• Air Toxics
• CMAQ
• Conformity
• It All Adds Up

• Laws and Regulations
• Policy and Guidance
• Research and Analysis
• Training

FHWA and EPA National Near-Road Study Las Vegas

4 Site Selection Methodology and Site Selection Criteria

4.1 Methods

The site selection process consisted of a series of seven steps as shown in Table 6 and Figure 12. Each of these steps has varying degrees of complexity due to "real-world" issues. The first step, determining site selection criteria (Table 7), had been developed by FHWA prior to the site selection process and documented in the monitoring protocol ². The follow-on steps (Table 6) include: (1) determine site selection criteria; (2) develop list of candidate sites and supporting information; (3) apply site selection filter ("coarse" and "fine"), (4) site visit; (5) select candidate site(s) via team discussion; (6) obtain site access permission(s); and (7) implement site logistics.

Table 6. Site Selection Process Steps.

Figure 12. Decision Process Schematic.

Table 7. Selection Considerations and Settlement Agreement Criteria.

4.2 Candidate Site Listing

The purpose of any site selection process is to gather and analyze sufficient data that would lead one to draw informed conclusions regarding the selection of the most appropriate site for the monitoring that will be performed in Las Vegas, NV.

A list of candidate sites was developed using the monitoring protocol's site selection criteria. We used geographic information system (GIS) data, tools and techniques and on-site visits by project team members as a means of developing supporting information regarding each potential site. The Nevada Department of Transportation (NDOT) provided annual average daily traffic (AADT) counts and their associated spatial coordinate locations. Other types of spatial data (e.g., street network) were downloaded from the Clark County GIS web site as well as other relevant web sites. Non-spatial data (meteorological data) were downloaded from the National Climatic Data Center for Las Vegas, NV. ArcGIS 9.2 was used to create the maps for the site selection process. WRPLOT View by Lakes Environmental was used to create wind rose plots for the meteorological data. In addition, site visits provided information not readily available elsewhere or provided information not easily gained from site maps.

Initially, the targeted site to conduct the monitoring in Las Vegas, NV was at the O.K. Adcock Elementary School. This was one of the three schools named in the settlement agreement. However, after further investigation and analysis it became apparent that a more formalized site selection process to either confirm or deny the suitability of this site for our project would be required. During this process, the O.K. Adcock Elementary School became increasingly less acceptable as the optimal site owing to the presence of large noise barriers (> 15 feet in height), very poor quality wind rose data (prevailing winds channeled down roadway as opposed to across), lack of access for site installation (no access at roadside due to noise barriers) and the roadway being below grade. In essence, the O.K. Adcock site is on top of an urban canyon. Thus, it became necessary to expand our search for a more optimal site through a more formal process.

Stepping through the process, a list of 19 of the 22 sites was developed (See Table 8). Three additional sites were added during an on-site visit to Las Vegas during the Spring of 2007. This list contains sites that are located along interstate or U.S. highways, State highways or major streets that would be of interest to a project of this nature. At this point, the list also included sites that might be below certain minimum requirements (e.g., AADT < 100,000).

Table 8. Table of All Sites Considered.

4.3 Application of Coarse Site Selection Filter

A series of joint-agency team meetings were held which resulted in reducing the number of sites from a list of 22 to approximately four sites. ^14,15 This step involved a review of the full table of candidate sites and eliminating sites that obviously did not meet minimum site criteria requirements (Table 7). After applying the site selection criteria as a set of "filters" we eliminated most candidate sites. For example, the first filter eliminated sites with low AADT (i.e., < 100,000). Further, the presence of extensive noise barriers eliminated additional sites. Other filters, complex geometric design or lack of available traffic volume data eliminated additional sites. Other criteria of interest, while not explicitly stated in the monitoring protocol, include restricted downwind sampling, presence of confounding air pollutant sources and site access (administrative and physical).

4.4 Ground Truthing

During the first site visit to Las Vegas, all sites considered to be primary sites were visited, i.e., Sites 6, 11, 1 and 15. However, these sites were considered to be unsuitable based on previously unforeseen factors that had not been readily obvious from our earlier analysis. These factors included the roadway being above/below grade and presence of noise barriers. Site 19, which was also visited, was deemed to be unsuitable for several reasons: confounding winds, adjacent railroad (main line) and the presence of confounding air pollutant sources (nearby vehicle scrappage plants). Moreover, the Nevada Department of Transportation (NDOT) staff provided information indicating that this location would be in the path of an upcoming "design-build" highway project thus making this location unsuitable.

During this site visit the remaining, available candidate sites were also eliminated. It was also during this visit that three additional sites were added to the list of candidate sites (Table 8). These additional sites were added based on "real-world" observations at these locations and how well these locations met the site selection criteria. It is important to note that it is not always possible to determine the most suitable sites without an actual site visit. Thus, sites may be added or deleted based on factors not previously known. In this case, these additional candidate sites were Sites 20, 21 and 22. Site 20 was deemed the most promising for several reasons: high AADT (> 190,000), lack of noise barriers, road being at or near grade, acceptable downwind sampling, and acceptable meteorology.

An important component of "ground truthing" during the site visit was to obtain information from local sources. For purposes of this project, both NDOT and Clark County Department of Air Quality and Environmental Management (DAQEM) staff provided information regarding local meteorological and road conditions that would have been difficult, if not impossible, to obtain otherwise. Too often, local resources are overlooked during a decision process such as this which can in turn lead to poor decisionmaking ^14,15.

4.5 Geospatial Tools

Historically, the use of spatial tools (e.g., GIS) in decision processes has been somewhat problematic in part due to the magnitude of the data required by a GIS, perception and reality of operating GIS software, and the level-of-knowledge required by end-users to manipulate data in a GIS. ^16,17 In the last 15 years, GIS data have become more readily available in both quantity and quality. Moreover, the realization of easy-to-use GIS software and GIS tools (in a Microsoft Windows® environment) has made implementation of GIS-based decision support tools more practical.

An example of the use of GIS as an environmental site selection decision support tool is its use in siting a landfill ¹⁸. Required typical data layers include: the location of suitable soils, wells, surface water sources, residential areas, schools, airports, roads, etc. From these data layers queries are formulated to provide the most suitable site(s). For example, a landfill should not be in the vicinity of an airport due to safety issues (i.e., aircraft striking birds) but would require suitable soil (i.e., soils with low permeability). A landfill should not be in the vicinity of wells or other water sources due to the issue of landfill leaching. Quantitative weighting criteria are associated with the siting criteria as well as elements of the data layers (e.g., certain types of soils would be more suitable than others and thus would have applicable quantitative values) ¹⁸.

Landfill siting by GIS has both similarities and differences to the near-road monitoring siting described by this report. Data layers relevant to near-road air quality monitoring are shown in Table 9. Numerous maps and based on near-road monitor site selection criteria (Table 7) were applied; advantages and disadvantages of each site were considered, with the ultimate decision that Site 20 would be the optimal site. The significant difference between air monitoring site selection process and the landfill example is that the former did not explicitly assign quantitative values to the selection criteria ^16,19. However, during the site visit and subsequent discussions quantitative values were implicitly assigned to the selection criteria. For example, sites with high AADT (> 150,000) were more highly "valued' than sites with a lower AADT; and sites without noise barriers were "valued" more highly than sites with noise barriers. Thus, the decision process may appear to be based on a high degree of subjectivity given the lack of assigning quantitative values to the selection criteria. The process of intra-team communication and application of our site selection criteria ^16,19 appropriately led to the selection of a site for this project based on numerous criteria.

ArcGIS 9.2 ²⁰ was used to create the maps used in the site selection process. Spatial data were downloaded from the Clark County GIS web site as well as other relevant web sites. Table 9 shows the sources of data used for this site selection process. It should be noted that the use of maps for the site selection process is only a tool in the site selection process. It is very important in this process to perform site visits, establish contacts with State/local transportation agencies and environmental agencies. Typically, these groups will be able to provide up-to-date information as to site conditions that may ultimately influence site selection decisions.

Table 9. Spatial and Non-Spatial Data Inputs.

4.6 Site Selection - Results and Discussion

Following the application of the selection criteria^14,15,21, the candidate sites were further prioritized during a series of team discussions between EPA and FHWA staff. The pros and cons of each site were discussed. Site 20 met or exceeded the AADT requirements (> 150,000), did not have noise barriers, did have acceptable downwind sampling and acceptable meteorology. An additional feature of this site was the availability of traffic count data. Traffic monitoring equipment was already installed and operating in the vicinity of this location.

Proper siting of downwind sampling locations was an important criterion for this project. Any location where proper siting of downwind sampling sites was restricted due to topology, existing structures, meteorology, or other conditions, excluded otherwise suitable sites for consideration and inclusion in this study. Meteorological conditions in Las Vegas were problematic as the city is located in a valley surrounded by mountains that channel the wind as shown in Figure 13. This channeling of wind presented technical challenges in site selection and achieving proper wind flow from the source to the detector (i.e., gas analyzers). As shown in Figure 14, the wind direction for Site 20 did have acceptable meteorological conditions. Site 20, on the other hand, is 1 km west of McCarran International Airport (nearby source) and is slightly below grade (modest cut) as shown in Figure 15. Another feature of note at this location was a spur line of the Union Pacific Railroad (UPRR). This is a commodity line that runs approximately 12 miles from Las Vegas, NV to Henderson, NV and passes this site twice a day (once in the morning and returns in the afternoon).

A construction project, in the vicinity of Site 20, involved the conversion of the inner shoulders and median to express lanes (lanes in either direction of travel). Based on information from the NDOT design engineer, there was minimal impact to our monitoring project for the following reasons: (1) construction involved the addition of center lanes-our project was on the shoulder behind a guardrail to the east of any construction activity; (2) the segment of roadway carries > 200,000 vehicles per day - diesel emissions from construction equipment were overwhelmed by the sheer volume of the on-highway vehicles; and (3) construction vehicles were not operated in "front of monitoring station" 24-hours per day-construction activity occurred along a 5-6 mile stretch of the freeway. Therefore, while not perfect, Site 20 was still a viable site.

Right-of-way (ROW) access was needed from both Nevada DOT and the UPRR. Other property owners in the vicinity of our location were reluctant to permit access to their property. Property owners may be reluctant to grant ROW access for a variety of reasons, including liability, financial issues, or suspicion of government activities. Therefore, access to any given property was not guaranteed. and researchers must be prepared for a long involved access process.

Figure 13. Las Vegas Topographic and Meteorological Conditions.

Source: Adapted from ²²

Figure 14. Wind Roses for Site 20 (McCarran International Airport).

Figure 15. Photos of Site 20.

The project participants recognized that no perfect air monitoring site was possible; trade-offs were a factor of the Las Vegas study and would be a factor with almost any other environmental study conducted within any other city. It was a question of balancing benefits with risks and costs. The selection was further complicated by external constraints and drivers. The principal constraint was the legal mandates of the Settlement Agreement, especially the data that must be derived pursuant to the monitoring protocol. Few, if any, design decisions can be made exclusively from a single perspective. These decisions can be visualized as attractions within a force field. If the factors are evenly distributed and weighted, the diagram might appear as that in Figure 16a and b. But, as a given differential force increases, that factor will progressively drive the decision. In the present case study, the decision is most directly influenced by legal requirements, but also needs to be scientifically credible and economically feasible (Figure 16a).

Figure 16a. Decision Force Field - Equal Weight Factors. (b). Decision Force Field - Unequal Weight Factors. Source: Adapted from D. Vallero and C. Brasier. (2008). Sustainable Design: The Science of Sustainability and Green Engineering. John Wiley & Sons, Inc. Hoboken, NJ.

As shown in Figure 16a, a number of factors have nearly equal weight in a design decision, Figure 16b indicates a decision that is most strongly influenced by legal constraints and drivers. For example, the stronger the influence of a factor (e.g., high AADT), the greater the decision will be drawn to that perspective. If the monitoring protocol is somewhat ambiguous, a number of alternatives are available, costs are flexible, and scientific credibility is minimally impacted, the design has a relatively large degree of latitude and elasticity. Note that all factors drive the decision, but that the monitoring protocol and other legal instruments have the greatest influence on the decision.

There is also the question of the best use of resources for this project. For example, a site could be chosen that would call for additional monitoring (and concomitantly additional resources) to overcome certain physical constraints (i.e., above/below grade, noise barriers); or, a site could be chosen that has some other issue such as low AADT or where traffic monitoring equipment would have to be installed. Some sites that would otherwise be favorable are near open areas, such as desert land that is prone to fugitive dust (common in unpaved or un-irrigated areas in Las Vegas).

Site 20 had high AADT (206,000 AADT for 2006), no noise barriers, meteorological and traffic data availability, manageable site logistics including right-of-way access, "clean geometric design", and favorable wind direction. Clean geometric design was defined as a facility that did not impede the effective data collection of MSATs and PM_2.5 ². For example, a clean geometric design would be a site that did not include multiple on/off ramps, interchanges, or other complicating facility characteristics. Of the disadvantages, Site 20 had a modest "cut" at that location and only for a short distance. The roadway passed under a railroad bridge (Figure 15) and returned to at/near grade conditions. Thus, this site was much more suitable than the urban canyon situation that exists with the O.K. Adcock site. In addition, McCarran International Airport, a source of vehicle and aircraft emissions, was approximately 1 km due east of Site 20; however the predominant winds at this location generally maintain the monitoring site upwind of the airport. The wind blows predominately from the southwest quadrant until approximately 3:00 pm. Beginning at about 3:00 pm in the afternoon, the winds shift and become more variable. In the late afternoon and early evening (6-9:00 pm), the winds are predominately from the east and northeast. Then, after 9:00 pm, the winds shift again into a more variable pattern. After midnight the winds return to blow predominately from the southwest. The wind roses shown in Figure 14 have been developed from meteorological data downloaded from the National Climatic Data Center, Years 1977 - 2007.

4.7 Site Logistics

Site logistics included, but were not limited to, obtaining site access permissions, gaining access to electrical power, communications connectivity, county/city permits, or arranging for security fencing. Site logistics, while not explicitly included in the monitoring protocol, was mission critical. Any location where site logistics, was restricted or prohibited either due to administrative or physical issues, was highly problematic and eliminated a site from further development. For this specific project, obtaining site access permissions, obtaining the proper electrical feed, communications connectivity and being able to establish security fencing was vital to the project.

Electrical and communications connectivity is also a very challenging activity. Utility companies have a multitude of requirements for obtaining their services. This is a very involved process that requires interactions with utility companies as well as local (i.e., county or city) inspections departments. Implementation of site logistics may require more time than is needed to obtain ambient air measurements for a given project.

4.8 Site Selection Summary/Conclusions

Site 20 was the site of choice with the most advantages and fewest disadvantages compared to other monitoring sites that have been considered. Site 20 had high AADT (206,000 AADT for 2006), no noise barriers, meteorological and traffic data availability, manageable site logistics including ROW access, and favorable wind direction. We attempted to develop the site somewhat further south than our final location. This would have been the more ideal location as the 20-meter roadside site would have been an at-grade site. However, the property owners would not agree to allow site access.

Meetings, teleconferences, site visits, and written reportsare key activities to any site selection process. These activities ensure that all interested parties are aware of the selection process. Moreover, this ensures that the pros and cons of each site are thoroughly considered and discussed and trade-offs among the various sites are weighed. Throughout this process, trade-offs will and do occur. For example, an ideal site for the air quality modelers (e.g., complex terrain) is not necessarily an ideal site from the perspective of the field researchers (e.g., less complex terrain, site access, etc.). For this project, a group consensus was reached, culminating in a written report for FHWA and U.S. EPA management.