Introduction
Mercer County is located in central New Jersey. The county contains New Jersey’s state capital, Trenton, and has a population of 364,883 according to the US Census Bureau[1]. Mercer contains a handful of college campuses, including Princeton University in Princeton, Rider University in Lawrenceville, Mercer County Community College in West Windsor, and Thomas Edison State College in Trenton.
Below is a map of the county:
[2]
Mercer ranks 80th our of over 3000 counties the United States in per capita income according to the Bureau of Economic Analysis[3]. Some of the larger employers include the universities, Bristol-Meyers Squibb, Church & Dwight Company, McGraw Hill Companies, and Educational Testing Services.[4] The county has three train stations on the Northeast Corridor Line that runs from Washington D.C. through New Jersey and New York up to Boston: Trenton Transit Center, Hamilton Station, and Princeton Junction.
General Strategies and Initial Networks
Urban—Trenton
For urban areas within Mercer such as Trenton, my strategy was to build interconnected concentric cycles to create a grid. The natural grid network of streets made the network design relatively simple, with stations at street crossings and guideway along the roads. I think this system worked particularly well in Trenton, as the city contains a diverse mix of productions and attractions, and, in general, the concentric cycles allow for direct trips that only incur an additional “around the block” cost. See below for a picture of the Trenton initial network:
Interchanges played a key role in the urban design, serving as ways to move between the cycles and provide two-way guideway in some areas. The stations’ coverages tend to overlap only because the density of Trenton productions and attractions requires nearly full coverage in order to be effective.
Suburban/University—Princeton University
My main strategy in planning the suburban and university networks was to minimize the invasiveness of the system. In general, schools and neighborhoods pay close attention to atmosphere and PRT guideway all over campus and the surrounding areas would be the last structuring that these places would want in a PRT network. However, by completely avoiding the suburbs and colleges, we neglect an important part of the population and a very high number of trips. Thus, I attempted to balance the issues of serving trips while minimizing invasiveness in planning out these networks.
In order to minimize intrusion, I decided to use two-way guideway in a handful of instances. I think colleges and neighborhoods would be happier with a few instances of two-way guideway versus lines that run all the way through the campus or surrounding areas. See below for a closer look example (from the old Princeton Network):
I chose to create the first initial network at Princeton University and its surrounding areas. The University alone has a student population of around 7,500 and a staff of around 5,000. However, the network design is mainly with the staff in mind, as college students tend to be frugal and everything on Princeton’s campus seems to be an easy walk away. The only semblance of PRT currently exists on the campus of West Virginia University in Morgantown, WV, so another college town seems like an appropriate place to begin. See the next page for a snapshot of the Princeton Network.
I made a few changes to the initial Princeton Network, with the goal of reducing travel time for short line of sight trips across campus. On a second consideration, these changes would be imperative to the creation of a network in Princeton. In my old network, I tried to keep campus guideway as close to zero as possible. However, in doing so, there were a few situations where short line of sight trips looked to be a fair amount longer because the guideway went in a loop around campus. Thus, if adjacent stations were not connected by guideway in the right direction, one would have to ride the entire loop to go from one to another. By not making these trips convenient (across campus for instance), we neglect a large portion of trips within the University that could provide significant revenue
In this new network, I decided to take advantage of the existing on-campus streets Elm and the street south of Dillon Gym, to provide the desired accessibility. This design would probably not align with the University’s ideal system, but by serving more trips and demonstrating higher profitability of this initial network , we can make the push for expansion earlier. And, with this additional revenue, we might be able to make a profit-sharing agreement with the University such that we would be allowed to place some guideway on roads on campus.
Rural
I did not design an initial rural network, as I mainly used these to interconnect the previous two network designs. My basic strategy for rural areas was to make loops with one-way guideway, and use two-way guideway only to connect loops or minimize invasiveness. For an illustration, see the next page:
General Statistics
Below is a table of the general network statistics. The dataset from last year slotted “recreation trips” under “non-student patron trips”, and for consistency I decided to do the same. One area of improvement that can be made to a proposal for PRT in Mercer county would be to split the non-student patron trips into recreation, dining, shopping, etc. The main concept that I changed since the earlier report had to do with home trip ends. According to the old dataset, only 200,000 people resided in Mercer County, so I took a look at the placemark file, and evenly distributed the population throughout the Census blocks to get the number to the 360,000 threshold. Hopefully the 2010 Census data can be used to get more accurate results. The new results are below:
Total Trip (Ends) ServedServes / Possible / % Service
Home Trip Ends / 1,306,572 / 1,451,448 / 90.0%
Work Trip Ends / 393,840 / 425,649 / 92.5%
Recreation Trip Ends / - / - / -
Transport Trip Ends / 39,000 / 39,000 / 100.0%
Student trip Ends / 111,656 / 113,410 / 98.5%
(Non-Student) Patrons trip Ends / 294,020 / 324,334 / 90.7%
Total / 2,145,088 / 2,353,841 / 91.1%
On the next page is a table containing the population statistics for Mercer as well as a population distribution for New Jersey as of the 2000 Census. The county has about 364,000 people, so the 425,649 work trips indicates a (total) employment percentage of 46.4%, assuming that each person works away from home and 2.5 trip ends are at work. The breakdown is between male and female:
New Jersey Population Distribution, 2000 Census0-4 / 288,085 / 3.42% / 275,700 / 3.28%
5-10 / 309,563 / 3.68% / 294,966 / 3.51%
10-15 / 302,708 / 3.60% / 287,869 / 3.42%
15-19 / 271,020 / 3.22% / 254,196 / 3.02%
20-24 / 244,628 / 2.91% / 235,451 / 2.80%
25-29 / 272,873 / 3.24% / 272,044 / 3.23%
30-34 / 319,031 / 3.79% / 325,092 / 3.86%
35-39 / 360,230 / 4.28% / 367,694 / 4.37%
40-44 / 348,061 / 4.14% / 359,121 / 4.27%
45-49 / 297,845 / 3.54% / 313,512 / 3.73%
50-54 / 263,357 / 3.13% / 284,184 / 3.38%
55-59 / 202,559 / 2.41% / 220,779 / 2.62%
60-64 / 156,073 / 1.85% / 174,573 / 2.07%
65-69 / 132,558 / 1.58% / 160,638 / 1.91%
70-74 / 121,639 / 1.45% / 159,834 / 1.90%
75-79 / 95,560 / 1.14% / 144,571 / 1.72%
80-84 / 58,291 / 0.69% / 104,046 / 1.24%
85+ / 38,732 / 0.46% / 97,267 / 1.16%
Total under 25 / 32.9%
Total over 65 / 13.3%
Sum / 46.1%
100%-Sum / 53.9%
10% Unemployment / 48.5%
As it turns out, around 48% of the population should be making work trips, so the Mercer numbers slightly low-ball actual employment. This is with fairly liberal assumptions— everyone from 5-25 makes school trips only and everyone over 65 (retirees) makes no work trips. This result makes a lot of sense, as the employment data in the placemark file is less than complete. For instance, I did a search for a handful of popular fast-food chains (McDonald’s, Wendy’s, and Burger King), and not a single one appeared in the file. The placemarks are fine for the large employers and census data, but if someone could find a way to improve the employment data, that would do a great deal in improving the bid for a Mercer PRT system. Given the spotty data, I think the number of work trips is fairly realistic.
The 0.89 ratio of patrons trip ends per day to the county population seems reasonable, as on the average each person might make an non-work, non-home, outside trip every day or every other day. The trip ends per person per day seems slightly on the low side at 6.45, given that the 0-5 and 75+ populations are assumed to take no trips. However, once these factors are considered, the average person has about 7.6 trip ends per day, which seems reasonable.
Population StatisticsPopulation / 364,883
Possible # Trip Ends Per Person Per Day / 6.45
Persons between 5-75 / 310,151
Possible # Trip Ends Per 5-75 Person Per Day / 7.59
Patrons Trips Per Day : Population / 0.89
The interchanges were assumed to incur no construction cost, leading to a close to 1:1 ratio of interchanges to stations.
Network StatisticsTotal Stations / 491
Total Interchanges / 456
Total Miles of Guideway / 478.85
An image of the entire network is included on the following page.
Entire Network
Further Analysis
Below is a chart of the total trips served by station in Mercer County. The chart has a heavy tail, as only a pair of stations serve more than 25,000 individuals per day. By far most stations serve 5,000 people or less. See below the chart for descriptions of the two most populous stations.
Trenton (serves 35,106 people per day) – This station is located in probably the densest residential and commercial area in Mercer County, as it contains Thomas Edison State College (~7,000 people, counted at three trips each).
Princeton Junction (serves 27,609 people per day) – NJTransit Data indicate that 9,000 boardings (each way, so 18,000 total) occur at the Princeton Junction Transit Center on a typical weekday, and residential/commercial establishments make up the rest.
Finances
In terms of financial feasibility, the Mercer County Network seems reasonable with the base line assumptions. At a fare of $3, average trip length of 5 miles, and using a linear mode split discount factor, the network makes an annual profit of $178 million after paying the annual operating expenses and cost of capital. Because Mercer is a very wealthy county, we could also consider a higher fare than the base $3, in which case Mercer would turn even higher profits.
With the changes made to the Census population data to hit the correct Mercer County population, the profitability swung favorably:
Mercer Networks StatisticsStations / Interchanges / Miles of Guideway Length / Total Trip ends served / Total Trips / Peak hour Trips / Fleet size / Average trip Length / Average Vehicle Occupancy / Fare / Vehicle Operating Costs
(#) / (#) / (miles) / (per day) / (per day) / (per day) / (#) / (Miles) / (Trips/vehicle)
491 / 456 / 478.85 / 2,046,802 / 851,997 / 127,799 / 14,058 / 5 / 2 / $3.00 / $0.20
Most of these numbers can be tweaked to test what-if scenarios for running a Mercer PRT Network, and there is a fare adjustment graph on the following page. A P&L analysis is included in the table below. After amortizing the capital cost in debt (at an interest of 8%) and including the yearly operating costs, the system attains a profit of $178 million:
Mercer Basic Costs, RevenueCapital Costs / Annual Recurring Costs / Annual Revenue / P&L
Stations / Guideway / Vehicles / Total / Cost of Capital / Maintenance / Operating / Total / Fare / Station lease and naming rights / Total
(M$) / (M$) / (M$) / (M$) / (M$) / (M$) / (M$) / (M$) / (M$) / (M$)
$982 / $2,394 / $1,406 / $4,782 / $383 / $96 / $128 / $606 / $767 / $18 / $784 / $178
One key item to keep in mind is the size of this network’s coverage. In the attempt to reach 90% coverage, I had to build stations in zones that contained fewer and fewer trips. I would guess that there is most likely an optimum profitability level of coverage for profitability between 30% and 80%. However, most importantly, the network does satisfy the overall goal of the NJ PRT system of profitability and 90%+ service.
Another potential adjustment to the model might be in charging a different fare. We assumed $3 for an average trip length of five miles—the chart below depicts yearly profit when varying fares and given completely inelastic demand. With the changes made to the population statistics, the break-even fare dropped a dollar to about $2/ride.