On the last half day of the meeting, the peer review panel spent about one-and-one-half hours in an executive session, closed to all other participants of the meeting. The reason for this closed session was to allow panel members to speak freely and openly among themselves while developing formal recommendations. This section details those panel recommendations.
The panel provided guidance for updating model inputs and observed data, including socioeconomic, land-use, and economic data, household travel survey data, and observed transit and traffic data.
The panel agreed that SRTC does not need to develop a land-use or economic model. However, they do need to develop consistent procedures across jurisdictions for developing economic, land-use, and socioeconomic forecasts. SRTC should seek out expert opinion to ensure that the developed procedures are valid. The panelists encouraged SRTC to make use of the State of Washington’s purchased REMI data, but cautioned SRTC to make sure the data units are consistent with what is needed for the SRTC model. SRTC should regularly review local jurisdiction comprehensive plans to ensure that SRTC’s land-use forecasts match-up to local land-use forecasts. SRTC should improve its methodology for developing economic forecasts. It is state-of-the-practice to develop economic forecasts independently and before population forecasts and then base population forecasts off of the economic forecasts. SRTC’s current practice of basing employment forecasts directly off of population forecasts based on historical trends in the data may work for near term forecasts but not for long-term forecasts. In the short term, the ratio of population and employment may be held constant, but this ratio may vary substantially 20 or 30 years later due to several factors, such as population aging, productivity improvement, economy structure change, which are all in the domain of economic forecasting.
The panel recommended that SRTC move forward with collection of new household travel survey data using a reputable and experienced consultant. They suggested that SRTC should collaborate with local universities on data collection to ensure that University travel is appropriately surveyed. They also suggested that SRTC review the travel behavior, such as trip rates and lengths, from the existing 2005 household travel survey. They should review trends between Year 2005 to 2016 in trip rates and sociodemographic characteristics from ACS, NHTS and from cities similar to Spokane. They should also review Spokane’s current socioeconomic profile, transportation system, and spatial distribution of population and employment to see the region has changed substantially from 2005. Some aspects of travel patterns, such as trip rates, revealed by household travel surveys are largely driven by demographics and socioeconomics. The others, such as mode share and trip length, are also affected by the changes of transportation system and spatial distribution of population and employment. Therefore, if the regional demographics and socioeconomics have not changed substantially since the last survey, and the population and employment growth has occurred in the same manner as in the past, SRTC may not need a new survey. If they are comfortable that trip rates and lengths have not changed significantly over the past ten years, then they could update the travel demand model using existing survey data in conjunction with pursuing new household travel survey data collection. However, if SRTC believes travel behavior has changed substantially, then they should collect household travel survey data first before performing significant updates of the model.
The panel encouraged continued collection of transit on-board survey data and APC deployment, as that data is an essential resource for model validation. They also suggested to incorporate periodic tablet-based surveys into SRTC’s data collection plan.
The panel assumed that the current observed traffic count data collection was adequate, but stressed the importance of these counts for model validation. They did suggest to make sure that when a Bluetooth survey is conducted for OD analysis that SRTC gets hold of the data before it is thrown out. They also suggested looking into private aggregate speed data providers.
In response to SRTC’s request for guidance on model improvements the peer review panel made recommendations on each component of the travel demand model.
The peer review panel made several recommendations to SRTC to improve their trip generation and trip distribution models:
The peer review panel recommended continued use of a nested logit model for mode choice application. They recommended that bike and walk should be split into separate modes, but there is no need to assign these trips. The light rail placeholder should be removed. With regard to the transit mode, the panel assured SRTC that locally adjusted asserted transit parameters are acceptable and to make sure to review the alternative specific constants when reviewing household travel survey weights. The panel also suggested that captive ridership is hard to express in mode choice utility functions, and the model might do better estimating ridership levels in other ways (e.g. with STOPS). A failure to include captivity may cause too great a sensitivity to service improvements, since captives are, by definition, relatively insensitive to LOS parameters. One way to capture transit captivity is to base transit captivity on the number of zero-auto households (there is a much lower percentage of captive riders than households without cars), then allocate this captive percent directly to the transit system (if there is access). Then add choice riders as necessary. A good check on the captive/choice ridership ratio is the transfer percentage, since captive riders transfer at a higher rate than choice riders.
The peer review panel recommended several improvements to traffic assignment model including the following:
The peer review panel suggested that SRTC continue to use headway-based assignment. However, they did suggest to consider a targeted investigation of the region’s bus schedules to see where a headway-based model may fail to adequately measure transfer time. They encouraged SRTC to investigate the best process for revising the transit assignment time periods. They also noted that SRTC should check to make sure that wait time is capped at 30 minutes.
The peer review panel made several additional recommendations to SRTC related to improvements of their travel demand model:
