roadways. The annual datasets are adjusted to correct for changes in linear measures for that year.

There will always be uncertainty in spatial accuracy in locating linear events using a method that relies on a current GIS dataset to map historical linear referenced data. The best way to initially locate historical data is to have a separate view of the LRS for each year of data, and to try to achieve a similar view of the LRS between the linear referenced data and GIS network for each historical year, that is, historical GIS data of the road network with an attached route system that represents the linear referenced data for each year’s dataset.

It becomes a challenge for agencies to develop procedures and methodology for GIS to adopt for the accurate representation of the road network over the life of the system. Each HSIS State has adopted a strategy for mapping historical data, as reported in the HSIS State GIS Survey. The HSIS States were also asked how accidents have been mapped for historical reference. The States’ response is provided in table 3.

Table 2. HSIS States’ methods for mapping historical accident data.

HSIS State DOTs

In Mapping Accidents, How is Historical Data Referenced?

California

Expect to begin mapping accidents next year.

Illinois

Annual static view of network and LRS, tied to link/node base.

Maine

Crash locations are brought up to current LRS as changes in network occur.

Michigan

Annual static view of network and LRS.

Minnesota

Records map against current routes.

North Carolina

Have not set standard.

Utah

Geographical coordinates.

Washington

Annual static view of network and LRS.

 

It may be that all historical data cannot be confidently located, however the key is to plan for the future and use old data as best as one can. For those States that implement a data warehouse approach to their LRS or linear referenced data, spatially enabling the data warehouse will provide a solution to historical data reference by generating coordinate locations for linear referenced data within the data warehouse.

Two Case Studies

For some time, FHWA has been promoting the use of GIS capabilities in highway safety analysis. Now, as a pilot for integrating GIS in highway safety analysis, HSIS has acquired the associated spatial datasets for Washington State and Maine. This will allow FHWA and HSRC to develop new prototype tools and to incorporate that GIS data into its efforts to support system wide analysis. These two case studies of integrating HSIS data with GIS spatial data from the corresponding states provides examples of the successes and problems with developing GIS linkage to the LRS.

Maine Case Study

Maine Department of Transportation (MeDOT) relies on TINIS (Transportation Integrated Network Information System) to bring together data for accidents, roadway inventory, bridges, railroads, and project history/maintenance type information, and to support their LRS. Recently, MeDOT, with the assistance of GIS/Trans, implemented the TIDE (Transportation Information for Decision Enhancement) system as their data warehouse to integrate their legacy systems with GIS, and augment their LRS to provide new system-wide access and capabilities. (MeDOT, ASHTO, 1998)

One of the many benefits of TIDE is in the area of historical data referencing of accident locations. Through the use of static segmentation, as referred to by Maine, MeDOT is able to manage crash data coordinates on a periodic weekly basis through segmentation of all linear referenced data for changes that occur in the LRS for that time period. The TIDE goal is to provide a transactional database for analysis capabilities through static segmentation. This process provides a solution to the historical data referencing of accident data location by assigning an accurate crash location coordinate position in its own time frame.

The FHWA HSIS uses data already collected by HSIS States for the management of the highway system for the study of highway safety. In HSIS, crash, roadway, traffic volume, and interchange data files are maintained for the state of Maine for years since 1985, representing 22,000 roadway miles, and an average of 38,000 accidents per year. HSIS annually receives the selected HSIS States on a periodic basis having a two-year lag, not anticipating current year data for an additional year.

Maine recently implemented a change in their data standards to meet the increasing size of the MeDOT LN LRS. For Maine, Link-ID is a key item for routes, and is defined as a composite field made up of Beginning Node-ID plus Ending Node-ID. LNs are fairly stable in Maine, but over the years new Links were created requiring additional Node-IDs to be added. The 4-digit number used for Node-ID eventually proved inadequate, and in 1999, prior to implementing TIDE, an additional digit was added to the Node-IDs. This change was implemented in TINIS and migrated to TIDE, the source of MeDOT GIS route system made available to HSIS.

In developing HSIS GIS capabilities for use with Maine data, the MeDOT TIDE route system was acquired for use in the linkage to HSIS data. The seemingly small change to Maine Node-ID had a large impact on HSIS. Now linking TIDE GIS data with existing HSIS data using the LRS was impossible, although not impossible to correct. The "key" field in the GIS is the Link-ID, a composite Node-ID. For HSIS to link to Maine GIS, it has had to accelerate the changes taken place within the Maine systems and upgrade its twelve years of Maine data to accommodate the new Node-ID field format, which affected all tables within HSIS.

Small considerations in standards, such as discrepancy in route naming convention will have a large impact on developing GIS linkage to safety data for highway safety analysis. We anticipate all States will encounter problems in developing GIS and integrating data, and can use the experiences gained through the Maine case study. In summary of the Maine case study, we found even with a well-managed GIS, such as Maine, solutions will need to be found for existing incompatibilities. The following conditions and situations, both advantageous and problematic, were encountered in the Maine case study:

Washington State Case Study

Washington State DOT (WSDOT) relies on the Transportation Information and Planning Support (TRIPS) system to bring together data for accidents, roadway inventory, bridges, curve/grade/features, roadway crossings, and roadside facilities, special-use lane information, railroad grade crossing index, and traffic data to support their LRS. In HSIS, crash, roadway, traffic volume, curve/grade, and interchange data files are maintained for WSDOT for years since 1993, representing 8,400 roadway miles, and an average of 35,000 accidents per year.

Washington State has invested much in developing their GIS road network and route system data, developing GIS route systems at two scales of resolution, 1:500,000-scale (good for small-scale mapping), and a higher resolution GIS road network based on 1:24,000-scale maps. The WSDOT GIS route systems contain measures based on the Washington TRIPS system State Routes and Accumulated Route Mileage (ARM), a RMP LRS. Although both route systems contain the same Route-ID and similar ARM values they must be treated differently in the linkage with the LRS.

In developing HSIS GIS capabilities for use with Washington data, HSIS acquired the use of the WSDOT GIS route system for linkage to HSIS. Working with the two GIS road networks it was easily discerned that road features are depicted differently, as would be expected. At the 1:500,000-scale (small-scale mapping) a highway interchange containing ramps and collectors is generalized as a simple intersection of mainline routes as depicted in figure 6B. This generalization and reduction of detail is adequate for that scale of mapping, but the lack of ramp features degrades the GIS linkage to the LRS, not being able to link to ramps and collectors. At the 1:24,000-scale, the roadway has been modeled differently with additional feature detail. Ramp features and divided roadways are present. This spatial data model represents a closer representation of the LRS depiction of the roadway, where each lane of travel is represented as a route with an increasing or decreasing direction and ramps are represented as other routes.

The HSIS effort to map Washington accidents and roadway inventory data (referred to as WSDOT Roadlog data) represents the first attempt, by anyone, at using the WSDOT GIS route data for that purpose, although Washington HSIS data has been used frequently by HSIS and others for research and analyses. Linkage for Washington State to map accident and roadway inventory data was accomplished through a clear understanding of the linear referenced data model deployed by WSDOT. Linkage with HSIS accident and roadway inventory data to 1996 WSDOT 1:500,000-scale route data was achieved with a success rate of 89 percent and 86 percent for accidents and roadway inventory data, respectively. When broken down by road type, a 99 percent success was achieved for linkage of mainline roadway inventory data to the 1:500,000-scale route data. This linkage would be equivalent to mapping to road centerline. Table 4 represents the success of mapping WSDOT Roadlog data to 1:500,000-scale road network. The 3% difference from mapping all data and mainline data is explained as ramps not having a representation in the GIS at that scale.

Table 3. Summary of success in mapping Washington State Roadlog data to 1:500,000-scale route system.

All Data

Mainline Data

Data Set

Roadlog

Miles

% Mapped

Data Set

Mainline

Miles

% Mapped

1993

8,583

7,314

85.2%

1993

7,265

7,207

99.2%

1994

8,659

7,321

84.5%

1994

7,265

7,209

99.2%

1995

8,352

7,317

87.6%

1995

7,265

7,204

99.2%

1996

8,397

7,240

86.2%

1996

7,265

7,122

98.0%

Average

8,498

7,298

85.9%

Average

7,265

7,186

98.9%

Washington data mapped at the 1:24,000-scale (large-scale mapping) presented several challenges in terms of geographic division of data and functional dependency. First, the WSDOT route systems were developed independently for the 39 state counties. This provided a technical challenge when working with HSIS data, maintained on a statewide basis by year. Secondly, the large-scale mapping permitted greater feature resolution, less generalization, and depicted ramps, collectors and divided highways not shown in the smaller-scale mapping. The large-scale GIS data model was developed to represent the WSDOT LRS data model.

In the WSDOT TRIPS system Accumulated Route Mileage (ARM) values are used. The ARM values are computed from the "State Route Mile Post" (SRMP) equations, which contain measures for increasing and decreasing directions on the roadway. For undivided highways the mileage would be computed as the same, but for divided highways the increasing and decreasing side of the same route section can have different ARM values. WSDOT was able to represent the TRIPS LRS data model in the GIS spatial data model by developing separate route systems within the same GIS: increasing routes, decreasing routes and ramps (as another route system). This preserved the functional dependency inherent in the TRIPS data for direction of route-measure.

Taking advantage of a little used crash data variable in HSIS for direction of crash impact, the functional dependency requirements was satisfied in mapping accidents using the large-scale road network. Crash locations were mapped with great success. The larger-scale mapping allowed better accuracy in mapping events for divided road and interchange by mapping crash data to the correct side of the roadway for divided highways and by permitting ramp-related crash data to be mapped. However, both the multiple route systems and the geographic division of the data by county jurisdiction complicated the data processing, so scripts were developed to assist in data processing and analysis. The results for the 1:24,000-scale route data are that linkage with HSIS accident data was achieved for 87 percent of the available accidents. When broken down by road type, a 99 percent success was achieved for linkage of the 1:24,000-scale route data to accidents that occurred on mainline roads.

Table 5 summarizes the success in mapping accident data to 1:24,000-scale maps. The 1.8 percent difference in mapping all accident data and mainline data is attributed to incomplete and inaccurate representation of mapping interchange ramps and couplets. As previously mentioned in the case represented in table 4, ramps and couplets are not modeled in the 1:500,000-scale route data. In the WSDOT route-naming convention for the ramp or couplet Route-ID, ramps and couplets use the mainline Route-ID and mainline ARM values, at the mileage where the feature begins along the mainline road, as the prefix and suffix, respectively. It is suspected that many more couplets are not being represented, but this naming convention sets up a situation where ramp Route-IDs can easily change, as represented in the following case. If a mainline road is re-stripped and the locations of the ramp or couplet were changed by as much as one-hundredth of a mi, or 53.8 ft (16.1 m), the associated ID for the ramp or couplet would also change. This is in affect a realignment of the ramp or couplet, and the new ramp or couplet Route-ID designation creates the potential to break the linkage for that feature and the remaining data in the LRS.

Table 4. Summary of success in mapping Washington State accident data to 1:24,000-scale route system.

All Data

Mainline Data

Data Set

Records

Accidents

% Mapped

Data Set

Records

Accidents

% Mapped

1993

33,837

32,972

97.4%

1993

30,315

30,017

99.0%

1994

36,784

35,806

97.3%

1994

32,933

32,525

98.8%

1995

38,935

37,660

96.7%

1995

34,711

34,284

98.8%

1996

42,141

40,801

96.8%

1996

37,737

37,365

99.0%

Average

37,924

36,810

97.1%

Average

33,924

33,548

98.9%

 

A complete understanding of the LRMs in use will provide easier development of GIS linkage to safety data for highway safety analysis. We anticipate all states will encounter problems in developing GIS and developing a spatial data model compatible with the linear referenced data model, and can use the experiences gained through the Washington State case study. In summary, we found that even with a well-developed GIS, such as Washington State’s, understanding of the LRMs and LRS will need to be gained for GIS development. The following conditions and situations, both advantageous and problematic, were encountered in the Washington State case study:

Summary and Conclusions

Summary

This report was written to discuss GIS safety integration in terms that can be understood by both the safety engineers and the GIS specialists and to develop the dialogue in a manner understood by all. This report discussed the benefits of GIS technology and the potential GIS has to offer the safety engineer for safety analyses, and how GIS capabilities can be extended through the use of safety-related data, such as crash and roadway inventory data. It provided ideas for developing safety management systems or simply improving safety analysis through the use of GIS-based safety analysis tools. Capabilities and requirements of existing GIS-based safety analysis tools were illustrated. This report suggests capabilities of GIS for safety analysis, and provided conclusions drawn from working with state crash and road inventory data and ways safety engineers and others can better approach GIS development for safety analyses.

It was found that the number and variation of LRM implementation complicates GIS implementation for safety analyses. Of the HSIS states surveyed, no single LRM implementation is standard among the HSIS States, and for some States more than one LRM is used. Two case studies were presented for MeDOT and WSDOT in the implementation of GIS in data integration for safety analyses. This report also outlined typical problems that users will have in implementing GIS for safety analyses. This report discussed other issues and considerations safety engineers and GIS specialists should be aware of in GIS implementation for safety analyses, such as standards and attribute coding issues, understanding of resolution and generalization of spatial data, and how the relationship of scale accuracy and precision relate to linear referenced data. These issues can be summarized as:

Conclusion

A shared understanding of the needs of GIS-based safety analysis requirements is essential for all groups that desire to implement GIS for safety analysis. These case studies serve to illustrate a simple point: that understanding the LRM and the LRS data serve to build a GIS. The case studies reinforce the need for established standards that make it easier for GIS to come together. Hopefully this report will aid the safety engineer and GIS specialist in implementing GIS for safety analyses by helping in the following ways:


Acknowledgments

Thanks to Mike Griffith, FHWA; Forrest Council and David Harkey, HSRC; and, Bobby Harris, GIS/Trans, Ltd.

Appendix A: List of Abbreviations

AML Arc Macro Language

ARM Accumulated Route Mileage

CGIA Center for Geographic Information Analysis

CRM County Route Milepoint

DOT Department of Transportation

FHWA Federal Highway Administration

GIS Geographic Information Systems

HSIS Highway Safety Information System

HSRC Highway Safety Research Center

LLRS Linear Location Referencing System (also LRS)

LN Link-Node

LRM Linear Referencing Method

LRS Linear Referencing System (also LLRS)

MeDOT Maine Department of Transportation

MPO Metropolitan Planning Organization

NCDOT North Carolina Department of Transportation

RPC Remote Procedure Call

RMP Route Mile Point

RRPO Route Reference Post Offset

SA Street Address

SRMP State Route Mile Post

TIDE Transportation Information for Decision Enhancement

TINIS Transportation Integrated Network Information System

TRIPS WSDOT Transportation Information and Planning Support system

WSDOT Washington State Department of Transportation

Appendix B: Selected Terms and Definitions[2]

Cartesian Coordinates: A two dimensional x,y location of a point on a plane (planar) in relation to two intersecting straight lines, called axes. If the axes are perpendicular to each other, the coordinates are rectangular; if not, they are oblique. The X-axis measures the horizontal distance and the Y-axis measures the vertical distance from the origin point of intersection. An x,y coordinate defines every point on the plane. Relative measures of distance, area and direction are constant throughout the Cartesian coordinate plane. (Minnesota Department of Transportation, 1992a).

Conflation: A process by which two digital maps, usually of the same area at different points in time, or two different thematic maps of the same area, may be matched and merged into one through geometrical and rotational transformations. (Association for Geographic Information, the AGI GIS dictionary, http://www.agi.org.uk/pag-es/dict-ion/dict-agi.htm).

Coordinate: Pairs of numbers expressing horizontal distances along orthogonal axes; alternatively, triplets of numbers measuring horizontal and vertical distances. (STDS Part 1, 1992).

Any of a set of numbers used in specifying the location of a point or position. Sources: Webster’s Dictionary, as modified by David Fletcher (1991). (Wisconsin Department of Transportation LCM Manual, 1995).

Coordinate System: A framework used to define the position of a point, line, curve, or plane and derivative map features within a two- or three-dimensional space. (Antenucci et. al., 1991).

A reference system for defining points in space or on a particular surface by means of distances or angles or both with relation to designated map projection, datum, one or more standard parallels and a central meridian. (Minnesota Department of Transportation, 1992a).

Datum: A set of parameters and control points used to accurately define the three-dimensional shape of the Earth (e.g. as an ellipsoid). The corresponding datum is the basis for a planar coordinate system. (Minnesota Department of Transportation, 1992a).

A reference surface for horizontal or vertical measurements. Sources: Brinker/Wolf (Wisconsin Department of Transportation LCM Manual, 1995).

A base reference level for the third dimension of elevation for the earth’s surface. A datum can depend on the ellipsoid, the earth model, and the definition of sea level. (Clarke, 1997).

Dynamic Segmentation: Dynamic segmentation of lineal spatial objects provides a means by which new point or line objects can be created by relating the distance-referenced attributes with a manageable set of distance-referenced linear objects. Dynamic segmentation removes the need for a set of spatial objects for each attribute. Spatial objects and distance referencing of routes are used to create attribute-based spatial objects as needed. (Dueker and Vrana, 1992).

A method of referencing attribute data on demand, based on a variable segmentation of a single route or network structure. (Viggen, 1994).

Divided highways: A divided highway is a roadway where the opposing directions are separated by a median that restricts movement between the two directional roadbeds. Note that some GIS installations consider highways to be divided only if the scale of the map and the size of the median are such that the two roadbeds can be mapped separately.

Generalization: A reduction of detail and a transformation of cartographic data into a representation at a reduced scale. Sources: Weibel & Buttenfield, 1988 GIS/LIS Proceedings, p352. (Wisconsin Department of Transportation LCM Manual, 1995).

The process of moving from one map scale to a smaller (less detailed) scale, changing the form of the features by simplification, and so on. (Clarke, 1997).

Global Positioning System: A satellite based navigational system allowing the determination of any point on the earth's surface with a high degree of accuracy given a suitable GPS receiver. The US Department of Defense owns the network of satellites, and as such, the accuracy of the signal has in the past been intentionally degraded for non-US military users. The error introduced into the signal is known as selective availability. Error in the accuracy of GPS derived positions can also be introduced through the nature of local conditions, for example, multipath. These errors can be greatly reduced using a technique known as differential GPS. (Modified from Association for Geographic Information, http://www.agi.org.uk/.)

Traverse: A method of surveying in which lengths and directions of lines between points on the earth are obtained by or from field measurements, and used in determining positions of the points.

Linear Feature: A geographic feature that can be represented by a line or set of lines, for example, rivers, roads, and electric and telecommunication networks can all be represented as linear features.

Linear Referencing: Process of identifying location(s) on a network or specific link in a network by specifying a start position, direction and distance. (Viggen, 1994).

Linear Location Referencing Method: A mechanism for finding and stating the location of an unknown point along a network by referencing it to a known point. Note: This is a modification of the definition provided by Deighton and Blake (1993).

All linear referencing methods consist of traversals and associated traversal reference points that together provide a set of known points, a metric, and direction for referencing the locations of unknown points. No attributes are assigned to linear referencing methods. (Vonderohe and Hepworth, 1996).

Linear Location Referencing Systems: The total set of procedures for determining and retaining a record of specific points along a linear feature. The system includes the location reference method(s) together with the procedures for storing, maintaining, and retrieving location information about points and segments on the highways. (NCHRP Synthesis 21, 1974).

Milepoint: The name given to the numerical value of the mileage displacement from a base point to any location. (NCHRP Synthesis 21, 1974)

Mileage: A given distance expressed in miles.

Mileage marker: See milepost.

Mileage measure: See mileage.

Milepost: One of a series of posts or markers set along a highway or other thoroughfare to indicate distance in miles. (Academic Press Dictionary of Science and Technology, http://www.harcourt.com/dictionary/)

A physical entity, ordinarily a sign, placed beside a highway and containing a number that indicates the mileage to that point from some zero point on the highway. (NCHRP Synthesis 21, 1974)

Reference Markers: Physical objects along roads that may or may not have a simple relationship to the length of roads and that form control points with a route and MP measure. (Dueker and Vrana, 1992)

Reference Point: A fixed identifiable feature, such as an intersection, railroad crossing, or bridge, from which a location can be measured or referenced. (NCHRP Synthesis 21, 1974)

Reference Post: A physical entity, ordinarily a sign, placed beside a highway and containing a number that does not reflect a milepoint, but is an identification number for the point of location of the post. The identification number is associated with the actual milepoint of the location in office records. (NCHRP Synthesis 21, 1974)

Scale: Definition 1: The proportion between two sets of dimensions. Sources: Webster’s Dictionary as modified by David Fletcher (1991).

Definition 2: In relation to maps: "The best scale for your map depends on the resolution of the original data, as well as the level of detail you want your map to include. For example, a 0.25 in on a 1:250,000-scale map represents approximately 1.0 mi (640 acres) on the ground. But a 0.25 in on a 1:63,360 scale map represents 0.25 mi (160 acres)." Sources: Understanding GIS (Rev 6), Esri, Lesson 9, page 13. (Wisconsin Department of Transportation LCM Manual, 1995).

State Plane Coordinate System (SPCS): The plane-rectangular coordinate systems developed by the U.S. Coast and Geodetic Survey (now known as the National Geodetic Survey or NGS), one for each state in the United States, for use in defining positions of geodetic stations. Each state is covered by one or more zones, over each of which is placed a grid imposed upon a conformal map projection. Zones having limited north-south dimension and indefinite east-west extent have the Lambert conformal conic map projection with two standard parallels as the base for the state plane coordinate system. Zones for which this sequence is reversed (i.e. limited east-west dimension and indefinite north-south extent) have the transverse Mercator projection as the base. (Minnesota Department of Transportation, 1992a).

Endnotes

[1] The U.S. Geological Survey publishes accuracy standards for Digital Line Graph (DLG) data. As applied to the U.S. Geological Survey 7.5-minute quadrangle topographic map, the horizontal accuracy standard requires that the positions of 90 percent of all points tested must be accurate within one-fiftieth of an inch (0.05 cm) on the map. At 1:24,000-scale, one-fiftieth of an inch is 40 ft (12.2 m).

[2] Most terms and definitions have been taken from the BTS Resource Guide Glossary with credit in parenthesis to the contributing person or organization.

References

Breyer, Jerome P., P.E., Tools to Identify Safety Issues for a Corridor Safety Improvement Program, Proceedings of the Transportation Research Board (TRB) 79th Annual Meeting, January 9, 2000.

FHWA-RD-99-081, GIS-Based Crash Referencing and Analysis System, U.S. Department of Transportation, Federal Highway Administration.

FHWA-RD-00-033, HSIS – Highway Safety Information System: The essential information system for making informed decisions about highway safety, U.S. Department of Transportation, Federal Highway Administration.

Dueker, Kenneth J. and Ric Vrana, Dynamic Segmentation Revisited: A Milepoint Linear Data Model, GIS-T Symposium Proceedings, AASHTO, 1992. (Also Journal of the Urban and Regional Information Systems Association, Volume 4, Number 2.)

GIS Safety Analysis Tools V1.0, CD-ROM, U.S. Department of Transportation, Federal Highway Administration, 1999.

Highway Location Reference Methods: Synthesis of Highway Practice, Transportation Research Board, National Academy of Sciences, Washington, D.C., 1974.

HSIS State GIS Review, University of North Carolina, Highway Safety Research Center, Prepared by GIS/Trans, Ltd., October 1999.

Linear Reference Practitioner’s Handbook, prepared by GIS/Trans, Ltd. for the Federal Highway Administration, Draft November 20, 1997.

Miller, John S., The Unique Analytical Capabilities Geographic Information Systems can Offer the Traffic Safety Community, Proceedings of the Transportation Research Board (TRB) 79th Annual Meeting, January 9, 2000.

Resource Guide on the Implementation of Linear Referencing Systems in Geographic Information Systems, Bureau of Transportation Statistics, U.S. Department of Transportation, CD-ROM, 1999.

TIDE: A GIS-linked Data Warehousing Approach for Building an Integrated Transportation Information Environment, ASSHTO GIS-T ’98, June 3, 1998.

Author Information

Richard C. Smith
Senior Analyst/Programmer
GIS/Trans, Ltd.
8555 16th St., Suite 700
Silver Spring, MD 20910
Phone: 301-495-0217
Fax: 301-495-0219
Email: rsmith@gistrans.com