RoadRight: An open-access decision support system for prioritizing roads for repair or decommissioning

The U.S. Forest Service manages 380,000 miles of roads that were largely built to extract natural resources.  They cause many types of environmental damage, such as habitat fragmentation and siltation of streams. The agency has a mandate to crumbling-roaddecommission many of the now unnecessary roads.  In this time of scarce funding, effective prioritizing of which ones to decommission or repair is essential.  Our collaborative has developed a GIS-based process for systematically helping with prioritization. For more background and overview, please see the Summary Document.

The RoadRight Decision Support System (DSS) uses a hierarchical multi-criteria framework to identify the road segments that seem to be high priority for decommissioning or repair.  First, all roads Road-Right-Multi-Criteria-Hierarchy_v1are classified as either candidates for decommissioning or for repair based on several social criteria, such as if the road leads to any point of interest, is necessary to protect communities from fire, etc.  Then it takes into account 17 input criteria (such as road grade, number of stream crossings, passing through critical species habitat, etc.) to rank the “risks” of each road.  RoadRight runs within ArcGIS and is being released as an open-access software to be used, developed, and/or referenced by organizations and agencies interested in restoration.  For more details, please see the Summary Document or the User Guide.

Download the Model

Request the model, or contact us with any comments:

About the Model

Principles of RoadRight

Rather than operate in the mode of competitive and self-promoting scientific research, The Landscape Collaborative stresses teamwork, and has the following principles for this system, as it does for all of its systems:

  • Open Access: The full system “code” (ArcGIS ModelBuilder models and Python scripts) is released under the General Public License.
  • Transparency: The system strives to be clear and understandable, with all models and scripts viewable by the GIS analyst.
  • Flexibility: The end-users have the flexibility to change the weights among criteria to meet their specific needs.
  • Expandability: The system is designed to allow inclusion of additional criteria over time.

A decision support system is, as its name indicates, simply a tool to support good decision-making. It provides a means of compiling and measuring information that stakeholders supply. It does not make decisions.


Development of this system is a collaborative venture. The RoadRight project was conceived and initiated by Matthew Dietz of The Wilderness Society, and modeled manually for two criteria by Jennifer Boggs of The Wilderness Society. John Gallo of The Wilderness Society suggested that the process be automated in ModelBuilder, and assisted Jennifer Boggs in the creation of version 1.0. He then created version 2.0, which included additional ecological criteria. Feedback on model development was provided by U.S. Forest Service staff and incorporated into the model at several stages, including vital contributions from Don Yasuda, Regional Analyst in the Pacific Southwest Region. Kurt Menke of Bird’s Eye View GIS and Hugh Irwin, then with Southern Appalachian Forest Coalition, ran early versions of the system and provided evaluation and troubleshooting. Recently Kurt Menke evaluated the model for accuracy and contributed to the User Manual. Anyone can provide future contributions in accordance to the open-access terms and conditions in the model’s zip folder.

System Requirements

The RoadRight toolbox requires an ArcInfo-level license copy of ArcGIS 10.0 with a Spatial Analyst extension.  It can also run with the less expensive ArcGIS license called ArcView, but will need a small amount of additional programming. Here are the minimum specifications published by Esri for a computer running ArcGIS 10.0.

Note: It is likely that it will run well in ArcGIS version 10.1. However, it has not yet been adequately tested on that platform.


Version 2.2
  • Release: April 16, 2013
  • Released for Inyo, Sierra, and Sequoia NF
  • Improvements
    • Fixed a bug in the isolation index
    • Other improvements that Kurt can mention at some point
Version 2.1.1
  • Release: Sept 21, 2012
  • For Shasta Trinity Forest. (see the above Forest Specific Notes)
  • Improvements
    • Makes the spatial referecne of the .mxd datalayers a parameter or model 1000d rather than hardcoded.
    • Makes the creation and population of the KW_Value field a pre-processing condition, due to inconsistencies of SSURGO
    • The ShastaTip Data Pre-Processing Model was created, and is also on the Downloads page, in a separate zipfile.
    • Still needs to rename the substitutions as per the forest specific notes.
    • Tip: this is how you make text of a different color  (see wiki markup tab)
Version 2.1
  • Release: June 28, 2012
  • Improved outputs
    • A layer file for each criterion,
      • parameter: display the layers in either a 3 or 5 class output.
      • parameter: make summary layers fo just the roads deemed eligible for decommissioning, those that are deemed ineligible, or all roads
    • The 30 m segment raw output grids are now in the outputs geodatabase.
Version 2.0
  • Release: May 31, 2012
  • Summarizes cells to roads before normalizing the solution set of roads
  • Provides a drop-down menu of which type of normalization to use on all criteria.
  • Provides a separate drop-down menu for which type of normalizationto use on the final criterion output
  • Provides a new normalization technique: “mean to mid value without zeros”  (see user guide
Version 1.3.4
  • Release: April 15, 2012, without an update to the User Guide, classified as a beta release.
  • Two inputs geodatabases provided, one is a small subset of data for faster processing/testing.
    • (sia is a copy or rna and not true, just used for testing)
  • Bug fixed to allow model 2 to run through automatically rather than having to run the first several steps individually.
Version 1.3.3

Internal release

Version 1.3.2
  • Release: March 27, 2012, without an update to the User Guide; classified as a beta release
  • Segmented the problematic “1000” model into four steps to improve performance and ease any troubleshooting
  • The pre-processing model (described in Appendix C of the User Guide) for the five forests is available upon request, and will be released in the next major release.
Version 1.3.1
  • Released March 12, 2012
  • Road ID improvement
    • now all roads in the roads input file have their Road ID (and all other original fields) so the final result can be queried by this attribute.
  • partial fix to the <null> values bug.
    • version 1.2 had a preponderonce of <Null> values in the output tables.  This was due to the cell size of the raster that helps convert raster data to vector data.
      • new cell size is 7.5 m per side, so there are a handful or road segments that are less than 7.5 m long that get <null> values.
  • Bug fixed: road polyline to raster
    • in the first two sub-models of Model 2, the polyline to raster command was outputting a cell size relative to the project extent, now it is set to 30 m
  • Appendix C added to user guide (tips on how to pre-process your data
Version 1.2

Some of the changes compared to version 1.1 are as follows:

  • Key Species Habitats (see below diagram), currently comprised of a weighted sum between
    • Spotted Owl PACs
    • Goshawk PACs
    • Critical Species Habitat
  • Several other weighted sums have been added as flexible parameters
    • Important Designations
    • Important Territory Index
    • Road Topography
  • Improved output
    • Three shapefile/table combinations are now created:
      • All Roads
      • Just roads eligible for decommissioning according to the screen”
      • Just roads not eligible, and hence, candidates for mitigation
  • Optional inputs
    • all 12 inputs that feed into the Important Territory Index are now “optional”; if you do not have one or several the model will still run.
  • Improved Workflow
    • Now just need to run the prep tool, and then the all models combined tool.
    • Tips for running it in Windows 7 environment are in the Collaboratory; still need to be added to the user guide.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s