Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 28
16 I N N O VAT I O N S I N T R AV E L D E M A N D M O D E L I N G , V O L U M E 1
spreading model component that was used was trans- San Francisco tour model with transit boardings in San
ferred from the Metropolitan Transportation Commis- Francisco, which may overestimate from the original
sion (MTC) model. Some issues had to be addressed in model. These differences resulted in a reduction in transit
transferring the model and expanding it for different trip trips in the tour model compared with the regional model
purposes. The peak-spreading model has been updated and a corresponding increase in driving alone and walking.
based on the results of FHWA-sponsored research on This difference was a calibration issue more than a differ-
integrating time-of-day models with activity-based mod- ence in the models. Tour-based models and trip-based mod-
els. Due to limited resources, a traditional aggregate els are both validated to observed data. The differences
assignment was used for trip assignment rather than a that were identified in the validation process related to dif-
microsimulation assignment. ferences in the underlying data sets, not the models.
· The model was developed for SFCTA. To maintain · There are differences in the outputs from the two
consistency with the regional model, the MTC's regional models for the forecast year. A comparison of the trip
trip-based model was used and integrated with the tour- tables for 2030 highlights one of these differences. Most
based model for San Francisco County. This approach of the differences by districts are small. The San Fran-
presents some limitations on the cross-county move- cisco tour model shows a larger increase in trips in the
ment. Stated preferences surveys were used to collect suburban district and a drop in trips in the intradown-
data on crowding and reliability and the impacts of these town district. The MTC trip-based model shows more
two features on mode choice. Equilibrium measures of growth in trips to the downtown district and more
time were estimated for commuters and noncommuters growth in intradowntown trips.
to higher and lower levels of crowding and reliability. · In terms of mode share, both models show an
Although there were significant effects from these mea- increase in drive-alone trips for 2030. The growth in sub-
sures, the results were not intuitive to the transit board- urban portions of the county, which do not have good
ing data available at the time. As a result, they were transit access, may account for this increase in drive-
taken out of the model. alone trips. The MTC trip-based model shows a more
· The model validation process required significant significant drop in walk trips, while the San Francisco
resources. A variety of traditional data sets were used for tour model has a more significant decline in walk-to-
validation purposes. Validation was conducted for each transit trips. In the San Francisco model, walking is inte-
model component separately. Additional validation was grated as part of many different types of tours that
conducted by comparing the model to the trip-based people make during a day. As a trip-based model, the
regional model for each model component. The compar- MTC model does not have this feature. Increases in trip
ison to the four-step model was conducted for both the distances impact the number of walk trips.
base year and the forecast year. The comparison, which · The San Francisco tour model has been used for a
was conducted for San Francisco County residents, number of different applications. The model has been
included all the input data, the assumptions, and the well received by technical personnel, policy makers, and
model output for the base year and the forecast year. other groups. The model has been used for both tradi-
Because of the limitations in the trip-based model, which tional planning studies, as well as projects utilizing the
produced only trips and not tours, the comparisons were tour-based features. A disaggregate equity analysis was
made at the trip level. conducted to examine possible unintended consequences
· The trip generation comparison included examin- of countywide improvements being considered in the 30-
ing the trip rates per household for different trip pur- year plan. The analysis focused primarily on two factors:
poses. The other, non-home-based trip categories were mobility, as measured by total travel time for a group or
overestimated in the San Francisco tour-based model, total transit travel time for a group, and accessibility, as
while the work and school trips were underestimated. measured by the total amount of employment that could
These differences appear to be the result of using estima- be reached within 30 minutes of a zone or the total
tions based on two different surveys, rather than the amount of retail that could be reached within 30 minutes
models. A comparison of the district-to-district trip table of a zone. The different groups examined in the analysis
summary showed very little difference in the two mod- were households with no automobile available, low-
els. One of the noticeable, although not significant, dif- income households, female-headed households with chil-
ferences was in the higher percentage of trips in the San dren, and single-parent households. The no-automobile
Francisco CBD zone for the San Francisco tour model. It households and the low-income households received
appears that this difference also results from the under- most of the benefits from the countywide plan because
lying survey data set and not the models. the improvements focused primarily on the transit sys-
· The mode share components in both models were tem. Female-headed households with children received
estimated based on the same data sets. The differences in few benefits from the plan. It may be that these house-
mode share appear to be a by-product of calibrating the holds are not making trips by transit.
