|
Air quality analysis of transportation: Is it important to model the land use effects of transportation scenarios? by Caroline
J. Rodier Department
of Environmental Science and Policy University
of California One
Shields Avenue, Davis 95616 (530)
757-2791 & fax (530) 752-3350 John
E. Abraham Department
of Civil Engineering University
of Calgary 2500
University Drive NW Calgary,
Alberta CANADA
T2N 1N4 (403)
230-5897 & fax (403) 282-7026 and Robert
A. Johnston Department
of Environmental Science and Policy University
of California One
Shields Avenue, Davis 95616 (530)
582-0700 & fax (530) 582-0707 Submitted
for presentation at the TRB annual conference in 2000 and
for publication in the Transportation Research Record July
29, 1999 TABLE
OF CONTENTS ABSTRACT Within the last decade, U.S. legislation and regulations have provided a mandate for planning agencies to analyze the relationship between land use and transportation decisions. The Transportation Equity Act for the 21st Century (TEA-21) of 1998 requires that transportation planning consider the effects of transportation policy decisions on land use and economic development. The U.S. Environmental Protection Agency’s (EPA) conformity regulation for the Clean Air Act Amendments of 1990 (CAAA) require a logical correspondence between future regional land use projections and transportation plans in serious or worse nonattainment regions (40 CFR 93.122(b)(1)(iii)). The CAAA also allows for the evaluation of land use policies that may reduce vehicle travel and emissions. Decisions in recent litigation have also endorsed the importance of analyzing the land use and transportation interaction. In the San Francisco Bay Area ( Sierra Club v. Metropolitan Transportation Commission, 1990), the district court required the Metropolitan Transportation Commission to account for the effect of transportation decisions on land use in its regional travel and emissions analyses. More recently, a U.S. District Court case in the Chicago region held that the National Environmental Policy Act requires the consideration of land development changes when a new freeway segment is analyzed. These decisions suggests that, unless regional planning agencies represent the land use and transportation interaction, it is not unreasonable to expect an increase in suits aimed at slowing or halting new highways. The history of urban development and transportation technology seems to provide clear evidence for the land use and transportation interaction. When rail first appeared in cities the central areas around stations grew dramatically. As new technology reduced travel time and costs to more areas, as with the streetcar and then the automobile, cities began to decentralize until low-density suburban development became the norm for the modern metropolitan area. However, recently in the U.S. some have questioned the current significance of the relationship between land use and transportation and its effect on travel patterns and vehicle emissions (1). The results of some U.S. studies in which regional travel demand models have been integrated with land use models indicate that the effects are small (2,3). Many question those results and argue that the land use models used in those studies were far less than state-of-the-art (4). A recent Academy of Science Panel concluded that the addition of highway capacity can have a “decentralizing effect on urban development” (1, pg. 222). THE SACRAMENTO REGIONThe Sacramento region is located in the central valley of Northern California. In 1995, the region was estimated to have a total population of 1.8 million and total employment of about 700,000. Population is expected to grow annually at a rate of 1.9% to 2015 and employment is expected to grow annually at a rate of 2.2% to 2015 (7). Average household income in 1995 was about $63,000 dollars. In the past, the employment base of the Sacramento region has been largely government and agriculture; however, more recently there has been a rapid expansion of high technology manufacturing. The residential and employment densities of the region can be characterized as medium to low. Current mode shares for home based work trips are approximately 76% drive alone, 17% carpool, 3% transit, 2% walk, and 2% bike. | TOP | METHODSAn Integrated Land Use and Transportation Model: MEPLANFrom the mid-1960s to the late 1970s, research in the Martin Centre at Cambridge University, England spawned a family of interactive land use and transportation models known as the Martin Centre Model (8). One of the models developed from this structure is Marcial Echenique’s software package know as MEPLAN (9). The MEPLAN model integrates three economic models: (1) an input-output or social accounting model, (2) a random utility model of location choice integrated with the social accounting model in a way that is similar to, but more general than, Lowry’s 1961 model, and (3) a rent-density function based on Alonso’s 1964 theory of urban land markets or bid-rent theory (8). The input-output model uses exogenous basic demand in each time period to generate endogenous economic activities (population and nonbasic employment). The random utility model of location choice spatially allocates basic employment, residents, and nonbasic employment in a series of iterations until the land markets equilibrate. The result is the estimation of the amount and location of population and employment, land use, rents, and flows of economic activities between locations (e.g., from home to work). These flows are then transformed into person trips and truck movements between origin-destination pairs. The model is “quasi-dynamic” in that the time and monetary costs of travel from the transportation model are fed back to the land use model in the next time period and the amount of development in each zone between time periods is a function of the prices and arrangement of activities in the previous time period. Abraham and Hunt describe in detail the calibration and structure of the Sacramento MEPLAN model (10). For this study, a trip assignment model with separate A.M. and P.M. peak (both 3 hour peak) and off-peak periods was added to the Sacramento MEPLAN model. An Urban Transportation Planning Model: SACMETThe standard Urban Transportation Planning (UTP) model was developed in the late 1960s and early 1970s to support large-scale regional transportation studies in the U.S. and to determine the need for additional roadway lanes or segments to relieve traffic congestion. State-of-the-practice versions of the UTP model, like SACMET, have adapted it to better address air quality problems by improving the models’ representation of congestion, travel modes, auto ownership, land use variables, and time and cost variables. DKS Associates developed the SACMET model for the Sacramento Area Council of Governments (SACOG) with a 1991 travel behavior survey conducted in the region (11). Some of the key features of this model include: (1) full model iteration on level of service variables; (2) auto ownership and trip generation steps with accessibility variables; (3) a joint destination and mode choice model for work trips; (4) mode choice models with separate walk and bike modes, walk and drive access to transit modes, and carpool modes; (5) mode choice models with land use, travel time and monetary costs, and household attribute variables; (6) all mode choice equations in logit form; and (7) a trip assignment step with separate A.M. and P.M. peak (both 3 hour and 1 hour peak) and off-peak periods. | TOP | The California Department of Transportation's Direct Travel Impact Model 2 (DTIM2) and the California Air Resources Board's EMFAC7F emission factors were used in the emissions analysis. The outputs from SACMET and MEPLAN used in the emissions analysis include the results of traffic network assignment for each trip purpose by each time period (A.M. peak, P.M. peak, and off-peak). SACOG provided regional coldstart and hotstart coefficients for each hour in a twenty-four hour summer period. | TOP | SCENARIOSThe scenario descriptions apply to the scenarios modeled by both MEPLAN and SACMET except when differences are identified. All the transportation network improvements are made in the year 2005 for the MEPLAN scenarios and thus land use is affected in 2010 and 2015. Trend. The trend scenario represents a financially conservative expansion of the Sacramento region’s transportation system and serves as a point of comparison for the other scenarios. This scenario includes a relatively modest number of road-widening projects, new major roads, one freeway HOV lane segment, and a limited extension of light rail. High Occupancy Vehicle (HOV). The high occupancy vehicle (HOV) lane scenario represents an significant expansion of the Sacramento region’s HOV lane system to encourage the use of carpools and reduce traffic congestion and emissions. HOV lanes are increased from 26 to 179 lane miles. Light Rail and Pricing. The light rail transit scenario represents an extensive expansion of the Sacramento region’s light rail system to encourage the use of public transit and to reduce traffic congestion and emissions. In total, these expansions include approximately 75 new track miles of light rail. In the MEPLAN scenario, auto pricing policies are added to this scenario. These policies include a 30% increase in the operating cost of private vehicles (to simulate a gas tax) and a CBD parking tax representing an average surcharge of $4 for work trips and $1 for other trips. In SACMET, an auto pricing only scenario is simulated, which includes a parking scenario in which home-based work trips are charged $5/day for workplace parking (in the CBD zones as well as a number of zones projected to have relatively high employment densities in the future). In addition, a five cent VMT tax on all trips and a peak period charge of an extra 10 cents per mile for home-based work trips is imposed. Transit Oriented Development (TOD). Transit oriented developments or TODs describe mixed-use centers with a relatively high intensity mix of shopping, jobs, and housing located along light rail transit lines. In TODs, many activities are within walk or bike distances, and high quality transit service is readily available for activities that are outside the TOD. The increased densities in the TOD scenarios were modeled differently in the MEPLAN than in the SACMET model. The method of simulation in each model illustrates some of their respective strengths and weaknesses. The MEPLAN model is theoretically comprehensive, representing land markets with endogenous prices and market clearing in each period. As a result, MEPLAN can simulate such policies as, for example, the release of zoning density caps near to rail stations, tax benefits for infill development, and land development fees on raw-land projects near the urban edge. In this MEPLAN simulation, increased densities in the TODs are achieved through land subsidies of 5% of expenditures in the year 2000 on land rent in the TOD zones. The subsidies are offset by 30% land rent surcharges in other zones so that regionwide the effect is revenue neutral. SACMET does not have a land use model and thus cannot simulate large-scale land use policies such as land subsidies and taxes. However, SACMET has small zones, detailed travel networks, and includes zone-based walk and bike accessibility variables. Thus, in the SACMET simulation, increases densities in the TODs are achieved by manually adjusting zonal land use. Growth in households, retail employment, and non-retail employment from 1995 to 2015 in the outer zones (farther than 3 miles from the light rail lines) is moved to the zones in the TODs. Average densities in the TODs are assumed to be a total of 15 households per acre, 10 retail employees per acre, and 20 non-retail employees per acre. These densities are based on areas in the Sacramento region that are considered to be TOD prototypes. To reflect the improved walk and bike environment of the TODs, the pedestrian environmental factor (walk and bike accessibility variable) is increased in all the TOD zones. In general, the TOD densities in SACMET are greater than in MEPLAN. This is because the MEPLAN simulation could not match the SACMET TOD densities with a reasonable subsidy and taxation policy. In
this scenario, transit frequencies in the light rail network are doubled
and advanced transit information systems (ATIS) and local paratransit
service are added. ATIS for transit takes the form of pre-trip transit
service information. Transit users are assumed to access real time transit
scheduling information through 100 kiosks located at transit stations and
workplaces, the telephone, the Internet, and cable television.
Paratransit service provides an additional transit option to
travelers in TOD zones to connect travelers to light rail transit.
In MEPLAN, the value of wait time is reduced by a factor of three
to represent ATIS and the access time to transit in the TOD zone is
reduced by 3 minutes to represent paratransit service.
In SACMET, the maximum initial wait times for all transit service
is reduced to five minutes to represent ATIS and paratransit service is
simulated by adding new bus only routes with short direct routes between
TOD zones in the transit network. In the trend scenario, land development from 1990 to 2015 occurs north, east, and south of the City of Sacramento. There is limited land development in Yolo county because of exclusive agricultural zoning in the county. Over time, households and employment tend to locate primarily in existing build-up areas northeast, east, and immediately south of the CBD. In general, households and employment location tends to follow land development; however, density is increased in some zones. The land use results for the other scenarios are discussed in comparison to the trend scenario. Roadway expansion in the HOV scenario allows industry to locate further away from the households that it serves and employs. Compared to the trend scenario, the distant eastern zones that include the cities of Auburn and Folsom lose commercial employment and become more like “bedroom communities.” As a result of increased roadway capacity, retail activity can shift from local commercial to more remote zones where “big-box” retailing is likely to occur. In the light rail and pricing scenario, the parking charges in the CBD result in a loss of employment as businesses relocate to nearby zones to avoid the parking charges. There is also again in households because commercial activities are no longer willing to outbid residential activities. The increased mobility over short distances in central zones allows for a greater separation between households and employment. This is similar to what was found in the HOV lane scenario, but the effect is much smaller in this scenario and only occurs in the most central zones where light rail service is very good. The land subsidies and taxes in the TOD scenario have a dramatic effect on development. Almost all of the employment is attracted to zones with land subsidies and many zones that do not have light rail service lose employment (i.e., has lower growth over time). Households are also attracted to the subsidized zones, but to a lesser degree than employment. The rents in the subsidized zones go up and the rents in the taxed zones go down. This is because activities bid against each other to locate on the subsidized land. Hence most of the subsidies and taxes ultimately flow to the landowners. | TOP | In the HOV scenario, the greater distance between the home and workplace tends to increase HOV and transit use. This is because travelers take advantage of the faster travel times provided by the HOV lanes by carpooling or using commuter buses that are also allowed on the HOV lanes. The daily mode share projections for the 2015 MEPLAN scenarios are presented in Table 1. As a result, drive alone mode share is reduced somewhat. The walk and bike mode share is reduced in this scenario because the jobs and housing balance has been degraded compared to the trend scenario. These mode shifts produce a small reduction in auto trips, a significant increase in VMT, and an increase in mean travel speed compared to the trend scenario. Daily vehicle travel projections for the 2015 scenarios are presented in Table 2. Note that daily travel results in MEPLAN includes 18 hours of travel (3 hours for AM and PM peak and 12 hours for off-peak). In the light rail and pricing scenario, there is an increase in mobility over short distances in central zones where light rail service is very good compared to the trend (however, this increase is less dramatic than in the HOV lane scenario). The greater separation of home and work, the availability of high quality rail service, and the increase in auto operating costs serve to increase transit mode share significantly and to reduce drive alone mode share. There is an increase in the shared ride mode share in this scenario (even greater than in the HOV lane scenario) because ride sharing allows the cost of travel to be shared. Walk and bike mode share also increases. The mode shifts produce a decrease in auto trips, a significant decrease in VMT, and a slight increase in mean travel speed compared to the trend scenario. Increased densities and a better mix of households and employment in the TOD scenario produce dramatic increases in transit mode share and significant increases in walk and bike mode share. TODs make transit use quicker and cheaper and thus drive alone and shared ride mode shares are significantly reduced. Auto trips and VMT are also significantly reduced. Mean travel speed is increased slightly. TABLE 1 Daily Mode Share Projections for 2015 MEPLAN Scenarios.
a
Figures
in
parentheses are percentage change from the trend. TABLE 2 Daily Vehiclea Travel Projections for 2015 MEPLAN Scenarios.
a
Vehicle
trips include drive alone and HOV mode only. b
Figures
in
parentheses are percentage change from the trend | TOP | In
general, the MEPLAN emissions results follow the travel results described
above. The HOV scenario
increases emissions compared to the trend scenario.
The TOD scenario provides the greatest increase in emission
reductions followed by the light rail and pricing scenario.
The daily emission results for the MEPLAN scenarios are presented
in Table 3. TABLE 3 Daily Emissions (tons) Projections for 2015 MEPLAN Scenarios.
a
Figures
in
parentheses are percentage change from the trend. | TOP | COMPARISON
OF MEPLAN RESULTS TO SACMET’S The mode share, the daily vehicle travel, and emissions projections for the SACMET scenarios are presented in Tables 4-6. As described above, the scenarios simulated in SACMET are somewhat different from the scenarios simulated in MEPLAN. The results of the pricing scenarios in SACMET can be roughly compare to MEPLAN’s light rail and pricing scenario. A light rail only scenario simulated with SACMET made only minor changes in mode share and vehicle travel projections compared to the trend. In general, the TOD scenario has a much greater intensity of household and employment location in the TODs in the SACMET scenarios compared to the MEPLAN scenarios. An approximate comparison between the results of the two models on roughly comparable scenarios may be suggestive of the contribution that the simulation of the land use and transportation interaction in transportation policy scenarios might make. As discussed above, we have not controlled for the differences in the travel models between MEPLAN and SACMET and thus in this study we cannot isolate the effects of the differences in the travel model from the contribution of the representation of the land use and transportation interaction. The mode share results in Table 4 show that MEPLAN predicts a high transit mode share that does SACMET. This is because MEPLAN counts unlinked transit trips while SACMET counts linked transit trips. In addition, MEPLAN omits short trips that tend to be non-transit trips. Auto trips, VMT, and travel speed (in Table 5) are also lower compared to SACMET. These differences can be explained in part by differences in total population and employment, the location of population and employment, and the travel models. Note also the MEPLAN simulates travel over an 18 hour period as opposed to the 24 hour period in SACMET. However, in addition, MEPLAN (like most integrated land use and transportation models) represents intrazonal trips but uses large zones and a sketch networks. SACMET uses small zones and a detailed network and thus can represent all the trips and VMT in the system. As a result, for most emission types, the total emissions for a scenario is much lower in the MEPLAN scenario compared to the SACMET scenario (see Table 6). The
rank ordering of the scenarios with respect to travel and emissions
results is not altered between the MEPLAN and SACMET scenarios.
However, the magnitude of change from the policy scenarios compared
to the trend scenarios is significantly greater in the MEPLAN scenarios
compared to the SACMET scenarios. While acknowledging the limitations of
the comparison in this study, this finding does suggest that that the land
use and transportation interaction represented in MEPLAN may make a
significant contribution to these results. Table
4 Daily Mode Share
Projections for the 2015 SACMET Scenarios.
a figures in parentheses are percentage change from the trend scenario. Table
5 Daily Vehicle Travel
Projections 2015 SACMET Scenarios.
a
figures in parentheses are percentage change from the trend scenario. Table
6 Daily Emissions
(tons) Projections for the 2015 SACMET Scenarios.
a figures in parentheses
are percentage change from the trend scenario. | TOP | The comparative analysis of the MEPLAN and SACMET simulations of low- emission scenarios in the Sacramento region provide a number of examples that illustrate the advantages of a more theoretically comprehensive model in providing insights that can direct heuristic policy development. However, these strengths can be offset by limited representation of geographic detail. In the simulation of the TOD scenario, the MEPLAN model represents land markets with endogenous prices and market clearing in each period and thus allowed the development of a revenue neutral land use policy that would increase densities along light rail lines in the regions. In SACMET, the land use densities of existing areas in the region that are considered to be prototype TODs were used to guide the manual adjustment of zonal land uses. This suggests that SACMET TOD densities could not be achieved by taxes or subsidies only. Strict growth controls would also be needed. With respect to the TOD scenarios, the SACMET model’s strength lies in its detailed representation of zones, transportation networks, and zonal walk and bike accessibility variables. Smaller zones allow more accurate representation of TODs. Pedestrian environmental factors in the model allow for the representation of improved walk and bike facilities in the TODs. Detailed transportation networks allow for accurate representation of the quality of available travel modes. MEPLAN does not represent geographic detail at this scale. However,
a key finding of this study is that the magnitude of change from the
policy scenarios compared to the trend scenarios is significantly greater
in the MEPLAN scenarios compared to the SACMET scenarios.
While acknowledging the limitations of the comparison in this study
(i.e., the differences in the travel models are not controlled), this
finding does suggest that that the land use and transportation interaction
represented in MEPLAN may make a significant contribution to these
results. If the land use and
transportation interaction is a significant variable in travel and
emissions results, then a model like SACMET that did not represent this
interaction would provide inaccurate results.
In future research we plan to control for the travel model in
MEPLAN and thus better isolate the contribution of the land use component
of the model. Over periods of
time longer than 25 years, it is likely that the inaccuracies due to lack
of geographic detail would be less significant than the failure to
represent the relationship between land use and transportation. (1)
Transportation Research Board (1995).
Expanding Metropolitan Highways: Implications for Air Quality and
Energy Use, Special Report. 245. National
Research Council, Washington, D. C. (2) ABAG (1991) Assessing the Future: A Sensitivity Analysis of Highway and Road Improvements on Growth in the San Francisco Bay Area. Working Paper 91-4, Oakland, CA. (3) Putman, S. (1993) Sensitivity Tests with Employment and Household Location Models. Presented at the 3rd International Conference on Computers in Urban Planning and Urban Management, Georgia Institute of Technology, July. (4) Replogle, M. in Transportation Research Board (1995) Expanding Metropolitan Highways: Implications for Air Quality and Energy Use, Special Report. 245. National Research Council, Washington, D. C. (5) Wegener M. (1994) Operational Urban Models, Journal of the American Planning Association, Vol. 60, No.1, Winter, pp. 17-29. (6) Hunt J.D. and Echenique M.H. (1993) Experiences in the application of the MEPLAN framework for land use and transportation interaction modeling. Proceedings of the 4th National Conference on the Application of Transportation Planning Methods, Daytona Beach, Florida, USA, May, pp 723-754. (7) Sacramento Area Council of Governments (1996) 1995 Regional Housing, Population, & Employment Projections, Documentation and Analysis. Sacramento, CA, February. (8) Simmonds, D. (1995) Available Methods for Land-Use/Transportation Modelling. David Simmonds Consultancy. (9) Echenique, M. H. Urban and regional studies at the Martin Centre: its origin, its present, its future. Environment and Planning B, Vol. 21, 1994, pp. 517-533. (10) Abraham, J. E. and Hunt, J.D. (1998) Calibrating the MEPLAN model of Sacramento. Transportation Research Board Annual Conference, Washington DC, January 1998, paper 980649 (available on the pre-print CD-ROM) (11) DKS & Associates. Model development and User Reference Report. Sacramento, CA, October 1994. (12)
Morgan M. and Henrion M. (1990). Uncertainty.
Cambridge University Press, Cambridge, MA. (13) Abraham and Hunt, 2000, Parameter Estimation Strategies for Large Scale Urban Models, submitted for presentation at the January 2000 annual meeting of the Transportation Research Board, Washington DC. | TOP | |