Case Study:

Portland, Oregon

Methodology

User Benefits

The STEAM Model

The Surface Transportation Efficiency Analysis Model (STEAM) was developed by FHWA in 1997. Full documentation of the model can be obtained from the FHWA's STEAM web site. The primary objective of STEAM is to provide a framework for estimating the network-wide impacts of multimodal transportation alternatives. STEAM accepts inputs from four-step travel demand models, including trip tables, network and centroid files, and time and cost change files. STEAM then computes the following measures of effectiveness:

• Benefits and costs to transportation users;

• Annualized cost to public agencies;

• Change in criteria pollutant emissions;

• Change in energy use;

• Change in noise and other external costs;

• Change in fatal, injury, and property damage only accidents;

• Revenue transfers due to toll or fare changes; and

• Monetary value of user benefits, costs, emissions, noise, and accidents.

STEAM operates in a Windows environment. As shown in Figure 3, STEAM contains four modules:

1. A User Interface Module, to specify parameters and file names, conduct runs, etc.
2. A Highway Network Analysis Module, which reads a file containing vehicular traffic volumes, highway segment lengths, highway capacities, and other link data and produces zone-to-zone travel times and distances based on minimum time paths through the network.
3. A Trip Table Analysis Module, which produces estimates of user benefits based on a comparison of Base Case and Improvement Case conditions. It also produces estimates of emissions, noise costs, accident costs, energy consumption, and other external costs associated with highway use.
4. An Evaluation Summary Module, which calculates net annual worth and a benefit-cost ratio for the improvement under consideration. It also provides summary information on individual benefit and cost items.

Figure 3. Overview of STEAM

A noteworthy feature of STEAM is that it post-processes traffic assignment outputs from conventional four-step planning models. This provides more accurate highway travel speeds, especially under congested conditions, compared to many four-step models. The speed models in STEAM are calibrated separately for freeways and signalized arterials, under both peak and off-peak conditions, based on hour-by-hour simulations of traffic volumes and queuing for facilities with different levels of congestion. The speed models account for:

• Delays due to incidents, using data on the frequency, severity, and duration of incidents;

• Peak spreading that occurs when facilities become more congested;

• Day-to-day variations in traffic; and

• Decreases in highway capacity observed after demand volumes exceed capacity.

As a result, speeds and related values, such as vehicle-hours of travel, are likely to differ compared to those estimated using the regional travel demand model. In cases in which local agencies prefer to use their own estimated speeds, STEAM can still be used to calculate user benefits and other performance indicators.

STEAM includes other useful features to help develop more accurate estimates of user benefits. For example:

• The model allows consideration of induced demand. This is done using a default or user-specified elasticity of travel with respect to travel time. Benefits to new users as well as to existing users are estimated on each link in the network using the concept of consumer surplus (see the STEAM documentation for details). The consumer surplus concept is also used to estimate changes in benefits for users who shift routes, modes, or origins/destinations.
• The model can perform risk analysis to clearly describe the level of uncertainty in the results of the analysis. STEAM accomplishes this by allowing ranges (probability distributions) to be attached to the forecasts of each input variable. The risk analysis can help assess which input assumptions have significant impacts on the outputs and which do not. Figure 4 shows sample output for a cost-benefit analysis. The cumulative probability chart in Figure 4 indicates that there is a roughly 70 percent probability (1 - 0.3) that the benefit-cost ratio is greater than one, and a roughly 40 percent probability (1 - 0.6) that the benefit-cost ratio is greater than two.

Figure 4. STEAM Risk Analysis Results: Cumulative Benefit-Cost Ratio

Application of STEAM

In the Portland I-5 freight study, STEAM was used to calculate user benefits based on the difference between the base case and each alternative investment strategy. STEAM was run using inputs provided by the Metro travel model and the default STEAM parameters for value of time, fuel consumption rates, accident rates, etc. Metro also provided local data to override these defaults in some cases.

The STEAM results used in the study included monetary equivalents for the change in four user benefit components: travel time, accidents, non-fuel operating costs, and fuel costs. STEAM generated an estimate for each component for each of the three peak periods and each commodity/vehicle type. These were then converted into 24-hour values to estimate overall benefits.

Some of the specific procedures involved in calculating user benefits include:

In-Vehicle Travel Time Savings. The value of time to vehicle occupants was obtained from the U.S. Department of Transportation guidance (U.S. DOT, 1997). In the Portland study, the value of time was expanded to include value per vehicle and value of inventory. The values of travel time used in the Portland study are summarized in Table 1.

Table 1. Value of One Hour of Travel Time (1995 Dollars)

To compute the inventory costs for five-axle combination trucks, an hourly discount rate was computed and multiplied by the value of a composite average shipment. The discount rate selected was 9.8 percent, equal to the average prime bank lending rate in 1995 plus one percent. Dividing this rate by the number of hours in a year produces an hourly discount rate is 0.0033 percent. The average payload of a five-axle combination is about 35,000 pounds. In 1993, the average value of commodities shipped by truck was $1.35 (on a ton-mile weighted basis). (Source: U.S. Bureau of the Census, 1992 Census of Transportation, 1993 Commodity Flow Survey, U.S. Government Printing Office, Washington, D.C., 1996. Values from Table 6 data on value, tons and ton-miles by distance shipped.) Inflating to 1995 dollars using the GDP deflator and multiplying by the average payload produces an average payload value of roughly $50,000. The resulting time value of the average payload is approximately $0.60 per hour (ignoring any costs for spoilage and depreciation over time). Payload for four-axle combination trucks is lower than for five-axle combination trucks, but the value of the cargo probably is higher. Consequently, the value per shipment was assumed to be the same for both types of trucks.

To compute average value per person, the sum of value per vehicle and value of inventory is divided by the average vehicle occupancy and added to the in-vehicle value per person.

Source: Cambridge Systematics, Inc., U.S. Department of Transportation, and Highway Economic Requirements System (HERS).

STEAM calculates the change in vehicle travel time by running trips from Metro's trip tables through the EMME/2 networks, comparing the base network results to those of each alternative network. Total travel time is based on the speed and distance traveled by each trip through the EMME/2 network, summed for all trips. Congested speeds and link distances come directly from EMME/2, and STEAM determines the minimum time path.

An important assumption built into the Metro projections of on-the-clock travel, which in this study consisted of only heavy and medium truck trips, is the absence of additional truck trips caused by induced travel. Unlike commute or other trip purposes, Metro assumes that businesses produce the amount of on-the-clock trips necessary to operate efficiently and do not consider the amount of congestion, delay, accidents or other per vehicle-mile costs as a constraint to their trip-making.

Vehicle Operating Costs. Vehicle operating costs in STEAM are a combination of two components: fuel costs and non-fuel costs. Fuel costs depend on both speed and VMT. STEAM uses a series of fuel consumption rates at different speeds in addition to average fuel cost as key inputs to this calculation. For each trip, STEAM sums the cost of fuel used on each and every link of the trip, based on the distance of each link and the congested speed on that link. Non-fuel costs are VMT-dependent costs associated with operating a vehicle. These account for oil consumption and maintenance costs. STEAM simply multiplies a cost factor by the VMT of each trip, summing all trips for total non-fuel costs.

Accident Costs. STEAM determines accident costs based on VMT and the facility-based accident rate. For each trip, the product of the length, accident rate, and accident cost on each link is added up for all links on a trip and for all trips. Costs per accident are provided for fatal, injury, and property damage only accidents.

[TOP]