Planning Freight-Efficient Land Uses: Methodology, Strategies, and Tools (2022)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

224 10.1 The City of Paris’s Logistics Land-Use Plan Land-use and transportation agencies in most urban and metropolitan areas do not suffi- ciently consider freight in their regional planning. Without government intervention, logistics facilities tend to locate outside of urban areas to avoid high land prices and pushback by local communities, leading to environmental and systemic inefficiencies when making deliveries to cities. To combat the excessive externalities produced when warehouses are located far from urban cores, the public sector should foster FELUs. The logistics land-use plan enacted by the City of Paris provides insights for cities looking to improve the efficiency of their urban logistics. By identifying locations for urban logistics facilities and supporting their development, Paris has paved the way for the development of an efficient urban freight system. In addition, Paris has also faced challenges that may be instruc- tive for others following similar paths. Logistics freight land-use planning in Paris takes place on the local level, as it does in the United States, and the local municipalities do not necessarily adhere to the regional master plan. In fact, the Paris region is made up of 1,281 municipali- ties, which have their own say over whether to allow logistics facilities in their area (Dablanc and Nicol 2017). Often, as in the United States, these local municipalities do not consider the surrounding areas in their decisions, thus not necessarily making decisions that are best in the regional context. A key reason why Paris has advanced its logistics planning is because of a partnership, which started in 2006, involving both the public and private sector (Debrie and Heitz 2016). The partner- ship involves annual meetings between stakeholders, including government agencies, planning agencies, and private freight personnel. This led to an operational charter in 2013 with 80 partners (Dablanc and Nicol 2017). The charter resulted in a list of 16 projects to improve logistics in the Paris region. One major task of Paris’s logistics planning is to support urban facilities by assisting developers in obtaining urban land at strategic locations. Without this support, the developers of logistics facilities cannot afford to locate in urban areas and instead have to locate wherever land is cheaper and available. To support FELUs, Paris’s regional plans identify land parcels in strategic locations to reserve for future logistics facilities, whether it be private or public land. This ensures that there can be sufficient infrastructure to meet current and future demand for urban freight. The 1994 Master Plan was the first regional plan in Paris to substantially consider the freight network (Raimbault et al. 2018). More recently, Paris is experimenting with a new framework for urban logistics, as shown in Figure 24. This framework consists of larger multimodal sites (shown as larger rectangles) and smaller distribution centers (shown as smaller rectangles). On the edges of urban areas, C H A P T E R 1 0 Case Studies

Case Studies 225   these large multimodal sites are being developed as logistics hotels. ese sites are multistory, mixed-use complexes that can include multimodal logistics terminals, oce space, and housing (Dablanc 2017a). e purpose of these logistics facilities is to take in large shipments by rail, water, or large truck, and send out smaller vehicles to service the urban core. For example, the northernmost large rectangle in Figure 24 represents the site of the now-operational Chapelle International Logistics Hotel, a facility including a rail terminal as well as oces and housing (Liu and Dablanc 2017). It was estimated that this site would reduce up to 700 truck trips per day and 50% of pollution generated by freight activity for its tenants (Liu and Dablanc 2017). Another experiment involves the development of smaller logistics areas throughout the urban region, oen converted from underground parking structures (Diziain et al. 2012a). ese facilities encourage the use of clean transport modes and cargo consolidation. Since the open- ing of the Beaugrenelle underground facility in 2013, the facility is reported to have reduced the environmental impact produced by its tenants, including a signicant emissions reduction, a 52% decrease of VMT, and a decrease in noise pollution (IFSTTAR 2018). e concept for this type of facility is shown in Figure 25. ese facilities are hidden from the public, are utilized by small freight vehicles, and can be multistory. Within the City of Paris, zoning ordinance plans in 2006 and 2016 were fundamental in the development of urban logistics sites. e 2006 zoning ordinance, shown in Figure 26, mapped areas that were dedicated to logistics developments and marked the locations of potential inter- modal facilities, eight of which are along railways, and 14 of which are along the Seine River (shown as stars). e goal of these facilities is to reduce reliance on truck trac in the city by accepting freight from rail and water modes and utilizing other environmentally friendly freight modes for the last-mile deliveries (Diziain et al. 2012b). e 2016 zoning plan further emphasized the development of urban logistics (Dablanc 2016b). It is worth mentioning that the City of Paris has Source: (Dablanc and Nicol 2017) 3000 ft Bulk Urban distribution Last mile Legend: Multimodal logistics platforms Cross-docking places Commodity flow: Figure 24. Proposed framework for urban logistics.

226 Planning Freight-Efficient Land Uses: Methodology, Strategies, and Tools Source: (IFSTTAR 2018) Figure 25. Concept for the Beaugrenelle logistics facility. Notre Dame Cathedral Source: (Diziain et al. 2012a) Figure 26. 2006 zoning ordinance of Paris.

Case Studies 227   allocated prime urban spaces for logistics facilities. As shown in the figure, one of these facilities is very near the Notre Dame Cathedral. Even more than simply identifying urban land suitable for logistics, Paris is supporting the development of urban logistics in other ways. Public agencies can indirectly financially sup- port the development of logistics hotels through site cleanup and infrastructure development prior to selling land to a private developer (Raimbault et al. 2018). As the proposed facilities are new concepts for urban freight and not well established, another way to support their development is by conducting exploratory studies into the feasibility of these urban logis- tics sites to inform developers about the concept and the expected operating costs (Dablanc 2017a). Beyond what is discussed here, Paris has implemented other transportation initiatives, such as emissions regulations for freight vehicles and mandatory nighttime deliveries for large trucks (Dablanc 2017b). The initiatives in Paris reflect some of the best practices for incorporating logistics into land- use planning. Urban planners can consider the case of Paris to find opportunities for how to incorporate the needs of urban freight into land-use planning. Like the Paris region, other met- ropolitan areas can develop working groups with local stakeholders to identify sources of prob- lems, define goals, and make commitments to achieving more efficient urban freight systems. Public agencies should consider how they can have a role in supporting urban logistics, such as by identifying optimal locations for logistics facilities throughout the region. Public agencies can lead the process of altering land-use regulations and reviewing building permits to allow for these developments; for example, Paris allowed parking structures, which have no windows, to become workplaces for logistics employees (Dablanc 2017a). In their regional plans, they should consider what kinds of facilities are best for their cities, and whether a similar framework to what Paris is establishing (using the logistics hotels and distribution centers) might be ideal. In addition, regional agencies should set up partnerships between municipalities to encourage unity between regional- and local-level land-use planning. 10.2 Cali, Colombia: Logistics Sprawl This case illustrates the consequences of a lack of coordination between the land-use policies enacted by municipal governments in the same metropolitan area. The analyses focus on the metropolitan area of Cali, Colombia. This important metropolitan area, located in the southwest of Colombia, encompasses five municipalities, three of them—Cali, Yumbo, and Jamundí—are relevant to the discussion: the City of Cali, with 2.5 million residents (2020), the third most populous city in the country; and Yumbo and Jamundí with 131,645 and 132,572 residents, respectively (Departmento Administrativo Nacional de Estadística 2011). The Cali metropolitan area is strategically located with easy access to the Buenaventura Port at the Pacific Ocean; and is traversed by the Pan-American Highway, which connects Colombia with North and South America. Of great relevance to this discussion are the physical constraints presented by the West Andes Mountains to the West, and the Cauca River and its flood plains to the East. As a result, any growth in the Cali area would have to take place either to the north toward Yumbo, or to the south toward Jamundí. Until the 1940s Cali grew slowly and the population remained below 100,000 residents. Then, the population grew exponentially, almost doubling every 10 years. Cali became highly indus- trialized. Some of the biggest manufacturers in Colombia were located in Cali, thus attracted skilled and unskilled workers. During the 1990s, although the rate of growth slowed, the popu- lation exceeded 2 million in 2010. As a result, land values experienced large increases, which significantly impacted the location of manufacturing, logistics activities as well as residential development.

228 Planning Freight-Efcient Land Uses: Methodology, Strategies, and Tools In response to the increasing land costs, the inherent challenges associated with urban opera- tions, and the tax incentives oered by City of Yumbo, manufacturers and logistics operators started to relocate from Cali to Yumbo. In the 1940s, Yumbo had around 100 small to medium industry establishments. By the 1960s, the largest companies in Colombia had located in the Cali-Yumbo corridor. In the 1990s, this corridor became the second most important corridor in Colombia, and the third most important producer of consumer goods (Alcaldía de Yumbo 2020). By 2010, Yumbo had around 2,000 companies, with 200 large multinationals (La Republica 2013). e pattern of development in Yumbo, which centered around a narrow strip of land at both sides of the Pan-American Highway, led to a very inecient use of the land, where the only option for newcomers was to nd land further north, next to the Pan-American Highway. At the same time, since the housing sector in Cali could not provide sucient aordable housing to accommodate the population growth, people started moving south to Jamundí, seeking better options. is was made possible by the actions of Jamundí’s city leaders who saw urban development as the way to improve the city’s nances. As a result, Jamundí’s population grew from 44,438 inhabitants in 1985 to 132,572 in 2000 (Departmento Administrativo Nacional de Estadística 2011). Further complicating matters, seeking to expand to accommodate the increasing population, numerous schools and universities located on the Cali-Jamundí corridor. e net result was the addition of tens of thousands of trips to the already congested corridor. e satellite pictures in Figure 27 show the development in the Cali metropolitan area between 1969 and 2016. As people started moving south of Cali to Jamundí, and manufacturers and logistics operators kept expanding north of Cali to Yumbo, congestion signicantly worsened as a consequence of (1) the commuter trac from Jamundí to Yumbo during the peak trac hours; (2) the trac of trucks transporting the solid waste generated in the metropolitan region to a treatment facility located 27 miles north of Cali; and (3) the trac of freight vehicles trans- porting supplies from Yumbo to Cali and Jamundí. Regarding the third item, Cali’s ocials have 2016 JAMUNDI CALI YUMBOYUMBO 1986 1996 2006 JAMUNDI JAMUNDI JAMUNDI CALI YUMBO YUMBO 1969 CALI YUMBO Urban Expansion Urban Expansion Urban Expansion Urban Expansion Urban Expansion CALI CALI JAMUNDI Industrial/ Logistics Growth Industrial/ Logistics Growth Industrial/ Logistics Growth Industrial/ Logistics Growth Industrial/ Logistics Growth Source: (2020 Google) Figure 27. Expansion of Cali toward Jamundí and Yumbo.

Case Studies 229   considered banning the freight thru traffic which, ironically, was created directly or indirectly by their predecessors’ failures to implement a metropolitan vision for the Cali region. Essentially, for their own seemingly sensible reasons, the city leaders of Yumbo and Jamundí pursued development agendas that complemented each other to aggravate the problem. Yumbo leaders took advantage of the city’s ideal location to export products to Bogotá, the largest market in the country, and to international markets through the Buenaventura Port. Jamundí’s leaders con- cluded that the city’s future was in becoming a satellite city of Cali. Throughout this period, there is no evidence that Cali’s leaders were aware of the implications of these trends. Although the first planning attempts began in the 1940s, it took many years for these plans to be approved, and most of them were never fully executed (Vinasco Martínez 2016). In 2000, the City of Cali formulated its first land-use plan (Alcaldía de Santiago de Cali 2019). However, the plan does not establish coordination mechanisms with the other municipal governments in the metropolitan area. From the standpoint of FELUs, the consequences of this failure to create a regional vision for development are clear. As in the case of the Port of New York, poorly thought out decisions left both cities worse off. Had the City of Cali retained some manufacturing and logistics companies, Cali would have avoided some of the massive congestion created by the long journeys back and forth from Yumbo to the south of the metropolitan area. Moreover, some of that logistics capacity could have been used for last-mile deliveries to the urban core. Had the City of Yumbo conducted minimal land-use planning, it could have achieved its goal of becoming a manufacturing and logistics hub while using the land more efficiently. For instance, had they fostered develop- ment along the arch defined by Cali’s northern city limit—instead of along the Pan-American Highway—they could have preserved prime land along the highway for more productive uses. See Figure 28. As shown in the figure, manufacturing sites, warehouses and distribution centers in Yumbo tend to be located along to the Pan-American Highway, frequently with direct access to the highway. Buffer zone Consumer-oriented businesses Industrial park Industrial/ Logistics park Source: (2020 Google) (a) Yumbo, Colombia (b) Clifton Park, New York Figure 28. Comparison of industrial area in Yumbo, Colombia, and Clifton Park, New York, United States.

230 Planning Freight-Efficient Land Uses: Methodology, Strategies, and Tools As a result, prime land that could have been used for consumer-oriented businesses that need the visibility provided by the highway, has been used by industrial and logistics activities that could have been located further back, without materially impacting their operations. The example of Route 9 in Clifton Park, New York provides a contrasting example. In this case, the industrial activities take place behind a buffer zone that separates them from the consumer-oriented busi- nesses, located next to the highway. Instead of direct access to the highway, the freight traffic enters and exits the industrial park through an intersection at Route 9, specifically designed to handle heavy trucks. This case study shows the importance of incorporating FELU principles in land-use planning. As in the case of the Port of New York, the lack of efforts to integrate freight activity into the urban and metropolitan fabric produced detrimental effects on the local communities. 10.3 Economic Geography Measures for Selected U.S. Metropolitan Areas Understanding the spatial distribution of freight activity is a first step for land-use planners in establishing successful FELU initiatives. This case study uses the tools explained in Chapter 6 to describe spatial distribution of economic activities (i.e., measures of centrality and spread of economic activities, in a number of metropolitan areas). The selected MSAs represent different regions, economies, and urban forms in the United States: New York MSA (New York-Newark- Bridgeport, NY-NJ-CT-PA); Los Angeles MSA (Los Angeles-Long Beach-Anaheim, CA); Houston MSA (Houston-The Woodlands-Sugar Land, TX); Washington, DC, MSA (Washington-Arlington- Alexandria, DC-VA-MD-WV); New Orleans MSA (New Orleans-Metairie, LA); and Albany MSA (Albany-Schenectady-Troy, NY). Table 15 shows a summary of key economic statistics for the year 2016, most of which were previously discussed in Chapter 5. 10.3.1 Brief Economic Descriptions of Selected MSAs New York MSA The New York MSA is the largest metropolitan area in the United States. It includes counties from four states: New York, New Jersey, Connecticut, and Pennsylvania. It also includes the principal cities of New York, Newark, and Jersey City. It is highly populated, with more than 20.2 million inhabitants in an area of 8,203 square miles. Its GDP per capita ($82,000) surpasses the national average of $58,000 (World Bank 2020). It houses large headquarters and employs more than 8 million people. It is also an important freight hub, with the Port of New York and New Jersey, reported in 2018, as the country’s third largest in imports to the United States. N ew Y or k M SA L os A ng el es M SA H ou st on M SA W as hi ng to n D .C . M SA N ew O rl ea ns M SA A lb an y M SA Population 20,275.18 13,328.26 6,798.01 6,150.68 1,271.20 882.80 Area (miles2) 8,292.60 4,851.00 8,266.00 6,246.90 3,202.80 2,811.70 Density (Pop/miles2) 2,444.97 2,747.53 822.41 984.60 396.90 313.97 GDP per capita (USD) 82.01 74.76 69.48 82.85 60.27 59.88 Establishments 550.19 357.56 134.04 143.57 25.22 22.38 Employment 8,032.24 5,501.09 2,633.62 2,604.51 483.97 375.91 B2B FTG 2,065.57 1,448.16 615.70 462.04 121.86 106.78 B2C Deliveries 3,041.28 1,999.24 815.76 738.08 152.54 97.11 B2C FTG 608.26 399.85 163.15 147.62 30.51 19.42 Table 15. MSA summary statistics (in thousands).

Case Studies 231   The port handled 3.6 million imported TEUs and a total of 5.2 million TEUs in 2018 (Logistics Management 2019). The metropolitan area also has two of the top 15 busiest airports in the United States, John F. Kennedy International Airport, ranked sixth, and Newark Liberty Inter- national Airport ranked 14th (World Atlas 2018). Within the New York MSA, there are multiple freight rail yards operated by CSX and Norfolk Southern. From all the establishments and employ- ment, 43.3% and 41.9%, respectively, belong to the FIS; 40% to 50% of these establishments are mainly retail stores, hotels, and restaurants, which generate more than 60% of all freight traffic. Los Angeles MSA The Los Angeles MSA is the second largest metropolitan area in the United States, which includes major cities such as Los Angeles, Long Beach, Anaheim, Irvine, Pasadena, and others. Los Angeles is also the home of Hollywood with its mega entertainment industry that included more than 13,000 establishments and a total of 161,862 employees in 2011 (Kleinhenz et al. 2012). Besides the entertainment industry, Los Angeles also includes the headquarters of multiple large corporations. The MSA includes the two largest seaports in the United States according to the imported TEUs in 2018 (World Atlas 2018). The Port of Los Angeles handled 4.8 million loaded imported TEUs (Port of Los Angeles 2020), followed by the Port of Long Beach, which handled 4 million imported TEUs (Port of Long Beach 2020). The Port of Los Angeles also includes rail facilities that are used to transport cargo into and from the port. The second busiest airport in the nation, the Los Angeles International Airport, moves a significant amount of freight, handling a total of 2.1 million tons of air freight in 2019 (Los Angeles World Airports - LAWA 2020). FIS establishments in the MSA are estimated to be 40.8% of all establishments, for 48.2% of the total employment. Retail and wholesale stores constitute 48% of these FIS establishments, and they generate around 55% of all freight traffic. Houston MSA The Houston MSA is the fifth most populous in the United States, with more than 6.7 million residents in 2016. The MSA consists of nine counties, including Harris, Fort Bend, and Galveston. It is home to the headquarters of 20 of the top 500 corporations in the United States (Greater Houston Partnership Research 2017). Houston MSA has two ports that rank in the top 50 ports in the United States with respect to annual tonnage according to the U.S. Army Corps of Engineers: the Port of Houston (ranked second) and the Port of Texas city (ranked 15th) (Texas Depart- ment of Transportation 2020). The MSA’s airports have an important share of freight trans- portation of the entire region. For example, Bush Intercontinental Airport which is one of the main airports in Houston, moved around half a million tons of cargo in 2018 (Transport Topics 2019). On the rail side, the MSA has a major railyard (Englewood Yard) that is used by freight trains and operated by Union Pacific. In this MSA, the FIS constitute 53% of its employment and 44.7% of the establishments; retail stores are almost a third (30.5%) of the FIS establishments, and they also generate almost a third (32.7%) of the total freight trips generated. Washington, DC, MSA Washington, DC, is the capital of the United States. Its MSA had a population of more than 6.15 million in 2016 in an area of 6,247 square miles. It includes counties within the states of Virginia, Maryland, and West Virginia, besides Washington, DC. Washington, DC, MSA reported about 2.6 million employees in 2016 and a GDP per capita similar to New York MSA of $82,000 in 2016 (Bureau of Economic Analysis 2018). There are two major airports in this MSA: Washington Reagan National Airport and Washington Dulles International Airport, the latter handles about 300,000 tons of cargo per year and is within an overnight drive from 56% of the U.S. population (Metropolitan Washington Airports Authority 2020). Washington, DC, MSA has a service-inclined economy where 64.3% of its employment and 66.3% of its

232 Planning Freight-Efficient Land Uses: Methodology, Strategies, and Tools establishments are in SIS, consistent with the large number of governmental and other service- related establishments within the U.S. capital. New Orleans MSA The New Orleans MSA is located in the southeastern part of the United States near the Gulf of Mexico, and includes the City of New Orleans, which is well-known for its Mardi Gras festival as well as its vibrant live-music scene all year long. It had a population of more than 1.27 million in 2016, and its GDP per capita of $60,272 is comparable to the national average. This MSA includes the Port of New Orleans, which is a multimodal port that provides transfers between rail, river, and road. The port is the only deep-water container port in the State of Louisiana, and it handled more than half a million TEUs in 2018, which was the fifth year in a row it surpassed the volume of 500,000 TEUs (Port of New Orleans 2019; 2020). Also, the port is the only deep-water port that is served by CSX, Union Pacific, BNSF, Norfolk Southern, Kansas City Southern, and Canadian National, through the New Orleans Public Belt Railroad (NOPB). NOPB is considered to be the fifth largest rail gateway in the United States (New Orleans Public Belt 2020). Similar to Houston MSA, New Orleans MSA has 53% of its employment in FIS, while 45.8% of the establishments are in FIS. Retail stores and accommodation and food establishments constitute more than 58% of the FIS establishments, generating around 45% of the total freight trips. Albany MSA The Albany MSA includes the counties of Albany, Rensselaer, Saratoga, Schenectady, and Schoharie. It includes the Capital Region of New York State, Albany, which is home to many public-sector agencies. The population of this MSA was estimated to be 883,000 in 2016, living in an area of 2,812 square miles; the total number of employees was around 376,000 in 2016. The major ports in this region include the Port of Coeymans; the Selkirk rail yard operated by CSX; the Port of Albany, which reported the handling of a total of 327,500 tons for multiple com- modities in 2018 (Port of Albany 2018); and the Albany International Airport, which includes an air cargo facility of 53,000 square feet used by major carriers such as FedEx and UPS (Albany International Airport 2020). Albany MSA has nearly 46% of FIS employment and establish- ments; the construction industry in Albany MSA includes around 24% of the FIS establishments and generates more than 21% of the total freight trips. 10.3.2 Spatial Distribution of Economic Activities The analyses are based on employment data at a zip code level collected from the Census Bureau CPB data (U.S. Census Bureau 2013). Two methodologies are used to analyze the spatial distribution of economic activities. To determine the measures of centrality, the team calculated three metrics: employment, employment density, and the interaction index. To measure spread of economic activities, the team calculated a weighted physical distance of truck trips that uses the FTG models from NCFRP Report 37 (Ortúzar and Willumsen 2011; Holguín-Veras et al. 2017b), and Fratar II to calculate trip distribution of flows (Ortúzar and Willumsen 2011). Measures of Centrality The identification of economic pole(s) is key to understanding spatial structures of economic activities in a region. Rather than relying on a qualitative investigation, this case study uses three data-driven measures of centrality: employment, employment density, and this interaction index. The interaction index, developed by the research team, is inspired by the gravity model to quantify spatial interactions of activities. It is proportional to the product of employment of two zones and inversely proportional to the travel distance between them. To identify economic pole(s), zip codes in an MSA are ranked by descending order of each measure. Then, the top zip codes that make up 10% of the MSA’s total employment are defined as economic centers.

Case Studies 233   The team selected the threshold of 10% after experimenting with other thresholds and con- cluding that 10% provided the most consistent results. In addition, the average distance between the zip codes identified as part of the economic pole(s) are shown. Figure 29 shows identified economic poles based on these three measures. The corresponding average distances between the zip codes identified as part of the economic pole(s) are shown MSA Employment Employment Density Interaction Index Lo s A ng el es -L on g Be ac h- A na he im W as hi ng to n, D C - A rl in gt on A le xa nd ri a A lb an y- Sc he ne ct ad y- Tr oy N ew Y or k- N ew ar k- Je rs ey C ity Figure 29. Economic poles for selected MSAs.

234 Planning Freight-Efficient Land Uses: Methodology, Strategies, and Tools in Table 16. Los Angeles MSA exhibits the most dispersed urban form, with the longest average distance between economic poles for all measures. In contrast, Washington, DC, MSA and New York MSA show relatively monocentric urban forms. Moreover, the results suggest that no single measure is perfect for identifying economic poles. For instance, in Washington, DC, MSA, the interaction index approach captures large retailers and firms in Northern Virginia while the employment density ignores this economic pole. Each measure (employment, employ- ment density, and the interaction index) has unique characteristics to identify the economic centers. The employment-based approach is able to identify large zip codes with large employ- ment but fails to identify the high-density core (which includes numerous small zip codes with relatively large numbers of employment). The interaction index approach is also supported by a strong theoretical background. The results also highlight the importance of the local economic geography when establishing land-use policies. Approaches to reduce negative externalities caused by freight in a compact city should be different from those for a dispersed city. For instance, the connections between economic centers are important in dispersed cities, such as Los Angeles MSA, while addressing highly concentrated negative externalities is important in compact cities, such as New York H ou st on -T he W oo dl an ds - Su ga r La nd N ew O rl ea ns -M et ai ri e Figure 29. (Continued). N ew Y or k M SA Lo s A ng el es M SA H ou st on M SA W as hi ng to n D .C . M SA N ew O rl ea ns M SA A lb an y M SA # of selected zip codes 6.00 9.00 4.00 5.00 2.00 1.00 Average distance 1.30 34.40 11.00 16.40 8.20 N/A # of selected zip codes 6.00 20.00 6.00 7.00 2.00 5.00 Average distance 1.30 15.20 5.00 1.70 1.80 8.00 # of selected zip codes 6.00 11.00 5.00 6.00 2.00 1.00 Average distance 1.30 20.70 5.70 4.90 1.80 N/A Employment Employment density Interaction Index Table 16. Average distances between selected zip codes.

Case Studies 235   MSA. The team’s conclusion is that the interaction index and employment density produce the most consistent results, though the interaction index has the edge because of its ability to consider the role of physical separation. Measures of Spread The physical separation between key agents involved in supply chains has a direct impact on both the VMT and externalities in the network. Goods are transported within and across industry sectors resulting in a constant flow of production and attraction of cargo across the metropolitan areas. Consequently, land-use distribution and location of establishments and employment play a critical role in the distances trucks must travel to match supply and demand. To measure the average distance of truck trips, the authors propose a weighted physical distance of truck trips. The weights are based on the number of trips between each industry pair from a trip dis- tribution model. The Fratar II methodology was used to calculate the trip distribution for the analyses (Ortúzar and Willumsen 2011). Instead of trips produced and attracted, the numbers of establishments by industry sector for the origin and destination locations were used as the productions and attractions. The average distance between industry sectors is weighted based on the resulting distribution matrix of the trip distribution. The methodology assumes that (1) the amount of freight produced and attracted is proportional to the establishment, consistent with the freight generation models from NCFRP Report 37 (Holguín-Veras et al. 2017b); (2) the freight flows between origin locations and destination locations are proportional to the amounts produced and attracted at these locations, by industry sectors; and (3) the appropriate metric of physical separation is the weighted distance for the resulting freight origin-destination matrix (weighted by the flows). Table 17 exhibits the results of the weighted travel distance, which is computed for the com- binations of industry sectors that are technically feasible, between the following industry sectors: Albany New Orleans DC Houston LA NYC 13,283,824 20,118,063 5,384.11 6,926.38 881,551 3,480.05 58.99 1,263,526 3,779.38 61.48 6,091,560 6,764.25 82.25 6,664,187 10,914.38 104.47 73.38 83.22 23.033 23.356 33.531 27.678 21.291 22.338 31.315 28.566 21.142 21.114** 29.879** 28.462 Warehouses* to Warehouses* 20.657** 22.999 31.536 26.626** 25.203*** 23.544*** 35.418*** 28.644 23.594 22.415 33.386 29.569*** Manufacturing to Accom + Food 23.475 21.133 32.030 29.480 FIS to FIS 22.727 21.864 34.153 30.903 SIS to SIS 20.425 19.917 27.476 28.395 38.52%FIS Average Distance / Square Root of Area Population 2015 Area (sq mi) Square Root of Area (mi) Distance between Industry Sectors Manufacturing to Retail Manufacturing to Manufacturing Manufacturing to Warehouses* Warehouses* to Accom + Food Warehouses* to Retail 32.655 31.858 31.598 30.644** 34.554*** 33.869 33.637 34.747 33.155 41.75% 28.147 29.786 29.437 27.978** 28.276 29.868*** 29.515 29.454 30.834 40.14%29.58%41.53%35.57% Notes: (*) Includes all establishments in NAICS 42, 484, and 493. (**) Shortest average distance for the MSA. (***) Longest average distance for the MSA. Table 17. Results measures of spread.

236 Planning Freight-Efficient Land Uses: Methodology, Strategies, and Tools (1) Manufacturing, (2) Transportation and Warehousing, (3) Retail Trade, and (4) Accommo- dation and Food Services. In addition, the last row of the table illustrates a dimensionless index—average distance of the trip for all FIS divided by the square root of the area—which physically represents the distance of the freight trip with respect to the size of the MSA. Last, for each MSA, the longest and the shortest average distances are highlighted. The results illustrate a tradeoff between an MSA’s urban form and the efficiency of its supply chain. This shows that as the concentration of the MSA increases, so do the challenges faced by the logistics activities. Cities like New York City or Washington, DC—which are monocentric— have the longest distances between key industry sectors in the supply chain. On average, these MSAs have an even longer distance between industry sectors than dispersed urban form cities such as Los Angeles. In the majority of the MSAs (four out of six), the pair of industry sectors with the minimum weighted distance are warehouses to warehouses, and the maximum distance is from manufacturing to manufacturing sites. In most of the MSAs, the distances between FIS are longer than the distances between SIS. Last, the index of average distance of the trip divided by the square root of the area illustrates how monocentric and dispersed cities travel in propor- tion to a similar percentage of its MSA. For example, the weighted distance between FIS in New Orleans is 21.86 miles and in Houston 30.90 miles. Nonetheless, when accounting for the area of the MSA, the trips between FIS in New Orleans cover up to 35% of the square root of the MSA area. In contrast, in Houston, the average distances represent 29.5% of the square root of the MSA area. The Houston MSA case deserves further discussion. In the case of Houston, the average distances between these stages is approximately 29 miles, which is relatively high compared with other MSAs. However, the ratio of average distance over the square root area of the MSA is 30%, which is the smallest. This does not mean that the Houston MSA has the most compact supply chains, because the ratio is also influenced by Houston’s total area. The issue is that, because of the lack of land-use planning, there are no restrictions to urban development. Thus, Houston’s urban footprint is significantly larger than other comparable cities. Since the ratio uses the square root of the MSA area as a normalization value, the ratio for Houston is lower than it would be otherwise. This case study shows two complementary analyses that help policymakers or land-use planners interested in understanding local economic geography. The results indicate that the use of a single measure of centrality cannot holistically explain the urban form. In addition, the results seem to suggest that monocentric urban forms tend to drive logistics activity to the outskirts of the city, increasing the average distance of the freight trips. The measures of spread developed and tested here are useful for benchmarks to both gain insight into the physical separation between the key stages of supply chains, and to compare alternative land-use configu- rations to identify which ones would lead to the most compact supply chains. 10.4 A Simulation of the Impacts of Facility Location in the New York State Capital Region An establishment’s location has important implications for logistics efficiency. This is partic- ularly important for distribution centers and other logistics facilities because of their role in the distribution of goods from manufacturers and suppliers to local commercial establishments and households. As the demand for freight and the number of internet deliveries to households continues to increase, the location of these facilities will be of even greater importance in the near future. To plan for freight efficiency, policymakers must be aware of the systemwide impacts of the location of these facilities and how land-use policies can influence the locations of freight facilities for the better.

Case Studies 237   The purpose of this application of the BMS-FELU is to provide insights on the effects that facility location could produce in urban areas. The BMS-FELU—first developed as part of the New York City Off-Hour Delivery project—simulates freight tours, together with the series of stops where a single freight vehicle makes pickups or deliveries, based on information about a given geographical area. This analysis focuses on the New York State Capital Region—the counties of Albany, Rensselaer, Saratoga, and Schenectady that surround the City of Albany. First, the team simulated the addition of a new large distribution center in response to increased demand for freight ship- ments to and from establishments in the Wholesale Trade sector (NAICS 42). The impacts of operating this distribution center in two locations were estimated and compared. Figure 30 shows the two locations: Amsterdam, New York (35 miles from downtown Albany with numerous logistics facilities), and Colonie, New York (6 miles from downtown Albany with a sizable number of distribution centers). Second, the team simulated the case of relocating a distribution center from Amsterdam to Colonie in response to land-use changes and policies incentivizing the densification of freight activity toward urban areas. The baseline condition used in this analysis reflects the current freight activity in the Capital Region. The base case was developed using the number of freight vehicle trips arriving and departing each zip code from the FASTGS using the 2017 Zip Code Business Patterns dataset. The inputs also considered the freight flows and vehicle trips through gateways to the Capital Region; which include five rail gateways, three river ports, one airport, five interstate highways, and three state routes. The baseline conditions consist of 62,300 freight trips each day within the region, accounting for approximately 1.3 million freight VMT. At the echelon level, gateways to large establishments account for 10% of the trips of the MSA while it produces 15% of the VMT. This echelon considers all the freight entering the MSA from different gateways, with 80% corresponding to highways. These trips are assumed to be point-to-point and often utilize large trucks. The large to large echelon accounts for 40% of the trips of the MSA and 39% of the VMT. It considers the trips going from large establishments to other large establishments, Downtown Albany Figure 30. Distribution center locations considered in the scenarios.

238 Planning Freight-Efficient Land Uses: Methodology, Strategies, and Tools which usually can be done in either large or small trucks and include multiple stops. The large to small echelon accounts for 33% of trips and 32% of the VMT, while the small to small echelon accounts for 16% of trips and 14% of VMT. In the case of the new distribution center, its effects were estimated separately for each location by adding 160 freight trips arriving to the distribution center and 260 freight trips leaving the distribution center to make deliveries elsewhere. The new distribution center, in both locations, produces a total of 1,300 additional stops, assuming an average of five stops per tour. In the relocation case, 160 and 260 freight trips that were entering and leaving Amsterdam, respectively, in the baseline conditions were removed from Amsterdam and added to Colonie. Thus, the relocation case does not consider an increase in demand for freight, because the same distribution center is simply being relocated. In all cases (baseline conditions, new distribution center in Amsterdam, new distribution center in Colonie, and relocation from Amsterdam to Colonie), the simulation results were averaged over 10 runs and aggregated to compute perfor- mance metrics including the number of freight tours and the VMT, both of which are related to the congestion and the pollution created by freight traffic. Table 18 shows a summary of the three scenarios and their FTG in both locations, Amsterdam and Colonie. The performance metrics obtained for the different scenarios were in line with the expec- tations. From the perspective of freight efficiency, the Colonie location, closer to downtown Albany, is better than the Amsterdam location. Locating the new distribution center in Amsterdam would increase the freight VMT across the entire MSA by about 2.7% (over 34,300 additional freight vehicle miles per day), while the same size distribution center in Colonie would increase VMT by 2.2% (around 28,000 additional miles), saving over 6,300 miles per day by locating the new distribution center in Colonie rather than in Amsterdam. If a distribution center is just relocated from Amsterdam to Colonie, the total VMT of the baseline conditions decreases by around 4,200 miles per day. The results at the echelon level were in line with the expectations as well. Figure 31 depicts the differential effects of VMT in comparison with the baseline conditions, for NAICS 42 only. For the echelon of gateways to large, the results correspond to the freight activity from all gate- ways of the MSA to the large establishments of NAICS 42. For large to large and large to small, the results correspond to the large establishments of NAICS 42 delivering to large or small establishments, respectively, of all sectors of the MSA. Because the distribution center is a large establishment, the most direct impact occurs in the echelons included in Figure 31. Overall, adding a distribution center increases the VMT regardless of the location due to increasing the number of trips. In the case of locating a new distribution center in Amsterdam, the highest increase of VMT occurs in the large to large echelon. The same pattern occurs if the distribution center is located in Colonie, with less pronounced differences among echelons. At all echelons, the impacts of locating the distribution center in Amsterdam are higher than the alternative location. The end result is that Amsterdam as a location alternative creates more VMT than Colonie across the entire MSA. This increase in VMT occurs mostly in the large to large Scenario FTG in Amsterdam FTG in Colonie Additional distribution center in Amsterdam Baseline + (160, 260) Baseline Additional distribution center in Colonie Baseline Baseline + (160, 260) Relocation of distribution center from Amsterdam to Colonie Baseline - (160, 260) Baseline + (160, 260) Note: The numbers in parentheses represent the freight trip attraction and freight trip production, respectively, generated by the distribution center. Table 18. Summary of the FTG for the scenarios considered.

Case Studies 239   -2% -1% 0% 1% 2% 3% 4% 5% Gateways to Large Large to Large Large to Small 3.09% 4.08% 3.15% 2.72% 2.89% 2.78% 0.47% -1.02% -0.32% Pe rc en t c ha ng e in fr ei gh t v eh ic le m ile s tr av el ed d ue to v ar io us c as es DC in Amsterdam DC in Colonie Relocation of DC Figure 31. Differential effects on freight VMT for NAICS 42. echelon, mainly because Colonie is closer to major delivery areas than many existing distribu- tion centers in the Capital Region. Only the case of relocating a distribution center yields a reduction in VMT. The results show that the impacts differ among echelons. First, as Amsterdam is on average closer to all the gateways than Colonie, the longer trips entail higher VMT in gateways to large for the reloca- tion case. Second, as all establishments receiving trips from large establishments of NAICS 42 are closer to Colonie than they are to Amsterdam, the shorter trips yield VMT savings in large to large and large to small echelons. The reason for the large to large echelon having the highest impacts in all the cases is that it accounts for approximately four times more VMT than the remaining echelons. The results also show the impacts on trip distribution patterns. Figure 32 shows the vehicle trips for a 1-day simulation between the large establishments after the relocation of the distribu- tion center, for both locations. The maps show the gateways (LTGs, such as ports and highways), and the trips that depart from the zip code in Amsterdam (left hand side of the figure) and Colonie (right hand side). Intrazonal trips, or trips that start and end within the same zip code, are not shown. Many of the freight vehicle trips for NAICS 42 that leave Amsterdam are directed toward zip codes near Colonie. The trips departing from Colonie have more diverse destina- tions, including Saratoga (12866), Latham (12110), Clifton Park (12065), and Troy (12180). Although these results provide insight into the impact of location decisions involving a distri- bution center, the freight VMT is not the only metric that affects freight efficiency. Freight efficiency seeks to minimize the social cost of freight activity, which includes the private cost of land ownership and operating freight vehicles and the externalities produced. Location affects each of these costs. In general, there is a tradeoff between minimizing the costs from freight vehicle travel (by locating closer to the major delivery areas) and minimizing the impacts on local communities (by locating farther from the city center). In this context, it is crucial to identify and implement complementary initiatives aimed at mitigating, or eliminating altogether, the negative effects on local communities. Anticipating these externalities and putting in place mitigation strategies facilitate progress toward FELUs. As illustrated in this application, the BMS-FELU can help policymakers understand the potential impact of their land-use planning choices and help them choose policies and programs that have a positive effect on freight efficiency. The BMS-FELU can be applied to a variety of sce- narios or initiatives (see Chapter 7 for an overview of land-use initiatives) by allowing users to adjust (1) freight supply and demand at each establishment and at each freight gateway, (2) how industry sectors interact, and (3) parameters to construct routes and delivery tours.

240 Planning Freight-Efcient Land Uses: Methodology, Strategies, and Tools Figure 32. Trip distribution for large establishments after relocation of distribution center in Amsterdam (left) and Colonie (right).